EPA/600/3-89/061 C
                                           July 1989
          Direct/Delayed Response  Project:
   Future Effects of Long-Term Sulfur Deposition
             on Surface Water Chemistry
in the Northeast and Southern Blue Ridge Province
         Volume HI:  Level III Analyses and
                  Summary of Results
                            by

   M. R. Church, K. W. Thornton, P. W. Shaffer, D. L Stevens, B. P. Rochelle,
      G. R. Holdren, M. G. Johnson, J. J. Lee, R. S. Turner, D. L. Cassell,
      D. A. Lammers, W. G. Campbell, C. I. Liff, C. C. Brandt, L. H. Liegel,
       G. D. Bishop, D. C. Mortenson, S. M. Pierson, D. D. Schmoyer
                      A Contribution to the
            National Acid Precipitation Assessment Program
              U.S. Environmental Protection Agency
    Office of Research and Development, Washington, DC 20460
    Environmental Research Laboratory, Corvallis, Oregon 97333

-------
                                        NOTICE
The information in this document has been funded wholly (or in part) by the U.S. Environmental
Protection Agency.  It has been subjected to the Agency's peer and administrative review, and it has
been approved for publication as an EPA document.  Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.

-------
                                      CONTENTS
SECTION
Notice	  jj
Tables  	   Xjj
Figures	   xx
Plates	   xxix
Contributors	   xxxj
Acknowledgments	  xxxiii

1    EXECUTIVE SUMMARY	                   1
     1.1  INTRODUCTION	  1
          1.1.1  Project Background	  1
          1.1.2  Primary Objectives	  2
          1.1.3  Study Regions  	  2
          1.1.4 Time Frames of Concern  	  2
     1.2 PROCESSES OF ACIDIFICATION   	'.'.'.'.'.'.'.'.'.'.  4
          1.2.1 Sulfur Retention	  4
          1.2.2  Base Cation Supply	  4
     1.3 GENERAL APPROACH  	'.'.'.',  5
          1.3.1 Soil Survey	  5
          1.3.2 Other Regional Datasets	  7
          1.3.3 Scenarios of Atmospheric Deposition  .	  7
          1.3.4  Data Analysis   	                        7
     1.4 RESULTS 	  8
          1.4.1  Retention of Atmospherically Deposited Sulfur	  8
               1.4.1.1  Current Retention  	  8
               1.4.1.2  Projected Retention	  8
          1.4.2 Base Cation Supply	10
               1.4.2.1  Current Control	10
               1.4.2.2  Future Effects	10
          1.4.3 Integrated Effects on Surface Water ANC 	12
               1.4.3.1  Northeast Lakes  	12
               1.4.3.2  Southern Blue Ridge Province	      15
     1.5  SUMMARY DISCUSSION	18
     1.6  REFERENCES	i 18

2    INTRODUCTION TO THE DIRECT/DELAYED RESPONSE PROJECT	       23
     2.1  PROJECT BACKGROUND	23
     2.2  PRIMARY OBJECTIVES  	                           24
     2.3  STUDY REGIONS  	24
     2.4  TIME FRAMES OF CONCERN  	  27
     2.5  PROJECT PARTICIPANTS	27
     2.6  REPORTING  	27

3    PROCESSES OF ACIDIFICATION  	           29
     3.1  INTRODUCTION	29
     3.2  FOCUS OF THE DIRECT/DELAYED RESPONSE PROJECT  	     30
     3.3  SULFUR RETENTION  PROCESSES	30
          3.3.1 Introduction  	30
          3.3.2 Inputs  	31

-------
                                   CONTENTS (Continued)
          3.3.3 Controls on Sulfate Mobility within Forest/Soil Systems	32
                3.3.3.1  Precipitation/Dissolution of Secondary Sulfate Minerals	32
                3.3.3.2  Sulfate Reduction in Soils and Sediments	32
                3.3.3.3  Plant Uptake	34
                3.3.3.4  Retention as Soil Organic Sulfur 	34
                3.3.3.5  Sulfate Adsorption by Soils	35
          3.3.4 Models of Sulfur Retention	37
          3.3.5 Summary	38
     3.4 BASE CATION SUPPLY PROCESSES	39
          3.4.1 Introduction	39
          3.4.2 Factors Affecting Base Cation Availability  	42
                3.4.2.1  Mineral Weathering  	42
                3.4.2.2  Cation Exchange Processes	45
          3.4.3 Modelling  Cation Supply Processes  	47
                3.4.3.1  Modelling Weathering	47
                3.4.3.2  Modelling Cation Exchange Processes	48

4    PROJECT APPROACH	49
     4.1 INTRODUCTION	49
     4.2 SOIL SURVEY	49
          4.2.1 Watershed Selection	49
          4.2.2 Watershed Mapping	49
          4.2.3 Sample Class Definition	51
          4.2.4 Soil Sampling	.51
          4.2.5 Sample Analysis	51
          4.2.6 Database Management  	51
     4.3 OTHER REGIONAL DATASETS      	51
          4.3.1 Atmospheric Deposition	52
          4.3.2 Runoff Depth	52
     4.4 DATA ANALYSIS  	52
          4.4.1 Level I Analyses	53
          4.4.2 Level II  Analyses	53
          4.4.3 Level III Analyses	53
          4.4.4 Integration of Results  	54
          4.4.5 Use of a Geographic Information System	54

5    DATA SOURCES AND DESCRIPTIONS  	55
     5.1 INTRODUCTION	55
     5.2 STUDY SITE SELECTION  	55
          5.2.1 Site Selection Procedures	55
          5.2.2 Eastern Lake Survey Phase I Design  	55
          5.2.3 Pilot Stream Survey Design	58
          5.2.4 DDRP Target Population	58
                5.2.4.1  Northeast Lake Selection  	58
                5.2.4.2  Southern Blue Ridge Province Stream Selection   	60
                5.2.4.3  Final DDRP Target Populations  	   82
     5.3 NSWS LAKE AND STREAM DATA  	   82
          5.3.1 Lakes in the Northeast Region	   82
                5.3.1.1  Lake Hydrologic Type  	   82
                5.3.1.2  Fall Index Sampling	   82
                5.3.1.3  Chemistry of DDRP Lakes	   89
                                             IV

-------
                             CONTENTS (Continued)
     5.3.2  Streams in the Southern Blue Ridge Province Region	  91
          5.3.2.1  Spring Baseflow Index Sampling  	  91
          5.3.2.2  Chemistry of DDRP Stream Reaches  	  93
5.4 MAPPING PROCEDURES AND DATABASES   	  93
     5.4.1 Northeast Mapping	  95
          5.4.1.1  Soils	  95
          5.4.1.2  Depth to Bedrock 	  99
          5.4.1.3  Forest Cover Type   	101
          5.4.1.4  Bedrock Geology    	101
          5.4.1.5  Quality Assurance   	101
          5.4.1.6  Land Use/Wetlands   	105
          5.4.1.7  Geographic Information Systems Data Entry 	118
     5.4.2 Southern Blue Ridge Province Mapping	132
          5.4.2.1  Soils   	134
          5.4.2.2  Depth to Bedrock	137
          5.4.2.3  Forest Cover Type/Land Use	137
          5.4.2.4  Bedrock Geology          	137
          5.4.2.5  Drainage 	139
          5.4.2.6  Quality Assurance    	139
          5.4.2.7  Land Use/Wetlands	142
          5.4.2.8  Geographic Information Systems Data Entry 	143
5.5 SOIL SAMPLING PROCEDURES AND DATABASES	146
     5.5.1 Development/Description of Sampling Classes	147
          5.5.1.1  Rationale/Need for Sampling Classes   	147
          5.5.1.2  Approach Used for Sampling Class Development  	147
          5.5.1.3  Description  of Sampling Classes   	148
     5.5.2 Selection of Sampling Sites	;	150
          5.5.2.1  Routine Samples  	150
          5.5.2.2  Samples on Special Interest Watersheds  	155
     5.5.3  Soil Sampling	155
          5.5.3.1  Soil Sampling Procedures  	156
          5.5.3.2  Quality Assurance/Quality Control of Sampling   	156
     5.5.4  Physical and Chemical Analyses  	157
          5.5.4.1  Preparation Laboratories	157
          5.5.4.2  Analytical Laboratories	159
     5.5.5  Database Management	167
          5.5.5.1  Database Structure	172
          5.5.5.2  Database Operations	174
     5.5.6  Data Summary  	178
          5.5.6.1  Summary of Sampling Class Data	178
          5.5.6.2  Cumulative  Distribution Functions  	178
5.6 DEPOSITION DATA  	178
     5.6.1  Time Horizons of Interest 	190
          5.6.1.1  Current Deposition	190
          5.6.1.2  Future Deposition 	190
     5.6.2  Temporal Resolution	190
          5.6.2.1  Level I Analyses  	190
          5.6.2.2  Level II Analyses	190
          5.6.2.3  Level III Analyses	190

-------
                                   CONTENTS (Continued)
                                                                                      Page
           5.6.3 Data Acquisition/Generation	192
                5.6.3.1  Level III Analyses - Typical Year Deposition Dataset	192
                5.6.3.2  Level I  and II Analyses - Long-Term Annual Average
                        Deposition Dataset	208
           5.6.4 Deposition Datasets Used in DDRP Analyses	                  224
     5.7  HYDROLOGIC DATA	224
           5.7.1 Runoff	224
                5.7.1.1  Data Sources	224
                5.7A.2  Runoff  Interpolation Methods	224
                5.7.1.3  Uncertainty Estimates	227
           5.7.2 Derived Hydrologic Parameters	227
                5.7.2.1  TOPMODEL 	.228
                5.7.2.2  Soil Contact (Darcy's Law)    	231
                5.7.2.3  Mapped Hydrologic Indices      	234

6    REGIONAL POPULATION ESTIMATION	                           242
     6.1  INTRODUCTION	     242
     6.2  PROCEDURE	 .  242
           6.2.1 Use of Variable Probability Samples	242
           6.2.2 Estimation Procedures for Population Means	243
           6.2.3 Estimators of Variance	244
           6.2.4 Estimator of Cumulative Distribution Function	245
     6.3 UNCERTAINTY ESTIMATES  	245
     6.4  APPLICABILITY  	246

7    WATERSHED SULFUR RETENTION	                        247
     7.1  INTRODUCTION	247
     7.2  RETENTION IN LAKES AND WETLANDS	248
     7.2.1  Introduction   	248
           7.2.2 Approach	249
           7.2.3 Results	251
     7.3  WATERSHED SULFUR RETENTION	253
           7.3.1  Methods	253
                7.3.1.1  Input/Output Calculation	253
                7.3.1.2  Data Sources   	255
           7.3.2 Uncertainty Estimates	255
                7.3.2.1  Introduction  	255
                7.3.2.2  Individual  Variable Uncertainties	255
                7.3.2.3  Uncertainty Calculation - Monte Carlo Analysis  	260
           7.3.3 Internal Sources of Sulfur	262
                7.3.3.1  Introduction/Approach  	262
                7.3.3.2  Bedrock Geology	 .  . 662
                7.3.3.3  Upper Limit Steady-State Sulfate Concentration   	265
           7.3.4 Results and Discussion  	268
                7.3.4.1  Northeast   	271
                7.3.4.2  Mid-Appalachians  	279
                7.3.4.3  Southern Blue Ridge Province	280
                7.3.4.4  Conclusions	280
                                            vi

-------
                             CONTENTS (Continued)
LEVEL I STATISTICAL ANALYSES	285
8.1  INTRODUCTION	285
     8.1.1  Approach	285
     8.1.2  Statistical Methods	286
8.2   RELATIONSHIPS BETWEEN  ATMOSPHERIC  DEPOSITION  AND  SURFACE
     WATER CHEMISTRY   	291
     8.2.1  Introduction	291
     8.2.2  Approach	291
     8.2.3  Results and Discussion	292
          8.2.3.1 Northeast	292
          8.2.3.2 Southern Blue Ridge Province	292
          8.2.3.3 Summary	292
8.3  DERIVED HYDROLOGIC PARAMETERS	295
     8.3.1  Soil Contact (Darcv's Law)  	295
          8.3.1.1 Introduction  	295
          8.3.1.2 Results and Discussion  	299
     8.3.2  Geomorphic/Hvdroloqic Parameters  	302
          8.3.2.1 Introduction  	302
          8.3.2.2 Results and Discussion  	310
     8.3.3  TOPMODEL Parameters 	316
          8.3.3.1 Introduction  	317
          8.3.3.2 Results and Discussion  	317
          8.3.3.3 Summary	326
8.4  MAPPED BEDROCK GEOLOGY	326
     8.4.1  DDRP Bedrock Sensitivity Scale	 327
     8.4.2  Results	328
          8.4.2.1 Sulfate and Percent Retention	332
          8.4.2.2 Sum of Base Cations, ANC, and pH  	335
     8.4.3  Summary   	336
8.5  MAPPED LAND USE/VEGETATION	337
     8.5.1  Introduction  	337
     8.5.2  Data Sources   	337
     8.5.3  Statistical Methods	338
     8.5.4  Sulfate and Percent Sulfur Retention  	338
          8.5.4.1 Northeast  	338
          8.5.4.2 Southern Blue Ridge Province	347
          8.5.4.3 Regional Comparisons	347
     8.5.5  ANC. Ca plus Mg. and pH  	347
          8.5.5.1 Northeast	347
          8.5.5.2 Southern Blue Ridge Province	349
          8.5.5.3 Regional Comparisons	349
     8.5.6  Summary and Conclusions	351
8.6  MAPPED SOILS	351
     8.6.1  Introduction  	351
     8.6.2  Approach	352
     8.6.3  Sulfate and Sulfur Retention 	354
          8.6.3.1 Northeast	360
          8.6.3.2 Southern Blue Ridge Province	362
          8.6.3.3 Regional Comparisons	365
                                      vii

-------
                             CONTENTS (Continued)
     8.6.4  ANC. Ca plus Mq. and pH  	367
          8.6.4.1  Northeast	367
          8.6.4.2  Southern Blue Ridge Province	369
          8.6.4.3  Regional Comparisons	377
     8.6.5  Summary and Conclusions	378
8.7 ANALYSES OF DEPTH  TO BEDROCK	379
     8.7.1  Introduction  	379
     8.7.2  Approach	379
     8.7.3  Sulfate and Percent Sulfur Retention  	381
          8.7.3.1  Northeast	381
          8.7.3.2  Southern Blue Ridge Province	381
          8.7.3.3  Comparison of Regions  	381
     8.7.4  ANC. Ca plus Ma  and pH	385
          8.7.4.1  Southern Blue Ridge Province	385
          8.7.4.2  Comparison of Regions  	386
     8.7.5  Summary and Conclusions	386
8.8 INTEGRATED ANALYSIS OF ALL MAPPED VARIABLES	388
     8.8.1  Introduction  	388
     8.8.2  Approach	388
     8.8.3  Sulfate and Sulfur Retention	388
          8.8.3.1  Northeast	388
          8.8.3.2  Southern Blue Ridge Province	390
          8.8.3.3  Regional Comparisons	392
     8.8.4  ANC. Ca plus Mq. and pH  	393
          8.8.4.1  Northeast	393
          8.8.4.2  Southern Blue Ridge Province	395
          8.8.4.3  Regional Comparisons	398
     8.8.5  Summary and Conclusions	398
8.9 SOIL PHYSICAL AND CHEMICAL CHARACTERISTICS	399
     8.9.1  Introduction  	399
     8.9.2  Approach	399
          8.9.2.1  Statistical Methods  	400
     8.9.3  Aggregation of Soil Data	402
          8.9.3.1  Introduction  	402
          8.9.3.2  Aggregation of Soil Data  	403
          8.9.3.3  Assessment of the DDRP Aggregation Approach	404
          8.9.3.4  Estimation of Watershed  Effect  	406
          8.9.3.5  Evaluation of Watershed  Effect  	407
     8.9.4  Regional Soil Characterization	407
     8.9.5  Sulfate and Sulfur Retention  . .  .	413
          8.9.5.1  Northeast	418
          8.9.5.2  Southern Blue Ridge Province	421
     8.9.6  Ca plus Mq  (SOBC). ANC. and pH   	421
          8.9.6.1  Northeast	421
          8.9.6.2  Southern Blue Ridge Province	425
     8.9.7  Evaluation of Alternative Aggregation Schemes  	426
     8.9.8  Summary and Conclusions	426
          8.9.8.1  Alternative Aggregation Schemes	426
          8.9.8.2  Sulfate and Sulfur Retention	429
          8.9.8.3  Ca plus Mg (SOBC), ANC, and pH  	429
     8.9.9  Summary Conclusions  	430
                                       viii

-------
                                  CONTENTS (Continued)
     8.10 EVALUATION  OF ASSOCIATIONS  BETWEEN  WATERSHED ATTRIBUTES  AND
          SURFACE WATER CHEMISTRY  	430
          8.10.1  Introduction	430
          8.10.2  Approach	431
          8.10.3  Regional Characterization of Watershed Attributes	431
               8.10.3.1  Northeast Subregions	431
               8.10.3.2 Northeast and Southern Blue Ridge Providence	 435
          8.10.4  Sulfate and Sulfur Retention	 436
               8.10.4.1  Northeast  	436
               8.10.4.2 Southern Blue Ridge Province   	436
          8.10.5  Ca plus Ma (SOBC). ANC.  and pH	437
               8.10.5.1  Northeast  	437
               8.10.5.2 Southern Blue Ridge Province   	437
          8.10.6  Summary and Conclusions  	450
               8.10.6.1  Sulfate and Sulfur Retention 	450
               8.10.6.2 Ca plus Mg (SOBC), ANC, and pH   	450
          8.10.7  Summary Conclusions  	450

9    LEVEL II ANALYSES - SINGLE FACTOR  RESPONSE TIME ESTIMATES  	452
     9.1  INTRODUCTION	452
     9.2  EFFECTS OF SULFATE ADSORPTION ON WATERSHED SULFUR RESPONSE TIME  . 453
          9.2.1  Introduction  	453
          9.2.2  Section Objectives	454
          9.2.3  Approach	455
               9.2.3.1  Model Description    	455
               9.2.3.2  Data Sources	456
               9.2.3.3  Model Assumptions and Limitations	456
               9.2.3.4  Adsorption Data  	458
               9.2.3.5  Evaluation of Aggregated Data and Model Outputs  	461
               9.2.3.6  Target Populations for Model Projections   	462
          9.2.4  Results	464
               9.2.4.1  Comparison of Northeast and Southern Blue Ridge Province Isotherm
                       Variables  	464
               9.2.4.2  Model Results - Northeastern United States  	466
               9.2.4.3  Model Results - Southern Blue Ridge  Province   	479
               9.2.4.4  Uncertainty Analyses and Alternative Aggregation Approaches	493
               9.2.4.5  Summary of Results  from the Southern Blue Ridge Province   	.501
          9.2.5  Summary  Comments on Level II Sulfate Analyses	502
          9.2.6  Conclusions  	504
     9.3  EFFECT OF CATION EXCHANGE AND WEATHERING ON SYSTEM RESPONSE .... 506
          9.3.1  Introduction  	506
               9.3.1.1  Level II Hypotheses	506
               9.3.1.2  Approach	509
          9.3.2  Descriptions of  Models	512
               9.3.2.1  Reuss Model	512
               9.3.2.2  Bloom-Grigal Model   	'.	527
          9.3.3  Model Forecasts  	533
               9.3.3.1  Reuss Model  	535
               9.3.3.2  Bloom-Grigal Model	577
          9.3.4  Comparison of the Bloom-Grigal and Reuss Model Projections	605
          9.3.5  Summary  and Conclusions	612
                                           IX

-------
                                  CONTENTS (Continued)
10   LEVEL III ANALYSES - DYNAMIC WATERSHED MODELLING  	618
     10.1  INTRODUCTION	618
     10.2  DYNAMIC WATERSHED MODELS	620
          10.2.1 Enhanced Trickle Down (ETD) Model	. . . .	622
          10.2.2 Integrated Lake-Watershed Acidification Study (ILWAS) Model	627
          10.2.3 Model of Acidification of Groundwater in Catchments (MAGIC)	628
     10.3  OPERATIONAL ASSUMPTIONS   	629
     10.4  WATERSHED PRIORITIZATION	629
          10.4.1 Northeast	629
          10.4.2 Southern Blue Ridge Province	632
          10.4.3 Effects of Prioritization on Inclusion Probabilities	  632
     10.5  MODELLING DATASETS  	634
          10.5.1 Meteorological/Deposition Data	634
          10.5.2 DDRP Runoff Estimation	634
               10.5.2.1  Annual Runoff	634
               10.5.2.2 Monthly Runoff 	,	635
          10.5.3 Morphometry  	636
          10.5.4 Soils	636
          10.5.5 Surface Water Chemistry	637
          10.5.6 Other Data	637
          10.5.7 Chloride Imbalance	637
     10.6  GENERAL APPROACH  	639
     10.7  MODEL CALIBRATION	642
          10.7.1 Special Interest Watersheds  	642
               10.7.1.1  Northeast	643
               10.7.1.2 Southern Blue Ridge Province	643
          10.7.2 General Calibration Approach	644
          10.7.3 Calibration of the Enhanced Trickle Down Model	644
          10.7.4 Calibration of the Integrated Lake-Watershed Acidification Model  	647
          10.7.5 Calibration of the Model of Acidification of Groundwater in Catchments	650
          10.7.6 Calibration/Confirmation Results	652
     10.8  MODEL SENSITIVITY ANALYSES  	656
          10.8.1 General Approach	657
          10.8.2 Sensitivity Results	667
     10.9  REGIONAL PROJECTIONS REFINEMENT	658
          10.9.1 Enhanced Trickle Down	658
          10.9.2 Integrated Lake-Watershed Acidification Study	659
          10.9.3 Model of Acidification of Groundwater in Catchments	659
          10.9.4 DDRP Watershed Calibrations	661
               10.9.4.1  Integrated Lake-Watershed Acidification Study	661
               10.9.4.2  MAGIC	664
               10.9.4.3  Southern Blue Ridge Province	664
     10.10  MODEL  PROJECTIONS	668
          10.10.1 General  Approach	668
          10.10.2 Forecast Uncertainty  	672
               10.10.2.1 Watershed Selection	672
               10.10.2.2  Uncertainty Estimation  Approaches	673
               10.10.2.3  Relationship Among Approaches  	674
               10.10.2.4 Confidence Intervals	678

-------
                                 CONTENTS (continued)
     10.11  POPULATION ESTIMATION AND REGIONAL FORECASTS  	678
          10.11.1 Northeast Regional Projections  	678
               10.11.1.1  Target Population Projections Using MAGIC	678
               10.11.1.2  Target Population Projections Using MAGIC and ETD  	687
               10.11.1.3  Restricted Target Population Projections Using All Three Models . . .796
          10.11.2 Southern Blue Ridge Province	723
               10.11.2.1  Target Population Projections Using MAGIC	723
               10.11.2.2  Restricted Target Population Projections Using ILWAS and
                        MAGIC	749
          10.11.3  Regional Comparisons  	765
               10.11.3.1  Northeastern Projections of Sulfate Steady State  	765
               10.11.3.2  Southern Blue Ridge Province Projections of Sulfate
                        Steady State   	771
               10.11.3.3  ANC and Base Cation Dynamics  	771
     10.12  DISCUSSION   	790
          10.12.1  Future Projections of Changes in Acid-Base Surface Water Chemistry  ....  790
          10.12.2  Rate of Future Change	790
               10.12.2.1  Northeast	790
               10.12.2.2  Southern Blue Ridge Province	792
          10.12.3  Uncertainties and Implications for  Future  Changes in Surface Water
                  Acid-Base Chemistry	795
               10.12.3.1  Deposition Inputs	795
               10.12.3.2  Watershed Processes	797
     10.13  CONCLUSIONS FROM LEVEL III ANALYSES   	799

11   SUMMARY OF RESULTS	801
     11.1  RETENTION OF ATMOSPHERICALLY DEPOSITED SULFUR  	801
          11.1.1  Current Retention	801
          11.1.2  Projected Retention  	801
     11.2   BASE CATION SUPPLY  	805
          11.2.1  Current Control 	805
          11.2.2  Future Effects 	805
     11.3 INTEGRATED EFFECTS ON SURFACE WATER ANC	806
          11.3.1  Northeast Lakes	807
          11.3.2  Southern Blue Ridge Province  	814
     11.4  SUMMARY DISCUSSION	820

12   REFERENCES 	823

13   GLOSSARY	856
     13.1  ABBREVIATIONS AND SYMBOLS 	856
          13.1.1  Abbreviations	856
          13.1.2  Symbols	858
     13.2  DEFINITIONS	862

APPENDICES	888
                                           xi

-------
                                          TABLES
1-1.    Lakes in the NE Projected to Have ANC Values <0 and <50 ^eq L'1
       for Constant and Decreased Sulfur Deposition	  14
1-2.    SBRP Stream Reaches Projected to Have ANC Values <0 and <50 /jeq L'1
       for Constant and Increased Sulfur Deposition  	  17

3-1.    Major Rock Forming Minerals and Their Relative Reactivities  	  44

5-1.    Sampling Structure for Phase I, Region 1 (Northeast), Eastern Lake Survey  	  57
5-2.    Sample Structure for the Direct/Delayed Response Project -Northeastern Sample	  61
5-3.    ANC Group, Lake Identification, ELS-I Phase I ANC, Weight and Inclusion
       Probabilities for the Direct/Delayed Response Project Northeast Sample Watersheds ...  62
5-4.    Lake Identification and Name, and State and Latitudinal/Longitudinal Location
       of the Northeast Sample Watersheds	  66
5-5.    Lake Identification and Name, Sorted by State ~ Northeast Sample Watersheds	  69
5-6.    Stream  Identification, Weight, and Inclusion Probabilities for the Southern
       Blue Ridge Province Direct/Delayed Response Project Sample Watersheds  	  78
5-7.    Stream  Identification and Name, and State and Latitudinal/Longitudinal Location
       of the Southern Blue Ridge Province Sample Watersheds	  79
5-8.    Stream  Identification and Name, Sorted by State - Southern Blue Ridge Province
       Sample Watersheds	  80
5-9.    DDRP Reclassification  of Northeastern Lakes Classified  as "Seepage" or "Closed"
       by the NSWS  	  83
5-10.   Depth-to-Bedrock Classes and Corresponding Level of Confidence	   100
5-11.   Interpretation Codes for Northeast Map Overlays - Land Use/Land Cover,
       Wetlands, and Beaver Activity	   106
5-12.   Northeast Watersheds Studied for Independent Field Check of Land Use and
       Wetland Photointerpretations	   109
5-13.   Chi-Square Test for General Land Use Categories	   110
5-14.   Comparison of Field Check (Matched) General Land Use  Determinations with
       Office Photointerpretations  	   111
5-15.   Chi-Square Test for Detailed Wetland Categories	   113
5-16.   Comparison of Field Check (Matched) Detailed Wetland Determinations with
       Office Photointerpretations  	   114
5-17.   Comparison of Beaver Dam Number, Breached and Unbreached Status,
       and Lodges, Identified via Field Check and Office  Photointerpretation Methods	   115
5-18.   Aggregated Land Use  Data for  Northeast Watersheds	   117
5-19.   Watershed No. 1E1062 Soil Mapping Units   	   130
5-20.   Land Use Codes Used as Map Symbols   	   138
5-21.   Percent Land Use Data for Southern Blue Ridge Province Watersheds  	   144
5-22.   Laboratory Analysis of DDRP Soil Samples	   158
5-23.   Analytical Variables Measured in the DDRP  Soil Survey   	   160
5-24.   Data Quality Objectives for Detectability and Analytical Within-Batch Precision  	   163
5-25.   Detection Limits for Contract Requirements, Instrument Readings,
       and System-Wide Measurement in the Northeast   	   165
5-26.   Detection Limits for the Contract Requirements, Instrument Readings,
       and System-wide Measurement in the Southern Blue Ridge Province  	   166
5-27.   Attainment of Data Quality Objectives by the analytical laboratories as
       determined from blind  audit samples for the Northeast	   168
5-28.   Attainment of Data Quality Objectives by the Analytical Laboratories as Determined
       from Blind Audit  Samples for the  Southern Blue Ridge Province	   170
                                            XII

-------
                                    TABLES (Continued)
5-29.   Quality Assurance and Quality Control Checks Applied to Each Data Batch  	   176
5-30.   Medians of Pedon-Aggregated Values of Soil Variables for the DDRP
       Regions and Subregions	   189
5-31.   Monthly Values of Leaf Area Index Used to Apportion Annual Dry Deposition to
       Monthly Values	   202
5-32.   Ratios of Coarse-to-Fine Particle Dry Deposition	   205
5-33.   Ratios of Dry Deposition to Wet Deposition for DDRP Study Sites for the
       Typical Year Deposition Dataset	   207
5-34.   Deposition Datasets Used  in DDRP Analyses	   225
5-35.   DDRP texture classes and saturated hydraulic conductivity (K) for the NE
       study systems	   229
5-36.   SCS slope classifications	   233
5-37.   Mapped and calculated geomorphic parameters collected for the NE study sites. ....   236
5-38.   Mapped and calculated geomorphic parameters collected for the SBRP  study sites.  . .   240

7-1.    Summary of Computed Sulfur Retention by In-lake Reduction for Lake
       Systems in the Eastern United States	   250
7-2.    Intensively Studied Sites Used in Surface Water Chemistry Uncertainty Analysis	   257
7-3.    Summary Statistics on Percent Differences Between  Flow-weighted Average
       Annual Sulfate Concentration and the Fall/Spring Flow-weighted Averages	   261
7-4.    Bedrock Geology Maps Used in the DDRP Internal Sources of Sulfur
       Bedrock Geology Analyses	   263
7-5.    Potential for Sulfur Contribution by Geologic Type	   264
7-6.    Summary of Watersheds (by ELS and NSS Subregion) Dropped Due
       to Suspected Internal Sources of Sulfur Identified  by Steady-State Analysis	   270
7-7.    Percent Sulfur Retention -  Summary Statistics by Region	   273
7-8.    Summary of Sulfur Retention Status and of Watershed Variables
       Contributing to Sulfur Retention for 42 Watersheds in the Northeastern United States .   278

8-1.    Surface Water Chemistry and Percent Sulfur Retention Summary Statistics
       for the Northeastern DDRP Sample of 145 Lake Watersheds	   287
8-2.    Surface Water Chemistry and Percent Sulfur Retention Summary Statistics
       for the DDRP Sample of 35 SBRP Stream Watersheds	   288
8-3.    Summary Statistics for Wet and  Dry Deposition on the DDRP Sample
       of 145 Northeastern Lake Watersheds  	   289
8-4.    Summary Statistics for Wet and  Dry Deposition on the DDRP Sample of 35
       SBRP Stream Watersheds	   290
8-5.    Results  of Regressions Relating Surface  Water Chemistry to Atmospheric  Deposition
       in the Northeast Region	   293
8-6.    Results  of Regressions Relating Surface  Water Chemistry to Atmospheric  Deposition
       in the Southern Blue Ridge Province  	   294
8-7.    Estimated Population-Weighted Summary Statistics on the Darcy's Law Estimates
       of Flow  Rate and the Index of Flow Relative  to Runoff 	   296
8-8.    Estimated Population-Weighted Summary Statistics for Northeastern Geomorphic/
       Hydrologic Parameters   	   303
8-9.    Estimated Population-Weighted Summary Statistics for Southern Blue Ridge
       Province Hydrologic/Geomorphic Parameters	 .   304
8-10.   Mapped  and Calculated Geomorphic Parameters Collected for
       the Northeastern Study Sites  	   305
8-11.   Mapped  and Calculated Geomorphic Parameters Collected for the SBRP Study Sites .   308
8-12.   Stratification Based on Sulfur Deposition   	   311
                                            XIII

-------
                                    TABLES (Continued)
                                                                                       Page
8-13.    Results of Stepwise Regression Relating Surface Water Chemistry
        versus Geomorphic/Hydrologic Parameters for the Entire NE	  312
8-14.    Stepwise Regression Equations for Surface Water Chemistry and Hydrologic/
        Geomorphic Parameters Based on Sulfur Deposition Stratification	  313
8-15.    Results  of  Stepwise   Regression  Relating  Surface   Water  Chemistry   and
        Geomorphic/Hydrologic Parameters for the SBRP	  314
8-16.    Population-Weighted Summary Statistics for ln(a/KbTanB) for the Northeast	  318
8-17.    Population-Weighted Summary Statistics for ln(a/TanB) for the Southern Blue
        Ridge Province  	  319
8-18.    Spearman's  Correlation  Coefficients  Between  ln(a/Kt>TanB)   and Surface  Water
        Chemistry	  320
8-19.    Pearson's Correlation Coefficients Between ln(a/TanB) and NSS Pilot Chemistry	  325
8-20.    Tabulation of the Generic Bedrock Types Used to Classify the Mapped Units
        Identified on State Map Legends	  329
8-21.    Tabulation of the Generic Bedrock Types Used to Classify the Mapped Units
        Identified on State Map Legends	  330
8-22.    Regional and Subregional Statistics for the Bedrock Sensitivity Code Variables  	  331
8-23.    Results of Regressions of Surface Water Chemistry on Bedrock Sensitivity
        Code Statistics and Deposition Estimates for Northeast	  333
8-24.    Results for SBRP of Regressions of Surface Water Chemistry on Bedrock
        Sensitivity Code Statistics and  Deposition Estimates	  334
8-25.    Land Use and Other Environmental Variables Related to Surface Water
        Chemistry of  Northeastern Lakes	  339
8-26.    Factor Loadings for First 13 Principal Components after Varimax Rotation of
        the Correlation  Matrix of Land Use and other Environmental Variables for
        Northeastern  Lakes  	  340
8-27.    Interpretation of the First 13 Principal Components After Varimax Rotation of the
        Correlation Matrix of  Land Use and Other Environmental Variables for Northeastern
        Lakes	  342
8-28.    Land Use and Other Environmental Variables Related to Surface Water Chemistry of
        Southern Blue Ridge Province Streams	  343
8-29.    Composition of First  11 Principal Component Analysis (PCA) Factors Land
        Use and Other  Environmental Variables Related to Surface Water Chemistry
        of Southern Blue Ridge Province Streams	  344
8-30.    Interpretation of  the First  11  Principal  Components of  Land  Use  and  Other
        Environmental Variables for Southern Blue Ridge Province Streams  	  345
8-31.    Results of Regressions Relating Surface Water Chemistry of Northeastern Lakes
        to Land Use and Other Environmental Data	  346
8-32.    Results of Regressions Relating Sulfate and Percent Sulfur Retention of
        Southern Blue Ridge Province Streams to Land Use Data   	  348
8-33.    Results of Regressions Relating ANC, Ca plus Mg, and pH of Southern Blue
        Ridge Province Streams to Land Use Data  	  350
8-34.    Summary Statistics for Percent Area Distribution of the 38 Soil Sampling
        Classes and the 4  Miscellaneous Land Areas on the DDRP Sample of 145
        NE Lake Watersheds	  355
8-35.    Summary Statistics for the Percent Area Distribution  of the 38 Soil Sampling
        Classes and the 4  Miscellaneous Land Areas in the GIS 40-ft Contour on the
        DDRP Sample of 145 NE  Lake Watersheds	  356
8-36.    Summary Statistics for the Percent Area Distribution  of the 38 Soil Sampling
        Classes and the 4  Miscellaneous Land Areas in the Combined  GIS  Buffers  on the
        DDRP Sample of 145 NE  Lake Watersheds	  357
                                            XIV

-------
                                   TABLES (Continued)
                                                                                    Page

8-37.   Summary Statistics for the Percent Area Distribution of the 12 Soil Sampling
       Classes and the 2 Miscellaneous Land Areas on the DDRP Sample  of 35 SBRP
       Stream Watersheds	358
8-38.   Summary Statistics for the Percent Area Distribution of the 12 Soil Sampling
       Classes and the 2 Miscellaneous Land Areas in the 100-Meter Linear GIS Buffer
       on the DDRP Sample of 35 SBRP Stream Watersheds  	  359
8-39.   Lake Sulfate and Percent S Retention Regression Models Developed for NE Lakes
       Using Deposition, Mapped Soils and Miscellaneous Land Areas as Candidate
       Independent Variables	361
8-40.   Regression Models of Sulfate in SBRP Streams, Developed Using Deposition,
       Mapped Soils and Miscellaneous Land Areas as Candidate
       Independent Variables	  363
8-41.   Regression Models of Percent Sulfur Retention In SBRP Stream Watersheds
       Developed Using Deposition,  Mapped Soils, and Miscellaneous Land Areas as
       Candidate Independent Variables	  366
8-42.   Lake ANC and the Sum of Lake Calcium and Magnesium Regression Models
       Developed for NE Lakes Using Deposition, Mapped Soils, and Miscellaneous Land
       Areas as  Candidate Independent Variables	  368
8-43.   Lake pH Regression Models Developed for NE Lakes Using Deposition,
       Mapped Soils, and Miscellaneous  Land Areas as Candidate
       Independent Variables	  370
8-44.   Regression Models of ANC in SBRP Stream Watersheds, Developed Using
       Deposition, Mapped Soils, and Miscellaneous Land Areas as Candidate
       Independent Variables	  372
8-45.   Regression Models of Calcium Plus Magnesium in SBRP Streams, Developed
       Using Deposition, Mapped Soils, and Miscellaneous Land Areas as  a Candidate
       Independent Variables	  373
8-46.   Regression Models of SOBC  in SBRP Streams,  Developed Using  Deposition,
       Mapped Soils, and Miscellaneous  Land Areas as Candidate
       Independent Variables	  375
8-47.   Regression Models of Stream pH  in SBRP Streams, Developed Using Deposition,
       Mapped Soils, and Miscellaneous  Land Areas as Candidate
       Independent Variables	  376
8-48.   Depth-to-Bedrock Classes for the  Northeast and the Southern Blue  Ridge Province  . .  380
8-49.   Regional  and Subregional Statistics for Percentage of Watershed  Coverage of the
       Depth-to-Bedrock Classes  	  382
8-50.   Results for NE of Regressions of Surface Water Chemistry on Depth-to-Bedrock
       Classes and Deposition Estimates	  384
8-51.   Results for SBRP of Regressions of Surface Water Chemistry on Depth-to-Bedrock
       Classes and Deposition Estimates	  387
8-52.   Regression Models of Surface Water Sulfate and Sulfur Retention in the
       NE Lake  Watersheds  	  389
8-53.   Regression Models of Surface Water Sulfate and Sulfur Retention in the SBRP
       Stream Watersheds   	  392
8-54.   Regression Models of Surface Water ANC, Ca plus  Mg, and pH  in the  NE  Lake
       Watersheds	  394
8-55.   Regression Models of Surface Water ANC, Ca plus  Mg, and pH  in the  SBRP
       Stream Watersheds   	  397
8-56.   Standard Deviations Within and Among Northeast Sampling Classes Estimated
       from B Master Horizon Data	  405
8-57.   Means and Standard Deviations of Soil Characteristics by Aggregation
       Method and Region   	  408
                                            xv

-------
                                    TABLES (Continued)
8-58.   Population Means and Standard Errors for Selected Variables, by Subregion/Region
       and Aggregation (Watershed Adjusted Data)	  411
8-59.   Non-parametric Correlations Between Lake Chemistry Variables and Selected Soil
       Properties for the NE DDRP Watersheds	  414
8-60.   Non-parametric Correlations Between Stream Chemistry Variables and Selected
       Soil Properties for the SBRP  DDRP Watersheds	  416
8-61.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream Sulfate
       Concentrations Versus Soil Physical and Chemical Properties  	  419
8-62.   Results of Stepwise Multiple Regressions for DDRP Watershed Sulfur Retention
       Versus Soil Physical and Chemical Properties  	  420
8-63.   Results of Stepwise Multiple Regressions  for DDRP Lake Calcium plus Magnesium
       Concentrations and Stream Sum of Base Cation Concentrations Versus Soil Physical
       and Chemical  Properties	  422
8-64.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream ANC
       Versus Soil Physical and Chemical Properties	 .  423
8-65.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream pH
       Versus Soil Physical and Chemical Properties  	  424
8-66.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream ANC
       Versus Unadjusted and Watershed Adjusted Soil Physical and Chemical Properties   . .  427
8-67.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream Sulfate
       Versus Unadjusted and Watershed Adjusted Soil Physical and Chemical Properties   . .  428
8-68.   Population Means and Standard Errors for Selected Variables, by Subregion/
       Region and Aggregation	  432
8-69.   Non-parametric Correlations Between Lake Chemistry Variables and Selected
       Watershed Attributes for the  NE DDRP Watersheds  	  438
8-70.   Non-parametric Correlations Between Stream Chemistry Variables and Selected
       Watershed Attributes for the SBRP DDRP Watersheds   	  442
8-71.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream Sulfate
       Concentration  Versus Watershed Attributes	  445
8-72.   Results of Stepwise Multiple Regressions for DDRP Watershed Sulfur Retention
       Versus Watershed Attributes	  446
8-73.   Results of Stepwise Multiple Regressions for DDRP Lake 'Calcium Plus Magnesium
       Concentrations and Stream Sum of Base Cations Versus Watershed Attributes  	  447
8-74.   Results of Stepwise Multiple Regressions for  DDRP Lake and Stream ANC  Versus
       Watershed Attributes 	  448
8-75.   Results of Stepwise Multiple Regressions for DDRP Lake and Stream Air Equilibrated
       pH Versus Watershed Attributes  	  449

9-1.    Comparison of Summary  Data for Sulfate Adsorption Isotherm Data for Soils in the
       Northeastern United States and Southern Blue Ridge Province	  465
9-2.    Summary Statistics for Modelled Changes in  Sulfate  Concentration,  Percent Sulfur
       Retention, and Delta Sulfate for Northeast Watersheds Using  Long-Term Average
       Deposition Data.  .	  470
9-3.    Summary Statistics for Modelled Changes  in Sulfate  Concentration, Percent Sulfur
       Retention, and Delta Sulfate for Northeast Watersheds  Using Typical Year
       Deposition Data	  471
9-4.    Comparison of Measured  and Modelled Base Year (1985) Sulfate  Data for SBRP
       Watersheds, Using Long-Term Average Deposition Data	  482
9-5.    Comparison of Modelled Rates of Increase for  [SO42~] in DDRP Watersheds in the
       SBRP with Measured Rates of Increase in Watersheds in the Blue Ridge and
       Adjoining Appalachians	  484
                                            XVI

-------
                                     TABLES (Continued)
 9-6.    Summary Statistics for Modelled Changes in Sulfate Concentration, Percent Sulfur
        Retention, and Delta Sulfate for Watersheds in the Southern Blue Ridge Province, Using
        Long-Term Average Deposition Data	  488
 9-7.    Summary Statistics for Modelled Changes in Sulfate Concentration, Percent Sulfur
        Retention, and Delta Sulfate for Watersheds in the Southern Blue Ridge Province,
        Using Typical Year Deposition Data	•   439
 9-8.    Summary Comparison of Watershed Sulfur Status and Model Forecasts in the
        Northeastern United States and Southern Blue Ridge Province	  503
 9-9.    List of the Chemical Species and Reactions Considered Within the Reuss
        Model Framework	  515
 9-10.   Effect of pCO2 on  Changes Projected to Occur in Surface Water ANC Values at
        50 and 100 Years  Using the Reuss Model.   Deposition  Used in the Model is LTA  . . .  524
 9-11.   List of Input Data for the  Bloom-Grigal Soil  Acidification Model	  534
 9-12.   Summary Statistics for the Population Estimates of Current ANC Conditions for
        Lakes in the NE Region for Five Different Deposition or Soils Aggregation
        Schemes   	  541
 9-13.   Descriptive Statistics of the Population Estimates for Changes
        in Lake Water ANC for Systems in the NE   	  546
 9-14.   Summary Statistics Comparing the Projections Regarding Changes in Surface
        Water ANC Values Obtained Using Different Soils Aggregation Schemes   	  549
 9-15.   Summary Statistics of the Differences Between the Population Estimates for
        Future ANC Projections Made Using the Constant Level  and Ramped
        Deposition Scenarios	  550
 9-16.   Summary Statistics for the Population Estimates of Current ANC Conditions for Stream
        Reaches in the SBRP for Four Different Deposition  Scenarios   	  552
 9-17.   Descriptive Statistics  of the Population Estimates for Changes in Stream Reach
        ANC Values for Systems in the SBRP	   555
 9-18.   Summary Statistics of the Differences Between the  Population Estimates for Future
        ANC Projections Made Using the Constant  Level and Ramped Deposition Scenarios for
        Stream  Reaches in the SBRP 	   559
 9-19.   Summary Statistics of the Projected Changes in Soil Base Saturations in the NE
        Region, Obtained Using the Different Deposition Scenarios or Soil
        Aggregation Schemes	   562
 9-20.   Summary Statistics of the Projected Changes in Soil pH  in the NE Region, Obtained
        Using the Different  Deposition Scenarios or Soil Aggregation Schemes	   563
 9-21.   Summary Statistics of the Projected  Changes in Soil Base Saturations in the SBRP,
        Obtained Using the Different Deposition Scenarios	   571
 9-22.   Summary Statistics of the Projected Changes in Soil pH  in the  SBRP, Obtained
        Using the Different  Deposition Scenarios	   572
 9-23.   Comparison of the  Changes in Soil Base Saturation and Soil pH that Are Projected to
        Occur in the NE and SBRP	   576
9-24.   Regionally Weighted Median Values of Initial Annual Deposition Inputs to the Bloom-
        Grigal Model for the Northeastern Region and the Southern Blue Ridge Province ....   579
9-25.   Regionally Weighted Median Values of Annual Initial Soil Chemical Values Input
        Into the Bloom-Grigal Model for the Northeastern Region and the Southern  Blue
        Ridge Province	   58-)
9-26.   Bloom-Grigal Model Regional Projections for the Change in Soil pH in the Northeastern
        United States. Organic Soil Horizons  Included	   585
                                            XVII

-------
                                    TABLES (Continued)
9-27.   Bloom-Grigal Model Regional Projections of the Change in Percent Base Saturation in
       the Northeastern United States.  Organic Soil Horizons Included	   587
9-28.   Bloom-Grigal Model Regional Projections of the Change in Soil pH in the Northeastern
       United States.  Organic Soil Horizons Included	   592
9-29.   Bloom-Grigal Model Regional Projections for the Change in Percent Base Saturation in
       the Northeastern United States.  Organic Soil Horizons Included  	   594
9-30.   Bloom-Grigal Model Regional Projections for the Change in Soil pH in the Southern
       Blue Ridge Province.  Organic Soil Horizons Included	   598
9-31.   Bloom-Grigal Model Regional Projections for the Change in Percent Base Saturation
       in the Southern Blue Ridge Province.  Organic Soil Horizons Included  	   600
9-32.   Summary  of the  Bloom-Grigal Projected  Changes  in Soil  pH  and  Percent Base
       Saturation in the NE and SBRP Under Constant LTA Deposition  	   603
9-33.   Comparison of the Results from the Reuss and Bloom-Grigal
       Models with Regard to the Magnitude of Changes in Soil pH and Base Saturation
       Projected in Soils of the NE   	   607
9-34.   Comparison of the Results from the Reuss and Bloom-Grigal  Models with Regard to
       the Magnitude of Changes in Soil  pH and Base Saturation Projected in Soils of the
       SBRP	   613

10-1.   Major Processes Incorporated in the Dynamic Model  Codes  	   621
10-2.   Meteorological Data Required by the Dynamics Model Codes   	   623
10-3.   Chemical Constituents in Wet and Dry Deposition Considered by the MAGIC, ETD, and
       ILWAS Codes   	   624
10-4.   Chemical Constituents  Included in Soil Solutions
       and Surface Water for the MAGIC,  ETD,  and ILWAS Codes   	   625
10-5.   Definitions of Acid Neutralizing Capacity  (ANC) Used  by the MAGIC,  ETD,
       and ILWAS Codes (Brackets indicate concentration in molar or molal units, and R',
       R", and  R'" represent mono-, di-, and triprotic organic acids, respectively.)  ANC
       Simulated by All Three Models is Equivalent to the Modified Gran ANC  	   626
10-6.   Level III  Operational Assumptions	   630
10-7.   Comparison of Calibration/Confirmation RMSE for Woods Lake Among ETD, ILWAS, and
       MAGIC Models, with the Standard Error  of the Observations	   653
10-8.   Comparison of Calibration/Confirmation RMSE for Panther Lake Among ETD,
       ILWAS, and MAGIC Models,  with the Standard Error of the Observations  	   654
10-9.   Comparison of  Calibration RMSE for Clear Pond Among ETD, ILWAS, and  MAGIC
       Models,  with the Standard Error of the Observations	   655
10-10.  Percent  Change in RMSE for MAGIC and ETD for a Ten Percent Change in Parameter
       Values   	   658
10-11.  Watersheds, by Priority Class, for Which Calibration Criteria Were Not Achieved	   671
10-12.  Deposition Variations Used in Input Uncertainly Analyses  	   675
10-13.  Target Populations for Modelling Comparisons and Population Attributes   	   679
10-14.  Descriptive Statistics of Projected ANC,  Sulfate,  pH,  Calcium Plus Magnesium, and
       Percent  Sulfur Retention for NE Lakes in Priority Classes A -1 Using  MAGIC for Both
       Current  and Decreased Deposition	   682
10-15.  Change  in Median ANC and Sulfate Concentrations Over a 40-Year Period as a
       Function of the  Initial ELS-Phase I or NSS Pilot Survey ANC Groups  . . . .	   690
10-16.  Descriptive Statistics of Projected ANC, Sulfate, and Percent Sulfur Retention for NE
       Lakes in Priority Classes A - E Using MAGIC and ETD for Both Current and Decreased
       Deposition  	   797
                                           XVIII

-------
                                    TABLES (Continued)
10-17.  Descriptive Statistics for Projected ANC, Sulfate, Percent Sulfur Retention,
       and Calcium Plus Magnesium for NE Lakes in Priority Classes A and B Using
       ETD, ILWAS, and MAGIC for Both Current and Decreased Deposition	
10-18.  Descriptive Statistics of Projected ANC, Sulfate, and Percent Sulfur Retention, and
       Calcium and Magnesium for SBRP Streams in Priority Classes A -E Using MAGIC for
       Both Current and Increased Deposition	
10-19.  Descriptive Statistics of Projected ANC, Sulfate, Percent Sulfur Retention,
       and Calcium Plus Magnesium for SBRP Streams in Priority Classes A and B Using
       ILWAS and MAGIC for Both Current and Increased Deposition	
10-20.  Effects of Critical Assumptions on Projected Rates of Change	
11-1.   Weighted Median Projected Change  in ANC at 50  Years  for Northeastern DDRP
       Lakes	
11-2.   Lakes in the NE Projected to Have ANC Values <0 and <50 /L/eq  L1 for
       Constant and Decreased Sulfur Deposition	
11-3.   Weighted Median Projected Change in ANC at 50 Years for DDRP SBRP
       Stream Reaches  	
11-4.   SBRP Stream Reaches Projected to Have ANC Values <0 and <50 jueq L'1 for
       Constant and Increased Sulfur Deposition	
 Page



  716


  744


,  756
,  896

.  809

.  812

.  816

.  819
                                            XIX

-------
                                          FIGURES
1-1.    Steps of the Direct/Delayed Response Project (DDRP) approach	   6

2-1.    Activities of the Aquatic Effects Research Program within the National Acid
       Precipitation Assessment Program	  25

3-1.    Diagram of sulfur cycle in forest ecosystems	  33
3-2.    Diagram of terrestrial base cation cycle	  41

4-1.    Steps of the Direct/Delayed Response Project (DDRP) approach	  50

5-1.    Representation of the point frame sampling procedure for selecting NSS
       Stage I  reaches	  59
5-2.    DDRP site locations for Subregion 1A	  72
5-3.    DDRP site locations for Subregion 1B	  73
5-4.    DDRP site locations for Subregion 1C	  74
5-5.    DDRP site locations for Subregion 1D	  75
5-6.    DDRP site locations for Subregion 1E	  76
5-7.    The pH-ANC relationship for (A) lakes of the ELS Phase I sampling in the Northeast
       and (B) DDRP study lakes in the Northeast	  90
5-8.    The pH-ANC relationship for samples with ANC <400 jueq L taken at the downstream
       nodes of stream reaches sampled in the NSS	  94
5-9.    Location of Northeast field check sites and other DDRP watersheds	  108
5-10.   Example of digitization log sheet	  125
5-11.   Example of attribute entry log  sheet	  126
5-12.   Definition of soil sampling classes for the DDRP Soil Survey in the Northeast	  149
5-13.   Definition of soil sampling classes for the DDRP Soil Survey in the Southern
       Blue Ridge Province	  151
5-14.   Selection of watersheds for sampling	  152
5-15.   Selection of starting points for sampling	  153
5-16.   Field selection of a sampling point for sampling class on a watershed	  154
5-17.   Major steps and datasets from the DDRP database	  173
5-18.   Calculation percentage of regional or subregional area in each soil sampling	  179
5-19.   Relative areas of sampling classes in the Northeast subregions	  180
5-20.   Relative areas of sampling classes in the entire Northeast and Southern
       Blue Region Province	  181
5-21.   Aggregated soil variables for individual pedons in the Northeast	  182
5-22.   Aggregated soil variables for individual pedons in the Southern Blue Ridge Province.  .  184
5-23.   Calculation of cumulative distribution function for a soil variable in  a region
       or subregion	  186
5-24.   Cumulative distribution functions for pedon aggregated soil  variables for the
       Northeast and the Southern Blue Ridge Province	  187
5-25.   Sulfur deposition scenarios for the NE and SBRP for Level II and III Analyses  	  191
5-26.   Example of average annual runoff map for 1951 -80  	  226
5-27.   Flow chart of Darcy's Law soil contact calculation as applied to the DDRP
       study sites	<,	  235

7-1.    Estimated  percent sulfur retention by in-lake processes  in drainage lakes
       in ELS Region  1 (northeastern United States)	  252
7-2.    Percent sulfur retention for intensively studied sites in the United States and
       Canada  relative to the southern extent of the Wisconsinan glaciation	  254
                                             xx

-------
                                     FIGURES (Continued)
7-3.     Model of flow-weighted average concentration calculations for Biscuit Brook	  259
7-4.     Flow chart for the determination of internal sources of sulfur using the
        steady-state sulfate concentration	  267
7-5.     Scatter plot of the Monte Carlo calculated standard deviation versus the
        calculated  mean  [SO42"]SS 	  269
7-6.     Comparison of percent sulfur retention calculated using (A) modified-LTA
        deposition  and (B) modified-LTA deposition adjusted with a 20 percent increase
        in dry deposition	  272
7-7.     Population-weighted distribution of projected percent sulfur retention
        (upper and lower bounds for 90 percent confidence interval):  (A) Northeast;
        (B) Mid-Appalachians,  and (C) Southern Blue Ridge Province	  274
7-8.     Supplemental watersheds mapped for special  evaluation of sulfur retention	  276
7-9.     Population-weighted distributions of projected percent sulfur retention, with
        upper and lower  bounds for 90 percent confidence intervals, for additional NSS
        subregions:  (A) Southern Appalachian Plateau, (B) Mid-Atlantic Coastal Plain,
        (C) Catskills/Poconos,  and (D) Piedmont	  281
7-10.    Combination regional population-weighted distributions of projected percent
        sulfur retention, with upper and lower bounds  for 90 percent confidence intervals,
        for the Northeast, Mid-Appalachians, and Southern Blue Ridge Province  	  282

8-1.     Distribution of estimated contact rate using Darcy's Law calculation	  297
8-2.     Distribution of index of contact using Darcy's Law calculation	  298
8-3.     Scatter plot of ANC versus contact rate calculated using Darcy's Law	  300
8-4.     Scatter plot of ANC versus index of soil contact calculated using Darcy's Law	  301
8-5.     Scatter plot of ANC versus ln(a/KbTanB)	  321
8-6.     Scatter plot of Ca plus Mg versus  ln(a/KbTanB)	  322
8-7.     Scatter plot of pH versus ln(a/KbTanB)	  323
8-8.     Data and regression model  development flow diagrams	  353
8-9.     Model development procedure	  401
8-10.    Histograms of unadjusted and adjusted watershed means for selected SBRP soils
        variables	  409
8-11.    The mean  pH ±  2 standard errors for the SBRP watersheds estimated using the
        common aggregation and the watershed effects adjusted aggregation the lack of
        variation among the common aggregation values	  410

9-1.     Schematic  diagram of  extended Langmuir isotherm fitted to data points from
        laboratory  soil analysis	  .  459
9-2.     Comparison of measured lake (NE) or stream  (SBRP) sulfate concentration with
        computed  soil solution concentration	  462
9-3.     Historic deposition inputs and modelled output for soils in a representative
        watershed  in the  northeastern United States	  466
9-4.     Schematic  of surface water  response to changes  in sulfur inputs	  467
9-5.     Comparison of measured, modelled and steady-state sulfate for Northeast lake
        systems in 1984	  472
9-6.     Projected changes in percent sulfur retention and sulfate concentration for
        soils in northeastern lake systems  at 10, 20, 50 and 100 years	  474
9-7.     Box-and-whisker plots  showing changes in sulfate concentration, percent sulfur
        retention, and change  in sulfate concentration for soils in northeastern lake
        watersheds, using long-term average deposition data	  475
                                             xxi

-------
                                     FIGURES (Continued)
9-8.    Box-and-whisker plots showing changes in sulfate concentration, percent sulfur
       retention, and change in sulfate concentration for soils in northeastern lake
       watersheds, using Typical Years deposition data	  476
9-9.    Projected time to steady-state concentration for sulfate in northeastern lakes (A)
       at current deposition and (B) after end of decreasing  input in ramp scenario	  478
9-10.   Historic deposition inputs and modelled output for soils in stream systems in the
       Southern Blue Ridge Province	  480
9-11.   Comparison of measured, modelled, and steady-state sulfate for stream  systems in
       the Southern Blue Ridge Province in 1985  	  483
9-12.   Comparison of forecasts based on two sulfur deposition datasets for soils in SBRP
       watersheds	  485
9-13.   Projected changes in percent sulfur retention and in sulfate concentration for stream
       systems in the Southern Blue Ridge Province at 0, 20, 50, 100 and 140  years	  487
9-14.   Box and whisker plots showing changes in sulfate concentration, percent sulfur
       retention, and change in sulfate concentration for soils in watersheds of the Southern
       Blue Ridge Province. Data are shown for Typical Year deposition data  	  490
9-15.   Box and whisker plots showing changes in sulfate concentration, percent sulfur
       retention, and change in sulfate concentration for soils in watersheds of the Southern
       Blue Ridge Province.  Data are shown for Typical Year deposition data	  491
9-16.   Projected time to 95 percent of steady-state sulfur concentration of Southern
       Blue Ridge Province stream systems	  492
9-17.   Comparison of model simulation results for DDRP Southern Blue Ridge
       watersheds	  495
9-18.   Projected base year sulfate concentration with upper and lower bounds  for 90
       percent confidence intervals for Southern Blue Ridge Province watersheds	  496
9-19.   Projected time to sulfur  steady state with upper and lower bounds for 90
       percent confidence intervals in Southern Blue Ridge Province watersheds	  497
9-20.   Effects of data aggregation on simulated watershed sulfur response for soils
       in DDRP watersheds of  the Southern Blue Ridge Province	  499
9-21.   Evaluation of alternate soil aggregation procedures for soils in SBRP watersheds. .   . .  500
9-22.   Schematic diagram of the principal process involved in the cycling of base
       cations in surficial environments	, .  513
9-23.   Plot of the log of the activity  of AI3+ vs.  soil solution pH for individual soil
       samples collected for DDRP	  518
9-24.   Plot of the log of the selectivity coefficient for the calcium-aluminum exchange
       reaction  vs. the measured base saturation in A/E horizons in the NE	  520
9-25.   Histograms of the (unweighted for the population estimates) projected
       present-day ANC values for lakes in the NE	  521
9-26.   Histograms of the (unweighted for the population estimates)  projected,  present-day
       ANC values for lakes in the NE	  523
9-27.   Flow diagram  for the one-box Bloom-Grigal soil simulation model	  529
9-28.   Cumulative distribution of projected, present-day ANC values for lakes  in the study
       population in the NE as projected using Reuss cation exchange model	538
9-29.   Scatter plot of the projected, present-day ANC values for lakes in the NE,  obtained
       using the Reuss model vs. observed (ELS) values	  539
9-30.   Scatter plot of the present-day lake ANC values projected using the Reuss model in
       conjunction with the Watershed-Based Aggregation (WBA) soils data vs. observed (ELS)
       ANC values	  542
9-31.   Cumulative distribution of the projected surface water ANC values projected for the
       study population of lakes in 50 years in the NE	  544
9-32.   Cumulative distribution of the projected surface water ANC values projected for the
       study population of lakes in 100 years in the NE	  545

                                             xxii

-------
                                    FIGURES (Continued)
9-33.   Schematic illustration of the titration-like behavior Displayed by soils in response to
       constant loadings of acidic deposition	  547
9-34.   Cumulative distribution of projected present-day ANC values for stream reaches
       in the study population in the SBRP,  as projections using Reuss's cation exchange
       model	  551
9-35.   Scatter plot of the projected present-day ANC values for stream reaches in the SBRP,
       obtained using the Reuss model, vs.  observed (NSS) values	  553
9-36.   Cumulative distribution of projected  changes (at  50 years)  in surface water ANC
       obtained using the Reuss model for stream reaches in the SBRP	  556
9-37.   Cumulative distribution of projected  changes (at 100  years) in surface water ANC
       obtained using the Reuss model for stream reaches In the SBRP	  557
9-38.   Comparison of measured vs. calculated soil pH values for the 580 aggregated master
       horizons in the NE	  561
9-39.   Cumulative distribution of projected (a) base saturations and (b) soil pH values for soils
       in NE. Projections made using the Reuss model	  564
9-40.   Cumulative distribution of projected (a) base saturations and (b) soil pH values for soils
       in the NE. Projections were made using the Reuss model	  565
9-41.   Plot  of  the  measured  (ELS) ANC values for  lakes  in the  NE vs. the estimated,
       watershed-level,base saturations for mineral horizons in those watersheds	  566
9-42.   Plot of the changes in surface water ANC values at (a) 20, (b) 50, and (c) 100 years
       as projected by the Reuss model vs. the estimated, present-day,  watershed-level base
       saturations for mineral horizons in those watersheds	  567
9-43.   Plot of the projected changes  in soil  base saturations vs. he observed, present-day,
       aggregated base saturations for mineral horizons in the NE. The projections were made
       with the Reuss model	  568
9-44.   Cumulative frequencies of changes in (a) soil base saturation and (b) soil pH for the
       population of soils in the SBRP	  573
9-45.   Cumulative frequencies of changes in (a) soil base saturation and (b) soil pH for the
       population of soils in the SBRP..   .	  574
9-46.   Cumulative distributions  of aggregate initial soil pH and percent base saturation in
       the NE and SBRP, with and without organic horizons	  582
9-47.   RegionafCDFs "of the projected change in the pH of soils on NE lake watersheds under
       constant and ramp  down (30 percent 4-) deposition scenarios after 20, 50,  and 100
       years of LTA, LTA-rbc, and LTA-zbc deposition.  Organic horizons included.    ......  583
9-48.   Regional CDFs of the projected change in  the percent base saturation of soils
       on NE lake watersheds under constant and ramp down (30 percent 4-) deposition
       scenarios after 20, 50, and  100 years of LTA, LTA-rbc,  and  LTA-zbc deposition.
       Organic horizons included	  584
9-49.   Regional CDFs of the projected change in the pH of soils on NE lake watersheds under
       constant and ramp down (30% 4.) deposition scenarios after 20,  50, and 100 years of
       LTA, LTA-rbc, and LTA-zbc deposition.  Organic horizons are excluded	  590
9-50.   Regional CDFs of the projected change in the percent base saturation of soils on NE
       lake watersheds under constant and ramp down (30% l) deposition scenarios after 20,
       50,  and  100 years of LTA,  LTA-rbc, and LTA-zbc deposition.  Organic  horizons
       excluded.  	  591
9-51.   Regional CDFs of the projected change in the pH of soils on SBRP stream watersheds
       under constant and ramp up (20% t)  deposition scenarios after 20, 50, 100,  and 200
       years of LTA, LTA-rbc, and LTA-zbc deposition.  Organic horizons included.   *	  597
9-52.   Regional CDFs of the projected change in the percent base saturation of soils on SBRP
       stream watersheds under constant and ramp up (20% t) deposition scenarios after 20,
       50,  100, and 200 years  of LTA, LTA-rbc, and LTA-zbc  deposition.  Organic  horizons
       included	  598

                                            xxiii

-------
                                    FIGURES (Continued)
9-53.   Cumulative distributions of changes in  soil base saturation for the population of
       watersheds in the NE	  608
9-54.   Cumulative distributions of changes in soil pH for the population of watersheds
       in the NE	  609
9-55.   Scatter diagrams of the projected changes in base saturation for individual
       systems (not population weighted) in the NE obtained from the  Reuss and
       Bloom-Grigal models	  610
9-56.   Scatter diagrams of the  projected changes in  soil  pH for individual systems  (not
       population weighted) in the NE obtained from the Reuss and Bloom-Grigal models.  .  .  611
9-57.   Cumulative distributions of changes in soil base  saturation for the population of
       watersheds in the SBRP	  .  614
9-58.   Cumulative distributions of changes in soil pH for the population of watersheds
       in the SBRP	  615

10-1.   Modelling priority decision tree:  Northeast	  631
10-2.   Modelling priority decision tree: Southern Blue Ridge Province	  633
10-3.   Decision tree used to identify watersheds with net chloride export and procedures for
       determining chloride imbalance	  638
10-4.   Approach used in performing long-term projections of future changes in surface water
       chemistry	  .  640
10-5.   Schematic of modelling approach for making long-term projections	  641
10-6.   Representation  of horizontal segmentation of Woods  Lake, NY, watershed for MAGIC
       and ETD	  645
10-7.   Representation  of vertical layers of Woods Lake  Basin for ETD	  646
10-8.   Representation  of horizontal segmentation of Woods  Lake  Basin for ILWAS	648
10-9.   Representation  of vertical layers of Woods Lake  Basin for ILWAS	  649
10-10.  Representation  of vertical layers of Woods Lake,  NY,  watershed  for MAGIC	  651
10-11.  Comparison of  population histograms for simulated  versus observed  (Eastern Lake
       Survey Phase I  1984 values) ANC for ILWAS and MAGIC	  662
10-12.  Comparison of  population histograms for simulated  versus observed  (Eastern Lake
       Survey -  Phase I 1984 values) sulfate concentrations for ILWAS and MAGIC, Priority
       Classes A and  6	  663
10-13.  Comparison of  population histograms for simulated  versus observed  (Eastern Lake
       Survey Phase  I 1984  values) ANC  and sulfate concentrations for MAGIC, Priority
       Classes A - E	  665
10-14.  Comparison of  population histograms for simulated  versus observed  (Eastern Lake
       Survey Phase I 1984 values ) ANC  and sulfate concentrations for MAGIC, Priority
       Classes A - 1	  666
10-15.  Comparison of population histograms for simulated versus observed (NSS Pilot Survey
       values) ANC, Priority Classes A and B using ILWAS and MAGIC	  667
10-16.  Comparison of population histograms for simulated versus observed (NSS Pilot Survey
       values) sulfate concentrations, Priority Classes A and B using ILWAS and MAGIC.  . .  .  677
10-17.  Comparison of population histograms for simulated versus observed (NSS Pilot Survey
       values) ANC and sulfate concentrations, Priority Classes A - E using MAGIC	  678
10-18.  Comparison of projection standard errors as  a function of ANC (top figure) and sulfate
       (bottom figure)  concentrations for the NE uncertainty  analysis watersheds using ETD
       and MAGIC	  685
10-19.  Projections  of ANC and sulfate concentrations for NE lakes, Priority Classes ~~—~~
       A - I, using MAGIC for  20, 50, and 100 years, under current deposition and a
       30 percent  decrease in deposition	  689
10-20.  pH projections for NE lakes, Priority Classes A -1, using MAGIC for 20, 50,
       and 100 years,  under current deposition and a 30 percent decrease in deposition.  .  .  692

                                            xxiv

-------
                                    FIGURES (Continued)
10-21.  Box and whisker plots of ANC distributions at 10-year intervals for NE
       Priority Classes A - I using MAGIC	  686
10-22.  Box and whisker plots of sulfate distributions at 10-year intervals for NE
       Priority Classes A - I using MAGIC	  687
10-23.  Box and whisker plots of pH distributions at 10-year intervals for NE
       Priority Classes A - I using MAGIC	  688
10-24.  Comparison of population histograms for ANC  under current levels of deposition
       and a 30 percent decrease in  deposition for NE lakes, Priority Classes A - I,
       using MAGIC	  691
10-25.  Comparison of population histograms for sulfate concentrations at current
       levels of deposition and a 30 percent decrease for NE lakes, Priority Classes
       A - I, using MAGIC	  692
10-26.  Comparison of MAGIC and ETD projections of  ANC for NE lakes, Priority
       Classes A - E, under current and decreased deposition	  693
10-27.  Comparison of MAGIC and ETD  projections of sulfate concentrations for NE lakes,
       Priority Classes A - E, under current and decreased deposition	  694
10-28.  Comparison of MAGIC and ETD projections of  pH for NE lakes, Priority
       Classes A -E,  under current and decreased deposition	  695
10-29.  Comparisons of projected change in ANC  under current and decreased
       deposition for NE Priority Classes A - E, using  ETD and MAGIC	  699
10-30.  Comparisons  of  projected change  in  sulfate concentrations  under  current and
       decreased deposition for NE Priority Classes A - E, using ETD and MAGIC	  700
10-31.  Comparisons of projected change in pH under current and decreased
       deposition for NE Priority Classes A - E, using  ETD and MAGIC	  701
10-32.  Box and whisker plots of ANC distributions projected using ETD  in 10-year
       intervals  for NE lakes, Priority Classes A -  E	  702
10-33.  Box and whisker plots of sulfate distributions projected using ETD in
       10-year intervals for NE lakes,  Priority Classes A - E	  703
10-34.  Box and whisker plots of pH projected using ETD in 10-year intervals for
       NE lakes, Priority Classes A - E	  704
10-35.  Box and whisker plots of ANC distributions in 10-year intervals using MAGIC
       for  NE lakes, Priority Classes A - E	  705
10-36.  Box and whisker plots of sulfate distributions in 10-year intervals using
       MAGIC for NE lakes, Priority Classes A - E	  706
10-37.  Box and whisker plots of pH in 10-year intervals using MAGIC for NE lakes,
       Priority Classes A - E	  707
10-38.  ETD ANC distributions at year 10 and year 50 for NE lakes, Priority
       Classes A -  E, under current and decreased deposition	  708
10-39.  MAGIC ANC distribution at year 10 and year 50 for NE lakes, Priority
       Classes A -  E, under current and decreased deposition	  709
10-40.  ETD sulfate distributions at year 10 and year 50 for NE lakes, Priority
       Classes A -  E, under current and decreased deposition	  710
10-41.  MAGIC sulfate distributions at year 10 and year 50 for NE lakes, Priority
       Classes A -  E, under current and decreased deposition	  711
10-42.  Comparison of ANC  projections using ETD, ILWAS, and MAGIC for NE lakes,
       Priority Classes A and B, under current and decreased deposition	  713
10-43.  Comparison of sulfate projections using ETD, ILWAS, and MAGIC for NE lakes,
       Priority Classes A and B, under current and decreased deposition	  714
10-44.  Comparison of pH projections  using ETD,  ILWAS,  and MAGIC for NE lakes,
       Priority Classes A and B, under current and decreased deposition	  715
                                            xxv

-------
                                    FIGURES (Continued)
10-45.  Comparison of ANC projections under current and decreased deposition for
       NE lakes, Priority Classes A and B,  at year 20 and year 50 using ETD, ILWAS,
       and MAGIC	  720
10-46.  Comparison of sulfate projections under current and decreased deposition for
       NE lakes, Priority Classes A and B,  at year 20 and year 50 using ETD, ILWAS,
       and MAGIC	  721
10-47.  Comparison of pH projections under current and decreased deposition for NE
       lakes, Priority Classes A and B, at year 20 and year 50 using ETD, ILWAS,
       and MAGIC	  722
10-48.  Box and whisker plots of ANC distributions in 10-year intervals projected
       using ETD for NE lakes, Priority Classes A and B	  724
10-49.  Box and whisker plots of ANC distributions in 10-year intervals projected
       using ILWAS for NE lakes, Priority Classes A and  B	  725
10-50.  Box and whisker plots of ANC distributions in 10-year intervals projected
       using MAGIC for NE lakes, Priority Classes A and B	  726
10-51.  Box and whisker plots of sulfate distributions in 10-year intervals projected
       using ETD for NE lakes, Priority Classes A and B	  727
10-52.  Box and whisker plots of sulfate distributions in 10-year intervals projected
       using ILWAS for NE lakes, Priority Classes A and  B.  	  728
10-53.  Box and whisker plots of sulfate distributions in 10-year intervals projected
       using MAGIC for NE lakes, Priority Classes A and B	  729
10-54.  Box and whisker plots of pH distributions in 10-year intervals projected
       using ETD for NE lakes, Priority Classes A and B.	  730
10-55.  Box and whisker plots of pH distributions in 10-year intervals projected
       using ILWAS for NE lakes, Priority Classes A and  B.  .	 .  731
10-56.  Box and whisker plots of pH distributions in 10-year intervals projected
       using MAGIC for NE lakes, Priority Classes A and B	  732
10-57.  ETD ANC population distributions at year 10 and year 50 for current and
       decreased deposition	733
10-58.  ILWAS ANC population distributions at year 10 and year 50 for current and
       decreased deposition	  734
10-59.  MAGIC  ANC population distributions at year 10 and year 50 for current and
       decreased deposition	  735
10-60.  ETD sulfate population distributions  at year 10 and year 50 for current and
       decreased deposition	  736
10-61.  ILWAS sulfate population distributions at year 10 and year 50 for current and
       decreased deposition	  737
10-62.  MAGIC  sulfate  population distributions at year  10 and year 50 for current and
       decreased deposition	  738
10-63.  MAGIC  ANC and sulfate projections for SBRP streams,  Priority Classes A - E,
       at year  20, year 50, year 100, and year 200 under current and increased
       deposition	  740
10-64.  MAGIC  pH projections for SBRP streams, Priority Classes A - E, at year 20,
       year 50, year 100, and year 200 under current and increased deposition	  742
10-65.  Box and whisker plots of ANC distributions in 10-year intervals projected
       using MAGIC for SBRP streams, Priority Classes A - E,  for current and
       increased deposition	  746
10-66.  Box and whisker plots of sulfate distributions in 10-year intervals projected
       using MAGIC for SBRP streams, Priority Classes A - E,  for current and
       increased deposition	  747
                                             xxvi

-------
                                    FIGURES (Continued)
10-67.  Box and whisker plots of pH distributions in 10-year intervals projected
        using MAGIC for SBRP streams, Priority Classes A - E, for current and
        increased deposition	   748
10-68.  MAGIC ANC population distributions at year 10 and  year 50 for current and
        increased deposition,  SBRP streams, Priority Classes A - E	   750
10-69.  MAGIC sulfate population distributions at year 10 and year 50 for current
        and increased deposition, SBRP streams, Priority Classes A - E	   751
10-70.  Comparison of ILWAS and  MAGIC projections for ANC at years 0, 20, and 50
        for SBRP streams, Priority Classes A and B, under current and increased deposition. .   753
10-71.  Comparison of ILWAS and  MAGIC projections for sulfate concentration at years
        0, 20, and  50 for SBRP streams, Priority Classes A and B, under  current
        and increased deposition	   754
10-72.  Comparison of ILWAS and  MAGIC projections for pH at years 0, 20, and  50 for
        SBRP streams, Priority Classes A and B, under current and  increased deposition.  .  . .   755
10-73.  Box and whisker plots for ANC distributions in 10-year intervals projected
        using ILWAS for SBRP streams, Priority Classes A and B, for current and
        increased deposition	   759
10-74.  Box and whisker plots for ANC distributions in 10-year intervals projected
        using MAGIC for SBRP streams,  Priority Classes A and B, for current and
        increased deposition	   760
10-75.  Box and whisker plots for sulfate distributions  in 10-year intervals
        projected using  ILWAS for SBRP streams, Priority Classes A and B, for current
        and increased deposition	   761
10-76.  Box and whisker plots for sulfate distributions  in 10-year intervals  projected
        using MAGIC for SBRP streams,  Priority Classes A and B, for current and
        increased deposition	   762
10-77.  Box and whisker plots for pH distributions in 10-year intervals projected
        using ILWAS for SBRP streams, Priority Classes A and B,  for current and
        increased deposition	   763
10-78.  Box and whisker plots for pH distributions in 10-year intervals projected
        using MAGIC for SBRP streams,  Priority Classes A and B, for current and
        increased deposition	   764
10-79.   ILWAS ANC population distributions at year 10 and year 50 for current and
        increased deposition,  SBRP Priority Class A and B streams	   766
10-80.   MAGIC ANC population distributions at year 10 and  year 50 for current and
        increased deposition,  SBRP Priority Class A and B streams	   767
10-81.   ILWAS sulfate population distributions at year 10 and year 50 for current  and
        increased deposition,  SBRP Priority Class A and B streams	   768
10-82.   MAGIC sulfate population distributions at year  10  and year 50 for current  and
        increased deposition,  SBRP Priority Class A and B streams	   769
10-83.   Comparison of projected sulfate versus sulfate steady-state concentrations
        using ETD, ILWAS, and MAGIC for NE lakes	   770
10-84.   Comparison of projected sulfate concentrations under decreased deposition
        with the current  sulfate steady-state concentrations using ETD; ILWAS, and
        MAGIC for  NE lakes	   772
10-85.   Comparison of projected sulfate concentrations between models for NE lakes
        after 50 years under current and decreased deposition	   773
10-86.   Comparison of projected sulfate versus sulfate steady-state concentrations
       for SBRP streams using ILWAS and MAGIC under both current and increased
        deposition	   774
10-87.   Comparison of projected ANC  between models in NE lakes after 50 years
        under current and decreased deposition	   775

                                            xxvii

-------
                                   FIGURES (Continued)
10-88.  Projected changes in ANC as a function of changes in sulfate for NE lakes
       using ETD, ILWAS, and MAGIC for current and decreased deposition	  776
10-89.  Comparison of pH - ANC relationship for each of the models	  777
10-90.  Comparison of projected pH values between models for NE lakes after 50 years
       under current and decreased deposition	  779
10-91.  Comparison  of projected changes in  calcium and  magnesium  versus changes in
       sulfate using ILWAS and MAGIC for NE lakes	  780
10-92.  Change  in median ANC, calcium and magnesium, and sulfate concentrations
       projected for NE lakes using MAGIC under current and decreased deposition. ......  781
10-93.  Comparison of the change in pH after 50 years as a function of the initial
       calibrated pH for MAGIC,  ETD and ILWAS  on northeastern lakes	  782
10-94.  Comparisons of projected ANC and sulfate concentrations and pH between
       ILWAS and MAGIC after 50 years for SBRP streams	  793
10-95.  Comparison of projected AANC and Asulfate relationships in SBRP Priority
       Class A  and  B streams using ILWAS and MAGIC	  785
10-96.  Comparison of projected AANC and Asulfate relationships and A(calcium and
       magnesium) and Asulfate relationships for  SBRP Priority Class A - E streams
       using MAGIC	  786
10-97.  Comparison of projected A(calcium and magnesium) and Asulfate relationships
       for SBRP Priority Class A and B streams using ILWAS and MAGIC	  787
10-98.  Change  in median ANC, calcium and magnesium, and sulfate concentrations
       projected for SBRP streams under current  and increased deposition using MAGIC.  .  .  788
10-99.  Comparison of the change in pH after 200 years  as  a function  of the initial
       calibrated pH for MAGIC on SBRP streams, Priority Classes A - E	  789
10-100. Comparison of projected MAGIC change in pH versus derived pH after 50 years
       for NE lakes	  793
                                          xxviii

-------
                                          PLATES
1-1.    Direct/Delayed Response Project study regions and sites	   3
1-2.    Sulfur retention and wet sulfate deposition for National Surface Water Survey
       subregions in the eastern United States	   9
1-3.    Changes in sulfur retention in the Southern Blue Ridge Province as projected
       by MAGIC for constant sulfur deposition	  11
1-4.    Change in median ANC of northeastern lakes at 50 years as projected by MAGIC  ....  13
1 -5.    Change in median ANC of Southern Blue Ridge  Province stream reaches at 50 years
       as projected by MAGIC  	  16

2-1.    Direct/Delayed Response Project study regions and sites 	  26

5-1.    Northeastern subregions and ANC map classes,  Eastern Lake Survey Phase I	  56
5-2.    ANC of DDRP  lakes by ANC group	  77
5-3.    DDRP stream reach study sites in the Southern Blue Ridge Province	  81
5-4.    Final DDRP classification of lake hydrologic type - Subregion 1A	  84
5-5.    Final DDRP classification of lake hydrologic type - Subregion 1B	  85
5-6.    Final DDRP classification of lake hydrologic type - Subregion 1C	  86
5-7.    Final DDRP classification of lake hydrologic type - Subregion 1D	  87
5-8.    Final DDRP classification of lake hydrologic type - Subregion 1E	  88
5-9.    Example of watershed soil map	  119
5-10.   Example of watershed vegetation map	  120
5-11.   Example of depth-to-bedrock map	  121
5-12.   Example of watershed land use  map	  122
5-13.   Example of watershed geology map	  123
5-14.   Example of 40-ft contour delineations on a 15' topographic  map	  131
5-15.   Example of combination buffer: (A)  stream and 30-m  linear buffer for streams, (B)
       wetlands and 30-m linear buffer for wetlands, (C) elevational buffer for lake, and (D)
       combination of all preceding buffers	  133
5-16.   ADS and NCDC sites linked with DDRP study sites for NE Subregion 1A	  194
5-17.   ADS and NCDC sites linked with DDRP study sites for NE Subregion 1B	  195
5-18.   ADS and NCDC sites linked with DDRP study sites for NE Subregion 1C	  196
5-19.   ADS and NCDC sites linked with DDRP study sites for NE Subregion 1D. .  .	  197
5-20.   ADS and NCDC sites linked with DDRP study sites for NE Subregion 1E	  198
5-21.   ADS and NCDC sites linked with DDRP study sites for the SBRP	  199
5-22.   DDRP study sites relative to distance from Atlantic Coast 	  204
5-23.   Pattern  of typical year sulfate deposition  for the DDRP NE study sites	 . . .  209
5-24.   Pattern  of typical year sulfate deposition  for the DDRP study sites in Subregion 1A. . .  210
5-25.   Pattern  of typical year sulfate deposition  for the DDRP study sites in Subregion 1 B. . .  211
5-26.   Pattern  of typical year sulfate deposition  for the DDRP study sites in Subregion 1C. . .  212
5-27.   Pattern  of typical year sulfate deposition  for the DDRP study sites in Subregion 1D. . .  213
5-28.   Pattern  of typical year sulfate deposition  for the DDRP study sites in Subregion 1E. . .  214
5-29.   Pattern  of typical year sulfate deposition  for the DDRP SBRP study sites	  215
5-30.   Pattern  of LTA sulfate deposition for the  DDRP NE study sites	  217
5-31.   Pattern  of LTA sulfate deposition for the  DDRP study sites in Subregion 1A	  218
5-32.   Pattern  of LTA sulfate deposition for the  DDRP study sites in Subregion 1B	  219
5-33.   Pattern  of LTA sulfate deposition for the  DDRP study sites in Subregion 1C	  220
5-34.   Pattern  of LTA sulfate deposition for the  DDRP study sites in Subregion 1D	  221
5-35.   Pattern  of LTA sulfate deposition for the  DDRP study sites in Subregion 1E	  222
5-36.   Pattern  of LTA sulfate deposition for the  DDRP SBRP study sites	  223
                                            XXIX

-------
                                    PLATES (Continued)
7-1.    Sulfur retention and wet sulfate deposition for National Surface Water Survey
       subregions in the eastern United States	  275
7-2.    Regional percent sulfur retention by major land resource area (MLRA) based
       on target populations (ELS and  NSS sites)	  283

11-1.   Sulfur retention and wet sulfate deposition for National Surface Water Survey
       subregions in the eastern United States	  802
11-2.   Changes in sulfur retention  in the Southern Blue Ridge Province as projected by
       MAGIC for constant sulfur deposition	  804
11-3.   Change in median ANC of northeastern lakes at 50 years as projected by MAGIC . . .  808
11-4.   ANCs of northeastern lakes versus time, as projected by MAGIC for constant sulfur
       deposition.	  810
11 -5.   ANCs of northeastern lakes versus time, as projected by MAGIC for decreased sulfur
       deposition	  811
11-6.   Changes in  median pH of northeastern lakes at 50 years as projected by MAGIC .  . .  813
11 -7.   Change in median ANC of Southern Blue Ridge Province stream reaches at 50 years
       as projected by MAGIC	  815
11 -8.   ANCs of Southern Blue Ridge Province stream reaches versus time, as  projected by
       MAGIC for constant sulfur deposition	  817
11 -9.   ANCs of Southern Blue Ridge Province stream reaches versus time, as  projected by
       MAGIC for increased sulfur deposition	  818
11-10.  Changes in  pH of SBRP stream reaches as  projected by MAGIC .  .	  821
11-11.  Changes in  pH of SBRP stream reaches as  projected by ILWAS	: . .  822
                                            xxx

-------
PRIMARY CONTRIBUTORS TO THE DDRP REPORT

     The Direct/Delayed Response Project and this Review Draft Report represent the efforts of many
scientists, technical and support staff.  The primary contributors to this report are noted here.

Section 1:  Executive Summary
     M. R. Church, U.S.  Environmental Protection Agency

Section 2:  Introduction
     M. R. Church, U.S.  Environmental Protection Agency

Section 3:  Processes of Acidification
     P. W. Shaffer, NSI Technology Services Corp.
     G. R. Holdren, NSI  Technology Services Corp.
     M. R. Church, U.S.  Environmental Protection Agency

Section 4:  Project Approach
     M. R. Church, U.S.  Environmental Protection Agency

Section 5:  Data Sources and Descriptions1
     L J. Blume, U.S. Environmental Protection Agency
     G. E. Byers, Lockheed Engineering and Sciences Co.
     W. G. Campbell, NSI Technology Services Corp.
     M. R. Church, U.S.  Environmental Protection Agency
     D. A. Lammers, U.S.D.A. Forest Service
     J. J. Lee, U.S. Environmental Protection Agency
     L. H. Liegel, U.S.D.A.  Forest Service
     D. C. Mortenson, NSI Technology Services Corp.
     C. J. Palmer, NSI Technology Services Corp.
     M. L. Papp, Lockheed Engineering and Sciences Co.
     B. P. Rochelle, NSI Technology Services Corp.
     D. D. Schmoyer, Martin Marietta Energy Systems, Inc.
     K. W. Thornton,  FTN & Associates, Ltd.
     R. S. Turner, Oak Ridge  National Laboratory
     R. D. Van Remortel, Lockheed Engineering and Sciences Co.

Section 6:  Regionalization of Analytical Results
     D. L. Stevens, Eastern Oregon State  University
     K. W. Thornton,  FTN & Associates, Ltd.

Section 7:  Watershed Sulfur Retention
     B. P. Rochelle, NSI Technology Services Corp.
     M. R. Church, U.S.  Environmental Protection Agency
     P. W. Shaffer, NSI Technology Services Corp.
     G. R. Holdren, NSI  Technology Services Corp.

Section 8:  Level I Statistical Analyses
     M. G. Johnson, NSI Technology Services Corp.
     R. S. Turner, Oak Ridge  National Laboratory
     D. L. Cassell, NSI Technology Services Corp.
     D. L. Stevens, Eastern Oregon State  University
     M. B. Adams, Automated Systems Group,  Inc.
     C. C. Brandt, Oak Ridge National Laboratory
     W. G. Campbell, NSI Technology Services Corp.
     M. R. Church, U.S.  Environmental Protection Agency
     G. R. Holdren, NSI  Technology Services Corp.
     L. H. Liegel, U.S.D.A.  Forest Service
                                             XXXI

-------
Section 8: Level I  Statistical Analyses (continued):
      B. P. Rochelle, NSI Technology Services Corp.
      P. F. Ryan,  University of Tennessee
      D. D. Schmoyer, Martin Marietta Energy Systems, Inc.
      P. W. Shaffer, NSI Technology Services Corp.
      D. A. Wolf, Martin Marietta Energy Systems, Inc.

Section 9:  Level II Single-Factor Time Estimates1
      G. R. Holdren,  NSI Technology Services Corp.
      M. G. Johnson, NSI Technology Services Corp.
      C. I. Liff, Utah State University
      P. W. Shaffer, NSI Technology Services Corp.

Section 10:  Level III  Dynamic Watershed Models
      K. W. Thornton, FTN & Associates, Ltd.
      D. L. Stevens, Eastern Oregon State University
      M. R. Church, U.S. Environmental  Protection Agency
      C. I. Liff, Utah State University
            Extramural Cooperators Providing Modelling Expertise and Support:
                  C.  C. Brandt, Oak Ridge National Laboratory
                  B. J. Cosby, University of Virginia
                  S. A. Gherini, Tetra-Tech, Inc.
                  G.  M. Hornberger, University of Virginia
                  M.  Lang, Tetra-Tech, Inc.
                  S. Lee, University of Iowa
                  R. K. Munson, Tetra-Tech, Inc.
                  R. M. Newton, Smith College
                  N.  P. Nikolaidis,  University of Connecticut
                  P. F. Ryan, University  of Tennessee
                  J. L. Schnoor, University of Iowa
                  R. S. Turner, Oak Ridge National Laboratory
                  D. M. Wolock, U.S. Geological Survey

Section 11:  Integration and Summary
      M. R. Church, U.S. Environmental  Protection Agency
      P. W. Shaffer, NSI Technology Services Corp.
2 Contributors to this section listed alphabetically
  Beginning on this line, remaining contributors listed alphabetically
                                               XXXII

-------
                                    ACKNOWLEDGMENTS

      The performance of this  portion of  the  Direct/Delayed  Response  Project  (DDRP)  and the
preparation of this report have required the efforts of hundreds of scientists and support personnel. We
acknowledge here a few of those persons who made particularly  outstanding contributions.  To all the
others who helped us,  but who are not named here, we also extend our sincere thanks.

      William Ruckleshaus led the way in calling for the initiation of the DDRP and  Lee Thomas showed
a continued and very patient interest in seeing that it was completed properly.  We thank them for their
foresight and leadership.

      Courtney Riordan and Gary Foley of the EPA Office of Research and Development  (ORD) provided
much encouragement and support for the Project throughout its development and  implementation. We
thank them for their appreciation of the technical complexity of the task.

      Rick Linthurst,  the first Director of the Aquatic Effects  Research  Program (AERP), played an
absolutely critical role in the development and nurturing of the Project during  its early years.  We  greatly
appreciate his early and continuing commitment to the DDRP.  Dan  McKenzie, as Director of the AERP,
provided important continuing support for the Project and we thank him for his efforts in helping guide
this phase of the Project to its conclusion.

      Tom Murphy, Laboratory Director for EPA's Environmental Research Laboratory-Corvallis (ERL-C),
and Ray Wilhour, Bob  Lackey and Spence Peterson, Branch Chiefs for ERL-C, have all supported the
Project and its staff from the first to the last.  We thank them for their support.

      Dwain Winters and Brian McLean from the Office of Air and Radiation at EPA-Headquarters provided
insight and suggestions for analyses  of particular relevance to questions of  Agency policy.  We thank
them  for their interest and assistance.

      Dixon  Landers, Technical Director of the National Surface Water Survey, Jay Messer, Technical
Director of the Pilot Stream Survey, and Phil Kaufmann, Technical Director of the National Stream  Survey
and their staffs all provided valuable help in interpreting and correctly using their surface  water chemistry
data.  We thank especially Tim Sullivan, Joe Eilers,  Jim Blick, Mark DeHaan,  Alan Herlihy and Mark
Mitch.

      Jim Omernik (EPA),  Andy Kinney (NSI)  and Andy Herstrom (NSI) provided many interesting hours
of instruction and discussion on the topics of physical geography and the proper use  and application
of Geographic Information Systems.  Our efforts in  these technical areas have certainly profited from
their valuable advice  and counsel.

      Bill Fallon  (ORD), Chuck Frank (EPA) and his staff, Linda Looney (EPA), and  Cindy Burgeson (NSI
Technology Services Corp.) all have provided much administrative assistance to help keep the  Project
moving in the right direction and at the pace required. We thank them all for their efforts and assistance.

      Many landowners and state and government agencies allowed us  to  map  and sample soils on
their properties.  We  thank them for permission to do so.
                                             xxxiii

-------
      The cooperation  of the U.S.  Department of Agriculture (USDA)  Soil Conservation  Service (SCS)
was  essential to the completion  of the DDRP Soil Survey. People in  the SCS state offices who were
responsible for mapping of DDRP watersheds and obtaining the soil descriptions and samples included
Ed Sautter, Roy Shook (Connecticut and Rhode Island); Gene Grice, Steve Hundley (Massachusetts);
Dick Babcock, Bob Joslin, Kenny LaFlamme (Maine); Sid Pilgrim,  Henry Mount (New Hampshire); Fred
Gilbert, Keith Wheeler,  Will Hanna (New York);  Garland  Lipscomb, George Martin  (Pennsylvania); Dave
Van  Houten (Vermont); Talbert Gerald, Bob Wilkes (Georgia); Horace Smith, Andy Goodwin  (North
Carolina); Darwin Newton,  David  Lewis (Tennessee); Niles McLoda (Virginia). In addition, more than 100
soil scientists were involved in mapping and sampling.

      Regional consistency and comparability was greatly assisted by the efforts of people at the SCS
National Technical  Centers, especially Oliver Rice, Ted Miller (Northeast) and Larry Ratliff (South). The
continuing support of DDRP activities by Milt Meyer, Ken Hinkley,  and Dick Arnold  of the SCS National
Office was extremely helpful.

      John  Warner, former  SCS  Assistant State Soil Scientist for New  York  was  the  Regional
Correlator/Coordinator  of the Soil Survey for both the Northeast and Mid-Appalachian Regions. Hubert
Byrd, former State  Soil  Scientist for North Carolina, served  as RCC for the SBRP Soil Survey.

      Elissa Levine and Harvey Luce (University of Connecticut),  Bill Waltman and Ray Bryant (Cornell
University), Cheryl Spencer and Ivan Fernandez (University of Maine), Steve Bodine and Peter Veneman
(University of Massachusetts), Bill Smith and Lee Norfleet  (Clemson University),  and Dave Litzke and
Marilew Bartling  (University of Tennessee) supervised the operation of  the soils preparation laboratories
for the DDRP Soil Survey.

      A large and dedicated staff at EPA's Environmental Monitoring  and Systems Laboratory-Las Vegas
(EMSL-LV) played an absolutely crucial role in  support of the DDRP Soil Survey.  Gareth Pearson and
Bob  Schonbrod  provided supervisory guidance for the  DDRP Soil Survey activities at EMSL-LV.  Lou
Blume (EPA)  served as Technical Monitor for the program  and was responsible for delivery  of verified
field, soil preparation laboratory,  and analytical  databases.  Lou Blume was responsible for contracting
and  management  of soil preparation laboratories and  analytical  laboratories and  for the delivery  of
operations reports, quality assurance reports, methods manuals and field sampling manuals for the Soil
Survey.  Mike Papp of Lockheed Engineering  and Sciences Corporation (LESC) was responsible for
delivery of verified field, soil preparation and  analytical databases for the Soil Survey.  Rick Van Remortel
(LESC) assisted  in  the  verification of the SBRP  analytical database and in the preparation of laboratory
operations and quality assurance reports.  Bill Cole (LESC)  was the Task Lead for the verification of the
analytical  database for the NE and  assisted in the preparation  of the  methods  manual and  quality
assurance report for the NE Soil Survey. Gerry Byers  (LESC) assisted in the preparation of methods
manuals  and quality assurance reports for the  NE and  SBRP.  Marilew Bartling (LESC)  served  as the
Task Lead for the verification of Soil Survey data for the SBRP, served as a manager of a soil preparation
laboratory for the SBRP Soil Survey and contributed to the  operations  and quality assurance reports for
the SBRP.  Rod  Slagle  (LESC) served as the DDRP soils database manager at EMSL-LV.  Steve Simon
and  Dan  Hillman (LESC)  assisted in methods  development and project implementation early in the
Project.  Craig  Palmer of  the Environmental Research  Center of the  University of Nevada-Las Vegas
provided invaluable technical assistance on  quality assurance of soils analytical data.
                                             xxxiv

-------
      Deborah Coffey (NSI) played a critical role in ensuring the quality of the watershed and soils data
gathered for the Project.  She either had a major responsibility for, or assisted in, the development of
data quality objectives,  field sampling  manuals,  laboratory methods manuals, field  operations reports,
field quality assurance reports and numerous other  facets of the Soil Survey.  We thank her for her
unswerving attention to  detail.  Jeff Kern (NSI) has also assisted in helping to assure the quality of field
and laboratory data.

      Other scientists who made major contributions to the design of the soil survey activities  included
Stan  Buol (North Carolina State University), John  Ferwerda (University of Maine-Orono),  Maurice
Mausbach (Soil Conservation  Service), Ben Hajek (Auburn  University),  John Reuss (Colorado State
University), Mark David  (University of Illinois), and Fred Kaisaki (Soil Conservation Service).

      Phil Arberg  (EPA), and Dave Williams (LESC)  of EMSL-LV  were responsible  for acquisition and
interpretation of aerial photography of the DDRP watersheds.

      Numerous extramural cooperators assisted in this Project.   Jack Cosby, George Hornberger, Pat
Ryan and David Wolock (University of Virginia), Jerry Schnoor, Tom Lee, Nikolaos  Nikolaidis, Konstantine
Georgakakos and  Harihar Rajaram (University of Iowa), Steve Gherini, Ron Munson  and Margaret Lang
(Tetra-Tech, Inc.),  Carl  Chen  and Louis Gomez (Systech,  Inc.)  all  assisted in watershed modelling
analyses.  Bob Newton  of Smith College assisted in gathering supplementary watershed data for use in
calibrating the models to the Special Interest lake/watersheds in the Adirondacks.  John Reuss and Mark
Walthall of Colorado State University and Tom Voice of Michigan State University performed investigations
of processes of base cation supply and sulfate adsorption, respectively, that assisted us in interpreting
our Soil Survey data and in modelling soil responses.  Warren Gebert, Bill Krug, David Graczyk and Greg
Allord of the U.S. Geological  Survey (Madison,  Wisconsin)  supplied  runoff data and maps that were
crucial to the Project.  Wayne Swank and Jack Waide of the USDA Forest Service cooperated with the
Project  in allowing us to use data gathered by the  Coweeta Hydrologic  Laboratory.  Jack Waide also
provided many insights into the workings of watersheds in the Southern Blue Ridge and in the application
of watershed simulation models.  Tony Olsen,  Sally Wampler and Jeanne Simpson of Battelle Pacific
Northwest Laboratories  provided a great deal of information on estimates of wet deposition to sites of
interest in the  Eastern United States.  Tony Olsen also assisted in  editing text describing analyses of the
wet deposition data.  Robin Dennis and Terry Clark of the EPA's Atmospheric and Exposure Assessment
Laboratory-Research Triangle Park and Steve Seilkop  of Analytical Services, Incorporated, provided key
information on estimates of atmospheric dry deposition. Steve Lindberg of Oak Ridge  National Laboratory
and  Bruce Hicks  and Tilden  Myers of the National Oceanographic and  Atmospheric Administration
provided considerable assistance in the form of discussions and preliminary data on rates of atmospheric
dry deposition. We thank all of these cooperators for their assistance.

      No project of the magnitude of the DDRP can  be  successfully  completed  without the assistance
of peer reviewers.  The DDRP benefitted  immensely from peer review comments all the way from its
inception to the completion of this report.

      The following scientists served as reviewers of the initial Review Draft Report:  David Grigal of the
University of Minnesota, Peter Chester,  R. Skeffington and D. Brown of the Central Electricity Generating
Board  (London),  Jerry  Elwood of Oak Ridge National  Laboratory, John Melack of the  University of
California - Santa Barbara, Phil Kaufmann of  Utah State  University,  Bruce Hicks of  the  National
Oceanographic and Atmospheric Administration,  Gary Stensland of the Illinois State  Water Survey, Jack

                                              xxxv

-------
Waide of the USDA Forest  Service,  David Lam of the National Water Research  Institute  (Burfington,
Ontario), Nils Christophersen of the Institute of Hydrology (Wallingford Oxon, Great Britain), Bill McFee
of Purdue University, Steve Norton of the University of Maine, Scott Overton of Oregon State University,
Ken Reckhow of Duke University, Dale Johnson of the Desert Research Institute (Reno, Nevada),  and
Gray Henderson of the University of Missouri.  We thank these scientists for their efforts in  reviewing a
long and complex document.  We especially thank Dave Grigal (Chairman), Jerry Elwood, John Melack
and Phil Kaufmann who served on the Overview Committee of reviewers. This report benefitted greatly
from the comments and constructive criticisms of all of these  reviewers.

      Numerous other scientists also served as reviewers over the  years  of individual  aspects of the
Project or of the Project as  a whole.  We thank them also for helping us to improve the quality of the
work that we performed.

      Dave  Marmorek, Mike Jones, Tim Webb and Dave Barnard of ESSA, Ltd. provided much valuable
assistance  in the  planning  of various phases  of the DDRP.   Their assistance in this planning was
invaluable.

      John  Berglund of InstaGraphics, Inc. prepared many of the figures that appear in this report.  We
thank him for the fine job  that he did.

      A majority of the word processing throughout the DDRP and, especially, for this report was done
by Carol Roberts  of NSI.   We thank Carol for her  many, many hours of diligent work and for her
forbearance in helping us in  our attempts to  get everything "exactly right".  Significant word processing
support was also provided by Laurie Ippoliti (NSI), Amy Vickland (NSI), Lana McDonald,  Rose Mary Hall
and Deborah Pettiford of Oak Ridge National Laboratory, and Eva Bushman and Suzanne Labbe of Action
Business Services.

      Penelope Kellar and Perry Suk of Kilkelly Environmental  Associates performed truly amazing tasks
in editing both the Review Draft and Final Draft of this report.  The job could not have been completed
on time without their efforts. Ann Hairston (NSI), Amy Vickland (NSI), Susan Christie (NSI) and Linda
Allison (ORNL) also provided important editorial assistance.

      The DDRP Technical Director sincerely thanks all of the Project staff and extramural  cooperators
for their unquenchable enthusiasm and dedication to seeing that this very tough job was  done correctly.
Good work  gang...thank you.
                                             xxxvi

-------
                                         SECTION 10

                   LEVEL III ANALYSES - DYNAMIC WATERSHED  MODELLING

10.1  INTRODUCTION

      Previous  sections have discussed (1) the  principal theories and basic processes  of acidification
(Section 3); (2) the relationship among atmospheric deposition, watershed attributes, and surface water
chemistry (Section 8); and (3) future changes that might occur in watershed sulfate adsorption and base
cation exchange (Section 9) for up to the next 200 years.  This section discusses the Level III Analyses -
the application  of dynamic watershed  models in  projecting future changes in surface water chemistry.

      Three terms are used to describe simulations of future change:

      •     Predict - to estimate some current or future condition within  specified confidence limits on
            the basis of analytical  procedures and historical or current observations.

            Forecast - to estimate the  probability of some future event or condition as a result of rational
            study and analysis of available data.

            Project - to estimate future possibilities based on rational  study and current conditions or
            trends.

The distinction among these three  terms and definitions is the  intended use of the model output.   Level
III  Analyses are defined, and intended to be used, as projections.

      Predictions typically are made to compare different scenarios, controls, or management options.
Predictions can be performed within specified confidence limits because of previous model evaluations,
testing,  applications, and  comparisons with  measured data  for a variety of  system  types.   Model
predictions  of various  surface water  attributes are legally required for  many  proposed management
strategies that range, for example, from examining potential alterations of hydrologic regimes due to land
use modifications to estimating mixing zones for effluent discharges to estimating  phytoplankton response
to nutrient reduction. Predictions generally are performed for short time periods (e.g., single events, parts
of a season, or a  few years) and  focus on before-after comparisons such as water quality before and
after wasteload reductions  or plankton biomass before  and after nutrient  reductions.

      Forecasts convey some estimate of the likelihood or probability  that various conditions or events
will  occur  in  the future.   Daily weather  forecasting,  with associated  probabilities of  showers,
thunderstorms,  etc., is a classic example of forecasting. This represents a short-term forecast.  Weather
forecasts also are made for annual or  decadal time frames.  Flood forecasts can be short term (daily or
weekly), but also are made for long-term events such as the probability  of 100-, 1000-,  and 1,000,000-
year events (NRC, 1988).

      Projections,  in contrast, are not accompanied  by estimates of the probability that any  of the
conditions or events might occur in  the future.   Projections can be  used   as a  basis  for relative

                                               618

-------
comparisons among various emission or deposition scenarios. While the probability that a scenario will
occur cannot be estimated, projections do provide a relative basis for comparing costs and beneficial or
deleterious effects associated with different control or management strategies.  This information generally
is relevant to policymakers and decisionmakers for evaluating different control strategies.  The models
in the Level HI analyses are being used in projecting, not in forecasting, the effects of alternative acidic
deposition scenarios on future changes in surface water acid-base chemistry.

      In Level III Analyses integrated, process-oriented watershed models are used to project long-term
changes (i.e., up to 100 years) in surface water chemistry as a function of current and alternative levels
of sulfur deposition. The watershed models integrate our current understanding of how various processes
and  mechanisms interact and  respond to  acidic deposition.   These mechanisms include soil-water
interactions (including soil-water contact time), sulfur retention, base cation exchange and replacement
of base cations through mineral weathering, and  other watershed  processes (e.g.,  vegetative uptake,
in-lake processes, organic interactions). However, the present study does not establish the adequacy of
the formulations that implement these processes,  the mode of spatial aggregation  of data,  and the
calibration approaches used for long-term acidification projections.

      The three watershed models  that have  been applied are the Enhanced  Trickle Down  (ETD),
Integrated  Lake-Watershed Acidification Study (ILWAS), and  Model of Acidification of Groundwater in
Catchments (MAGIC). The DDRP is an applied project and  has used existing techniques and models for
these analyses.  The use and application of these models to achieve the objectives of the DDRP was
approved by peer reviewers in accordance with the Agency's standard competitive funding process and
requirement for external review of environmental data collection  programs (Section 4.4.3).

      This  section presents

      •     dynamic watershed models used in the Level III Analyses,

      •     operational  assumptions of these analyses,

            watershed prioritization procedures,

            preparation of the modelling datasets (specifically identifying any differences required for Level
            ill Analyses compared to Level  I and II Analyses),

            general modelling approach,    \

      •     model calibration and  confirmation,

      •     model sensitivity analyses,

      •     regional projection refinements,

      •     model projection and uncertainty procedures,

                                               619

-------
            regional population estimates and uncertainties,

      •     regional comparisons and uncertainties, and

      •     discussion and conclusions.

 10.2  DYNAMIC WATERSHED MODELS

      Processes that influence the acid-base chemistry of surface water, and that were considered by the
 NAS Panel  (NAS,  1984), were described in Section 3.  Although these processes can be individually
 identified, discussed, and represented empirically, they are not independent and do not occur in isolation
 from other  processes.  The observed lake  or stream  response to acidic deposition represents the
 integrated response of many watershed and lake/stream processes controlling surface water chemistry.
 To project the future response of a lake  or stream to  acidic deposition, therefore,  requires dynamic
 watershed  models that incorporate and  integrate the important  processes controlling the  acid-base
 chemistry of surface water.

      Both dynamic and steady-state models can be used to project changes in surface water  chemistry
 as a function of changes in acidic deposition.  A dynamic watershed model, however, simulates the time
 trends of various lake, stream,  and watershed constituents, such as ANC, pH, sulfate, calcium, soil base
 saturation, and sulfate adsorption.  A steady-state model can project conditions at only one time in the
 future, the time at which steady state is achieved (i.e., ultimate constituent concentration or value),  and
 does not provide any indication of the changes that occurred between the initial conditions and steady-
 state condition or concentration,  it is the computation of concentrations and processes as a function of
 time that distinguishes dynamic models from steady-state models.

     Three dynamic watershed models were  used to project surface water chemistry for the next 50 to
 100 years both at current and  alternative levels of acidic deposition in the Northeast (NE):

            Enhanced Trickle Down (ETD)  (Nikolaidis et al., 1988;  Nikolaidis et al.,  1989)
            Integrated Lake-Watershed Acidification Study (ILWAS) (Chen et al., 1983; Gherini et al., 1985)
            Model  of Acidification for Groundwater in Catchments  (MAGIC) (Cosby et  al., 1985a,b,c)

Two of these three watershed models also are being used to project changes in surface water chemistry
in the Southern Blue Ridge Province (SBRP) - MAGIC  and  ILWAS.

     Although each model  incorporates the processes considered important in controlling  the acid-
base chemistry of surface water, process resolution and detail vary significantly among the models.  Some
of the processes included in the three models and their spatial/temporal resolution are compared in Table
10-1. The use of multiple models is important because:

     •     the level of detail  by which each process or mechanism  is represented varies  between
           models, reflecting the relative importance of each process in the systems for which the model
           was first developed;

                                              620

-------
Table  10-1.  Major Processes Incorporated in the Dynamic Model Codes
(Parentheses Indicate Limited Treatment of Process, and Dashes Indicate
Processes not  Included in a Code) (Jenne et al., 1989)
                                   MAGIC/TOPMODEL    ETD/PEN
ILWAS
Atmospheric Processes
- Dry deposition
- Wet deposition
Hvdrological Processes
- Snow sublimation
- Evapotranspiration
- Interception storage
- Snowmelt
- Overland flow
- Soil freezing
- Macropore flow
- Unsaturated subsurface flow
- Saturated subsurface flow
- Stream flow
- Lake stratification
- Lake ice formation
Geochemical Processes
- Carbonic acid chemistry
- Aluminum chemistry
- Organic acid chemistry
- Weathering
- Anion retention
- Cation exchange
Bioaeochemical Processes
- SO42' reduction in lake
- Nitrification in soil
- Nutrient uptake
- Canopy interactions
- Litter decay
- Root respiration
-"
X
X

_
X
(X)a
X
X
-
X
X
X
X
-
-

X
X
X
X
X
X

(X)'
(X)*
(X)b
(X)a
(X)a
(X)a

X
X

X
X
-
X
X
X
-
X
X
_
_
-

X
-
-
X
X
X

X
-
.
-
-
-

X
X

X
X
X
X
X
X
-
X
X
X
X
X

X
X
X
X
X
X

X
X
X
X
X
X
     Cosby, B.J. (written review comments, 1988) considers that canopy interactions and root decay
     and respiration are implicitly included in the MAGIC code by use of a dry deposition factor
     and by designation of CO2 partial pressure in soils and surface waters.

     Sulfate reduction, nitrification, and uptake of ions can be simulated with the MAGIC code by
     specifying uptake rates of SO4  and NH4+  for various hydrologic compartments.
                                                621

-------
           identification of similar key watershed parameters and processes in each model and their
           relationship to measured watershed characteristics provides greater confidence in conclusions
           about which factors influence the acid-base chemistry of surface water; and

      •    long-term data  sets for model validation/verification do not exist, so model accuracy and
           precision for long-term projections is unknown; however, similar projections of watershed
           responses among the models provides greater confidence in the  conclusions.

10.2.1  Enhanced Trickle Down  (ETD) Model

     The ETD is a lumped parameter model based on the concept of ANC mass  balance.  Various
chemical processes in the ETD model, such as mineral weathering, sulfate adsorption  and  desorption,
and cation exchange, are incorporated as either consuming or producing ANC (Schnoor and Stumm,
1985).  Rate expressions are used to describe mineral weathering, cation exchange, and sulfate reduction
reactions.  Equilibrium expressions are used to describe carbonic acid chemistry and  sulfate adsorption
and desorption.  ETD explicitly incorporates mineral weathering rate reactions and sulfate reduction but
does not explicitly incorporate chemical reactions involving aqueous aluminum, nitrate, or organic acid
chemistry,  although the ETD code does implicitly consider contributions to total acidity  from these
constituents.   Mass balance calculations are considered for ANC (equivalent to the modified  Gran ANC),
sulfate, and chloride, with chloride considered to be a conservative constituent.

     The Trickle Down model, a precursor to ETD, was formulated to perform assessments of the effects
of acidic deposition on a number of seepage lakes in the Upper Midwest.  The objective  of the modelling
effort was to provide a model with sufficient detail to calculate alkalinity concentrations in surface water,
soils, and ground water, but sufficiently simple to apply using a microcomputer with one master variable
(alkalinity) for acidic deposition assessments (Schnoor et al., 1984, 1986a). Enhanced Trickle Down was
modified to include sulfate  adsorption  and  desorption and improved hydrologic flowpaths (Nikolaidis,
1987).  The hydrologic submodel  simulates snowmelt, interflow, overland flow, groundwater flow, frost-
driven  processes, seepage,  and evapotranspiration (Nikolaidis et al., 1989). ETD is spatially partitioned
into three vertical components within the watershed:  soil, unsaturated  zone, and ground water. The
watershed  contributes to a lake compartment.  The lake and terrestrial  compartments are considered
areally homogeneous.  The temporal resolution of the  ETD output generally is daily.

     The meteorological and deposition input requirements for ETD are illustrated in Tables 10-2 and 10-
3.  The  chemical  constituents  projected in  soil solution and surface water are listed in  Table 10-4.
Because of the importance and pivotal role that ANC has in the projections of surface water  acidification
and chemical  improvement, the components  of the ANC calculations are shown for each of the three
models in  Table  10-5.  The  ANC calculation for ETD corresponds with the ANC calculation  for the
modified Gran titration method. These input  requirements and output constituents are contrasted with
those included in ILWAS and MAGIC.

     The ETD model was  originally applied to  Lakes  Clara and Vandercook in northcentral Wisconsin
(Lin and Schnoor, 1986),  as a joint effort by the U.S. EPA,  U.S. Geological Survey, the Wisconsin
Department of Natural Resources,  and the University of Iowa. ETD reproduced the seasonal and annual
changes  in water chemistry for the short periods of record on  these two lakes. ETD has since been

                                              622

-------
Table 10-2.  Meteorological Data Required  by the Dynamics Model Codes (from Jenne et
al., 1989)
Meteorological
   Data
MAGIC/TOPMODEL
ETD/PEN
                                                                               ILWAS
Interval for data
measurement

Precipitation

Relative humidity
or dewpoint

Min. air temperature

Max. air temperature

Ave. air temperature

Mean daylight hours

Cloud cover (fraction)

Atmospheric Pressure

Wind Speed
                                       Monthly8
                                       yearly
                                          m
                                          °C

                                          °C
                           Daily
                                                               mm
                                                             (unitless)
                                                               km day"1
                    Daily
                                                                                 cm
                                               °C

                                               °C
                                              (unitless)b

                                              mbars
                                              m sec"
                                                                                       "1
 TOPMODEL runs with a daily time step.

 Average values per month required.
                                             623

-------
Table 10-3. Chemical Constituents in Wet and Dry Deposition Considered by the
MAGIC, ETD, and ILWAS Codes (from Jenne et al.,  1989)
Constituent
                   MAGIC
                        ETD
                                 ILWAS
Wet
Dry3
Wet
Dry
Wet
Dry
S0x(g)
N0x(g)
H+
Al (total)
Ca2+
Mg2+
K+
Na+
NH+4
S042'
NOg"
cr
F
P043'
ANC
TOC°
TICd
H4Si04
Units
Interval


X
X
X
X
X
X
X
X
X




(X)b
(X)b

X
X
X
X
X
X
X
X
X




^eq L"1 /interval
monthly
or
yearly ave.

XXX
X
X
X
X
X
X
XXX
X
XXX

X
X X
X X
X
X
meq m"2 meq m"2 mg L"1
-daily- volume
monthly
X
X
X
X
X
X
X
X
X
X
X
X

X

X
X
X
mg m"3
weighted
average
  The MAGIC code requires that dry deposition be expressed by means of a
  dry deposition factor.

  Cosby, B.J. (written communication, 1988) considers that SOx(g) and NOx(g)
  are implicitly included by means of the dry deposition factor.

  Total organic carbon

  Total inorganic carbon
                                       624

-------
Table 10-4. Chemical Constituents Included in Soil
Solutions and Surface Water for the MAGIC, ETD,
and ILWAS Codes (from Jenne et al., 1989)
Chemical Constituent
ANC
Ca2+
Mg2+
K+
Na+
NH4+
H+
AI3+
AI(OH)n3-" (n=1 to 4)
AI(F)n3-" (n=1 to 6)
AI(S04)n3-" (n=1 to 2)
AI-R(a)
S042'
NCV
cr
F
H2PCV
H4SiO4(aq)
C02(g)
C02(aq)
H2C03(aq)
HC03-
cc/
HR'°, R"(b)
H2R"0, HR"', R"2-(b)
H3R'"0, H2R'"-(b)
HR'"2-, R'"3-
MAGIC ETD
X X
X
X
X
X
X
X X
X
X
X
X
X
X X
X
X X
X


X X
X X
X X
X X
X X

X


ILWAS
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X
X
X
X
X
X
X
X
X

X

  AI-Refers to the various organic complexes
  of aluminum.
 5 R', R", and R"1 refer to monoprotic, diprotic,
  and triprotic organic acids, respectively.
                                       625

-------
Table 10-5.  Definitions of Acid Neutralizing Capacity (ANC) Used by the MAGIC, ETD,
and ILWAS Codes (Brackets indicate concentration in molar or molal units, and R',
R", and R'" represent mono-, di-, and triprotic organic acids, respectively.)  ANC
Simulated by All Three Models is Equivalent to the Modified Gran ANC (from Jenne et
al., 1989)
Code
Units
                ANC Definition
MAGIC
ETDa
ILWAS
    L1
meq rrf
ANC =  [HCO3"] + 2[CO32'] +  [OK] +  [HR"~]


      + 2[R"2-]  + [AI(OH)4] -  [H+] - 3[AI3+]

      - 2[AI(OH)2+] - [AIOH2+]



ANC =  [HC03-] + 2[C032'] +  [OK] - [H+] + [R1']



ANC =  [HCO3~] + 2[CO32'] +  [OH'] +  [H2R"r]


      + 2[HR'"2-]  + SIR'"3"] +"[R'']

      + [AIOH2+]  + 2[AI(OH)2+] + 3[AI(OH)3°]

      + 4[AI(OH)4] + 3[AIR'"°] + [AIR'2+]

      + 2[AI(R')2+]
  The ETD code operates on the principle of ANC mass balance.
                                            626

-------
applied to several Adirondack lakes including Woods Lake, Panther Lake, and Clear Pond (see Appendix
A), other lakes in the Upper Midwest, and  several lakes in the Sierra Nevada mountains of California
(Nikolaidis et al., 1988, 1989; Lee et al., in press).

10.2.2 Integrated Lake-Watershed Acidification Study (ILWAS1 Model

      The ILWAS  model  is a  process-oriented  model  that  uses  both  equilibrium and rate-limited
expressions to describe mass balances for 18 chemical constituents (Table 10-4). The ILWAS algorithms
represent the effects of biogeochemical processes on surface water chemistry  (see Table  10-1).  Mass
balances for the major cations and anions and the effects of aqueous aluminum and  organic acids on
surface water chemistry are incorporated in the model.  The ILWAS model was formulated to simulate
the seasonal and long-term changes in water chemistry caused  by acidic deposition. As a result, ILWAS
has a strong assessment capability (Huckabee et al., 1989). The ILWAS model contains three modules:(1)
a  canopy module to simulate  forest canopy  interactions with both wet  and dry deposition,  (2)  a
hydrology and watershed soil module to route precipitation through the soil  horizons and simulate soil-
water physicochemical processes and biotic transformations, and (3) a lake  module to simulate aquatic
biochemical  reactions (Chen et al.,  1983; Gherini et al., 1985).

      The vertical  resolution  in the ILWAS code includes the  canopy,  a  snow component,  stream
segments, a lake component, and up to 10 soil layers for each subcatchment  in the watershed (Chen
et al., 1983).  The ILWAS model  can simulate up to 20 subcatchments and  associated stream segments.
For most DDRP watersheds, only one or two subcatchments were used. To  calculate the distribution of
water between flowpaths, the ILWAS model  uses various forms of the continuity equation, Darcy's law
for flow in unsaturated and saturated permeable  media, and Manning's equation,  Muskingum routing,  and
stage-flow relations for  surface waters (Chen  et  al., 1982, 1983).   The vertical  layers within each
subcatchment are assumed to be areally homogeneous.   The lake is vertically one-dimensional with up
to 80 vertical layers including snow and ice layers.  For DDRP application, the layer thickness was set
at 1 m so most lakes had between 3 and 7 layers. The temporal resolution of ILWAS output is generally
daily, but  more frequent output  can be obtained (with added computational effort and increased input
data).

      The meteorological and deposition input  requirements are shown in Tables 10-2 and  10-3.  The
output variables in the soil  solution and water chemistry are listed in Table  10-4.  The components of the
ANC calculation are shown in Table 10-5. The ANC simulated by ILWAS  is equivalent to  the modified
Gran ANC.

      The ILWAS model was developed to further the understanding of how atmospheric and terrestrial
acid-base  processes interact to produce observed surface water chemistry.  The model was developed
as part of the  Electric  Power Research Institute's (EPRI) Integrated  Lake-Watershed Study of three
Adirondack lakes, Woods,  Sagamore, and Panther (Chen et al., 1983; Gherini et al., 1985; Goldstein et
al., 1984). The model reproduced the seasonal and  annual changes in water chemistry in these three
lakes for the 5  years of record   (see Appendix A).  The model  has subsequently been applied to 25
watersheds in Wisconsin, Minnesota, North Carolina, Tennessee, Utah, and California (Chen et al., 1988;
Davis et al.,  1986; Gilbert et al., 1988; Greb et  al., 1987; Munson et al., 1987).  Regional assessments
                                             627

-------
have been conducted as part of the EPRI- funded Regional Integrated Lake Watershed Acidification Study
(RILWAS) and through other independent applications (Gherini et al., 1989).

10.2.3  Model of Acidification of Groundwater in Catchments (MAGIC)

      MAGIC is a lumped-parameter model of intermediate complexity, originally developed to project the
long-term effects (i.e., decades to centuries) of acidic deposition on surface water chemistry. One of the
model's principal assumptions is that a minimum number of critical processes in a watershed influence
the long-term response to acidic deposition.  The model simulates soil solution  chemistry and surface
water chemistry to project  the  monthly  or annual  average  concentrations of the  water chemistry
constituents listed in Table 10:4.  Hydrologic flow of water through soil  layers to the receiving system is
simulated using a separate hydrologic model, TOPMODEL (Hornberger et al., 1985).  TOPMODEL is a
topography-based, variable contributing area, catchment model adapted from the version of Beven and
Kirkby (1979). The model considers overland flow, macropore flow, drainage from the upper zone to the
lower zone and to the stream, and baseflow from the lower zone. Flow routing through the watershed
is provided from TOPMODEL to MAGIC,  a model  in which  both  equilibrium and  rate-controlled
expressions are used to represent geochemical  processes.  Mass balances for the major cations and
anions and the effects of aqueous aluminum and organic acid  species on ANC are incorporated in the
model.  The ANC simulated  by MAGIC is equivalent to the modified Gran ANC.  These processes are
listed  in Table 10-1.

      MAGIC represents the watershed with two soil-layer compartments.  These soil  layers can be
arranged vertically or horizontally to represent the vertical or horizontal movement, respectively, of water
through the soil.   A vertical configuration was  used  in the DDRP, and the soil  compartments were
assumed to be areally homogeneous. Annual output is the typical temporal resolution of the model, but
monthly output also can be obtained.

      The meteorological and depositional input requirements for MAGIC are shown in Tables 10-2 and
10-3.   The  output soil  solution  and water chemistry  constituents are shown in Table 10-4.  The
components included in  the ANC calculation are shown  in Table  10-5.

      MAGIC was originally formulated to be parsimonious in selecting  processes for inclusion and was
intended to be used as  a heuristic tool for understanding the influences of the selected  processes on
surface water acidification. The spatial/temporal formulations  in the model reflect the  intended use for
assessment and multiscenario evaluations. It was originally developed on two southeastern streams but
has subsequently been applied to many watersheds in  the Southeast, lakes in the Adirondacks, and
watersheds in England, Scotland, Norway,  Finland, and Sweden (Cosby et al., 1985a,b, 1986a,b,c, 1987;
Lepisto  et al., 1988;  Musgrove  et al.,  1987;  Neal et  al.,  1986;  Whitehead et al., in  press).  MAGIC
reproduced the annual changes in water chemistry for these systems for the short period of available
record.  It also has been used for a  regional assessment of Norwegian lakes using the Norwegian lake
resurvey data (Cosby et  al., 1987;  Hornberger et al., 1987a,b).
                                             628

-------
10.3  OPERATIONAL ASSUMPTIONS

      There are several operational assumptions associated both with DDRP and the individual models
(Table 10-6). These assumptions underlie the DDRP Level III Analyses in toto.  Each of the models has
specific assumptions in addition to those made for the DDRP.  These specific assumptions, summarized
by Jenne et al. (1989),  are described in  more detail by the authors and developers of the models (Chen
et al., 1983; Cosby et al., 1985a,b,c; Gherini et al., 1985; Nikolaidis,  1987; Nikolaidis et al., 1988, 1989).

10.4  WATERSHED PRIORITIZATION

      The general approach for selecting the DDRP watersheds was  described in Section 5.2.  This
section presents the approach for prioritizing watersheds for Level III Analyses in the NE and SBRP.

10.4.1  Northeast

      Developing a priority order for performing the watershed  calibrations and forecasts permitted early
comparisons among model  outcomes, identified data problems of general concern  to all three models
as the problems developed, and  permitted joint resolution of these problems  by all modelling  groups.
The priority ordering ensured that problems encountered with regard to the watershed classes of highest
interest or greatest concern could be  addressed early  in the projection period, so  that if additional
projections were  precluded  due to time or manpower  constraints,  projections for  the  highest priority
systems would be completed by all three modelling groups.

      A decision tree was developed for the watersheds in the NE (Figure 10-1).  The decision tree was
based on several criteria including previous calibration and projections,  internal sources of sulfur,
hydrologic type, sulfur retention,  chloride status,  and  ANC [based  on values from the Eastern Lake
Survey  Phase I (ELS-I), Linthurst et al.,  1986a)].   These criteria were  used to rank the watersheds  in
descending  order of priority  with the highest priority given to Class A watersheds and the lowest priority
to Class I  watersheds.  The number of lakes in each priority class (A -1) is shown on the right-hand side
of the priority class box.

      Class  A watersheds are those that previously had been investigated as part of an  internal EPA
evaluation for the Administrator. Two of  the models previously had been calibrated on these watersheds,
so minimal  problems  were anticipated  in  recalibration with  aggregated soils  data  and  site-specific
deposition data. Watersheds with unequivocal internal  sources of sulfate confound the effects of sulfur
deposition on surface waters, so these systems were ranked lowest priority.  Systems with no inlets or
outlets, i.e.,  seepage lakes (see Section  5.3), also confound interpretation of sulfur deposition effects on
surface water chemistry because  of internal alkalinity generation; these systems  also were ranked as a
lower priority class for projections.  Although drainage lakes with  long residence times  (i.e., > 1  yr) also
can  have significant  internal sulfate reduction,  the estimated  median  hydraulic  residence  time for
northeastern lakes was 0.20 yr so internal alkalinity generation for the DDRP lakes was  not considered
to be a confounding factor.  Watersheds that appear to be currently retaining sulfur were judged to be
higher priority than watersheds that appear to be at or near sulfur steady  state, because  of the  potential
for acidification as  sulfate (acting as a  mobile  anion)  depletes soil  base cations.  Many  northeastern
watersheds are influenced by the application of road salt (calcium chloride, magnesium chloride, sodium
                                               629

-------
Table 10-6.  Level III Operational Assumptions
1.     Index sample water chemistry from the NSWS provides an index of chronically acidic systems and
      systems with low ANC that are susceptible to acidic deposition.

2.     Index soil data from the DDRP Soil Survey adequately characterize watershed attributes influencing
      surface water chemistry.

3.     Projections of future acidification consider  primarily chronic acidification.  Episodic acidification is
      considered in the EPA Episodic Response Project.

4.     Surface water acidification is a sulfur-driven process.  Sulfur is assumed to be the primary acidifying
      agent in acidic deposition.  Eastern deciduous forests generally are nitrogen-limited (Likens et al.,
      1977; Swank and Crossley, 1988) so there  is low export of nitrate.  In addition, annual nitrate
      deposition exceeds annual  ammonium deposition in the eastern United States (Kulp, 1987) and
      nitrate has a slight alkalizing effect in the watershed (Lee and Schnoor, 1988).

5.     The watershed processes controlling the effects of sulfur deposition on  surface waters are sulfate
      adsorption and desorption and base cation depletion and resupply through mineral weathering and
      exchange.

6.     The effects of organic acids on acid-base chemistry are constant through time and independent
      of sulfate.

7.     These major processes are known well enough to be incorporated in the projection  models used
      in the DDRP.

8.     Current watershed attributes and conditions (e.g., climate, land use, basin characteristics) will remain
      relatively constant over the next 50 years.

9.     Long-term projections using models are plausible and are the only feasible approach for evaluating
      the  regional, long-term effects  of sulfur deposition scenarios on surface water chemistry.

10.   "Typical" year projections are  not intended to represent future forecasts of water chemistry  but
      rather to provide a common basis for comparisons among deposition scenarios to assess potential
      changes in surface water chemistry.

11.   Acidification is reversible and the processes in the models are adequate  to describe both chemical
      acidification and chemical improvement.

12.   Physical and chemical processes are adequately considered in the Level III models.

13.   Uncertainty calculations provide estimates of relative error for long-term comparisons among models
      and deposition scenarios but are  not absolute error estimates.
                                               630

-------
C n = 145  J

  JE_
                                           *\"«-  ,,--\°«'vv
                                        ODELLING PRIORITY
                                       DECISION TREE: NE
             1 of the 10
          Special Interest
            watersheds

                YES
              NO
          Internal source
             of sulfur
                YES
              NO
                                 ^
           Drainage lake
           or reservoir
          Positive sulfur
            retention
Figure 10-1. Modelling priority decision tree:  Northeast.

                                       631

-------
chloride). The DDRP watersheds were screened to identify those systems for which the output chloride
was greater than the input from atmospheric sources.  Those watersheds with significant net chloride
export were given a lower priority.  Finally, those systems with initial ANC < 100 [leq L"1 were designated
higher priority than watersheds with ANC > 100  —  ' "1
      Ail three modelling groups followed this priority order when making projections.  The objective was
to complete analyses on at least the first 60 watersheds, which included those with ANC <100 peqL"1,
those that were, the  least disturbed with respect to road salt additions, those near sulfur steady state,
and those currently retaining  sulfur within the watershed (i.e., Classes A - C).

10.4.2  Southern Blue Ridge Province

      A decision tree also was developed for the SBRP watersheds (Figure 10-2) using criteria similar to
those for the NE with two exceptions:  watersheds previously were screened for internal sources of
sulfate, and none of the dynamic models was calibrated previously on SBRP watersheds. Therefore, the
first criterion for prioritization was whether  the watersheds are currently  retaining sulfur, followed  by
whether  chloride concentrations are less than 50 /*eq L"1  (indicating  little  impact or disturbance  by
roadsalting practices).  The chloride criterion was the same as that used for  northeastern lakes.  Those
systems with ANC < 100 /^eq L"1  [based on values from the National Stream Survey (NSS) Pilot Survey
(Kaufmann etal., 1988)] also were given higher priority than those with ANC > 100 ^eq L"1.  The number
of streams in each priority class is shown on the right-hand side of the priority class box (i.e.,  A - E).
Of the 35 total watersheds for  the SBRP, 25 were placed in the first two priority classes, which also
resulted in a restricted target population. This priority order was followed by the ILWAS and  MAGIC
modelling groups in performing projections for watersheds in at least the first two priority classes.  ETD
was not applied to streams, so  SBRP watersheds were not simulated using ETD.

10.4.3  Effects of Prioritization on Inclusion Probabilities

      Watersheds were ranked in priority order to minimize comparability problems among models in the
event that not every group could complete simulations on all 145 watersheds in the NE or 35 watersheds
in the SBRP.   This  prioritization  scheme does not affect the inclusion probability of any watershed.
Inclusion probabilities are based on the statistical design and the manner  by which the sample watersheds
were  selected from the population of  watersheds in the region (see Section 5.2.6).  Simulating only
selected  classes of watersheds, however, does affect the target population about which inferences can
be drawn. For example, if no watersheds with initial ANC > 100 peq L"1 are  included in the projection,
no inferences or conclusions  can  be drawn about the future response of this class  of systems to acidic
deposition scenarios.  Even though the original target population had ANC concentrations ranging from -
87 to  400 fj,eq L"1 the new target population about which inferences can be reached  refers only to that
portion of the original population with ANC concentrations ranging from  -53 to 100 ^eq L"1 .  The DDRP
target population for the NE that corresponds to these  Class A - C watersheds represents 1,851
watersheds compared with the full northeastern DDRP target population of 3,667 watersheds.  The first
two priority  groups  in the SBRP  also represent a  restricted target 'population of  1,051 watersheds
compared with  the full SBRP  DDRP target population of 1,531 watersheds.
                                              632

-------

( "=35
. I
Positive
sulfur
retention

>


I
NO

: ^DECISKDN TREE : SBRP'X;^
L^ ssss^ ssl %, * '•T'^s;^ ^ s ^ f'\ % *s^*S s^sS.ss.-.sssss^s.s^, ^ Y X v s s ^ ^ ^ ^ ^



is Priority ^
i ^ Class o'
V %-• s
v,"^ •. '•'"'•'••!

1 YES
[Cl] < 50
neq L-1

NO



k Priority^
s Class ^
\ s \ ^^^
•s.

1 YES
ANC < 100
neqL'1

?
NO



1 YES
I* Priority^
-^ Class ^
12
v. •• v,- •
%«&#• \-A:.

ANC < 200
jieq L-1
1 YES
s Priority ,-
Class s
13
••v
NO \p^°r|tyr ^
1 » X/ : .. f* }r>£+f+ X-.O ^S^^
^ '• LrlaSS ^ : p:
V> ^ ^ ^i ^ ^^ ^ :"• s "-^ -.
.^w^tiJH 	 -f-: , MnrlVtfiii

Figure 10-2.  Modelling priority decision tree: Southern Blue Ridge Province.
                                             633

-------
10.5  MODELLING DATASETS

      A major objective of the Level III Analyses was to ensure that all three modelling groups were given
the  same  datasets,  developed  using  identical  procedures  so  that  differences  among  model
projectionsreflected differences in model process formulations and not differences in input data.  Different
process formulations among the models requires that different averaging or aggregation procedures be
used to prepare model input and parameter data.  The source data provided to each modelling group
(e.g.,  meteorology, deposition, morphometry,  soils, and water chemistry) on which these procedures
operated, however, were  identical for each model to minimize problems of comparability among model
projections.

10.5.1  Meteorological/Deposition  Data

      The meteorological  and deposition data for both the NE and SBRP were discussed in Section 5.6,
with the exception  of daily  meteorological  data,  which  were  specific to the  Level  III  Analyses.
Meteorological data for daily temperature, dew point, pressure,  wind  speed, cloud cover, and solar
radiation also were required as model simulation input for ETD  and  ILWAS.  These data are not measured
at as many locations as daily precipitation  is.  Typical meteorological  year (TMY) data have been
produced for 238 locations across the United  States.  These  locations are usually in major cities.  Ten
different TMY sites were selected and  matched to each DDRP site, based on geographic location and
elevation. Temperature and dew point were adjusted to match 30-year normal temperatures for the period
1951-1980.   TMY temperature data also were adjusted to closely  match long-term monthly average
temperatures at the TMY site.  Hence, the monthly and daily temporal pattern for each TMY site was
representative of the long-term norm.  Because temperature is elevation dependent, the  TMY data were
adjusted to  match the annual 30-year normal at  a nearby site with an elevation comparable to the
National Climatic Data Center (NCDC) site assigned to a DDRP lake. The adjustment was additive based
on  the  difference between the  annual average  TMY  temperature  and  the annual 30-year normal
temperature.  A similar adjustment was applied to dew point.

      Watershed specific "typical year" meteorological and deposition data were provided to the modelling
groups for each watershed in the NE and SBRP.  These typical year data  were repeated year after year
for  50 years in performing the  watershed projection under current deposition  levels.   The altered
deposition scenarios for the NE and  SBRP followed a temporal sequence of current deposition levels for
the first  10  years of the projection,  altered  sulfur deposition  for the next  15  years  to  the  desired
percentage change relative to current deposition (30 percent decrease in the NE, 20  increase in the
SBRP), and then constant sulfur deposition at this altered level for the final 25 years.

10.5.2  DDRP Runoff Estimation

      Runoff is an important variable for the models used in  Level III Analyses.  The  DDRP study sites
are not gaged, so measured estimates of runoff were unavailable. A combination of techniques was used,
therefore, to obtain estimates of annual and monthly runoff for the northeastern and SBRP  watersheds.

10.5.2.1  Annual Runoff

      Annual runoff was  estimated  for each  of the 145  northeastern  and  35 SBRP watersheds, as
discussed in Section 5.7.  Long-term average annual runoff estimates were based on 1951-1980 records.
                                              634

-------
The annual runoff was partitioned into average monthly fractions for use in calibrating the hycfrofogfc
submodels.

10.5.2.2  Monthly Runoff

10.5.2.2.1  Northeast -

      Monthly runoff estimates were calculated for the NE  based on USGS long-term monthly averages.
USGS calculated a 30-year  average monthly runoff proportion  for a  12-month  period  (October  -
September) using 1951-1980 runoff data for stations that had complete records for the 30-year period and
did not have diversions or regulations (D.  Graczyk,  personal  communication).   The final database
contained runoff data for 134 USGS gaging stations.

      The USGS sites were linked then to the 145 DDRP study sites and 3 intensive study sites.  For the
NE, a "nearest neighbor" linkage was used with physiographic considerations included when appropriate
(R. Nusz, personal communication).   Using  the Geographic Information  System (GIS), a map was
prepared that depicted locations of USGS gaging stations and DDRP study sites.  A USGS station was
linked  to  each study site  based on station  proximity.   In  areas like  the White Mountains  of New
Hampshire, physiographic considerations (e.g., elevation and topography) also were included in the linking
process.   Physiographic data were obtained from Krug et al. (in press).  In the Adirondack Subregion
(Subregion 1A in the ELS-I), USGS station density was extremely sparse  relative to the number of ELS
sites.   In this area, a Thiessen  polygon weighting system was used to link the few USGS stations to the
ELS sites.  In many cases,  more than one DDRP site was linked to a single USGS station.

    Monthly  runoff  for each study site was calculated by applying the 12 monthly proportions for each
USGS station to the linked DDRP study sites. The average annual runoff value at each study site,
interpolated from the map of Krug et al. (in press), was multiplied by each of the 12 monthly proportions
to obtain 12 monthly runoff values (in inches)  for each site.

10.5.2.2.2  Southern Blue Ridge Province -

      Monthly runoff for the SBRP watersheds was estimated using a USGS database prepared  similarly
to the one for the NE.  The resulting database contains 30-year average monthly proportions (October -
September) for 41  USGS stations in the SBRP.

    The  USGS monthly proportion  data were linked to the interpolated  annual runoff at each site to
calculate a long-term monthly  runoff estimate for each of the 12 months for the water year.  The USGS
stations had limited spatial coverage of this region and did not  overlap consistently with the DDRP study
sites.   The GIS was used to link the USGS sites and DDRP study sites based  on topographic  features
and similar site characteristics. An average monthly proportion for each of the 12 months was calculated
for the USGS sites within the major land resource area  (MLRA) to  obtain a single file of 12  monthly
proportions.  For the SBRP, all but one of the watersheds were located in a single MLRA:

    Monthly  runoff for each DDRP study site was calculated by applying the single file of  12  monthly
proportions for each MLRA to the study sites located in the respective MLRAs. The average annual runoff
value at each study site, interpolated from the map of Krug et al. (in press), was multiplied  by each of
the 12 monthly proportions to obtain 12 monthly runoff values (inches) for each site.
                                              635

-------
10.5.3 Morphometry

      Basin, lake, and stream morphometry and  characteristics were discussed in Section 5.4.  These
data, obtained from the DDRP Soil Surveys and the NSWS for the NE and SBRP, were provided to each
modelling group for use in model calibration for each watershed in the NE and SBRP.

      Lake volume and stage-discharge empirical relationships  for the northeastern watersheds were
formulated using data  obtained  from the  ILWAS, RILWAS, ME, and VT  lakes for which bathymetric
information was available. These lakes were  partitioned by surface area and  relationships established
between  lake  volume and  lake  area.   Lake volume relationships were  improved  if  the  lakes were
partitioned by surface area  (i.e.,  surface area < 100 acres and surface area > 100 acres).  These
relationships are

           Volume (acre-ft)  = 4.486[Lake Area (acres)]1'382
                 for lake areas  < 100 acres, r2  = 0.87
           Volume (acre-ft)  = 5.670[Lake Area (acres)]1'227
                 for lake areas  > 100 acres, r2  = 0.96

      Empirical relationships also were established between  lake stage and discharge based  on data
available for ILWAS and RILWAS lakes. A relationship between discharge (Q), height of the lake spillway
(h) and lake depth was established for different classes of lakes based on their watershed areas. These
relationships, which were used in the ILWAS model, are

                 Q (cfs) =  2.694h (ft)3-538
                      for lakes with watershed  areas < 350 ha, r2 = 0.98
                 Q (cfs) =  0.897h (ft)5'279
                      for lakes with watershed  areas 350 - 600 ha, r2 = 0.98
                 Q (cfs) =  3.160h (ft)3'70
                      for lakes with watershed  areas 601 - 3000  ha, r2 = 0.96

Lake volume for the ETD and MAGIC simulations was assumed to be constant (i.e., inflow volume =
outflow volume  + evaporation volume + seepage volume) so a stage-discharge relationship was not
required.
10.5.4  Soils

      The soils data, discussed in Section  5.5, were aggregated (Sections 9.2.3.2 and 9.3.1.2)  and
provided to each modelling group on a model-specific basis.  The soils data used by each modelling
group were identical, but the aggregation procedures used by each  group were model  specific.  The
ILWAS modelling  group used unaggregated data.
                                              636

-------
10.5.5  Surface Water Chemistry

      As described in Section 5.3, surface water chemistry data were obtained from the NSWS and were
described in detail by Kanciruk et al. (1986a) and Messer et al. (1986a). Both 1984 ELS-I and 1986 ELS-
II data were provided for the northeastern watersheds, and NSS Pilot stream data were provided for the
SBRP watersheds for all  sampling periods in 1985.

10.5.6  Other Data

      Watershed data such as bedrock geology, land use,  vegetative cover, estimated depth to bedrock,
and other data (discussed in Sections 5.4 and 5.5.6) were provided to each of the modelling groups for
calibration of the individual  watersheds.  Because of different model formulations, the data were used
differently during  model calibration but the  information provided to each modelling group was identical.

10.5.7  Chloride Imbalance

      Preliminary mass balances for chloride indicated that chloride export exceeded deposition input for
a significant number of northeastern watersheds.  Road salt,  watershed disturbances,  and other factors
might account for these additional chloride inputs.  A decision tree approach (Figure 10-3) was used to
identify watersheds with net  chloride export and to correct  this imbalance.  First, a chloride concentration
below which sites  could be considered "unaffected" by any unusual sources was  investigated.  A
concentration of 50 /feq  L"1  was chosen following an examination of the data.  This concentration was
at the upper end of the range  in concentrations found  in "undisturbed"  (see below)  lakes.  Some
"disturbed" lakes had concentrations  below 50  fieq L"1,  but just because a  lake was classified  as
disturbed by our criteria does  not  mean that it actually was unusually or adversely  affected.   Rather,
disturbed simply implies that the site has the potential for  being affected because of the level of human
activity associated with its watershed.  "Unaffected" systems  were designated  as Class A (Figure 10-3),
to which 80 sites were assigned.

      Disturbance was based on a number of factors including location of roads in the watershed (e.g.,
proximity to  lakes,  streams, etc.),  as well  as the occurrence of  mines,  waste disposal  sites, urban
industrial  sites, or residential areas.  If a lake had chloride > 50 neq L"1 but was undisturbed, then its
distance from the coast was examined.  A distance of 50 km from the coast was selected as a cutoff
based on (1) plots of chloride vs. distance from the coast for undisturbed  sites,  (2) deposition  data (A.
Olsen, personal  communication),  and  (3) sea salt effects  relative to distance (R.  Dennis,  personal
communication).  Four undisturbed sites had chloride > 50 jueq L"1 but were within 50 km of the coast.
These sites were classified as Class B lakes with sea salt  influences only.  The chloride inputs  at these
sites were treated as sea salt inputs distributed uniformly across the watersheds.  The  anion inputs were
balanced  by cations  consistent with sea salt composition.  The occurrence  of undisturbed sites with
chloride  > 50 /ieq L"1 at distances greater than 50 km from the coast was  examined, and no such  sites
were  found in the DDRP dataset, resulting  in Class C.

      Disturbed sites then were examined relative to distance from the coast. Disturbed sites exceeding
distances of 50 km from the coast are probably  influenced only  by road salt  practices.  Sites  close to
                                               637

-------
                                          Cl STATUS
[Cl]>50
neq L'1
?
IYES
NO^

1 SUtUUDl
^^ v. % w> ™™ ™ „•. ™s ?x«».-x-« ijf

\/"" Not ^-^ o^O
" " "t --^i "r' «----x '5^A\
80
i^«-- ,,SX
     Disturbed
         YES
      Distance
      >50 km
   V-w
         YES
       Na:CI
       <0.8
   _
        YES
Roadsalt
influence
only

24

         Cl
      balanced,
-§  Distance
   > 50 km
Roadsalt
&
seasalt
influence

37

^   Na:CI
;   -i-;«w
^ balanced ?? ^ x ^ :r«
^byseasalt^ ^-^^-^JF,
^ ^ Sh-


221
                                            3
NO _

Seasalt " " % '^
•• f™» . „ , ' "•• %•.-,•,•. •
«™ s influence *v " " ."^
•• " only -> v5%°'
- :
'4fj
' NO
i
Cl
balanced ^ ,
;;byseasalts% ;
^ r\ •
Di
J36J
" N ss :
Figure 10-3. Decision tree used to identify watersheds with net chloride export and procedures for
determining chloride imbalance.
                                    638

-------
the coast might show both road salt and sea salt effects. Thirty-seven sites fell into the latter category
and 24 into the former.  For those sites apparently affected by road salt only, point source inputs should
be assumed. For the other sites, the inputs might be both "broadcast" and point source, but no method
for discrimination between these possibilities was available. Therefore, these sites were treated as having
point source inputs  of  salts directly to the  lake  rather than input sources spread evenly across the
watershed.

      The last factor to be examined  was the composition of the  salts.  The molar ratio of sodium
chloride in seawater  is 0.864 (Harvey, 1969).  Given the measurement uncertainty, a ratio less than 0.8
was used to screen lakes.  Only three of the remaining sites fell into this category, with 0.72 the lowest
ratio observed.  These low ratios might be due to uncertainty and, because of the difficulty in developing
a procedure for deciding which ions to use to balance the chloride, these sites also were balanced by
base cations of sea salt composition "added" directly to the lake.

10.6  GENERAL APPROACH

      The following general approach was used  in performing the long-term projections of future change
in surface water chemistry by each  of the modelling groups:

      •     Model calibration
            Sensitivity analyses
            Regional projection refinement
            Future projections
      •     Uncertainty analyses
      •     Regional population estimates

This approach, illustrated  in Figure  10-4,  is fully consistent with  the recommendations made by the
Environmental Engineering Committee of the EPA Science Advisory Board (EPA-SAB) on the use of
mathematical models by EPA for regulatory assessment and decision-making (EPA-SAB, 1988).

      All three models were calibrated to three watersheds in the NE - Woods Lake, Panther  Lake, and
Clear Pond.   In the SBRP,  MAGIC was calibrated to White Oak Run, VA, and  ILWAS  was calibrated to
Coweeta watershed 36.  These watersheds are discussed in the next section.  All three models performed
sensitivity analyses to determine those parameters and inputs to which the models were most sensitive.
Particular attention was  given to these parameters and inputs during model calibration  in preparation for
the long-term projections.   Sensitivity analyses are discussed in Section 10.9.  Intensive site calibration
and sensitivity analyses were conducted to document model behavior, to demonstrate that these models
can  predict  short-term watershed responses, and to identify areas for improvement in calibration and
projection procedures for the regional sets of watersheds. These refinements  are discussed in Section
10.10.

      The general approach for long-term projections followed by each of  the  modelling  groups  is
illustrated schematically in Figure 10-5.  The models were calibrated to  each of the DDRP  watersheds,
or some subset, in the Northeast and SBRP.  Long-term projections (i.e., for 50 years) were performed

                                              639

-------
           DYNAMIC MODELLING METHODOLOGY
Model Calibration
- Intensively studied
watersheds
- NE & SBRP
^
r
Sensitivity Analysis
- ETD
- ILWAS
- MAGIC
^
r
Refine Calibration/
Projection Approach
for DDRP Watersheds
- ETD
- MAGIC
i
r
Model Forecasts
Northeast SBRP
- Projection - Projection
- Uncertainty - Uncertainty
Estimate Estimate
i
r
Regional Population
Estimates
- Northeast
- SBRP
Figure 10-4. Approach used in performing long-term projections of future changes in surface water
chemistry.
                                 640

-------
                  Schematic Modelling Approach
  Measured
    Data
                    Data Preparation
Aggregation
  Function
   Population Mean
  Population Median
  Population Variance
     Uncertainty
Population Estimation
 Lumped Watershed
   Chatacteristics,
Soil Data, & Chemistry
                                            t
                                      Time Dependent
                                          Data
             Model Projections
                                                      o
                                                      D.
                                                      
-------
on the individual watersheds and the results presented as population estimates for the NE or SBRP.  The
population estimates were generated as indicated in Section 6.  Results from individual watersheds were
of interest only with respect to their representation of the target population.  Uncertainty analyses,
described in Section 10.11, were incorporated in the confidence intervals about the regional population
estimates. The following sections discuss each of these general topics in greater detail.

10.7  MODEL CALIBRATION

10.7.1 Special Interest Watersheds

      Three intensively studied watersheds were selected for model calibration in both the NE and SBRP.
Selecting multiple intensively studied watersheds for calibration was important for the following reasons:

      •     Watershed characteristics and parameter values vary from watershed to watershed, and thus
           a range of values can be simulated. For example, watersheds can be selected with varying
           combinations of sulfate adsorption, percent base saturation, depth of till and other watershed
           and lake attributes.
                                        \
           The relationship among  model  parameters and measured parameters can be examined
           because extensive information on watershed processes, watershed characteristics, and system
           responses is available and an intensive time-series record exists.

      •     The short-term behavior of the models on a variety of systems can be  evaluated.
      •    Comparable  results among the models simulating  a varying  combination of watershed
           processes and responses provides greater confidence in using them for long-term projections.

      The datasets for the  six watersheds were  each  subdivided  into  a calibration  dataset  and
confirmation dataset.  The calibration dataset was provided to each modelling group for use in calibrating
the model to the respective watershed.  The confirmation dataset was retained by Oak Ridge National
Laboratory until calibration was complete.  The confirmation dataset consisted only of the  model inputs
and not the lake or stream water chemistry record.  The modelling groups applied the calibrated models
to the confirmation datasets and then compared the predicted output to the observed water chemistry
record. For comparisons among models, calibration and confirmation root mean  square errors (RMSE)
were  computed for the following output variables:
              Instantaneous flow
              Cumulative flow
              Chloride
              Sulfate
              Gran alkalinity
              Calcium
              Magnesium
(m3 s'1)
(m yr'1)
     L'1)
                                              642

-------
             Potassium
             Total aluminum
             PH
10.7.1.1  Northeast

     The three northeastern intensively studied watersheds are Woods Lake, NY, Panther Lake, NY, and
Clear Pond, NY. Woods and Panther Lakes were EPRI ILWAS research sites (Chen et al., 1983; Goldstein
et al., 1984; Gherini et al., 1985). Clear Pond was an EPRI RILWAS site. All basin and lake morphometry,
soil chemistry, mineralogy, and hydrology data were obtained from EPRI  (Valentin! and Gherini, 1987; R.
Goldstein, personal communication).   Water chemistry data were collected approximately weekly during
the study periods. All three watersheds also were sampled during the DDRP Soil Survey, and these data
were provided to the modelling groups. The calibration and confirmation periods for these three sites
were
              Site
           Woods Lake
           Panther Lake
           Clear Pond
 Calibration
9/78 - 8/80
8/78 - 8/80
7/82 - 7/84
Confirmation
9/80 - 8/81
9/80 - 8/81
10.7.1.2  Southern Blue Ridge Province

     The three intensively monitored stream watershed sites in the SBRP are Coweeta watershed 34, NC,
Coweeta watershed 36,  NC,  and White  Oak Run,  VA.   Watershed  and stream morphometry, soil
chemistry, water chemistry, and historical site information were obtained for the Coweeta sites from the
USDA  Forest Service's  Coweeta  Hydrological  Laboratory (W.  Swank and  J.  Waide,  personal
communication) and for White Oak Run from B.  Cosby and  G. Hornberger (personal communication).
Water chemistry samples were collected approximately weekly during the study period.  These sites also
were sampled during the DDRP Soil Survey and these data provided to the modelling groups.   The
calibration and confirmation periods for these three sites were
                 Site
Calibration
6/82 - 5/86
6/73 - 5/79
1/80 - 12/82
Confirmation
6/73 - 5/82
6/79 - 5/86
1/83 - 12/84
                WS 34
                WS36
                WOR
     The period of record at the Coweeta sites permitted partitioning the datasets for the purposes of
both projecting and hindcasting.   Because  of time constraints,  the  Coweeta watersheds  were not
simulated. The ILWAS and MAGIC models will be calibrated on the Coweeta watersheds and the results
presented as part of the DDRP Mid-Appalachian report in mid-1990.  MAGIC was calibrated for White Oak
Run using data collected for period January 1980 to December 1984.  The MAGIC model was developed
using data from White Oak Run and Deep Run, VA (Cosby et al., 1985a).
                                             643

-------
10.7.2  General Calibration Approach

      The general approach for model  calibration was, first,  to  calibrate the hydrologic  submodel or
companion  model to the discharge records;  next, calibrate the chemical submodel or model to a
conservative constituent such as chloride; and finally to calibrate the watershed model to the suite of lake
or chemical concentrations simulated  by the model.   Calibration was an interactive  process.   The
hydrologic submodel can route flow through various soil horizons and still predict the observed stream
hydrograph  or lake discharge.   Calibration to a conservative constituent provides confidence that mass
balance is maintained in the models and also provides confidence in, and constraints for, the hydrologic
calibration.  If evapotranspiration, overland flow, subsurface flow, or other components of the hydrologic
budget were not properly calibrated, a flow balance might be achieved, but it is unlikely that the model
outputs would match the observed conservative constituent concentrations.  Calibrating the model to
additional nonconservative chemical constituents, such as anions (other than chloride) and cations, further
constrains the hydrologic flowpaths through various soil horizons, because the physical and chemical
attributes of the soils restrict the range of parameters for each compartment.  If the observed stream or
lake concentrations could hot  be  predicted within these ranges, the hydrologic  calibration was revised
to provide additional  flow through different soil  horizons or  along different flowpaths to achieve the
observed  water  chemistry  concentrations.   Calibration   of the  models  to  predict the  observed
concentrations of multiple constituents provided relatively restrictive constraints on calibration parameters.
Variables that could be calculated from  measured soil or lake attributes were incorporated directly into
the model without modification.  These variables included  constituents such  as soil cation exchange
capacity, exchangeable fractions of base cations,  base saturation, porosity, and lake hydraulic residence
time.  The sampling and measurement errors in  these variables were included in the uncertainty analyses.
The  mass-weighted mean or median values of the aggregated  variables were  used in the calibration
process.

10.7.3  Calibration of the Enhanced Trickle Down Model

      The ETD model represented the watershed  horizontally as a  homogeneous  catchment with no
subcatchments (Figure  10-6) and  vertically with  a snow compartment, soils,  unsaturated zone, and
saturated or groundwater compartment (Figure 10-7). Watershed data for ETD were lumped or aggregated
to provide average or weighted average values  for each of the soil  layers.  In the NE, the top soil
compartment represented the mass-weighted average conditions of the O, A, and upper B  horizons, the
unsaturated  zone was represented by mass-weighted averages of the lower B and upper C horizons, and
the groundwater compartment  was represented by mass-weighted averages for the lower C horizons.

      The processes represented  in the ETD model are shown in Table 10-1.   ETD calibrations were
achieved by decoupling the hydrologic, chloride, sulfate, and ANC-weathering submodels.  The hydrologic
calibrations were conducted first.  Next  the chloride submodel was calibrated.   Chloride was assumed
to be conservative and  was used  to evaluate  the hydrologic calibration.   Calibration was an iterative
process because of the coupling between hydrology and chemistry. The sulfate submodel was calibrated
next and the final calibration involved the ANC-weathering submodel.   The calibration of each submodel
was achieved by using a standardized optimization package, (DESIGN (Arora et  al.,  1985)  coupled with
ETD, and a trial-and-error procedure. The range  of parameters for calibration was input to IDESIGN.
                                              644

-------
                                     Woods Lake
               1000     0
                 I  I I  1 1  I
        1000    2000    3000 Feet
          i   	t	i
0
I   I
0.5
                                                 1 Kilometer
                                                        Approximate mean

                                                         declination 1979


Figure 10-6.  Representation of horizontal segmentation of Woods Lake, NY, watershed for MAGIC

and ETD.
                                         645

-------
            Prototype
Ocm
Model
      -''V,.  \ Litter- \os
   iiiiiiiiii-i!: B2hir'iii:i:;i;iii;
                                                                     Layer 1
             Upper
               till
               (C)
                                                                     Layer 2
             Lower
               till
               (C)
                          Layer 3
Figure 10-7. Representation of vertical layers of Woods Lake Basin for ETD.
                                           646

-------
A complete two-year simulation was performed at each iteration and a cost or penalty function evaluated.
1DESIGN  performs minimization  of  the cost function  using  the Fletcher-Reeves  algorithm  (gradient
method).   A priori bounds on physical quantities and  parameters were included as constraints in the
optimization process.  In all gradient methods when the objective function is of the least-squares type,
it is assumed  that the residuals  are  homoscedastic,  independent, and sufficiently small  to  assume
normality.  For the  DDRP simulations, the residuals were  not homoscedastic, which  resulted in the
IDESIGN optimizations being biased toward the extreme values of the residuals.  The  IDESIGN does,
however, bring the parameters  within reasonable range  of their optimal values.  Following the  IDESIGN
simulations, therefore, a trial-and-error method was employed to achieve optimal calibration. There were
three main guidelines used for  trial and error calibration:

      (1)   Obtain closure of cumulative flow or mass during  the entire calibration period.
      (2)   Capture the seasonal variability of the  state variables.
      (3)   Capture the peaks  and valleys of the daily flows and concentrations.

This calibration process was followed for each of the submodels.  The parameters for the  previous
submodel calibration were fixed during calibration of the succeeding submodel.  In some instances, this
was an iterative process. Calibration  of the sulfate and ANC submodels might indicate that the flowpaths
through the watershed would have to  be changed  to  match observed  water  chemistry and  maintain
parameter values within a reasonable range for the soils on that watershed. The hydrology submodel,
therefore,  would  be recalibrated  and the process  repeated.  Calibration was  an iterative  process of
constraining parameters to achieve an  optimal  calibration.  Additional  information  on the ETD model
calibration has  been presented  by Nikolaidis et al.  (1987).

10.7.4 Calibration of the Integrated Lake-Watershed Acidification Model

      In the ILWAS model the watershed was partitioned into a series of subcatchments to represent the
horizontal variation in the watershed (Figure 10-8) and a series of vertical layers to represent various soil
horizons (Figure 10-9).  Basin data are used quantitatively to characterize the system to be simulated and
delineate the  appropriate number of subcatchments.

      The ILWAS model requires specification of over 200 parameters, coefficients, and initial conditions
for model  calibration to represent the processes listed in Table  10-1. These values can be classified into
three  groups:   constants, measured values, and calibration parameters.   Constant  values include
thermodynamic constants or other factors that do  not  vary from watershed to watershed.  Measured
values included watershed area, base saturation, lake volume, and other attributes that were either directly
measured or calculated from  measured  data  at  a  specific  site  but  were not  varied  during  model
calibration. The third set of values was calibration parameters such as mineral weathering  rates, hydraulic
conductivity, nitrification rates, and other parameters that are not well known and  were modified during
calibration to match the observed watershed and lake constituent concentrations.  The general  rules for
calibration were:   calibrate the  system's hydrologic behavior before calibrating the  chemical behavior;
calibrate in the  same order as water flows  through the basin; and calibrate on an annual  basis first, then
seasonally, and finally to the instantaneous (daily) behavior.
                                              647

-------
                                       Woods Lake
                 1000      0
                   I .  i i i  i
1000    2000    3000 Feet
                        0           0.5           1 Kilometer
                        i  i  .  .   .  i   ,   i  iii
                                                           Approximate mean

                                                            declination 1979


Figure 10-8.  Representation of horizontal segmentation of Woods Lake Basin for ILWAS.
                                         648

-------
           Prototype
Ocm
Model
    SSftSfSfStfSfSSftf
        -"-^ organic
              "      "
       ,
-------
      The hydrologic calibration typically involves first  matching the annual cumulative lake/stream
discharge by adjusting the basin evapotranspiration coefficient.  Seasonal flow variations are matched
byvarying the seasonal evapotranspiration coefficient.  Flow through the watershed, both laterally and
vertically,  is adjusted by varying the  hydraulic conductivity to match  the instantaneous  discharge.
Chemical  calibration involves varying canopy, snowpack, and soil parameters to match the observed
surface and groundwater chemical concentrations.

      Although  there was a  significant number of parameters, and, therefore,  significant degrees of
freedom in selecting parameter values, only certain combinations of parameter values resulted in predicted
constituent concentrations that matched observed concentrations.  The interactions among parameters
and parameter combinations placed limitations on the number of feasible parameter combinations.  For
example, increasing the in-Iake nitrification rate coefficient will lead to decreased ammonium, increased
nitrate, decreased ANC, and decreased pH values.  These feedbacks provide a robust set of  constraints
for calibration.

      The calibration exercise involved identifying the set of  parameters that minimized the  differences
between the set of predicted versus observed constituent concentrations.  Additional information on the
ILWAS model calibration  has been presented by Munson et al. (1987).

10.7.5 Calibration of the Model of Acidification of Groundwater in Catchments

      MAGIC represents the horizontal dimension of the watershed  as  a homogeneous unit with no
subcatchments  (Figure 10-6) and the vertical dimension as two soil layers (Figure 10-10).  Watershed data
for MAGIC were lumped  or aggregated to provide average or weighted average values for each of the
soil layers.  The top soil  compartment  represented the mass-weighted average conditions of the A and
B  Master horizons  in both the NE and SBRP.   The lower soil compartment represented  the mass-
weighted average conditions in the C Master horizon in both regions.

      Projecting long-term effects of acidic deposition on surface water chemistry using MAGIC involves
coupling MAGIC with TOPMODEL (Cosby et al., 1985a,b,c).  Both models were  calibrated using an
optimization procedure  that  selected  parameters  so  that the  difference between  the observed and
predicted measurements was minimized.  The calibration exercise was a three-step process.  The first
step was to specify the model inputs such as precipitation, deposition (both wet and dry), an estimate
of historical inputs for the long-term chemical model, and fixed parameters or parameters whose values
correspond directly to (or can be computed directly from) field measurements, e.g., topographic variables
such  as  slope, aspect,  area.  This  approach, in effect, assigns all  of the  uncertainty associated with
sampling, aggregation, and intrinsic variability to the "adjustable"  parameters.   The adjustable  parameters
are those that are calibrated or scaled  to match observed field measurements.

      The second step  was the selection of optimal values for the adjustable  parameters.   These
adjustable parameters were selected using optimization.  The method of Rosenbrock (1960) was used.
Optimal values were determined by minimizing a loss function defined by  the sum of squared errors
between simulated and observed values of system state variables.  Different loss functions were used for
the hydrologic and chemical models.   The hydrologic model used daily stream flow volumes while the
chemical model used weekly lake outflow chemistry or observed soil chemistry.
                                              650

-------
Table 10-7. Comparison  of Calibration/Confirmation RMSE for Woods Lake Among ETD,
ILWAS, and MAGIC Models,  with the Standard Error of the Observations
Constituent3
Inst. Discharge
Chloride
Sulfate
Alkalinity
Calcium
Magnesium
Sodium
Potassium
Tot. Aluminum
Hydrogen

ETD
0.05
5.5
17.5
27.9






Calibration
ILWAS
0.09
3.3
17.3
31.4
6.6
1.8
4.4
1.9
7.9
8.1

MAGIC
0.05
1.9
11.4
17.9
16.6
6.7
6.5
3.2
3.5
1.3
Observed
SE

3.1
16.4
18.6
55.8
56.9
5.82
1.4
16.5
—

ETD
0.07
1.9
10.5
14.7






Confirmation
ILWAS MAGIC
0.07
3.8
16.9
16.4
6.9
2.9
3.2
1.7
6.2
6.9
  All units in /jeq L"1 except instantaneous discharge (m3 s"1) and total aluminum fag L"1).
  ILWAS was calibrated prior to the DDRP using all the data so the dataset could not be split for confirmation.
                                             653

-------
Table 10-8.  Comparison of Calibration/Confirmation RMSE for Panther Lake Among ETD,
ILWAS, and  MAGIC Models, with the Standard Error of the Observations

Constituent3
Inst. Discharge
Chloride
Sulfate
Alkalinity
Calcium
Magnesium
Sodium
Potassium
Tot. Aluminum
Hydrogen

ETD
0.03
5.1
11.3
82.6






Calibration
ILWAS
0.01
4.5
17.6
47.1
36.7
8.8
8.1
2.0
3.2
1.9

MAGIC
0.03
5.6
16.0
87.4
40.4
9.4
9.9
1.7
4.6
3.7
Observed
SE

4.3
14.0
71.0
150.3
154.8
8.7
1.7
11.1
—
Confirmation
ETD
0.04
2.4
11.7
70.0






ILWAS MAGIC
0.05
2.1
15.0
57.7
40.0
.8.4
9.9
2.8
2.1
1.4
  All units in peq l_~1 except instantaneous discharge (m3 s"1) and total aluminum
L'1).
  ILWAS was calibrated prior to the DDRP using all the data so the dataset could not be split for confirmation.
                                              654

-------
Table 10-9. Comparison of Calibration RMSE for Clear
Pond Among ETD, ILWAS, and MAGIC Models, with the
Standard Error of the Observations
Constituent3
                          Calibration
ETD  ILWAS   MAGIC
                         Observed
SE
Inst. Discharge
Chloride.
Sulfate
Alkalinity
Calcium
Magnesium
Sodium
Potassium
Tot. Aluminum
Hydrogen
0.16
2.4
8.9
18.6
23.6
5.0
6.3
1.0
1.1
1.0
0.03
1.4
10.6
17.9
21.1
4.7
5.0
0.7
1.2
0.2
0.15
4.7
9.5
18.6
21.2
4.7



1.6
9.7
18.8




a All units in fjeq L"1 except instantaneous discharge (m3 s"1) and
  total aluminum (fjg L"1)
                                            655

-------
snowmelt (Chen et al., 1983; Gherini et al., 1985). The volume averaging by the models conserves mass
but will result in lower predicted constituent concentrations if this snowmelt moves as a thin lens under
the ice (Gherini et al., 1985).  In the models, the higher/lower concentrations  in  this lens will be mixed
with the rest of the volume in the lake or that layer to compute the constituent concentrations.

      Although these models were (1) developed for different systems in different regions of the United
States,  (2) calibrated  independently  using  model-specific  procedures,  and (3)  run using different
computational time steps (daily versus monthly), the RMSEs for all constituents  are similar. Woods Lake,
a chronically acidic lake (ANC = -10 ;*eq L"1), Clear Pond, with an average annual ANC of approximately
100 A«eq L"1,  and Panther Lake, with an average annual ANC of approximately  150 /^eq L"1, span the
range of DDRP systems of interest in the NE, and the results indicate all three  models can predict acid-
base water chemistry with precision similar to that observed in the measured data for short-term  periods
of record. These results do not, however, necessarily ensure calibration for long-term projections. Long-
term calibrations can be achieved only with long-term data (Simons and  Lam,  1980).

      Comparisons  of  the time series of predicted  versus observed values for each of the model
applications to Woods Lake, Panther Lake,  and Clear Pond are described and shown  graphically  in
Appendix A.1.  The calibration and  confirmation exercise indicated the three models produced comparable
results for three watersheds with a range of watershed characteristics, from deep to shallow till depth,
and lake chemistry, from ANC concentrations of -40 to over 200  peq L"1.  Although, there is variability
for individual  daily values, the  models reproduce the flow-weighted average annual constituent values.
Average annual estimates, and the change in these estimates, represent the focus of the DDRP.

      This  calibration/confirmation exercise  and  calculation  of  RMSEs  is  consistent  with the
recommendations of the Environmental Engineering Committee of the Science Advisory Board for model
confirmation with field  data (EPA-SAB, 1988).   The  next step recommended for using environmental
models was to conduct sensitivity analyses (EPA-SAB, 1988).

10.8  MODEL SENSITIVITY ANALYSES

      Sensitivity analysis is a formalized procedure to identify the impact of changes in various model
components on model output.  Sensitivity analysis is an integral part of simulation experiments and model
applications.   Models  represent  aggregations  and simplifications  of  watershed  and soil  processes,
including physical, chemical, and  biological  processes.  Parsimony is introduced to represent  these
multiple processes by a single (or few) aggregated process(es) and a transfer coefficient or parameter.
Sensitivity analysis is an approach used to determine  if model output or system response is sensitive  or
responsive to small changes in these transfer coefficients.  Those parameters for which the model output
was sensitive received greater attention during  model calibration for long-term  projections.

      Sensitivity analyses were performed on the three intensively studied watersheds  in the NE.
Examining the sensitivity of model output over a short simulation period (e.g., 3-5 years) provides useful
information about model behavior;  however,  these  short-term analyses  might  not reveal  the full
dependence of the simulated  system response on these parameter values.  Certain parameters, for
example, might have little effect on  short-term model behavior but  might  be critical  for long-term
                                              656

-------
projections. This caveat must be considered in evaluating the model sensitivity analyses and the long-
term projections.

10.8.1  General Approach

      Sensitivity analyses were performed using the Woods and Panther Lakes and Clear Pond datasets
following the model calibration and confirmation exercises for the MAGIC model and those for ETD on
Woods and Panther Lakes.  The ETD sensitivity analyses were conducted as part  of a Ph.D. Thesis
(Nikolaidis, 1987).  The  Clear Pond dataset was  not available in time to  be included  in  this Thesis.
Sensitivity analyses were performed for  ILWAS prior to the  initiation of the  DDRP.  These qualitative
analyses for ILWAS are included in  Appendix A.1.  The classical approach of Tomovic (1963) was used
in performing sensitivity analyses on each of the models.  Each coefficient or parameter was individually
varied by ±10  percent with all other coefficients retaining their original, calibrated values.   The relative
change in model output RMSE for different variables or model components was noted to determine their
sensitivity to this parameter  change. If the increase in the RMSE was large, the model was considered
sensitive to this parameter.  RMSEs  permitted  a quantitative  estimate of  the  increase  in  variance
associated with each parameter.

      The optimization procedures used with ETD and MAGIC also indicated relative parameter sensitivity.
The response surface around an optimum parameter value was evaluated to determine if the  surface was
relatively flat  or steep.  A relatively flat  response surface  indicated several  parameter values could  be
selected without influencing the optimum system response.  A steep surface, however, indicated that small
changes in the  parameter value would affect the optimum system response or that the system response
would  be sensitive to that parameter.

      All three  modelling groups selected coefficients or parameters for processes expected to control
or strongly influence both hydrologic and water chemistry output variables, including hydrologic routing
parameters, sulfate adsorption, ion exchange, and  weathering  rate parameters.  Selection  of these
parameters was based on previous analyses and published results by each modelling group (Cosby et
al., 1985a,b,c; Gherini et al., 1985;  Nikolaidis et al.,  1988; Lee  et  al., in press; Georgakakos  et al., in
press).  The specific parameters evaluated are listed in Table 10-10.  This evaluation provided  an estimate
of the variability introduced by the parameter in the system response, and, therefore, an indication of the
range over which the parameter could vary without significantly altering the model output.

10.8.2  Sensitivity Results

      The parameters selected for sensitivity analyses are ranked in priority order from most sensitive to
least sensitive in Table 10-10. The effects of a ± 10 percent change in these parameters on the RMSE
for predicted  lake ANC concentrations also are listed with these parameters.

      Parameters related  to weathering and hydrologic transport processes,  in general, were sensitive in
each of the three models.   Parameters  related to  sulfate adsorption and bulk soil processes  such as
cation  exchange capacity were not particularly sensitive in any of the models.  Sulfate adsorption would
not be expected to be an important  process in the NE if lakes are near sulfate steady state.  The models
                                              657

-------
Table 10-10.  Percent Change in RMSE for MAGIC and ETD for a Ten Percent Change in Parameter
Values.  Parameters are Ranked in Descending Order of Sensitivity from  Left to Right
Factor
             Weathering       Capacity
                                           MAGIC
                                                       2-
SO4  Adsorp.     Hydrol.
Ion Exchange
Parameter41
Woods Lake
Alkalinity
Hydrogen ion
Panther Lake
Alkalinity
Hydrogen ion
Clear Pond
Alkalinity
Hydrogen ion

Factor
Weath.+ Weath

-2.1
0.5

0.0
0.0

4.3
0.0


2.1
-0.6

2.6
0.0

4.9
0.0

Weathering
.- Depth + Depth- EMax+

-1.0
0.0

0.0
0.0

0.1
0.0


2.1
-0.6

0.2
0.0

0.2
0.0

Ion Exchange

-1.0
0.0

0.0
0.0

-0.2
0.0
ETD
Snowmelt
EMax-

2.1
-0.6

0.2
0.0

0.1
0.0

Rate
PMAC+ PMAC-

1.1
-0.6

0.0
0.0

-0.1
0.0


-1.0
0.0

-0.1
0.0

-0.1
0.0

Lat./Vert.
Selec+ Selec-

0.0
0.0

0.0
0.0

0.0
0.0


1.1
0.0

0.1
0.0

0.0
0.0

Hydraul.Cond.
Parameter3    KH5+  KH5 -     RE+   RE-  KAPPA +  KAPPA-   KLAT3+  KLAT3-  KPERC3  +KPERC-
Woods Lake
Alkalinity         -7.0   9.4      -3.1   3.7       4.7   -3.8
Panther Lake
Alkalinity          4.6   3.9
                               6.2   -2.6     -3.8    3.1
                  3.1    2.2


                 -1.0   -0.9
  -2.5    8.9


   3.5   -2.2
 MAGIC Parameters
    Weath
    Depth
    EMax
    PMAC
    Selec
         Weathering rate for base cations (meq m"2 yr"1)
       = Estimated average depth to bedrock of the watersheds (m)
       = Maximum sulfate adsorption capacity (meq kg  )
       = Unsaturated zone channeling parameter
       = Specific base cation (e.g., Ca) to aluminum selectivity coefficient
  ETD Parameters
    KH5    = Hydrolysis rate constraint for water body (eq m"2 d)
             Ion exchange reaction rate coefficient (m /eq  d )
RE
KAPPA  = Snow melt rate (in d'1 °C'1)
KLAT3  = Lateral flow recession constraint for the soil compartment (I d"1)
KPERC3 = Vertical hydraulic conductivity for soil (md~1)
                                                 658

-------
all appeared to be robust to small changes in bulk soil properties, which also can be measured directly
in the field.

      The specific parameters that were sensitive for each model  differ because of different process
formulations among the models.  For  example,  a  10 percent change in weathering rates for MAGIC
resulted in a 2 to 5 percent change in the RMSE for predicted average annual ANC concentration.  A 10
percent change  in weathering parameters in  ETD resulted in a 4 to 9 percent change in the RMSE for
predicted average annual  ANC.  A 10  percent change in ILWAS weathering parameters resulted in a
minimal change  in the RMSE for predicted average annual ANC because the change was compensated
for by ion exchange in these short-term simulations (see Appendix A).

      Weathering rate parameters (which are calibration parameters) were not, and generally are not,
measured in the field.  These weathering rate parameters, however, are not completely unconstrained.
Weathering rates are constrained, in part, by cation-anion balances and ratios in surface waters and by
ranges observed in the literature for watersheds with similar geology, mineralogy, and soil and water
chemistry.

      Hydrologic parameters also are  constrained during  calibration.   Maintaining  mass  balance for
conservative constituents constrains evapotranspiration and runoff processes.  Calibration of the sulfate
adsorption and ion exchange submodels constrains lateral and  vertical hydraulic conductivity parameters
and, therefore, flowpaths through the watershed.

      While there are similarities in several sensitive parameters among models, process formulations are
different among  the three  models.  This is evident by the different parameters to which the  predicted
output is sensitive among  models.  There is not a 1:1  mapping between parameters  and processes for
any  of the models.  Mass balance, electroneutrality, and other requirements,  however, constrain the
parameter values for all models including  sensitive parameters. Similarities in calibration/confirmation
RMSEs  indicate parameter values for  all the processes can  be constrained by  watershed  and lake
attributes to achieve calibration within the range of observed values.

10.9  REGIONAL PROJECTIONS REFINEMENT

      The intensively studied watershed calibration/confirmation and sensitivity analysis exercises were
conducted to  evaluate model  behavior and  the calibration procedures.   These  exercises resulted in
improvements  and refinements in the calibration procedures used in the long-term DDRP  projections.
These refinements are discussed below.

10.9.1   Enhanced Trickle Down

      Calibration of the northeastern DDRP watersheds using ETD followed a similar procedure as that
used for the three intensively studied watersheds.  The optimization program, IDESIGN, was used initially
to optimize parameters followed by trial-and-error procedures. For the DDRP watersheds, the hydrologic
optimization focused  on  the  evaporation rate  for the lake  and  the lateral and  vertical  hydraulic
conductivities for the soil compartments.  Sensitivity analyses indicated the model output was sensitive
to these three hydrologic parameters. The watershed/soil physical and chemical attributes and lake water
chemistry were used to estimate values for the other  hydrologic parameters for each DDRP watershed,

                                              659

-------
chemistry were used to estimate values for the other hydrologic parameters for each DDRP watershed,
which were fixed during calibration. All of the chernistry parameters were optimized and calibrated for
each watershed.   ETD used the aggregated  soil  data discussed in Section  8.8.3 to determine the
watershed hydrologic and  biogeochemical parameter values (e.g.,  cation exchange capacity, sulfate
adsorption, base saturation) used in calibration.

     Initial conditions for each watershed simulation were set at the recalculated ELS-I ANC values and
the 1984 ELS - I value for other appropriate chemistry variables.  Subsequent comparisons of calibrated
versus  observed values in 1984 for ETD, therefore,  will be identical.

10.9.2   Integrated Lake-Watershed Acidification Study

     The ILWAS calibration procedure for the DDRP watersheds was similar to the calibration procedure
used for the  intensively studied watersheds.   The  primary  difference was  a reduced number  of
subcatchments for each of the DDRP watersheds. In general, only one or two subcatchments were used
to represent the DDRP watersheds. Individual soil pedon data, instead of aggregated soils  data, were
used to calibrate the soil parameters, similar to the procedure used  in calibrating these parameters on
the intensively studied watersheds.

10.9.3   Model  of Acidification  of Groundwater in Catchments

     The MAGIC calibration sequence was similar to that used for the intensively studied watersheds but
the procedure was refined and automated, where possible, for the DDRP watersheds. First, TOPMODEL
was calibrated using daily rainfall and monthly runoff to derive flow routing parameters for the two-layer
structure of MAGIC.   Next, MAGIC  was calibrated using annual  time  steps to simulate  average
volume-weighted lake chemical concentrations for comparison with the 1984 index lake chemistry values.
No calibration was attempted for chloride because  it was assumed to be a conservative ion  (except for
those northeastern watersheds in Priority Classes F  -1 where the chloride balance was completed by sea
salt correction). Sulfate was not calibrated in the  NE. The aggregated sulfate adsorption parameters
computed during  aggregation  of the soils  data were used directly in  MAGIC for the northeastern
watersheds.  Sulfate adsorption  parameters, however, were calibrated in the SBRP. The aggregated half-
saturation constant for sulfate adsorption was scaled by a constant factor for all catchments in the SBRP.
In the NE, the Baker et al.  (1986b) model and  coefficients for in-lake sulfate reduction were used.  This
model  computes  sulfate reduction, in  part,  based on theoretical hydraulic residence times in the lake.
No sulfate reduction was used in SBRP stream projections.

     Finally, base cation concentrations were calibrated using an optimization procedure based on the
Rosenbrock (1960) algorithm. The base cation calibration involved fitting the  results of long-term model
simulations to currently observed water and  soil base cation data  (i.e., target variables).   The target
variables were both  soil   exchangeable  fractions  (for  both  soil  compartments)  and  lake  index
concentrations of calcium, magnesium, sodium, and potassium.  The target variables comprised a vector
of measured values, all of which must be reproduced by the model for a successful calibration.  The use
of multiple,  simultaneous targets in an optimization procedure provided robust constraints on  model
calibration (Cosby et al., 1986a).
                                              660

-------
     Those physicochemical soil and surface water attributes measured in the field in the DDRP Soil
Surveys were considered "fixed" parameters in the model, and the measurements were used directly in
the models during the calibration procedure.  The maximum sulfate adsorption capacity and sulfate half-
saturation coefficient,  determined  for individual horizons and aggregated to the watershed, were used
directly in the model and were not calibrated. Base cation weathering rates and base cation exchange
selectivity coefficients for the soils were not directly measured and were used as "adjustable" or optimized
parameters in the calibration process.  The calibrations were performed on simulations run from 1844 to
1984 for the NE and 1845 to 1985 in the SBRP. The historical deposition sequence over this period was
estimated by scaling the present-day deposition provided in the  DDRP database to a reconstruction of
sulfur emissions for the NE or Southeast (OTA,  1984).  This scaling procedure has  been described  by
Cosby et al. (1985b).  After each simulation, the 1984 and 1985 simulated versus observed values were
compared; the adjustable parameters were modified as necessary to improve the relationship between
simulated and observed values; the simulation was re-run and the procedure repeated until no further
improvements in these relationships were achieved.

10.9.4 DDRP Watershed Calibrations

      The three models can be compared for the target population of lakes with ANC < 100 /j,eq L" ,
which corresponds to 495 northeastern lakes. The comparisons for ANC and sulfate for the three models
are shown as population histograms  in  Figures 10-11 and 10-12, respectively.  Estimated aluminum
concentrations were added to estimated MAGIC alkalinity concentrations so the MAGIC ANC projections
are consistent with  the ILWAS ANC estimates.  The ILWAS and  MAGIC models are calibrated on  base
cations and acid anions so ANC is a computed,  not a calibrated  value (e.g., ANC =  sum base cations -
sum acid anions).   The ETD histograms are  not discussed here  because ETD assumed the 1984 ELS-I
lake index chemistry value was the calibrated value for initiating the long-term forecasts.  Comparison
of the histograms for the calibrated 1984 ETD value versus the ELS-I value, therefore,  are nearly identical.
The discrepancies between the ELS-I distributions and the ETD values represent  lakes that were not
simulated by ETD in Classes A - E.

10.9.4.1  Integrated Lake-Watershed Acidification Study

      ILWAS was not calibrated on the very acidic lakes (i.e., ANC < -30 peq L"1) but generally matched
the ANC of moderately acidic lakes (ANC ~ -15 peq L"1) (Figure 10-11). In general, the calibrated ANC
for the low ANC lakes (0 < ANC < 75 peq L"1) was greater than the observed as evidenced by the
larger  number of calibrated lakes with higher ANC than observed in the DDRP Priority Class A-B target
population (Figure 10-11).

      Calibrated sulfate concentrations generally were underestimated for lakes with observed sulfate
concentrations less than 75 //eq  L"1 and overestimated for lakes with observed sulfate concentrations
between 75 and 125 neq L"1 (Figure 10-12).  Calibrated sulfate values were overestimated for lakes with
 sulfate concentrations greater than 125 jueq L"
                                              661

-------
    200i
  81501
    100-
  E
     50-
                Northeast Lakes
               Priority Class A - B
                 Model = Magic
             Deposition = Constant
    2001
     501
          -40
-15
10      35       60
          ANC(ueqL'1)
                Northeast Lakes
               Priority Class A - B
                 Model = ILWAS
             Deposition = Constant
110
135
                                                                              160
          -40      -15       10       35       60
                                      ANC(neqL-1)
                                  85
                                  110
         135
         160
Figure 10-11. Comparison of population histograms for simulated versus observed (Eastern Lake
Survey Phase I 1984 values) ANC for ILWAS and MAGIC.  ETD used the  ELS-I values as initial
model conditions, so the simulated values are nearly identical to the observed values.
                                        662

-------
  200-
J3150
CO
|5100
E

   50-
     Northeast Lakes
    Priority Class A - B
      Model = Magic
  Deposition = Constant
      30 40 50 60 70 80 90100110120130140150160170180190200210220230240250260270
  2001
§150-
5
$1001
1

^  50H
     Northeast Lakes
    Priority Class A - B
      Model = ILWAS
  Deposition = Constant
I
                                             -0-
      30 40 50 60 70 80 90 100110120130140150160170180190200210220230240250260270
Figure 10-12.  Comparison of population histograms for simulated versus observed (Eastern Lake
Survey - Phase I 1984 values) sulfate concentrations for ILWAS and MAGIC, Priority Classes A and
B.  ETD used the ELS-I values as initial conditions, so the simulated values are nearly identical
to the observed values.
                                         663

-------
10.9.4.2 MAGIC
10.9.4.2.1 Priority Classes A and B -
                                                                         "1
      MAGIC was not calibrated for the very acidic lakes (i.e., ANC < -30 peq L"1) but generally matched
the observed ANC for the moderately acidic lakes (ANC —15 peq L"1)  (Figure 10-11). Calibrated ANC
concentrations, in general, were consistently higher than observed ELS-I ANC concentrations as indicated
by the underestimated number of lakes with lower ANC and overestimates of the number of lakes with
higher ANC (Figure 10-11).

      The low sulfate lakes (e.g., SO42" < 50 /zeq L"1) were not represented in the MAGIC  calibration
(Figure 10-12).  The calibrated and observed sulfate concentrations were comparable for lakes with sulfate
concentrations between 75 and 115 /^eq L"1 (Figure 10-12). The calibrated sulfate concentrations typically
exceeded observed sulfate concentrations for lakes with observed sulfate concentrations greater
150 peq  L"1.

10.9.4.2.2  Priority Classes A - E -

      The calibrated ANC concentrations generally were higher than observed ANC concentrations for
the low ANC lakes (i.e., < 100 ^eq L"1) but were lower than  observed for moderate ANC lakes (i.e., 120-
175 fieq  L"1) (Figure 10-13).  The calibrated ANC concentrations for higher ANC lakes (i.e., > 175
    L"1) were similar to or greater than observed ANC concentrations.

      The calibrated sulfate concentrations were higher than observed for the low sulfate lakes (e.g., < 75
    L"1) and generally lower than observed sulfate concentrations for the moderate sulfate systems (i.e.,
          o             •<
75 <  SO4"  < 135 fieq L" ).  The calibrated sulfate concentrations generally exceeded observed sulfate
concentrations for the high sulfate lakes (Figure 10-13).             '

10.9.4.2.3  Priority Classes A - I -

      The calibrated ANC concentrations generally were higher than observed for the low ANC lakes (i.e.,
<100 /j,eq  L"1)  but similar for the higher ANC lakes  (Figure 10-14).  Calibrated sulfate concentrations
exhibited a varied pattern compared with the observed pattern that, in general, was slightly lower for low
sulfate lakes and slightly  higher for high sulfate lakes (Figure 10-14).

10.9.4.3  Southern Blue Ridge  Province

10.9.4.3.1  Priority Classes A and B -
      Calibrated versus observed ANC and sulfate concentrations are shown for the ILWAS and MAGIC
models in Figure 10-15.  The MAGIC calibrations overestimated the number of low ANC and high ANC
streams and underestimated the number of moderate (i.e., 75 to 125 /zeq L"1)  ANC streams (Figure 10-
15).  ILWAS-calibrated ANC was similar to observed ANC for the low ANC streams (e.g., <75 /zeq L"1)
                                              664

-------
    400-
  $300^
4-
0
1_
0

E

Z
    200-
    100-
                                   Northeast Lakes
                                  Priority Class A - E
                                    Model = Magic
                                Deposition = Constant
        -40 -15  10   35   60   85 110 135 160 185 210 235 260 285 310 335 360 385 410
                                      ANC(neqL-1)
    400i
    300
    2001

    100-
           nil
         30  40 50  60  70 80  90100110120130140150160170180190200210220230240250
Figure 10-13. Comparison of population histograms for simulated versus observed (Eastern Lake
Survey Phase I  1984 values) ANC and sulfate concentrations for MAGIC, Priority Classes A - E.
                                        665

-------
                                 Northeast Lakes
                                Priority Class A -1
                                  Model = Magic
                              Deposition = Constant
       -40  -15  10  35  60  85  110  135  160  185  210  235  260  285  310  335  360  385  410
                                    ANC(|ieqL-1)
(0

1_
Q)
E
z
700-
600-
500-
400-
300-
200-
100-
  0
                                           ID
iri
       30 40 50 60 70  80  90100110120130140150160170180190200210220240260280300
Figure 10-14. Comparison of population histograms for simulated versus observed (Eastern Lake
Survey Phase I 1984 values ) ANC and sulfate concentrations for MAGIC, Priority Classes A - I.
                                        666

-------
                              SBRP Stream Reaches
                                Priority Class A -B
                                  Model = Magic
                              Deposition = Constant
     2001
   w
   E 150
   (8
o 100
<5


Z  50-
      o-1—<-
                            i
                           J
                              w
ri

   I
   m
       -40    -15    10     35
                                    60     85     110

                                    ANCftieqL'1)
                    135    160    185   210
                              SBRP Stream Reaches
                                Priority Class A -B
                                 Model = ILWAS
                              Deposition = Constant
     2001
   (0
   E 150
   a
   £
   4->
   (0

   o 100 i
   &-
   0>
   XI
   E

   i  501
                                  i.
                                  I
         I
                                                       •
          -40    -15    10    35     60     85    110    135   160    185    210
                                    ANCOxeqL-1)

Figure 10-15.  Comparison  of population histograms for simulated versus observed  (NSS Pilot
Survey values) ANC,  Priority Classes A and B using ILWAS and MAGIC.
                                         667

-------
but overestimated the number of moderate ANC systems and underestimated the number of high ANC
streams (Figure 10-15).

      MAGIC underestimated the number of low sulfate streams and overestimated the number of higher
sulfate streams (Figure  10-16).   ILWAS also underestimated the  number  of low  sulfate streams and
overestimated the number of higher sulfate streams (Figure 10-16).

10.9.4.3.2  Priority Classes A - E -

      Calibrated versus observed ANC and sulfate concentrations for the MAGIC model in Priority Classes
A - E are shown in Figure 10-17.  MAGIC overestimated the number of low ANC streams but, in general,
the distribution  of calibrated ANC was similar to the observed ANC distribution. The calibrated sulfate
distribution for MAGIC underestimated the number of low sulfate streams (e.g., < 40 /*eq L~1) and slightly
overestimated the number of higher sulfate streams (Figure 10-17).

10.10  MODEL PROJECTIONS

10.10.1 General Approach

      The general approach for performing long-term projections of the effects  of sulfate  deposition on
surface water chemistry  over the next 50 years in the NE and SBRP is illustrated schematically in Figure
10-4.   The two simulated  deposition scenarios were illustrated previously  in Figure 5-25.  In the first
scenario in the NE, the  current deposition  rate at each individual site was  maintained over the  50-year
interval. The models then  projected changes in lake water chemistry over that 50 years.   In the  second
deposition  scenario, current deposition rates at  each site were  held constant for the  first 10 years,
decreased  by a total of 30 percent over the  next 15 years, and then held constant at this  reduced
deposition  rate for the next 25 years.  This  phasing corresponded with one possible scenario  of how
deposition  rates might  change if emissions controls  were implemented in the NE in  year 1 of the
simulation (OPPE,  personal communication).  Deposition rates have been declining in the NE over the
past 15 years and would decline further  if emissions controls were promulgated.  Current deposition rates
at each individual watershed in the SBRP also were simulated, but the alternative  deposition scenario was
an increase in deposition.   For the alternative scenario, deposition  rates were held  constant for the first
10 years,  increased  by  a  total of 20 percent over the next 15 years,  and then held constant at this
increased  rate for  the next 25 years (Figure 5-25).  Deposition rates are expected to increase in the
Southeast (OPPE, personal communication).

      Each model was calibrated on individual watersheds using the data sources indicated in  Section
10.5.  The projected change in surface water chemistry in the individual lake or stream was simulated for
the next 50 years  using the typical year meteorology and deposition data, discussed in Section 10.5.
Output for each model represents flow-weighted annual average constituent concentration.  The projected
ANC is defined similarly and is consistent among all three models.

      Not all watersheds in the NE and SBRP were simulated by  all three modelling groups.   For the
MAGIC model, there were some watersheds  for which the optimization and calibration criteria were not
satisfied.   Long-term projections for these watersheds, therefore, were not performed (Table 10-11). In
                                              668

-------
        4001
      w
      E 300
      
-------
                                  SBRP Stream Reaches
                                    Priority Class A -E
                                     Model = Magic
                                  Deposition = Constant
400-
tf\
E 300-
(8
2
CO
o 200-
Number
CD
CD CD





i i r
1


i
ft









^
SSSSSSSSSJ^SSSSiS^



1







g
i
D PHASE 1
B MAGIC Year 0


n ri i in • .

            -40-15  10  35  60  85110135160185210235260285310335360385410435460485510
         8001
                                  SBRP Stream Reaches
                                    Priority Class A -E
                                     Model = Magic
                                  Deposition = Constant
to
E 600-
(0
CO
o 400-
k.
E
i 200-
n-





Vm







D PHASE!
g MAGIC Year 0

•7JI
1
r— 1
h
	
i
i S%i v*% n
             10  20   30   40  50  60  70  80  90 100 110 120 130  140  150  160  170  180
Figure  10-17.  Comparison of population histograms for simulated versus observed (NSS Pilot
Survey values) ANC and sulfate concentrations, Priority Classes A - E using MAGIC.

                                         670

-------
Table 10-11.  Watersheds, by Priority Class, For Which Calibration Criteria
Were Not Achieved
Priority
Region Class
Northeast A
B


C

D
E
F


G




H



SBRP A

B

C


D
E

ETD
0
0


0

1E2-069
0
NA


NA




NA



NA
NA
NA

NA


NA
NA
Watershed
Model
ILWAS
0
0


NA

NA
NA
NA


NA




NA





2A07821
2A08803
NA


NA
NA
ID
MAGIC
1D2-027
1C1-068
1B3-056
1E1-106
1D3-002
1A2-004
0
1A2-058
1B1-043
1D3-003
1D2-094
1D1-067
1D3-029
1C2-054
1D2-049
1D1-031
1C3-055
1A3-028
1D2-036
1D1-068
2A0781 1
2A07816
0

2A07702
2A07803
2A08801
0
0
                               671

-------
the NE, optimizations criteria generally could not be achieved either because of chloride imbalances or
because cation inputs exceeded  outputs.  Time and funding constraints restricted the number of ETD
simulations  to  northeastern watersheds  in  Priority Classes A  - E (Figure  10-1).   Similar constraints
occurred with the ILWAS model; only watersheds in Priority Classes A and B were simulated in both the
NE and SBRP.  Optimization  and calibration criteria for some  northeastern watersheds also were not
satisfied for the ETD  and ILWAS  models.  These watersheds are listed in Table 10-11.  The individual
watersheds  simulated by each model in  each region are presented in Appendix A.2 and are listed by
NSWS lake  or stream identification number, name (if available), state, latitude and  longitude, and initial
NSWS ANC. Comparisons among models are made only for similar target populations, and the target
population is clearly identified  for each comparison.

      The results discussed below have been obtained by weighting the individual  watershed estimates
by the appropriate inclusion probability. Weighting by the appropriate inclusion probability is critical for
any analyses performed on these data.  The individual watershed estimates are of  interest only as they
relate to the distribution of the population attributes.

10.10.2  Forecast Uncertainty

      An integral part of all the analyses performed in the DDRP is  the estimate of error associated with
the analyses or projections.  Each modeling group conducted  error analyses on its respective model,
which were  incorporated in the confidence intervals about the population estimates.

10.10.2.1  Watershed Selection

      The computational time required to conduct uncertainty analyses on all simulated DDRP watersheds
would have  been prohibitive. Therefore, six northeastern watersheds were selected from Priority Classes
A  - D.  The  watersheds were sorted by the following criteria:

      •     previously simulated for the EPA Internal Staff Paper
            no  internal sulfur sources
            drainage lakes versus seepage lakes (drainage  lakes preferred)
            percent sulfur retention (positive sulfur retention preferred)
            watershed disturbance as indicated by chloride  balance (undisturbed watersheds preferred)
            ANC class (watershed/lake systems  with ANC < 100 >eq L"1 preferred).

     This sorting  emphasized (1) those systems considered likely to  show  a response, (2)  those for
which early  modelling output  might be available, and (3) those that provided a representative  cross-
section of potential watershed responses.  The six watersheds randomly selected for uncertainty analyses
were

      (1)    one watershed (1A3-048) from Class A - previously  simulated
      (2)    two watersheds (1A2-045, 1E1-111) from Class  B - low ANC, positive sulfur retention
      (3)    two watersheds (1A1-003, 1C2-035) from Class  C -  low ANC, negative  sulfur retention
      (4)    one watershed (1D3-025) from Class D - high ANC, positive sulfur retention
                                              672

-------
Characteristics of these six watersheds follow:
      Watershed ID

      1A3-048
      1A2-045
      1E1-111
      1A1-003
      1C2-035
      1D3-025
Priority Class

    A
    B
    B
    C
    C
    D
     ANC

     14.6
     13.2
     11.0
      -9.9
     73.6
    149.3
       % S Ret.    Soil Type
                      WA   LA     WA:LA
-99.0
4.0
12.2
-36.7
-15.6
14.6
Spodosols
Spodosols
Mixed
Mixed
Spodosols
—
228
168
80
96
215
57
54
26
24
13
11
8
4.7
6.4
3.4
7.5
19.0
7.4
These watersheds include a lake with negative ANC, systems with low ANC and relatively high ANC (i.e.,
74 and  150 /*eq L"1), watersheds in four of the five subregions, a distribution of watersheds across the
deposition  gradient, and watersheds selected from clusters representing the majority of watersheds with
ANC  <100 /*eq  L"1.   Uncertainty analyses conducted on  these watersheds were assumed  to  be
representative of the other watersheds within these priority classes.
      In the SBRP, five watersheds were randomly selected for uncertainty analyses.  The watersheds
were sorted in priority order, as illustrated in Figure 10-2.  Two watersheds were selected from Priority
Class A:  watersheds with ANC < 100 ^eq L"1 and chloride < 50  i*eq L"1  , and with positive sulfur
retention.  Two watersheds were selected from Priority Class B:  watersheds with ANC >100 n&q L"1 but
<200 /teq L"1 , chloride < 50 fj,eq L"1, and with positive sulfur retention.  One watershed was selected
from Priority Class D:  this watershed had ANC <  200 /*eq L"1 but  chloride  > 50 peq  L"1 ,  indicating
possible watershed disturbance.  This watershed also  had positive sulfur retention.  Characteristics of
these five watersheds follow:
      Watershed ID Priority Class
      2A07828
      2A08802
      2A08810
      2A08811

      2A07830
A
A
B
B
  ANC

  37.0
  71.0
114.4
 95.3

163.0
                       S Ret.   Soil Type
                                      WA
81.1
83.6
77.9
49.0

58.2
Acid crys., high org.
Acid crys., low org.
Acid crys., low org.
Acid crys.,/meta sedmt.,
low org.
Acid crys., low org.
19.1
 5.7
 4.9
 3.3

14.0
Uncertainty  analyses  conducted  on these watersheds were assumed  to be representative of other
watersheds or similar watersheds within these classes.

10.10.2.2 Uncertainty Estimation Approaches

      Three different approaches were used with the individual models to estimate the error associated
with the projections.  The three approaches were first order second moment analyses (ETD), first order
error analyses (ILWAS), and "fuzzy" optimization and multiple simulations (MAGIC).  These approaches
reflect differences in  model formulations, which required different uncertainty estimation  approaches.
Uncertainty estimates  were computed for both input and parameter error. However, parameter error and
input  error were computed separately for each of the models.
                                              673

-------
      Variance estimates were derived from the aggregated soils data and from the deposition monitoring
sites.  Variance algorithms were used with the soils aggregation procedures to obtain variance estimates
for the physical and chemical variables aggregated to the watershed level. These variance estimates were
used to scale the parametric variance associated with these different variables.  For example, hydraulic
conductivities were a function of soil porosity, sulfate half-saturation coefficients were a function of soil
sulfate adsorption,  and ion exchange selectivity coefficients were  a function  of soil cation  exchange
capacity and percent base saturation.  Estimating the effect of parametric uncertainty on model output
involved propagating this  range of watershed parameter values through the model and observing the
range in model output constituent concentrations or values.

      Wet deposition uncertainty estimates  were computed  by calculating  the variance for individual
chemical species over the period of record  at each Acid Deposition System (ADS) station used in the
NE or SBRP, and adding the variance associated with (1) kriging of precipitation to the individual NE  or
SBRP  sites  and (2) extrapolating  from the  nearest ADS site to the simulated watershed.  This latter
variance component was obtained using resampling or jackknifing procedures following random  deletion
on an ADS  site.  The wet deposition relative standard deviations are listed,  by  species, in Table 10-12.
Dry deposition estimates for these species were assumed to be ± 50 percent of the estimated annual dry
deposition values at each  individual site. Deposition uncertainty estimates were evaluated by varying the
deposition   (both  wet  and  dry)  consistently  up  or  consistently down  for  all  chemical   species.
Meteorological variability was not specifically investigated. The operational assumption is that typical year
projections  provide a common basis for comparisons among deposition scenarios to assess potential
changes in  surface water  chemistry.   Deposition uncertainty is an important part of these comparisons.
Meteorological  variance is implicitly incorporated in the  deposition uncertainty but there is no intent to
investigate interannual  or interdecadal variance in meteorology.

10.10.2.2.1   Enhanced Trickle Down -

      Input, initial  condition, and parameter errors were evaluated for  the ETD model using  first order
second moment analyses.  First order second moment analyses involve replacing the nonlinear model
with a  first  order  Taylor  series approximation for  error covariance propagation.  First order second
moment analyses includes not only the simultaneous effects of all state variables, inputs, and parameters
on each state variable,  but also the propagation of uncertainties in inputs, parameters, and state variables
(Lee et al.,  in  press).   Initial condition and parametric error were evaluated for the six  northeastern
watersheds  listed above.  Input (i.e., deposition) uncertainty was evaluated  only  for Kalers Pond (1E2-
063) (Georgakakos et al., in press).

10.10.2.2.2  integrated Lake-Watershed Acidification Study -

      First order error analysis was used to estimate uncertainty in the ILWAS model.  First order error
analysis is  similar to  sensitivity analysis.   Parameters or  input  variables were  subjected to small
perturbations and the change in selected output variables relative to the perturbation provided an  estimate
of the first derivatives.   The first derivative estimates were used as weights to propagate the parametric
or input variance  to an output  variance. This procedure is  similar to the  first order second  moment
analysis but the covariance matrix is not estimated with the first order  error analysis.
                                               674

-------
Table 10-12. Deposition Variations Used in Input
Uncertainty Analyses
Deposition
Chemical
Species
Sulfate
Nitrate
Chloride
Ammonium
Sodium
Potassium
Calcium
Magnesium
Hydrogen
Wet
% RSD
17.8
17.2
46.9
38.1
57.6
70.9
35.5
29.6
Used to
Dry
% RSD
50
50
50
50
50
50
50
50
complete charge balance
                                675

-------
10.10.2.2.3  Model of Acidification of Groundwater in Catchments -

      Uncertainty estimates for the MAGIC model were obtained using a fuzzy optimization procedure.
The fuzzy optimization procedure consisted of multiple calibrations of each catchment using perturbations
of the values of the fixed parameters to reflect the sampling and measurement error in these parameters.
These error estimates were obtained from the aggregated soils  data.  Each of the multiple calibrations
began with (1) a random selection of perturbed values of fixed parameters, and (2) a random selection
of the starting values of the adjustable parameters. The adjustable parameters then were optimized using
the Rosenbrock  algorithm to achieve a minimum  error fit to  the target variables.  Using the fuzzy
optimization on multiple  calibrations (i.e., an average of 7 calibrations for each DDRP catchment with a
minimum of 3 and a maximum of 10  calibrations/catchment), uncertainty  bands  of  maximum  and
minimum values were computed  for each  output  variable for each year.   These  uncertainty bands
encompass the range of variable values that were  simulated given the specified uncertainty  in  fixed
parameter values and measured  target  variables.   The difference  between  maximum  and  minimum
simulated values defines an uncertainty width about the simulated value arising from parametricuncertainty
for each output variable for each DDRP catchment.  The values of the uncertainty widths for each variable
in the calibration year were regressed against the simulated value of the variable in the calibration year
across all the DDRP catchments to derive a percentage  uncertainty value for each value, representative
of the region.

10.10.2.3  Relationship  Among Approaches

      Each of the approaches used for uncertainty analyses is appropriate for the model being used for
DDRP Level III projections.  MAGIC runs on a microcomputer,  for example, and was coded for the fuzzy
optimization analyses because computational time was not a consideration for MAGIC simulations.  This
approach was developed for the DDRP to incorporate both input and parameter uncertainty.  The ETD
model is intermediate in complexity and requires greater computational effort. The first order  second
moment analysis provides an estimate of simulation uncertainty  and permits partitioning this uncertainty
into input and parameter components (Lee et al., in press).  The ILWAS model has the greatest  number
and most complex set of formulations. The  ILWAS model, therefore, is not compatible with optimization
or first order second moment analyses but  is compatible with first  order error analyses,  which  provide
an estimate of simulation uncertainty.

      The relationship between the uncertainty estimates using the two different procedures for MAGIC
and ETD is shown for ANC and sulfate in Figure  10-18.  For both models, the magnitude of the standard
deviation was a function  of the ANC or sulfate concentration as indicated by the slope of regression line.
A multiplicative error term, therefore, was used in the uncertainty analyses.  The slope of the line for both
models was nearly  identical with the differences occurring in  the offset.  This offset  resulted in greater
uncertainty estimates for MAGIC than ETD.  The MAGIC simulations, however, incorporated both  input
and  parameter error while the ETD simulations (with  the exception of Kaler's Pond)  included only
parameter error.  Kaler's Pond included both input and parameter error and is comparable to the MAGIC
standard deviation estimates.  The offset between the two models represents the input error.  Because
of the similarity between MAGIC and ETD for Kaler's Pond, the MAGIC standard deviations were  used
to compute confidence intervals for both ETD and MAGIC.
                                              676

-------
            55
          o

          las i
          CD
          Q
          CO
          I15"
          CO   _
             -5
                                Northeast Lakes
                               Model Uncertainty
                                     ANC
                                Kalers Pond
                                   O
-20       0       20       40      60~

             ANC (ueq L'1) at 50 Years
                                                           80
                                                       o	ETD
                                                       P— MAGIC
             45
          •§30 ,
           CD
          Q   -]
          "S   -
           CO

          115^
          •*—•
          CO
                55
          —i	
          70
                               Northeast Lakes
                               Model Uncertainty
                                     SO4
85
100
115
                                                           130
                           SO4  (ueq L1) at 50 Years
                                                       o	 ETD
                                                       a	MAGIC
Figure 10-18. Comparison of projection standard errors as a function of ANC (top figure) and
sulfate (bottom figure) concentrations for the NE uncertainty analysis watersheds using ETD and
MAGIC.
                                       677

-------
     The estimates of parametric uncertainty using the ILWAS model also are consistent with the regional
estimates of both MAGIC and ETD (i.e., Kaler's Pond).  Input uncertainty estimates were computed by
all three  models using the individual uncertainty estimation procedures discussed above.  The input
uncertainty estimates computed for all three models  were on  the same order as the  parametric
uncertainty.  Parametric variance estimates, therefore, were doubled to obtain estimates of projection
uncertainty for the ILWAS  model.  The  procedures described in Section 6 were used to integrate this
projection error  with sampling  error and compute confidence intervals for the population  estimates
presented in the next section.

10.10.2.4 Confidence Intervals

      Upper and lower bounds  for a 90 percent confidence interval were computed for ANC, pH, and
sulfate projections from all  three  models and  for calcium  and  magnesium for MAGIC and  ILWAS
(Appendix A.3).  The confidence intervals were computed using the variance  estimator discussed in
Section 6. This variance estimator includes estimates of sampling and measurement error, parameter and
input error, and regional estimation error. Time constraints prevented the inclusion of confidence intervals
for the ILWAS SBRP projections. These figures will be incorporated in the Mid-Appalachian Report in mid-
1990.

10.11  POPULATION ESTIMATION AND REGIONAL FORECASTS

      Population  estimation procedures  were discussed in Section 6.  The   uncertainty estimation
procedures  were  discussed  in  the previous  section,  Section 10.10.   Comparisons among model
projections are presented only for comparable target populations.  The number of watersheds simulated
with each model differed; therefore, it  is critical that comparisons be made only among or between
models that simulated the same watersheds,  i.e., those that represent the same  target population. The
population estimates discussed below represent three  different target populations. The target population
discussed in each of the  following  sections  is clearly defined. These definitions are essential for the
proper interpretation of the results.  The models, target populations, and population attributes  for both
the NE and SBRP are  shown in Table 10-13.

10.11.1  Northeast Regional Projections

10.11.1.1  Target Population Projections Using MAGIC

      An estimated 3227 lakes in the target population in the NE were simulated using MAGIC, compared
to the DDRP total target population of 3667.  The smaller target population reflects an exclusion of lakes
for which MAGIC was unable to  satisfy the calibration criteria.  The simulated target population represents
Priority Classes A - I,  which includes  both  disturbed and undisturbed watersheds based on  chloride
concentrations (See Section 10.5.7), watersheds that had both positive and negative sulfur retention,  and
watersheds that had initial ELS-I ANC concentrations ranging from -53  to 392 /*eq L'1.  The MAGIC
simulations for this target population extend to 100-year projections. Projections using other models were
restricted to 50 years, because time and computational requirements prohibited longer simulations.  The
100-year time frame provides additional  insight into the cumulative effects of sulfur deposition on  changes
in surface water chemistry.

                                              678

-------
Table 10-13.  Target Populations for Modelling Comparisons and Population Attributes
Region
Priority
 Class
 Models
 Target
Population
Population
Attributes
Northeast
A and B
ETD.ILWAS
MAGIC
                                                    502
                ANC<100/jeq L'1
                (Int. Staff Paper)
A- E
A-l
ETD.MAGIC 1813
MAGIC 3227
ANC<400, Undisturbec
Watersheds
ANC<400,Represent-
ative of entire NE
population
Total DDRP
SBRP



Target Population (NE)
A and B

A- E


ILWAS,
MAGIC
MAGIC

3667
ANC<200,Undisturbed
567 Watersheds
1323 Representative of entire
SBRP population
Total DDRP Target Population (SBRP)
                                 1531
                                         679

-------
 10.11.1.1.1   Deposition scenarios -

      Projected changes in ANC and sulfate concentrations that might occur over a 100-year period,
assuming current and decreased deposition, are shown in Figure 10-19. The confidence limits about the
individual simulations are included in Appendix  A.3. Confidence intervals are not included on the figure
in order to increase the contrast among model projections or between deposition scenarios.

      The projected changes in median ANC concentrations over 100 years assuming either current and
decreased deposition were small (Figure 10-19). The median ANC concentration projected after 50 years
for constant deposition at current levels was 124 /*eq L"1 and for a 30 percent deposition decrease was
135 /zeq L"1, representing a difference of 11 /zeq  L"1  (Table 10-14).  The  change projected in  median
sulfate concentration after 50 years for current deposition was 99 jteq L"1 and for decreased deposition
was 71  peq L"1, representing a -28 /*eq L"1 difference.  The changes projected in median ANC over a
100-year period between current and decreased deposition were 121  versus 134 /*eq L"1, respectively,
or a difference  of 13 jteq L"1. The projected changes in median sulfate concentration after 100 years for
current  deposition and a 30 percent decrease in deposition were 99 versus 70 fj.eq L"1, respectively, or
a difference of  -29 ^eq L"1.  A small decline in ANC concentrations, (less than 1 /teq L"1) was indicated
between year 50 and  year 100 for both current  and decreased  deposition.   Projected calcium  and
magnesium concentrations also showed a small  but continual decline over the 100-year period under both
current  deposition and a 30 percent deposition  decrease (Table 10-14). Sulfate concentrations declined
during the initial 50-year period, asymptotically approaching steady state, and were projected to remain
essentially constant from year 50 to year 100 under current deposition and  to decrease slightly over this
same period  for decreased deposition. The confidence limits about the  projected CDFs represented a
projection error of about ± 36 //eq L"1 in ANC and ± 32 ^eq  L"1  in sulfate concentrations.  Both the
changes projected for 50 and 100 years,  assuming  different deposition scenarios,  were  within the
uncertainty bounds of the projections.

      The projections of the sulfate concentrations indicated that the watersheds would be near sulfate
steady-state after 50 years under current deposition with the median projected watershed  sulfur retention
of about 5 percent in both year 50 and year 100 (Table 10-14). The  interquartile range varied from about
1 to 11  percent in  both year 50 and 100, indicating the majority of the watersheds were projected to be
near sulfate  steady-state.  The median projected sulfur retention under decreased deposition for these
watersheds was nearly zero.  From year 50 to year 100, the projected median sulfur retention changed
from 0 to a slightly positive sulfur retention  (~5  percent) in the watersheds. The interquartile ranges for
the 50-  and  100-year periods were -7 to +8 percent and -1 to +9 percent, respectively.

      Projected changes in pH that might occur  over a 100-year period, assuming current and decreased
deposition are shown in Figure 10-20 and listed in Table 10-14.  The projected changes in pH over a 100
year  period  under constant and  decreased deposition were  small  (Figure  10-20).    The projected
differences between the median pH under constant and decrease deposition after 100 years were less
than 0.05 pH units (Table 10-14).  The projected differences in pH for the lower quartile after 100 years
under the two deposition regimes were less than 0.1 pH units, varying from 6.67 under current deposition
to 6.75  under decreased  deposition (Table 10-14).
                                              680

-------
         1.0
o
o.
£  0.6
Q.


       >
       5= 0.4
       S
       i
       o
         0.0
           -100
	Simulation Year 0
	 Constant Deposition
	Ramp Deposition
                  0     WO    200    300   400
                    ANC (jteq  L-i)
                                                                to
                              O 0.8
                              O

                              §-0.6
                              Q.

                              
-------
Table 10-14. Descriptive Statistics of Projected ANC, Sulfate, pH, Calcium Plus
Magnesium, and Percent Sulfur Retention for NE Lakes in Priority Classes A - I
Using MAGIC for Both Current and  Decreased Deposition
Year

MAGIC All.
Yr 0
Yr 20
Yr 50
Yr 100
MAGIC All,
Yr 0
Yr 20
Yr 50
Yr 100
MAGIC All.
YrO
Yr 20
Yr 50
Yr 100
MAGIC All.
Yr 0
Yr20
Yr 50
Yr 100
MAGIC All,
Yr 0
Yr20
Yr 50
Yr 100
Mean
X
ANC
151
151
149
146
SO,2'
1171
108
106
106
pH
6.01
6.02
5.99
5.90
Ca + Ma
229
221
217
213
Std.
Dev.
Min.
P_25 Median P_75
Max.
Current Deposition

114
114
114
113

48
44
43
43

0.60
0.60
0.62
0.66

122
121
120
120

-21
-21
-22
-22

50
47
46
45

4.47
4.49
4.48
4.47

41
40
38
37

70
70
67
62

71
70
68
66

6.73
6.73
6.71
6.67

128
124
121
118

126
126
124
121

111
101
99
99

6.97
6.98
6.98
6.96

197
190
186
182

223
224
224
216

152
147
142
140

7.22
7.22
7.22
7.21

296
288
284
280

416
417
414
409

246
221
215
214

7.49
7.49
7.49
7.48

560
544
540
531
% S Retention
-4
4
6
6
10
9
8
8
-25
-20
-18
-19
-11
-1
1
2
-3
3
5
5
4
10
11
11
19
25
26
27
                                                                  continued
                              682

-------
Table 10-14. (Continued)
Year
Mean
Std.
Dev.
Min.
P 25    Median   P 75
Max.
                         30% Decrease in Deposition
MAGIC
Yr 0
Yr 20
Yr 50
Yr 100
MAGIC
YrO
Yr20
Yr 50
Yr 100
MAGIC
YrO
Yr 20
Yr 50
Yr 100
MAGIC
Yr 0
Yr20
Yr 50
Yr 100
MAGIC
Yr 0
Yr 20
Yr 50
Yr 100
All. ANC
151
156
160
158
All, SO.2'
1171
99
80
77
All. DH
6.01
6.13
6.28
6.26
All, Ca + Ma
229
217
203
198

114
115
115
114

48
40
33
32

0.60
0.56
0.52
0.53

122
120
118
117

-21
-20
-f8
-19

50
44
35
34

4.47
4.52
4.58
4.58

41
39
35
32
All, % S Retention
-4
-10
-1
3
10
11
13
11
-25
-44
-44
-36
                                      70
                                      74
                                      76
                                      74
                                      71
                                      63
                                      52
                                      48
                                       6.73
                                       6.74
                                       6.77
                                       6.75
                                      128
                                      120
                                      113
                                      110
                                      -11
                                      -16
                                       -7
                                       -1
                                    126
                                    131
                                    135
                                    134
                                    111
                                     93
                                     71
                                     70
                                    223
                                    227
                                    231
                                    231
                                    152
                                    132
                                    107
                                    103
                                    416
                                    425
                                    431
                                    428
                                    246
                                    202
                                    157
                                    153
                                      6.97
                                      6.99
                                      7.01
                                      7.01
                                      7.22
                                      7.23
                                      7.24
                                      7.24
                                      7.49
                                      7.50
                                      7.50
                                      7.50
                                    197
                                    184
                                    173
                                    170
                                     -3
                                    -10
                                      1
                                      4
                                    296
                                    286
                                    267
                                    268
                                      4
                                     -1
                                      8
                                      9
                                    560
                                    539
                                    529
                                    511
                                     19
                                     17
                                     24
                                     26
                             683

-------
                              10
                              0.8
                           5
                           Q.
                           O
                              0.6
                           := 0.4
                           co
                           o
                              0.0
                                          NE  Lakes

                                       Model  - MAGIC

                                    Priority  Class  = A  -  I

                                         Year •=  20
	Simulation Year 0
	Constant Deposition
	Ramp Deposition
                               4.0  4.5 5.0  SJS  3.0 6.5  7.0 7.5  8.0

                                              pH
                                          Year -  50
                              to





                           O 0.8

                           •_
                           o


                           £ 0.6




                           §
                           V 0.4

                           JS
                           3



                           o
                              0.0

	
Simulation Year 0
Constant Deposition
Ramp Deposition
                               4.0  4.5  5.0  5.5  6.0 8.5  7.0  7.5 8.0

                                              pH
                                          Year -  100
                              to




                            O 0.8


                            o


                            8 o.e
                              0.4
                           o
                              0.0
 	Simulation Year 0
 	Constant Deposition
 	Ramp Deposition
                               4.0  45  5.0  SJ  8.0  t£ 7.0  7£  8.0

                                              PH
Figure 10-20.   pH projections for NE  lakes, Priority Classes A - I, using  MAGIC for  20, 50, and

100 years, under current deposition and a 30 percent decrease  in deposition.
                                                     684

-------
      Projections of the number of lakes currently not acidic that might become acidic in the next 50
years for current deposition were 87 (3 percent) lakes and 50 (2 percent) for a 30 percent deposition
decrease based on an estimated target population of 3227 lakes. Projections of the number of lakes
currently not acidic that might become acidic in the next 100 years under current deposition  and a 30
percent deposition decrease  were 100 (3 percent) and  50  (2 percent), respectively.  The number of
currently acidic lakes (i.e., 162 lakes in the target population with ANC <0 fieq L"1) that might chemically
improve (increase in ANC)  under current -deposition levels and a 30 percent deposition reduction after
50 years were projected as  64 (39 percent)  and 125  (77 percent),  respectively.    The number of
currently acidic lakes that might chemically improve after  100 years at current and decreased deposition
levels was projected  as 52  (32 percent) and  113 (70  percent), respectively.  The percentages for
chemical improvement are based on the number of currently  acidic lakes estimated  in the target
population  (i.e., 162 lakes).

10.11.1.1.2 Rate of change of ANC, sulfate, and pH  over 100 years -

      The  projected changes in ANC, sulfate concentrations, and  pH over the  next 100 years are
displayed as box and whisker plots (Figures 10-21 through 10-23).  Box and whisker plots illustrate how
both the target population constituent median and interquartile range vary through time.

      The median ANC concentrations projected for current deposition changed from an initial calibrated
concentration of 126 jueq L"1  to 124  //eq  L"1 after 50 years and to 121  //eq L"1  after 100 years (Table
10-14).  For a 30 percent deposition decrease, the median ANC was projected to change from 126 neq
L"1 at year 0 to 135 /zeq L"1 at year 50, remaining  essentially unchanged over the next 50 years.  The
median calcium plus magnesium concentration decreased linearly over the entire 100-year simulation
period for current deposition  with a projected  decrease  from 197 to 182 neq L"1  (about 0.15 //eq L"1
yr1).  Median calcium plus magnesium concentrations also declined from 197 to 170 neq L"1 over the
100-year period for decreased deposition and from 197 ^eq  L"1 in year 0 to 173 peq L"1 in year 50
(approximately 0.5 peq L"1 yr"1 ).  The projected rate of change for the next 50 years decreased (less
than 0.1 //eq L"1 yr"1) but retained a negative slope.

      The projected change in median sulfate concentration for current deposition was an asymptotic
decrease toward  sulfate steady state (or to  a small positive retention due to in-lake  retention)  with
conceritrations near steady state after the first 10 years.  For the scenario of decreased deposition, the
projected change in the median sulfate concentration was -40  peq L"1 and was essentially complete by
year 50. The mean projected change in the median sulfate concentration over the subsequent 50 years
was slightly less than -2 /teq L"1.

      The median pH values changed from 6.97 to 6.96 under current deposition and from 6.97 to 7.01
under decreased deposition, a change of less  than 0.05  units (Table 10-14).   The  variance in  pH
remained relatively constant  through time (Figure 10-23).

      Neither the change in  ANC nor change in sulfate concentration was a function of the initial ELS-
I ANC, for either current deposition  or for a 30 percent decrease in deposition (Table 10-15).  Shifts in
the population distribution of median ANC and sulfate concentrations over the 40-year period from year

                                              685

-------
                                                              3rd Quartlo +
                                                              (15 x Interquartile Range)™
                                                              3rd Quart to

                                                              Mean

                                                              Median

                                                              IstQuartile

                                                              IstOuartile
                                                              (15 x Interquartile Range)"


                                                              "Not to exceed extreme value
                                                Constant
250-
200-
^150-
lj
|ioo-
o
< 50-
o-
•inn —



*


o
rr
















0
i
















8
ec
















§
OL
















O
*!•
^
















O
in
cc
















S
cc










                                                 Ramped
8
                                                         §
8
500-
400-
~300~
faoo-
o
< 100-
o-







£



'








£ §E §E §E §E £





















. 	 : —





















































Figure  10-21.  Box and whisker plots off ANC distributions at 10-year intervals for NE Priority
Classes A - I using MAGIC.
                                                 686

-------
                                                         3rdQuartile +
                                                         (15 x Interquartile Range)**

                                                         3rd Quart to


                                                         Mean

                                                         Median


                                                         lEtOuartte


                                                         IstQuartile
                                                         (1.5 x Interquartile Range)**



                                                         "Not to exceed extreme value
                                          Constant
                                                                 8

250-
200
f—
-* 150 —
I
50 —











































--

                                           Ramped
                        250— I
                        200
                     lj 150 —
                     cr
 ,
o
                           _,
                        100-1
                         50 —
                                                    DC
                                                    >-
                                                          EC
                                                                 EC
                                                                        s
                                                                        EC
Figure 10-22.  Box and whisker plots of sulfate distributions at 10-year intervals for NE  Priority
Classes A - I using MAGIC.
                                                   687

-------
                                                           3rd Quartile +
                                                           (1.5 X Interquartile Range}**
                                                           3rd Quartile

                                                           Mean

                                                           Median

                                                           1st Quartila

                                                           1st Quartile
                                                           (1.5 x Interquartlls Range)**


                                                           "Not to exceed extreme value
                                            Constant
                          8-1



c
c


3
C


C
T
0


2 ?
E 5


3
c


c
c
D

. 	 1
*
C





<
*
C
*


D
1-
C
»-





«
u
C


? 2
C DC
>
-


                         5-
                         8—I
                                             Ramped
                                             8
S
                                                                        O
                                                                        o



c

___<
c


t

..— '
c


c

. 1
c





c
>

1
c





c

'
c c
>


c


a:
--
••

                         5 —
Figure 10-23.   Box and whisker plots  of pH distributions at  10-year  intervals  for  NE Priority
Classes A - I using MAGIC.
                                                  688

-------
10 to year 50 indicate a relatively uniform change among ANC and sulfate intervals (Figures 10-24 and
10-25).  The 40-year interval was selected as the period for comparison  because the ramp change in
deposition did not occur until year 10; the First 10 years of all projections represent,  therefore, current
deposition levels. The change in ANC was smaller for acidic lakes for decreased deposition but similar
among the other three ANC groups (Table 10-15).

10.11.1.2  Target Population Projections Using  MAGIC and  ETD

      An estimated target population of 1920 lakes was simulated using both ETD and MAGIC.  These
lakes represent Priority Classes A - E (Figure 10-1), which have ANC concentrations ranging from -53 to
392 fj.eq L"1.  These priority classes  contains watersheds that  have both positive and negative sulfur
retention but, based on  chloride concentrations (Section 10.5.7), are relatively undisturbed.
10.11.1.2.1  Deposition scenarios -

      ETD and MAGIC projected similar changes in ANC, sulfate, and pH over the 50-year period for
both current deposition and a 30 percent deposition decrease (Figures 10-26 through 10-28).   Confidence
intervals for each of the projections are included in Appendix A.3.  For current deposition  levels, the
median ANC concentrations projected after 50 years using ETD and MAGIC were 74 and 110 neq L"1,
respectively (Table  10-16).   Under a 30 percent deposition decrease,  the median ANC concentrations
projected using ETD and MAGIC after 50 years were 85 and 119 fj.eq L"1, respectively. The differences
between the model  projections result primarily from the initially calibrated ANC  concentrations.  The
median calibrated ANC concentration for MAGIC was 116  /zeq L"1,  while the  median (ELS-l)  ANC
concentration assumed as the initial  model condition for ETD was 77 ^eq L"1.  The difference between
the ETD initial and 50-year  projected ANCs was 4 /*eq L"1  , similar to the 6 neq L"1 difference observed
for MAGIC (Table 10-16).   Similar differences between the initial and  50-year  projections occurred for
sulfate and pH.  These relatively minor discrepancies reflect differences in the calibration procedures for
both models (See Section 10.9) and are within the uncertainty bounds  on the projections.

      Results of within-model comparisons of the effects of alternative  deposition  scenarios on surface
water chemistry are shown in Figures 10-29 through 10-31.  Changes in median ANC projected for both
deposition scenarios after 50 years using either model were small and were  consistent with the MAGIC
projections discussed in the previous section.  For example, differences in the median ANC projected
after 50 years using ETD  versus MAGIC, between  current  deposition and a 30  percent  deposition
decrease were 74 neq L'1  versus  85 jueq  L'1 (+11 /zeq L"1 ) and 110 peq L'1 versus 119 ^eq L"1 (+9
neq L"1 ), respectively.  The differences in the median  sulfate concentrations  projected  after 50 years
using  ETD and MAGIC between current  deposition and a 30  percent deposition decrease were 100
versus 70 /teq L"1 (-30 neq L"1) and  91  versus 64 ^eq L"1 (-27 peq  L"1  ), respectively.

      The differences in the median  pH at the end of 50 years under current and  decreased deposition
for MAGIC were 6.92 versus 6.95 (+0.30) and for ETD were 6.80 versus 6.86 (+0.06), respectively. The
sulfate concentrations  projected using both models indicated the watersheds were near sulfate steady
state after 50 years.  The median sulfur retention for the watersheds projected using both models ranged
                                               689

-------
Table 10-15.  Change in  Median ANC and Sulfate Concentrations  Over a 40-Year Period as a
Function of the Initial ELS-Phase I or NSS Pilot Survey ANC Groups

NE
Priority Class AB
ETD, cons.
ETD, ramp
ILWAS, cons.
ILWAS, ramp
MAGIC, cons.
MAGIC, ramp
Priority Class A-E
ETD, cons.
ETD, ramp
MAGIC, cons.
MAGIC, ramp
Priority Class A-l
MAGIC, cons.
MAGIC, ramp
SBRP
Priority Class AB
ILWAS, cons.
ILWAS, ramp
MAGIC, cons.
MAGIC, ramp
Priority Class A-E
MAGIC, cons.
MAGIC, ramp

< 0


-0.4
7
-3
8
-0.6
5

-0.5
4
-2
5

-2
5


-
-
-
-

-
-
ANC
0 - 25


5
12
-3
10
0.4
13

1
9
-3
6

-2
10


-
-
-
-

-
-
(uea L"1)
25 - 100


3
14
-6
5
-3
6

2
14
-0.5
7

-1
10


-15
-16
-14
-21

-14
-21
Sulfate (uea L"1)
100 - 400


-
-
-
-
-
-

-4
11
0.5
7

-3
15


-6
-7
-14
-24

-24
-34
< 0


2
-44
5
-52
-5
-48

-5
-37
-0.5
-30

-0.5
-30


-
-
-
-

-
-
0 - 25


-1
-34
-
-21
-5
-37

-1
-31
-5
-21

-2
-33


-
-
-
-

-
-
25 - 100


-9
-24
-
-20
-6
-27

4
-25
-0.6
-27

-3
-30


37
53
27
44

27
44
100 - 400


-
-
-
-
-
-

-10
-28
-4
-20

-7
-34


25
33
31
47

31
47
                                       690

-------
                            Northeast Lakes
                           Priority Class A - I
                             Model = Magic
                         Deposition = Constant
CO
CD

-------
                                    Northeast Lakes
                                    Priority Class A -1
                                     Model = Magic
                                  Deposition = Constant
      V)
      CD
700-
600-
500-
400-

300-
200-
100-
r\ -





J





55


I
1

m
B5I

1
m


p.



i
1


i
-
i











YJ
n





1-1











r
\
iriru
              30 40 50 60  70  80 90100110120130140150160170180190200210220230
           [SOllOieqL-1)

         Northeast Lakes
        Priority Class A -1
          Model = Magic
Deposition = Ramped 30% Decrease
                                                                    O MAGIC Year 10
                                                                    H MAGIC Year 50
      %
700-
600-

500-
400-

300-
200-
100-
n .






fa
i
I
I

fit
X
I





r












i
—
I
6



\
!
r-l
I
	


I

|
—
«
a
»
JjJ
1

1
rflfi.flUn nn
              30 40 50 60 70  80 90100110120130140150160170180190200210220230
                                                                   D  MAGIC Year 10
                                                                   Zi  Year 50 Ramped
Figure 10-25.  Comparison of population histograms for sulfate concentrations at current levels of
deposition and a 30 percent decrease for NE lakes, Priority Classes A - I, using MAGIC.
                                         692

-------
         to
I

£ 0.6
0-
0>
S= 0.4
.2
I
O
         0.0
                     NE Lakes
              Priority Class  = A - E
               Deposition =  Constant
                     Year = 0
                0    100   200   300    400
                  ANC (|ieq L-I)
                                                                NE  Lakes
                                                          Priority Class = A -  E
                                                     Deposition  = Ramp 30% Decrease
                                                                Year -=  0
                                                        o 0.8
                                                        £ 0.6
                                                        o_
                                                        «
                                                        u

                                                        O
                                                          0.4
                                                          0.0
                                                           -100
                                                           0    100    200   300   400
                                                             ANC (jieq L-1)
         to
      O  0.8
         0.6
      O.

      CD
      O
        0.4
         0.2
        0.0
         -100
         Deposition = Constant
              Year -  20
                0    100   200   300    400
                  ANC (jieq L-I)
                                                           Deposition = Ramp 30% Decrease
                                                                      Year - 20
                                                          to
                                                  £ 0.6
                                                  Q.

                                                  o
                                                        o
                                                          OJ3
                                                           -100
                                                           0    100   200    300   400
                                                              ANC  (jieq  L-I)
         to
      _0  0.8

      O

      S.  0.6
      CL

      0)
         0.4
      O
               Deposition •= Constant
                    Year -  50
         °-°100    0    100    200   300   400
                  ANC (jieq L-1)
                                                     Deposition « Ramp 30% Decrease
                                                                Year - 50
                                                     to
                                                        O 0.8
                                                    0.6
                                                        7= 0.4
                                                        £
                                                        o
                                                            0     WO   200   300   400
                                                              ANC (jieq L-i)
Figure 10-26.   Comparison of MAGIC and ETD  projections of ANC for NE lakes, Priority Classes
A - E,  under current and decreased deposition.
                                                     693

-------
              10
           §"0.6
           a.
           I0-4
           3

           o
             0.0
                         NE Lakes
                   Priority Class =  A - E
                    Deposition « Constant
                         Year  = 0
                       100       200
                     ISO,*] (jieq L-1)
                                          300
               NE Lakes
         Priority Class  = A - E
   Deposition = Ramp  30% Decrease
               Year •= 0
   I0r
 O 0.8
~
 s
 Q.
 S 0.6
 0>
 >
 «= 0.4
                                                              0.0
             100       200
           [SO,*]  ((teq  L-1)
           0 0.8
           "S
           o
           o.
           8 0.6
           a.
           o
           >
           ~ 0.4
           s
           3

          o
             0.0
                   Deposition •=  Constant
                        Year  -  20
                       100       200
                     [S04*l (jieq L-i)
                                         300
   Deposition = Ramp 30% Decrease
              Year - 20
   tOr
j| 0.8
i
                                                           o
                                                              OJ>
  100
ISO,*]
                      200
                      L-1)
                               300
             10
          S 0.6
          Q.
          O
            0.2
            0.0
                   Deposition -= Constant
                        Year - SO
                       100       200       300
                     [SCVl  (|ieq  L-i)
   Deposition - Ramp 30% Decrease
              Year - 50
   I0r
                                                            O 0.8
                                                           ~S
  0.6
            100       200
           [SCV1 (|ieq L-i)
                               300
Figure 10-27.  Comparison of MAGIC and ETD projections of sulfate concentrations for NE lakes,
Priority Classes A - E,  under current  and  decreased  deposition.
                                                     694

-------
            1.0
_0  0.8

fe
O.
S  0.6
          a
          3


          O
            0.0
                        NE Lakes
                  Priority  Class  = A  -  E
                   Deposition =  Constant
                        Year = 0

	
Phase 1
MAGIC
ETD
             4.0  4.5  5.0 5.5  6.0  6.5 7.0  7.5 8.0
                            PH
                                                                  NE Lakes
                                                           Priority Class = A  -  E
                                                      Deposition = Ramp 30%  Decrease
                                                                  Year =  0
                                                             O 0.8
                                                             O

                                                             £ 0.6
                                                             Q.

                                                             a>
                                                   a
                                                   zj

                                                   O
                                                               0.0
                      l>

	
Phase 1
MAGIC
ETD
                                                       4.0  4.5  5.0  5.5 6.0  6.5 7.0  7.5  8.0
                                                                     PH
o
a.

i
            0.6
          O
                   Deposition
                        Year
                      Constant
                      20
             '4.0 4.5  5.0
                            6.0  6.5 7.0  IS  8.0
                            pH
   Deposition =  Ramp  30%  Decrease
              Year •=  20
3  0.8

o
Q.
S0.6


0)

5  0.4
RJ
                                                             O
                                                                0.0
                                                       4.0  4.5  5.0  5JS  6.0 6.5  7.0 7.5  8.0
                                                                      pH
                   Deposition  = Constant
                        Year - 50
              4.0  4.5 5.0  5.5  6.0 8.5  7.0  7.5  8.0
                                                       Deposition - Ramp  30%  Decrease
                                                                  Year -  50
                                                       to
                                                              O 0.8
                                                              D-
                                                              O
                                                              C
                                                              S


                                                              o
                                                       0.4
                                                                0.0
                                                        4.0 4.5  5.0  5.5  6.0  6.5 7.0  7.5  8.0
                                                                      PH
Figure 10-28. Comparison of MAGIC and ETD projections of pH for NE lakes, Priority Classes A
 E, under current and decreased deposition.
                                                       695

-------
from 3 to 6 percent for current deposition and from 2 to 4 percent for decreased deposition (Table 10-
16).  Although there was a range in this distribution of sulfur retention, the upper and lower quartile values
for current deposition ranged from -3  to 9 percent for ETD and 4 to 13 percent for MAGIC; under
decreased deposition, percent sulfur retention ranged from -4 to 7 for ETD and -1 to 10 for MAGIC after
50 years, indicating most of the watersheds were near sulfur steady state  (Table 10-16).

      Projection of the number of lakes not currently acidic that might become acidic in the next 50 years
for current deposition using ETD and MAGIC were 49 (3 percent) and  87 (5 percent), respectively. The
number of lakes currently not acidic that might become acidic for a 30 percent deposition decrease using
ETD and MAGIC were 37 (2 percent) and 50 (3 percent), respectively.  The number  of currently acidic
lakes that might chemically improve under current deposition after 50  years was projected by ETD and
MAGIC to be 52 (23 percent) and 64 (39 percent), respectively. Under a 30 percent deposition reduction,
the projections  of chemical improvement were estimated to be 103 lakes (46 percent) for ETD and 125
lakes (77 percent)  for MAGIC.

10.11.1.2.2.  Rate of change of ANC, sulfate,  and pH over 50 years -

      The changes in ANC and sulfate concentrations and in pH projected over the next 50 years using
both the ETD and MAGIC models are shown in Figures 10-32 through 10-37. The change in median ANC
projected using ETD and MAGIC under current deposition over the next 50 years was a total decrease
of -3.1 and -5.3 ^eq  L" , respectively (Table 10-16). The change in median sulfate concentrations using
ETD and MAGIC was -3.9 and -8.9 /*eq L"1, respectively, for current deposition. The  change in  median
pH using ETD and MAGIC under current deposition was -0.02 units for each model.  With  a 30  percent
deposition reduction, the ANC increase projected using ETD and MAGIC was from 77 to 85 (+8) and 116
to 119 (+4) jueq L"1, respectively,  similar to the change  projected for the larger target population in the
previous section. The decrease projected in sulfate concentrations with a 30 percent deposition decrease
using ETD and MAGIC  was from 104 to 70  (-34) and  164 (-36) /zeq L"1, respectively,  over  50 years.
These values are roughly equivalent to the measurement or projection error determined for sulfate. The
pH increase projected under decreased deposition using either model was less than +0.05  units  over 50
years. The variance in ETD projections, although larger than MAGIC projections, also remained relatively
constant through time (Figures 10-34 and Figure 10-37).

      The changes in ANC or sulfate concentration were not functions of the initial ELS-I concentrations
using either MAGIC or ETD for either deposition scenario (Table 10-15). Histograms of projected  change
in the population distribution  of median ANC and sulfate from year 10  to year 50 indicate a relatively
uniform change among ANC and sulfate intervals using both ETD and MAGIC (Figures  10-38 through 10-
41).  A slightly greater change in ANC was  projected for non-acidic lakes with decreased deposition.

10.11.1.3  Restricted Target Population Projections  Using All Three Models

      There were  an estimated 495 lakes  in the target population simulated  using  all three Level  III
models.  This target population  represents Priority Classes A and B (Figure 10-1).  Lakes in this target
                                              696

-------
Table 10-16.  Descriptive Statistics of Projected  ANC, Sulfate, and Percent Sulfur
Retention for NE Lakes in Priority Classes A - E Using  MAGIC and ETD for Both
Current and Decreased  Deposition
Model Mean
Std.
Dev.
Min.
P_25
Median
P_75
Max.
Current Deposition
MAGIC vs. ETD. ANC
Model Year 0
ETD 107
MAGIC 134
Model Year 20
ETD 106
MAGIC 134
Model Year 50
ETD 105
MAGIC 133

110
116

109
116

107
116

-53
-21

-52
-21

-52
-22

20
45

16
44

15
42

77
116

71
114

74
111

191
179

192
179

177
178

392
410

383
408

384
407
MAGIC vs. ETD. SO ~z
Model Year 0
ETD 106
MAGIC 107
Model Year 20
ETD 104
MAGIC 98
Model Year 50
ETD 103
MAGIC 97
MAGIC vs. ETD. oH
Model Year 0
ETD 5.61
MAGIC 5.78
Model Year 20
ETD 5.64
MAGIC 5.80
Model Year 50
ETD 5.62
MAGIC 5.77
MAGIC vs. ETD. % S
Model Year 0
ETD -5
MAGIC -1
Model Year 20
ETD 2
MAGIC 6
34
43

38
39

40
38


0.81
0.72

0.80
0.71

0.79
0.73
34
60

53
55

54
53


4.27
4.47

4.29
4.49

4.29
4.48
79
67

73
63

68
61


6.24
6.53

6.17
6.53

6.13
6.51
104
100

101
92

100
91


6.82
6.94

6.79
6.93

6.80
6.92
125
125

121
111

118
109


7.22
7.12

7.22
7.12

7.18
7.12
199
246

222
221

216
215


7.53
7.48

7.52
7.48

7.52
7.48
Retention

32
8

9
7

-97
-16

-24
-5

-19
-8

-4
2

-7
-1

2
5

18
4

7
11

69
19

30
24
                                                                  continued
                             697

-------
Table 10-16. (Continued)
Model Mean

Model Year 50
ETD 4
MAGIC 18
MAGIC vs. ETD. ANC
Model Year 0
ETD 107
MAGIC 134
Model Year 20
ETD 109
MAGIC 139
Model Year 50
ETD 112
MAGIC 143
MAGIC vs. ETD, SO f"
Model Year 0
ETD 106
MAGIC 107
Model Year 20
ETD 94
MAGIC 89
Model Year 50
ETD 74
MAGIC 71
MAGIC vs. ETD. DH
Model Year 0
ETD 5.61
MAGIC 5.78
Model Year 20
ETD 5.72
MAGIC 5.91
Model Year 50
ETD 5.82
MAGIC 6.08
Std.
Dev.
30%

8
6


110
116

108
116

107
116


34
43

34
36

29
29


0.81
0.72

0.76
0.67

0.71
0.61
Min.
Decrease

-17
-1


-53
-21

-47
-20

-40
-18


34
60

44
49

39
38


4.27
4.47

4.33
4.52

4.40
4.58
P_25
Median
P_75
Max.
in Deposition

-1
4


20
45

18
49

22
51


79
67

63
55

51
44


6.24
6.53

6.20
6.57

6.30
6.59

3
6


77
116

77
118

85
119


104
100

90
83

70
64


6.82
6.94

6.82
6.95

6.86
6.95

9
13


191
179

204
193

198
204


125
125

108
103

86
81


7.22
7.12

7.24
7.16

7.23
7.18

27
24


392
410

389
415

399
417


199
246

186
202

162
157


7.53
7.48

7.52
7.48

7.53
7.49
MAGIC vs ETD, % S Retention
Model Year 0
ETD -5
MAGIC -1
Model Year 20
ETD -10
MAGIC -6
Model Year 50
ETD 1
MAGIC 4

32
8

12
9

11
9

-97
-16

-50
-22

-50
-18

-19
-8

-17
-13

-4
-1

-7
-1

-9
-6

2
4

18
4

-3
-0

7
10

69
19

27
16

24
24
                              698

-------
                     NE Lakes
                  Model =  MAGIC
              Priority Class =  A  -  E
                     Year  = 20
c
_g
^
i_
o
Q.
O
        1.0 r
       0.8
       0.6
    *= 0.4
    (0
    u
    E

    O °-2
       o.o i—
        -100
                                                            1.0
                                                          O 0.8
                                                          O
                                                          Q.
                                                          O
                                                            0.6
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp Deposition
•-;= 0.4
as
3
E
O °-2
                                                                     NE  Lakes
                                                                  Model  =  MAGIC
                                                              Priority  Class = A  - E
                                                                     Year = 50
                0    too    200
                  ANC  (|ieq  L-
                                  soo
                                         400
                                                        o.o i—
                                                         -100
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp  Deposition
                                                                o     100    200   soo
                                                                  ANC (jieq  L-i)
                                                                                         400
       1.0 r
     o 0.8
     *3
     o
     Q.
     O

    £L
t:  0.6
     03

    •-P 0.4
     to
     3


    O 0.2
       O.O1—
        -100
                   Model =  ETD
                    Year =  20
                                                            1.0
O 0.8

O
Q.
S. 0.6
0.

CD

3= 0.4
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp Deposition
O
           o     100
             ANC
                            200    300
                              L-i)
                                         400
   0.2
                                                       O.O1—
                                                        -100
                                                                   Model
                                                                     Year
                         ETD
                         50
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp Deposition.
           0     100    200    300
              ANC (jxeq L"0
                                    400
Figure 10-29.  Comparisons of projected change in ANC under current and decreased deposition
for NE Priority Classes A - E, using ETD and MAGIC.
                                                699

-------
     o
        1.0
       0.8
       0.6
O
O.
O
oZ

Oi

*3 0.4
eg
3
E

O °-2
       0.0
                     NE Lakes
                  Model =  MAGIC
              Priority Class =  A  -  E
                     Year  = 20
 c
_o
V*
 1_
 o
 Q.
 o
                                                            to
                                                            0.8
                                                            0.6
                      	Simulation Year  0
                      	Constant Deposition
                      	Ramp Deposition
                                                         a.

                                                         03
                                                         «= 0.4
E
O
               100
             [SO/-]
                              200
                               L-i)
                                         300
                                                            0.0
                                                                     NE  Lakes
                                                                  Model  =  MAGIC
                                                               Priority  Class  = A - E
                                                                     Year =  50
     	Simulation Year 0
     	Constant Deposition
     	Ramp  Deposition
               100        200
             [SO,2'] (fieq L-1)
                                    300
       1.0
     O 0.8
     O
     Q.
     £ 0.6
    «= 0.4
    <0
    3
    E
    O 0.2
       0.0
                   Model =  ETD
                    Year =  20
                                                            1-Or
                                                          O 0.8
                                                     O
                                                     Q.
                                                     O
                                                            0.6
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp Deposition
               100
             [S042-]
                              200
                               L-1)
                                         300
O
>
3= 0.4
«
3


o
                                                            0.0
                                                                   Model  = ETD
                                                                     Year  = 50
     	Simulation Year 0
     	Constant Deposition
     	Ramp  Deposition
  100
[S042-]
                         200
                          L-1)
                                    300
Figure  10-30.   Comparisons of  projected change  in  sulfate  concentrations  under: current  and
decreased deposition for NE Priority Classes A - E, using ETD and MAGIC.
                                                700

-------
        1.0r
     O  0.8
     O
     Q.
     O
        0.6
Q.

-------
                                                          3rd Quarfile +
                                                          (1.5 x Interquartile Range)*"
                                                          3rd Quart te

                                                          Mean

                                                          Median

                                                          IstQuartle

                                                          IstQuartile
                                                          (1S x Interquartile Range)"


                                                          "Not to exceed extreme value
                                            Constant

                                                8      8
4OO —
350-
300-
25O —
-P 200-
i-i
f[ 150 —
so—
0 —
. -50-
-inn —


























































































                                            Ramped
                                                 O
                                                 ea
§
40O —
350 —
3OO —
25O —
15O —
1OO —
so—
o—
-SO —
-•(nn —












































































































Figure  10-32.  Box and whisker plots of ANC distributions projected using ETD in 10-year intervals
for NE  lakes,  Priority Classes A - E.
                                                   702

-------
                                                           3rd Quarfile +
                                                           (1.5 x Interquartile Range)**
                                                           SrdQuartle

                                                           Mean
                                                           Median

                                                           UtQuarffle

                                                           IstQuartle
                                                           (1.5 x Interquartile Range)**

                                                           **Not to exceed extreme value
                                            Constant
                         250-1



                         200-





                       .3.

                       0~ 100-



                          50-
                                  o
                                  DC
8
OC
CO
cc
                                        s
                                        rr
                                                                         oc
                                             Ramped
        o
        OC
250—   -"-
                         200—
                                                         CO
                                                         oc
                       S"
                       A
                         150-
                        "
                         10(
                          50-
                                                                 oc
                s
                oc
Figure 10-33.   Box  and  whisker plots of sulfate  distributions  projected  using ETD  in 10-year
intervals for NE lakes, Priority Classes A - E.
                                                     703

-------
                                                           3rd Quartite +
                                                           (1.5 x Interquartile Range)"

                                                           3rd Quartite

                                                           Mean

                                                           Median


                                                           IstQuartile

                                                           1st Quartite
                                                           (1.5 x Interquartile Range)"


                                                           "Not to exceed extreme value
                                             Constant

                                                o       <
                                                -
oc
        oc
                 o
                 CO
                 OC
                         oc
8
oc
Figure 10-34.  Box and whisker plots of pH projected using ETD in 10-year intervals for NE lakes,
Priority Classes A - E.
                                                    704

-------

                      o
                                                          3rd Quartile +
                                                          (1.5 x Interquartile Range)**
                                                          3rd Quartile
                                                          Mean
                                                          Median
                                                          1st Quartile
                                                          1st Quartile
                                                          (1.5 x Interquartile Range)**

                                                          "Not to exceed extreme value
                                           Constant
500 —

400-
300-
200-
100-
0—
inn







o
£

















o
8S



•












s
g=



'












o
CO
§E

















o
•^3-
£



•













o
m
§E










                                            Ramped
500-
400-
^ 300 —
i"
i

-------
                                                             3rdQuartile +
                                                             (1.5 x Interquartile Range)"
                                                             3rd Quart le

                                                             Mean

                                                             Median

                                                             1st Quart le

                                                             IstQuarffle
                                                             (1.5 x Interquartile Range)"


                                                             "Not to exceed extreme value
                            250-1
                            200-
                             150H
                           *
                             50-
                                     o
                                     OC
                                               Constant
•r-       C4      CT
cc.       cc      cc
>->->-
                                                10
                                                DC
                                                Ramped
                          cr
                          o

                          A
 250-



 200-



.150-



 100-



  50-



   0
                                     o
                                     CC
Figure 10-36.  Box and whisker plots of sulfate distributions in  10-year intervals using MAGIC for
NE lakes, Priority Classes A - E.
                                                    706

-------
                                                           3rd Quartite +
                                                           (1.5 x Interquartile Range)"
                                                           3rd Quartite

                                                           Mean

                                                           Median

                                                           1st Quartite

                                                           1st Quartila
                                                           (1.5 x Interquartile Range)**


                                                           "Not to exceed extreme value
                                             Constant



c
§


3
C


C
•F
C


3
C





?
[


3 S
c c
>

1
§
c





<
§


?
c


S
c


? e
c oc
>
-

                           7—   __    	     __     	
                           5-
                                              Ramped

                                              8      §
                          7—
                          5-
                                       tr
Figure  10-37.  Box and whisker plots of pH in 10-year intervals using MAGIC for NE lakes, Priority
Classes A - E.
                                                   707

-------
                               Northeast Lakes
                              Priority Class A - E
                                 Model = ETD
                             Deposition = Constant
500-
400-
co
CD
«j 300-
"o
CD
E 200-
100-
o-





Jl
1 	 pe^u 	 poo.





1
1

1






\

sssspsssssssss^

r

\


1 rl
i ™, !„ ,-, „„
1 1 P 0 .n. ,. If.
|:i_Li_i_i-i_i_li.







i

-40-15 10 35 60 85110135160185210235260285310335360385410
ANC^eqL-1) D
E3
ETD Year 10
ETD Year 50
                               Northeast Lakes
                              Priority Class A - E
                                 Model = ETD
                      Deposition = Ramped 30% Decrease
     500i
     400-

-40






-1


1
i
Kl
^
5






1

i
%
»
g
»
%
%
K
W
<&
ty\
0






3

1
i
i
&#
#
5






6

i
1
i
??
^
0






8

%
i
i
i
^
ft
5

r- 1




11

1
1
P
%
P
0


™ "Pr-P r
11 I I

yfa % Y% y/. % YJ
135 160 185 210 235 260 2
ANC^eqL'1)


1 F
l%l IS8i 1 !j!^ B33 \ \ w&
V& xa\ va Ws\ \ KM
85310335360385 410
O ETD Year 10
G9 Year 50 Ramped
Figure 10-38. ETD ANC distributions at year 10 and year 50 for NE lakes, Priority Classes A - E,
under current and decreased deposition.
                                         708

-------
                               Northeast Lakes
                              Priority Class A - E
                                Model = Magic
                            Deposition = Constant
 tn
 CD
    500i
    400-
    300-
 CD
 "i  200
    100-
                          ^
                                                                            .
-40-15  10  35  60 85 110135160185210235260285310335360385410
                          ANCOieqL'1)
                                                               D MAGIC Year 10
                                                               H MAGIC Year 50
                               Northeast Lakes
                              Priority Class A - E
                                Model = Magic
                      Deposition = Ramped 30% Decrease
 CO
 CD
    5001
    400-
    300-
"o
CD
e 200-
100-


1
p-1
%
g
V
%
V

\
ssssssss


BSSSSSSS3
1

\

sssssssssssss
I
-40-15 10 35 60 85110135160185
ANC(jieqL-




i
210235260



I
285
II
310

it
1

^
335360385410
D MAGIC Year 10
H Year 50 Ramped
Figure 10-39.  MAGIC ANC distribution at year 10 and year 50 for NE lakes, Priority Classes
A - E, under current and decreased deposition.
                                          709

-------
                                Northeast Lakes
                               Priority Class A - E
                                  Model = ETD
                              Deposition = Constant
500-
400-
(n
CD
«j SCO-
'S
i —
CD
E 200-
100-
o-



I




ijjjj


I




I
r-]

I
-




I


I
ta ffl




K p0 P par-pa

30 40 50 60 70 80 90 100110120130140150160170180190200210220230


.._._. 2-i .,.
lbU4J(n.eqL )
D ETD Year 10
E3 ETD Year 50
                                Northeast Lakes
                               Priority Class A - E
                                  Model = ETD
                       Deposition = Ramped 30% Decrease
     5001
400-
co
« SOD-
'S
CD
E 200-
100-
o-'

B?S




' — i 	
s
55

i



r



i
i








p

30 40 50 60









1
%

»
K
\
&




-

yy
i
%



t*x 	 , 	 !
»
K
—










lUi fl
f**j " ' r^ r I i i i i ' <
70 80 90 100110120130140150160170180190200210220230
[SOflOieqL-1) D ETD Year 10
tS Year 50 Ramped
Figure 10-40.  ETD sulfate distributions at year 10 and year 50 for NE lakes, Priority Classes
A - E,  under current and decreased deposition.
                                          710

-------
                              Northeast Lakes
                             Priority Class A - E
                               Model = Magic
                           Deposition = Constant
500-
400-
CO

-------
                                          "1
                                                                        "1
population had initial ELS-I ANC < 100 /*eq L", ranging from -43 to 86 ^eq L". The watersheds were
undisturbed, based on chloride concentrations (See Section  10.5.7), and in general had  positive sulfur
retention.

10.11.1.3.1  Deposition scenarios -

      All three models simulated comparable changes in ANC,  sulfate, and pH over the 50-year period
assuming  current deposition  or  a  30 percent  deposition decrease  (Figures  10-4,2  through 10-44).
Confidence intervals computed for each of the projections are included in Appendix A.3. Projections for
all three models were comparable at the lower ANC concentrations (i.e., ANC < 25 p.eq  L"1) but deviated
at higher ANC  concentrations.  Projections  from  MAGIC deviated  the  most at  the higher  ANC
concentrations but  were still  within the uncertainty  bounds  about the  projections  (Appendix A.3).
Projected ANC values  were  similar,  however,  among all three models for lakes with ANC in the lower
quartile of the population (Table 10-17). Lower quartile values of ANC projected using ETD, ILWAS, and
MAGIC with current deposition after 50 years were 2.6, 5.9, and 1 1 .9 /zeq L"1, respectively.  Lower quartile
values for sulfate projected using the ETD, ILWAS, and MAGIC models after 50 years of current deposition
were  70.4,  80.6, and  69.4 /*eq L"1,  respectively.  Assuming a 30 percent deposition  decrease, lower
quartile values of ANC  projected using ETD, ILWAS, and MAGIC after 50 years were 10.5,  19.6, and 25.1
neq L"1, respectively.  Sulfate concentrations projected under similar conditions using ETD, ILWAS, and
MAGIC were 51.7, 66.5, and 53.2 /*eq L"1, respectively.  These values were all within the  uncertainty
bounds for the projections (Appendix A.3).

      The projected pH values were similar at the higher pHs but deviated at  low pH with the greatest
difference between  ILWAS and the  other two  models.  The projected median  pH  after 50 years with
ETD,  ILWAS, and MAGIC under current and decreased deposition was 6.4, 6.1, 6.4 and 6.4, 6.5, 6.5,
respectively.  The projected  lower quartile pH values after 50 years with ETD, ILWAS, and MAGIC under
current and decreased deposition were 5.6, 4.9, 5.9 and 6.0, 5.6, 6.2, respectively. At the lower pHs, the
projected ILWAS pH was from 0.5 to 1.0 pH unit less than projected by the other two models.

      Changes in surface water chemistry under different deposition scenarios were compared within
models (Figures 10-45 through 10-47).   Differences in median  ANC concentrations  between current
deposition and a 30 percent  deposition decrease  projected using ETD, ILWAS, and MAGIC after 50 years
were  30.9 versus 31.5 (+0.6), 30.3 versus 48.9 (+18.6), and 64.9 versus 70.6 (+5.7) neq L"1, respectively.
Differences in median  sulfate  concentrations  between current  deposition  and a 30 percent deposition
decrease projected  using ETD, ILWAS, and MAGIC after 50 years were 106.5 versus 76.2 (-30.3), 103.6
versus 82.5 (-21.1), and 95.2 versus 67.6 (-27.6) /*eq L"1, respectively. Differences in median pH between
current and decreased deposition projected using ETD,  ILWAS, and MAGIC  after 50  years were 6.43
versus 6.44 (+0.01), 6.09 versus 6.53  (+0.04), and 6.40 versus 6.54 (+0.14), respectively.

      All three models indicated  northeastern  watersheds were near sulfate steady  state or near zero
percent net sulfur retention after 50  years for scenarios of either current or decreased deposition (Table
10-17).
                                               712

-------
            to
          o
          Q.
            0.6
         o
                       NE Lakes
                 Priority Class = A & B
                  Deposition •= Constant
                       Year = 0
                   0    100
                     ANC
                                   300
                                        400
                                          NE Lakes
                                    Priority Class * A & B
                               Deposition = Ramp 30% Decrease
                                          Year •= 0
                               to
                             a.
                             £ 0.6
                            a.
                                                          ~ 0.4
                                                          a
                                                             0.0
                                                              -wo
                                      0    100   200
                                        ANC (|ieq L-I)
            to
          O 0.8
          " 0.6
          a.
          o>
          £
          £
          Z3

          o
            0.0
             -100
                  Deposition = Constant
                       Year - 20
                   0    100   200   300
                     ANC (jieq L-I)
                                         400
                               Deposition  =  Ramp  30%  Decrease
                                          Year -  20
                               to
                                                           O 0.8
                             o
                             o.
                             I
                             u>
                                                             0.6
                            o
                               OJO>	
                                -100
	  Phase 1
	MAGIC
	ETC
...........  I.WAS
                                      0    100    200   300
                                        ANC (jieq L-I)
                                                                                          400
            to
            0.8
          £ 0.6
          Q.
          o
                  Deposition •= Constant
                       Year -  50
            o.o i—
             -100
	  Phase 1
	MAGIC
	ETD
......4».<  BWAS
                   0    100    200   300
                     ANC ((ieq L-I)
                                         400
                                Deposition - Ramp 30% Decrease
                                           Year - 50
                                to
                                                           O 0.8
                                                             0.6
                                                             0.4
                                                           J
                               0.0
                                -wo
                                       0     WO   200    300
                                         ANC (jieq  Li)
                                                            400
Figure 10-42.  Comparison of ANC projections using ETD, ILWAS, and MAGIC for NE lakes, Priority
Classes A and B, under current and decreased  deposition.
                                                      713

-------
            to
          S. 0.6
          o.
          S3 0.4
          a
          :a

          o
            0.0
                       NE Lakes
                 Priority Class •= A  & B
                  Deposition = Constant
                       Year - 0
                      WO       200
                    ISO,4-] (jieq L-1)
                                        300
             NE Lakes
        Priority Class = A  & B
   Deposition = Ramp 30% Decrease
             Year = 0
   to
                                                            0.6
i= 0.4
jg

|

O
                                                            0.0
             100
           ISO,*-]
                                                                               200
                                                                                        300
            to
          O 0.8

          o
          Q.
          8 0.6
          Q.

          CD

          5 0.4
          B

          I
            0.0
                  Deposition = Constant
                       Year -  20
   Deposition = Ramp 30%  Decrease
              Year  - 20
   to
 § 0.8

 o
 Q.
 £ 0.6
 D.
 a

 I

 O
                       WO       200
                     [SO,*] (jieq L-1)
                                                             OJ)
             100       200
           [S04»-]  (jieq  L-1)
                   Deposition  =  Constant
                        Year - 50
    Deposition -  Ramp 30% Decrease
              Year - 50
                       WO       200       SOO
                     ISO,*] (jieq LI)
Figure 10-43.   Comparison  of sulfate projections using ETD, ILWAS, and MAGIC for NE lakes,
Priority Classes A and B, under current and decreased deposition.
                                                    714

-------
              tOr
           5 0.8
           O
           o.
              0.6
              0'4
           o
              0.0
                         NE Lakes
                   Priority Class «= A & B
                    Deposition <= Constant
                         Year <= 0
               4.0 4£  5.0 SJS  6.0 6.5  7.0 7.5
                             pH
                                          NE  Lakes
                                    Priority Class  - A  & B
                               Deposition = Ramp 30% Decrease
                                          Year «=  0
                               to
                                                               0.6
                                                            I
                                                               0.0
                                                                4.0  4JS  5.0 5.5  6.0 6.5  7.0 7S  6.0
           O
                    Deposition
                         Year
Constant
20
               4.0  4.5  5,0  5.5  6.0 8.5  7.0  7.5  8.0
Deposition = Ramp 30% Decrease
           Year - 20
to
                                                             o
                                                             a.
                                                             8 0.6
                                                             a.


                                                             I-
                                                             3
                                                             O
                                                               OJO
                                                                4.0  4.5 SSI  &S  6.0  6.5  7.0  7J  8.0
           o
                    Deposition -  Constant
                         Year -  SO
               4.0  4.S 5.0  S.5 64)  6.5 7.0  7.5 (.0
                               Deposition •=  Ramp  30% Decrease
                                          Year -  50
                               to
                                                             8 0.6
                                                             a.
                                                             eg
                                                               0.4
                                                               0.2
                                                                OJ1
                                4.0  44! 6.0  5.5  6.0  6.5  7J»  7£  8.0
                                              pH
Figure 10-44.  Comparison of pH projections using ETD, ILWAS, and  MAGIC for NE lakes, Priority
Classes A and B, under current and decreased  deposition.
                                                      715

-------
Table 10-17. Descriptive Statistics  for  Projected ANC, Sulfate,  Percent  Sulfur
Retention, and Calcium Plus Magnesium for NE Lakes in Priority Classes A and B
Using ETD, ILWAS, and MAGIC for Both Current and Decreased Deposition
Model

All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Mean

ANC
0
32
44
57
20
30
43
57
50
30
39
56
so *-
o 1
90
118
114
20
110
118
105
50
110
118
103
BhL
0
5.55
5.07
5.39
20
5.50
5.04
5.41
50
5.48
5.01
5.40
Std.
Dev.



33
52
52

33
53
51

34
53
51


32
52
46

44
53
42

45
53
41


0.64
0.95
0.79

0.66
0.98
0.78

0.66
0.99
0.79
Min.
Current


-43
-67
-21

-49
-70
-21

-52
-74
-22


34
42
50

53
44
47

54
44
46


4.36
4.15
4.47

4.31
4.13
4.49

4.29
4.11
4.48
P_25
Deposition


6
8
15

1
8
14

3
6
12


67
83
78

70
81
71

70
81
69


5.83
4.93
5.83

5.55
4.86
5.95

5.63
4.91
5.94
Median



26
36
66

30
33
67

31
30
65


81
102
106

103
103
96

107
104
95


6.36
6.02
6.40

6.42
6.24
6.40

6.43
6.09
6.40
P_75



59
86
84

61
86
84

63
82
82


113
136
126

147
137
126

154
138
127


6.71
6.75
6.79

6.72
6.78
6.79

6.73
6.74
6.79
Max.



90
159
175

105
161
173

106
161
171


185
266
246

222
267
221

216
264
215


6.89
7.26
6.97

6.96
7.27
6.97

6.96
7.27
6.97
                                                                 continued
                             716

-------
Table 10-17. (Continued)

Model
All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC

Mean
Std.
Dev.

Min.

P_25

Median

P_75

Max.
% S Retention
0
20
1
1
20
6
1
9
50
7
1
11
ILWAS vs. MAGIC. Ca +
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
0
122
131
20
122
124
50
119
121

21
20
9

11
19
7

9
19
6
Mq

40
51

42
52

42
53

-17
-33
-14

-14
-30
-5

-7
-30
-1


45
41

43
40

41
38

6
-15
-9

2
-14
3

0
-14
6


102
99

103
90

102
86

27
0
3

4
3
10

7
3
11


119
122

116
114

111
111

33
12
7

12
12
14

12
11
17


141
145

143
139

143
135

69
63
19

30
64
21

27
63
21


204
281

220
280

211
280
Delta Ca+Mg
ILWAS
MAGIC

All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
-3
-6

ANC
0
32
44
57
20
33
50
61
50
38
56
64
5
3
30%


33
52
52

33
52
51

34
50
50
-11
-12
Decrease


-43
-67
-21

-44
-51
-20

-40
-40
-18
-7
-9
-3
-5
0
-4
8
-2
in Deposition


6
8
15

4
12
17

11
20
25


26
36
66

36
36
69

32
49
71


59
86
84

63
91
87

68
93
88


90
159
175

108
161
178

111
163
179
                                                                 continued
                             717

-------
Table 10-17. (Continued)
Model
All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
All models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
All Models.
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Model Year
ETD
ILWAS
MAGIC
Mean
SO,2'
0 1
90
118
114
20
gg
107
g7
50
80
gi
77
£H
0
5.55
5.07
5.39
20
5.59
5.18
5.52
50
5.67
5.37
5.67
% S Retention
0
20
1
1
20
-5
-12
-5
50
3
-9
5
Std.
Dev.


32
52
46

39
47
38

32
38
31


0.64
0.95
0.79

0.63
0.93
0.73

0.58
0.86
0.66


21
20
9

14
22
9

13
21
9
Min.


34
42
50

44
37
45

39
31
38


4.36
4.15
4.47

4.35
4.25
4.52

4.40
4.36
4.58


-17
-33
-14

-37
-51
-24

-38
-40
-11
P_25


67
83
78

62
76
66

52
67
53


5.83
4.93
5.83

5.68
5.11
6.11

6.00
5.57
6.17


6
-15
-g

-14
-26
-12

-3
-25
-1
Median


81
102
106

95
96
87

76
83
68


6.36
6.02
6.40

6.49
6.41
6.48

6.44
6.53
6.54


27
0
3

-6
-10
-4

4
-14
6
P_75


113
136
126

135
124
121

110
113
104


6.71
6.75
6.79

6.73
6.84
6.81

6.76
6.91
6.82


33
12
7

5
1
1

9
2
11
Max.


185
266
246

186
238
202

162
190
157


6.89
7.26
6.97

6.97
7.30
6.98

6.98
7.32
6.99


69
63
19

27
60
14

24
60
20
                                                                   continued
                             718

-------
Table 10-17. (Continued)
Model
Std.
Mean Dev. Min. P_25
Median P 75
Max.
ILWAS vs. MAGIC. Ca + Ma
Model Year 0
ILWAS
MAGIC
Model Year 20
ILWAS
MAGIC
Model Year 50
ILWAS
MAGIC
Delta Ca+Mg
ILWAS
MAGIC

122
131

120
121

112
108

-11
-18

40
51

42
52

41
53

7
g
                                45
                                41

                                41
                                39

                                35
                                35

                               -32
                               -54
102
 99

100
 85

 89
 74

-13
-23
119
122

115
112

106
 98

-11
-15
141
145

140
138

127
125

 -4
-13
204
281

213
279

201
274

  0
 -5
                             719

-------
                          NE  Lakes
                       Model  =  MAGIC
                   Priority  Class = A &  B
                         Year •= 20
             tO
           O 0.8
             0.6
          O
0.01	
 -100
              	Simulation Year 0
              	Constant Deposition
              	Ramp Deposition
                     0    100    200
                       ANC (jieej  L-
                                      300
                                            400
                                                                 NE  Lakes
                                                              Model  = MAGIC
                                                          Priority Class *  A &  B
                                                                 Year «= 50
                                                                  10
                                                                  0.8
                                                  o.
                                                  £ 0.6
                                                  0.
                                                                 0.0'—
                                                                  -MO
      	Simulation Year 0
      	Constant Deposition
      	Ramp  Deposition
                                                            0    100    200
                                                              ANC (jieq  L-
                                                                                          300
                                                                                                400
             to
           O 0.8
           o
           O.
           8 0.6
          Q.

           <0

          a 0.4
           ta
          O
             0.0"—
              -100
                        Model
                         Year
                    ETD
                    20
              	Simulation  Year 0
              	Constant Deposition
              —	Ramp Deposition
                     0     100   200    300
                       ANC (jieq L-1)
                                           400
                                                                  tOr
                                                  O 0.8
                                                  i.
                                                  o
                                                  Q.
                                                  £ 0.6
                                                  a.

                                                  
-------
           10
         O 0.8
S  0.6
o-


-------
            1.0
         O 0.8
 O

 £ 0.6
 0-

 
         §0.4
         2
         3

         o o-2
           0.0
                       Model  = ETD
                        Year - 20
	Simulation Year 0
---- Constant Deposition
	Ramp Deposition
             4.0  4.5  5.0 5.5  6.0  6.5  7.0  7.5  8.0
                            PH
                                                                   to
                                                                O 0.8
                                                                  0.6
                                                       S= 0.4
                                                       cj
                                                       3


                                                       O
                                                                  0.0
                                              Model  = ETD
                                               Year -  50
                       	Simulation Year  0
                       	Constant Deposition
                       	Ramp Deposition
                                    4.0  43  SJI  5.5 6.0  6.5  7.0  7.5  8.0
                                                   PH
            tOr
£ 0.8

|

£ 0.6
Q.
ffl
S 0.4
S

i
o o-2
           0.0
                      Model  - ILWAS
                        Year - 20
                                	Simulation Year 0
                                ---- Constant Deposition
                                —..... Ramp Deposition
             4.0  4.5 S.O 5.5  6.0  6.5  7.0  7.S  8.0
                            PH
                                                                   to
                                                                o
                                                                o.
                                                                £ 0.6
                                                                Q.
E
O
                                                                  0.0
                                             Model  - ILWAS
                                               Year -  50
                                                        	Simulation Year 0
                                                        	Constant Deposition
                                                        	Ramp  Deposition
                                    4.0  4.5  5.0  5.5 6.0  6.5  7.0  7.5  8.0
                                                   pH
Figure 10-47.  Comparison of pH projections under current and decreased deposition, for NE lakes,
Priority Classes A and  B, at year 20 and  year 50  using ETD,  ILWAS, and  MAGIC.
                                                          722

-------
      Projections of the number of lakes currently not acidic that might become acidic in the next 50
years under current deposition and a 30 percent decrease in deposition using ETD, ILWAS, and MAGIC
were 25 (5 percent), 74 (17 percent), 75 (17 percent) and  25 (5 percent), 25 (5 percent), and 50 (11
percent), respectively.  Projections of the number of currently acidic lakes that might chemically improve
under current deposition and a 30 percent deposition decrease  using ETD, ILWAS,  and MAGIC were
27(36 percent), 25 (32 percent), 0 (0 percent), and 52  (68  percent), 25  (32 percent), 13 (16 percent),
respectively.
10.11.1.3.2  Rate of change of ANC, sulfate, and  pH over 50 years -

      The changes in ANC, sulfate, and pH projected over the next 50 years using the three models are
shown in box and whisker plots (Figures 10-48 through 10-56).  The relative change in the median ANC
projected  using all three models and assuming current deposition levels was less than 0.1  yueq L"1 yr"1.
The rate of change of median ANC for a 30 percent  deposition decrease was about 0.3 /*eq L"1 yr"1 for
ILWAS and MAGIC  and remained less than  0.1 ^eq L"1 yr"1 for  ETD.  These rates,  while  three times
greater for ILWAS and  MAGIC, are still small and indicate little change in ANC over the 50-year period
under either deposition scenario.
      The rates of change projected for median sulfate concentrations under current deposition ranged
from -0.2 fieq L"1 yr'1 for MAGIC to less than 0.1 /zeq L"1 yr"1 for ILWAS to 0.4 ^teq  L"1 yr"1 for ETD.
Assuming a 30 percent deposition decrease, these rates of change in median sulfate  concentrations
changed sign and magnitude, varying from -0.1 /j.eq L"1 yr"1 for ETD to -0.4 ^eq L"1 yr "1 for ILWAS and
-0.8 /ieq L'1 yr"1 for MAGIC.
      The change in median pH projected over 50 years under current deposition ranged from 0.0 for
MAGIC to +0.07 for both ETD and MAGIC.  The change in median pH projected over 50 years under
decreased deposition ranged from 0.1 for ETD to  0.15 for MAGIC and 0.5 for ILWAS.  The variance in
pH was greatest for ILWAS and varied overtime (Figure 10-56).

      There also was no indication that the  rates of change in ANC or sulfate concentrations were
functions of the initial ELS-I  ANC concentration for either deposition scenario (Table 10-17).  Histograms
of projected change in median ANC and sulfate concentrations over 40 years using all  three models
indicate a relatively uniform change among lakes regardless of their initial ANC concentrations (Figures
10-57 through 10-62).

10.11.2  Southern Blue Ridge Province

10.11.2.1   Target Population Projections Using MAGIC
      An estimated 1323 streams in the SBRP target population were simulated using MAGIC. This target
population included both disturbed and  undisturbed watersheds based on chloride concentrations; all
watersheds had positive sulfur retention.  Three streams (which had NSS Pilot Survey ANC  > 400 jueq
L"1 ) were excluded subsequently from this target population, although they were simulated by MAGIC.
The MAGIC  projections indicated that ANC concentrations in these three systems essentially did not
                                              723

-------
                                                             3rd Quartile +
                                                             (1.5 x Interquartile Range)"
                                                             3rd Quart le

                                                             Mean

                                                             Median

                                                             1st Quartile

                                                             1st Quartile
                                                             (1 £ x Interquartile Range)"


                                                             "Not to exceed extreme value
                                              Constant
                            150-1
                                            cc
o       o
Ol       CT
CC       CC
>-       >-
                                                                   cc
100-
Vj 50-
cr
-50-


_




_!












^_




_J







	 —



                                               Ramped
                            150-1
                            100-
                         1

                             50-
                            -50-
                                    o
                                    CC
                                           cc
                                                   cc
                                                                   cc
                                                                           cc
Figure 10-48. Box and whisker plots of ANC distributions in 10-year intervals projected using ETD
for NE lakes, Priority Classes A and B.
                                                   724

-------
                                                              3rdQuartile +
                                                              (1.5 x Interquartile Range)**
                                                              3rd Quarfile

                                                              Mean

                                                              Median

                                                              IstQuartikj

                                                              lEtQuarfle
                                                              (1.5 x Interquartile Range)**


                                                              "Not to exceed extreme value
Constant

     S      8
                                                                           s
200 —
150—
	 IOC-
S'
.5: SO —
o—
-50-
-1OO —



































































                                               Ramped
20O—
150 —
100 —
50-
o—
-50 —
1OO —






























>-










>-









Figure  10-49.  Box and whisker plots of ANC distributions in 10-year intervals  projected using
ILWAS  for NE lakes, Priority Classes A and B.
                                                  725

-------
                          150-1
                          100—
                           50-
                            o—
                           -50-
                          -100-
                                                           3rdQuartilo-t-
                                                           (1 £ x Interquartile Range)"
                                                           3rd Quarfte

                                                           Mean

                                                           Median

                                                           I6tduartilo

                                                           IstQuartile
                                                           (1.5 x Interquartile Range)"


                                                           "Not to exceed extreme value
                                             Constant

                                                 8
                                                 oc
8
DC
                                              Ramped
                           150-1
                           100-
                         I
                        o
                            50-
                             0-
                           -50-
                          -100-
Figure  10-50.  Box and whisker plots of ANC distributions in 10-year intervals projected  using
MAGIC for NE lakes, Priority Classes A and B.
                                                   726

-------
                                                                 3fdQuart8e +
                                                                 (1.5 x Interquartile Range)"

                                                                 3rd Quart !e

                                                                 Mean

                                                                 Median


                                                                 IstQuartite

                                                                 IstQuart'le
                                                                 (1.5 x Interquartile Range)"



                                                                 "Not to exceed extreme value
                                                  Constant
                                                       s
                               250-
                               200-
                               isoH
                                50-
                                        DC
                                               CC.
        o
        CO

EC      CC
>-      >-
                                                                       CC
O
10

OC
                                                   Ramped
                               250-1
                               200-
                            -i 150-
                             y
                            .3.

                            O'IOOH
                                50 —
                                       o
                                       CC
                                               DC
O       O
CM       CO

EC       EC
>-       >-
                                                                       DC
to

-------
                                                                3rd Quartile +
                                                                (1.5 x Interquartile Range)"
                                                                3rd Quartile
                                                                Mean
                                                                Median
                                                                Ut Quartile
                                                                1st Quartile
                                                                (1.5 x Interquartile Range)"

                                                                "Not to exceed extreme value
                                                 Constant
                               250—1
                               200-
                               150-
                            g 100
                                50-
                                       o
                                       rr
                                               DC
                                                      DC
        S
        DC
                                                                      DC
                                                  Ramped
                              250—1
                              200 —
                            Jj 150 —
                            1
                            g'tooH
                               50-
                                       o
                                       ss
DC
        DC
                       DC
Figure 10-52.  Box and whisker plots of sulfate distributions in 10-year intervals projected using
ILWAS for NE lakes, Priority Classes A and  B.
                                                    728

-------
                                                               3rdQuartile +
                                                               (1.5 x Interquartile Range)"
                                                               SrdQuartile

                                                               Mean

                                                               Median

                                                               1st Qua/He

                                                               IstQuardle
                                                               (1.5 x Interquartile Range)**


                                                               "Not to exceed extreme value
                                                 Constant
                                                      s
S
                              250—1
                              200 —
                                              rr
                                                      cc
                            .3.
                               50 —
                                                 Ramped
                              250-1
                              200—
                            lj 150 —


                            I
                           c4
                            O 100 —
                            CO
                               50 —
                                                      cc
                                                              cc
Figure 10-53.  Box and whisker plots of sulfate distributions in 10-year intervals projected using
MAGIC for NE lakes, Priority Classes A and B.
                                                    729

-------
                                                            3rd Quartite +
                                                            (1.5 x Interquartile Range)**
                                                            3rd Quartite

                                                            Mean

                                                            Median

                                                            IstQuartile

                                                            IstQuartile
                                                            (1.5 x Interquartile Range)**


                                                            "Not to exceed extreme value
                                             Constant
                        56-
                          5—
                                                         o
                                                         co
                                                         oc
                                                                 cc
                                              Ramped
                          8—i
                          5-
                                        oc
                                                                 tr
                                                                         8
                                                                         DC
Figure 10-54.  Box and whisker plots of pH distributions in 10-year intervals projected using ETD
for NE lakes, Priority Classes A and B.
                                                   730

-------
                                                           3rd Quartile 4-
                                                           (1.5 x Interquartile Range)"
                                                           3rd Quartile
                                                           Mean
                                                           Median

                                                           1st Quartile
                                                           Istduartik)
                                                           (1.5 x Interquartile Range)"

                                                           "Not to exceed extreme value
                                             Constant
                       £<
                          5-
                                o
                                DC
                                        DC
         O
         CM
         DC
         §
         DC
         o
         in
         DC
                                             Ramped
                         8-1
                         5-
                               ?•
§=
§=
g
£
§=
        8
        rr
Figure 10-55.  Box and whisker plots of pH distributions in 10-year intervals projected using ILWAS
for NE lakes,  Priority Classes A and B.
                                                  731

-------
                                                            3rd Quartile +
                                                            (1.5 X Interquartile Range)**

                                                            3rd Quartile


                                                            Mean

                                                            Median


                                                            1st Quartile


                                                            1st Quartile
                                                            (1.5 x Interquartile Range)"



                                                            "Not to exceed extreme value
                                             Constant
                          7-
                          5-
                                o
                                oc
                                       CC
                                       >-
DC
       o
       CO

       OC
              rr
                          8—i
                          7-
                          5-
                                              Ramped
                                oc
       o
       eo
       DC
                                                                   O
                                                                   LO
                                                            DC
                                                                          O
                                                                          o
                                                                          DC
Figure 10-56.  Box and whisker plots of pH distributions in 10-year intervals projected using MAGIC
for NE lakes,  Priority Classes A and  B.
                                                    732

-------
        200
CO

I
CD
"E
•^
z
100-
                               Northeast Lakes
                             Priority Class A - B
                                Model = ETD
                            Deposition = Constant
                            I
              -40    -15    10     35    60     85
                                   ANC(jieqL-1)
                                                 110    135
                                                           160
                              Northeast Lakes
                            Priority Class A - B
                               Model = ETD
                     Deposition = Ramped 30% Decrease
     CO
     CD
    _
     CO
     05
    JD

     3
        200
        150-
        100-
         50-

                            i
             -40    -15     10    35    60     85

                                   ANC(neqL-')
                                                 110   135
                                                           160
                                                      D ETD Year 10
                                                      H Year 50 Ramped
Figure 10-57.  ETD ANC population distributions at year 10 and year 50 for current and decreased
deposition.
                                          733

-------
       OJ
       _*:
       03
       o>
       .a
       E
       3
          200
          150-
          100-
           50-
          200
          150-
       U)
      2
       03
          50-
                                 Northeast Lakes
                                Priority Class A - B
                                 Model = ILWAS
                              Deposition = Constant
                i
                I

 -40    -15    10     35     60
                     ANC(jieqL-1)

                  Northeast Lakes
                Priority Class A - B
                  Model = ILVtfAS
        Deposition = Ramped 30% Decrease
                                                  85    110    135   160
                                                             D ILWAS Year 10
                                                             H ILWAS Year 50
                              w.
               i

-40    -15     10    35     60     85    110    135    160

                    ANCOieqL-1)
                                                            D ILWAS Year 10
                                                            B Year 50 Ramped
Figure  10-58.   ILWAS ANC population distributions at year 10 and year 50 for current and
decreased deposition.
                                          734

-------
         co
         2
         cd
         JD
         E
            200
            150-
            100-
             50-
              0,
             2001
             150-
          in
          CD
         _
          CO
          °  100-
          
-------
                                   Northeast Lakes
                                  Priority Class A - B
                                     Model = ETD
                                Deposition = Constant
    s
    I
    z
       2001
       150"
       100"
        50'

_s
 v
                                                  a
           30 40 50 60 70 80 90100110120130140150160170180190200210220230240250260270
                                    [so n i
       2001
                                   Northeast Lakes
                                  Priority Class A - B
                                     Model = ETD
                          Deposition = Ramped 30% Decrease
       150-
0)
1
° 100-
3
e
50-
n .






I
1

\

^
1
1





-

—




I

r
I

I

\ n



1
1


1
r-1
J]

-


an n
           30 ,40 50 60 70 80 90 100110120130140150160170180190200210220230240250260270
Figure  10-60.   ETD sulfate  population distributions at  year  10 and year 50 for current  and
decreased deposition.
                                           736

-------
                                Northeast Lakes
                               Priority Class A - B
                                 Model = ILWAS
                              Deposition = Constant
<;uu
150
CO
Q)
03
_J
° 100-

-------
    200
    150
    100
 JO
                                Northeast Lakes
                               Priority Class A - B
                                 Model = Magic
                             Deposition = Constant

50'
o-




1



*




i
1


-
I





1
—


1


i
i

n
i

i\ Iki . n
       30 40  50 60 70 80 90 100110120130140150160170180190200210220230240250260270
   200
   150
£
CO
                               Northeast Lakes
                              Priority Class A - B
                                Model = Magic
                      Deposition = Ramped 30% Decrease
° 100-
o>
JO
|
2
50-





i


B
^
|




i

I

I



i

in
i


w\ , H n , n
                       P°pU'atJOn distributi°ns at year 10 and year 50 for current and
                                      738

-------
 change over the 200-year simulation.  Including these streams in the discussion distorts the scales of
 the ANC figures because two of these streams  had ANC concentrations  >  1000 /j,eq  L"1.   These
 projections apply only to streams in  the  SBRP target population and do  not necessarily represent
 southeastern stream responses.

 10.11.2.1.1  Deposition scenarios -

      There were significant changes in projected  ANC and sulfate concentrations and in pH over the
 200-year period assuming both current deposition and a 20 percent deposition increase  (Figures 10-63
 and 10-64). The 200-year time frame was selected to assess changes in surface water chemistry as the
 watersheds approach sulfate steady state.  The time frame is for comparative purposes only and does
 not represent expected changes over this time frame.  Median ANC was projected to decrease from 124
 to 78 (-46) fieq L"1 over 200 years under current deposition, and from 124 to  59 (-65) //eq L"1 assuming
 increased  deposition (Table  10-18).  This decrease  is greater than  the  uncertainty bounds  on  the
 projections.  The median sulfate concentration was projected to increase from 37 to 111  (+74)  /zeq L"1
 over the 200-year period under current deposition and from 37 to 133 (+96)  /zeq L"1 over the 200-year
 period assuming increased deposition (Table  10-18).  This increase also is greater than the  uncertainty
 bounds about the projections.  The median  pH was projected to decrease from 7.0 to 6.75 over the 200-
 year period under current deposition and from 7.0  to 6.6 with increased deposition.  The lower  quartile
 pH, however, was projected to decrease from 6.75 to  6.2 over 200 years under current deposition and
 from 6.75 to 5.3  under increased deposition.  A decrease also was projected for median calcium  plus
 magnesium concentration, with a projected  decrease from 115 to 105 (-10) /*eq L"1 for both current and
 increased deposition.

      Median sulfur  retention for the SBRP watersheds at year 0 is about 65  percent.  Median sulfur
 retention projected after 50 years was  about  26 percent and after 200 years was about 5 percent for
 both current and increased deposition.  The  lower and upper  quartile values ranged from  less than 1
 to about  10 percent after 200 years for  both  deposition scenarios,  indicating the watersheds were
 approaching sulfate steady state.

      Projections of  the  number  of streams  that might become acidic after 50,  100, and  200 years
 assuming current deposition were 129 (9 percent), 159 (11  percent), and 203  (14 percent), respectively.
 Projections of the number of streams that might become acidic after 50, 100, and 200 years assuming
 a 20  percent deposition  increase were 159  (11 percent), 159  (11  percent),  and 337  (24 percent),
respectively. For these estimates, the three streams with ANC > 400
population, which represented 1429 streams.
                                                                   L"  were included in the target
10.11.2.1.2  Rate of change of ANC, sulfate, and  pH over 200 years -

      The projected change in ANC and sulfate concentrations and in pH over the 200-year period for
both current and increased deposition are shown in box and whisker plots (Figures  10-65 through 10-
67).   The projected  rates  change in median  ANC  over 200 years assuming current  and increased
deposition were -0.23 peq L'1 yr'1 and -0.32 ^eq L'1  yr'1, respectively.  The relative changes in median
ANC projected to  occur for the first 50 years, from 50 to 100 years, and from 100 to 200 years were
                                              739

-------
        1.0 r
      o 0.8
      O
      Q.
      o
        0.6
     05

     «= 0.4
     ro
     3
     E

     O 0-2
        0.0
         -100
               SBRP  Stream Reaches
                   Model  =  MAGIC
               Priority Class =  A -  E
                     Year = 20
                    	Simulation Year 0
                  	Constant Deposition
                    	Ramp Deposition
                                                             1.0
                                                             0.8
                                                      o
                                                      a.
                                                      o
                                                             0.6
                                                     *3 0.4
                                                     TO
                                                          O
                0     100    200    300
                  ANC (jxeq L-i)
                                     400
   0.2
                                                        0.0
                                                               SBRP  Stream Reaches
                                                                   Model  =  MAGIC
                                                               Priority Class =  A  -  E
                                                                     Year = 20
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp  Deposition
                                                                    100         200
                                                                  tS042-] (jxeq L-1)
                                     300
                     Year
                         50
        1.0
     O  0.8
        0.6
o
a.
o
IX

05
     "~ 0.4
     a
     E
     O
       0.0'—
        -100
                                                             1.0
                                                          o  0.8
                                                          o
                                                          Q.
                                                          o
                                                            0.6
                  	Simulation Year 0
                  	Constant Deposition
                  	Ramp Deposition
•~ 0.4
f2
3
E

O °-2
           0     100
             ANC
                            200   300
                              L-i)
                                         400
                                                        0.0
                                                                          Year  = 50
                                                                            	Simulation Year 0
                                                                            	Constant Deposition
                                                                            	Ramp Deposition
  100
[SO,2']
                         200
                          L-1)
300
Figure 10-63. MAGIC ANC and sulfate projections for SBRP streams, Priority Classes A - E, at year
20, year 50, year 100, and year 200 under current and increased deposition.  (Continued).
                                                 740

-------
      o
      o
      d.
      o
          1.0r
         0.8
         0.6
      05

      V3 0.4
      (0
      O
         0.2
         o.o
          -100
                 SBRP  Stream Reaches
                    Model  = MAGIC
                Priority  Class  =  A  -  E
                      Year  =  100
                    	Simulation Year 0
                    	Constant Deposition
                    	Ramp Deposition
                 o     100    200    300
                    ANC  (jieq  L-i)
                                      400
                                                                1.0
                                                        .0 0.8
                                                        £•
                                                        O
                                                        Q.
                                                        E 0.6
                                                        05

                                                        *=  0.4
                                                        .S*
                                                        3
                                                        E
                                                            O
                                                               0.2
                                                           0.0
                                                                  SBRP Stream Reaches
                                                                      Model =  MAGIC
                                                                  Priority  Class  = A -  E
                                                                        Year =  100
                                                               	Simulation Year 0
                                                               	Constant Deposition
                                                               	Ramp Deposition
                                                           100
                                                         [S04*J
                                                                                  200
                                                                                             300
                     Year  = 200
         to
     O 0.8
        0.6
 O
 Q.
 O
(X

 
                                            ys  0.4
                                            to
                                            3
                                            E
                                           O  0.2
                                                          0.0
                                                                            Year =  200
                                                                              	Simulation  Year 0
                                                                              	Constant Deposition
                                                                              	Ramp Deposition
                                                                      100
                                                                    [SO,*-]
                                                                                      200
                                                                                                 300
Figure  10-63.  (Continued).
                                                   741

-------
                                SBRP Stream Reaches
                                    Model = MAGIC
                                Priority  Class  =  A  - E
                                       Year  =  20
                         1.0
O  0.8
                      o
                      CL
                      O
                      0)
                        0.6
                     ^ 0.4
                     ca
                     3
                     e
                     O 0.2-
                        0.0
                             	Year  0
                             	Constant
                             	Ramp
                          4.0  4.5  5.0   5.5  6.0  6.5   7.0  7.5  8.0
                                           pH
                                      Year  =  50
                                   Year 0
                                   Constant
                                   Ramp
                         4.0  4.5  5.0  5.5   6.0  6.5  7.0   7.5  8.0
                                           pH
Figure 10-64.  MAGIC pH projections for SBRP streams, Priority Classes A - E, at year 20, year 50,
year 100, and year 200 under current and increased deposition. (Continued).
                                           742

-------
                               SBRP  Stream  Reaches
                                   Model  =  MAGIC
                               Priority Class =  A  - E
                                     Year =  100
1.0


_
O 0.8

o
Q.
8 0.6
Q.

0)
>
+3 0.4
CO
ZJ
E

\r 	 r*

	 Constant .,-'

Kamp ?
fi
;' 1
!/
if
i ;

.. J
h .'•' 1
-: ' J
: ' /*

                     O
                        0.0
                         4.0  4.5  5.0  5.5  6.0   6.5  7.0  7.5  8.0
                                          PH
                                     Year = 200
                        1.0
                     O  0.8
                     o
                     Q.
                     o
                        0.6
                    Q_

                     0}

                    '^ 0.4
                     E
                    O 0.2
                       0.0
Year 0
Constant
Ramp
                                  -i	1	1—i—i	1	i	i
                         4.0  4.5  5.0  5.5  6.0  6.5  7.0   7.5  8.0
                                          pH
Figure 10-64.  (Continued).
                                           743

-------
Table 10-18.  Descriptive Statistics of Projected ANC, Sulfate, and Percent
Sulfur  Retention, and  Calcium and  Magnesium  for SBRP  Streams in Priority
Classes A-E Using MAGIC for Both Current and Increased Deposition
Model

MAGIC All.
Yr 0
Yr 20
Yr 50
Yr 100
Yr 200
MAGIC All.
Yr 0
Yr 20
Yr 50
Yr 100
Yr 200
MAGIC All.
Yr 0
Yr20
Yr 50
Yr 100
Yr 200
MAGIC AH.
Yr 0
Yr 20
Yr 50
Yr 100
Yr 200
MAGIC All.
Yr 0
Yr 20
Yr 50
Yr 100
Yr200
Mean

ANC
139
131
119
101
78
so.2-
4&
60
78
98
112
PH
6.87
6.68
6.11
5.68
5.52
Std.
Dev.


94
96
99
97
85

26
29
31
29
29

0.28
0.38
0.59
0.72
0.79
Min.
Current

20
10
-3
-13
-19

12
13
16
26
70

6.23
5.92
5.20
4.77
4.61
P_25
Deposition

71
63
52
40
19

29
35
57
85
87

6.76
6.71
6.63
6.50
6.19
Median


124
122
112
100
78

37
54
75
97
111

6.99
6.99
6.96
6.91
6.80
P_75


156
144
125
111
91

68
76
99
113
123

7.10
7.06
7.01
6.95
6.86
Max.

————___ ™ •_
510
507
509
466
371

99
118
144
173
209

7.60
7.59
7.59
7.55
7.46
% S Retention
59
49
34
18
7
Ca + Ma
131
136
139
141
129
21
23
24
17
8

73
73
74
79
74
24
18
5
1
-2

50
53
57
57
49
35
28
21
10
1

85
86
102
94
87
65
48
27
14
6

115
121
119
108
105
78
70
51
18
10

140
145
162
155
133
91
89
85
76
28

370
370
382
438
385
                                                          continued
                                  744

-------
 Table 10-18. (Continued)
 Model    Mean
Std.
Dev.
Min.
P 25   Median
                                                       P 75
                                           Max.
                          20% Increase in Deposition
 MAGIC All. ANC
 Yr 0       139       94       20       71
 Yr 20      128       96        8       62
 Yr 50      111      100      -10       46
 Yr 100      87       95      -18       29
 Yr 200      60       82      -22        0

 MAGIC All. SO/'
 Yr~0        48       26       12       29
 Yr 20       64       31       15       37
 Yr 50       94       38       19       68
 Yr 100     123       37       32      102
 Yr 200     137       33       97      105

 MAGIC All. DH
 Yr 0         6.87     0.28      6.23      6.76
 Yr 20        6.61      0.41      5.83      6.70
 Yr 50        5.79     0.70      4.86      6.57
 Yr 100       5.53     0.76      4.63      6.37
 Yr 200       5.28     0.91      4.52      5 42
                          124
                          121
                          105
                           83
                           59
                          37
                          57
                          92
                          125
                          133
                           6.99
                           6.99
                           6.93
                           6.82
                           6.68
                          156
                          141
                          115
                           92
                           73
                           68
                           81
                          118
                          146
                          148
                            7.10
                            7.05
                            6.97
                            6.88
                            6.77
                            510
                            507
                            507
                            443
                            344
                            99
                            128
                            175
                            228
                            251
                             7.60
                             7.59
                             7.59
                             7.54
                             7.43
MAGIC All. % S Retention
Yr 0        59       21
Yr 20       52       21
Yr 50       34       24
Yr100      15       16
Yr 200       5        6
MAGIC AH. Ca + Mo
Yr 0       131        73
Yr 20      137        73
Yr 50      145        75
Yr 100     147        84
Yr 200     130        75
        24
        25
         4
         0
         -2
        50
        54
        60
        60
        48
         35
         32
         21
          6
          0
         85
         86
        104
         98
         84
          65
          51
          26
          11
           5
         115
         122
         123
         109
         106
 78
 73
 51
 18
  9
140
145
171
174
135
 91
 89
 86
 76
 17
370
373
394
472
388
                                   745

-------
                                                           3rd Quart Se +

                                                           <1 £ x Interquartile Range}**

                                                           SidOuartl*


                                                           Mean


                                                           Median



                                                           1-



Rar

nped
o
cc

^








DC
^






CD CD
000
10 f- CJ
CC OC DC
>->•>-


= P=l
1 — 1 E3

Figure  10-65.   Box and  whisker plots of ANC  distributions in  10-year intervals  projected using

MAGIC for SBRP streams, Priority Classes A - E, for current and  increased deposition.
                                                  746

-------
                                                              SidQuartikn.
                                                              (1.S x Interquartile flange)"

                                                              SidOuartOe


                                                              Mean

                                                              Median


                                                              IstQuartile


                                                              Istduartile
                                                              (1.5 x Interquartile Range)"



                                                              "Not to exceed extreme value
                        200—,
                                       CC
                                                 Constant
                                               o
                                               CM

                                               cc
o
CO

CC
o
in
o
o
CM
                                                                             CC
~ 150-
J_i
f 100-
o-
03 50-
o—











J


















_!






_4






«J





E
q



•d




                                                 Ramped
                                                      §
                                       CC
                                              CC
                                                             CC
               S
               oc
                                                                             o
                                                                             o
                                                                             CC
                                                                                    CM
                       200-




                   ;=•>  150 —



                   g"
                   ^.  100—
Figure  10-66.   Box and whisker plots of sulfate distributions in 10-year intervals projected  using
MAGIC for SBRP streams, Priority Classes A - E, for current and increased deposition.
                                                  747

-------
                                     3rdQuartite +
                                     (1 £ x Inteiquartila Range)"
                                     3idOuart»a

                                     Mean

                                     Median

                                     IttOuartila

                                     IstQuartile
                                     (1 £ x Interquartile Range)"


                                     "Not to exceed extreme value
                       Constant
    8—1
    7—
   5—
                                 O
                                 cc
                                              n
                       Ramped
   8—|
   7—
5  6 —
   5 —
                      a
                                                   s
                                                   §=
                           ffl   C3
                                         s
                                                    §=
fosRspflm«
for SBRP streams,
         P'0tSA0f pc" ?istributio"s in 10-year intervals projected using MAGIC
     Classes A - E, for current and  increased deposition.
                           748

-------
 11, -12, and -23 /zeq L" ,  respectively.  These projections represent a relatively constant linear decrease
 in ANC over time assuming current deposition. Assuming a 20 percent deposition increase, the projected
 changes in median ANC for the first 50 years, from 50 to 100 years, and from 100 to 200 years were -19,
 -21, and -23 /zeq L" , respectively, indicating a constant relatively linear decrease for the first 100 years
 and then a slower  rate of change over the next 100 years.
      The projected change in median sulfate concentration varied over the 200-year simulation period
 with a relatively linear increase during the first 50 years, asymptotically approaching the 200-year sulfate
 concentration, 1 1 1 ^eq L"1 (Table 10-18). The increase for the first 50 years was from 37 to 75(+38)
 L' , for 50 to 100 years from 75 to 97 (+22) Meq L'1, and for 100 to 200 years from 97 to  111 (+14)
 L  under current deposition. The median sulfate projected for increased deposition was an increase from
                    "
 37 to 92 (+55) neq L" for the first 50 years; for 50 to 100 years, the increase was from 92 to 125 (+33)
 /*eq L"1; and for 100 to 200 years, the increase was from 125 to 133 (+8) ^eq L"1.

      Median pH values were relatively unchanged over  the  first 50  years  under either current or
 increased deposition and changed about -0.1  units over 100 years (Table 10-18). Over 200 years, the
 median pH changed -0.25 units under current deposition and -0.4 units with increased deposition. Lower
 quartile pH values were  projected to change about -0.15 units  in  50 years under either deposition
 scenario.  The lower quartile pHs changed -0.5 units in 100 years under current deposition and -0.6 units
 with increased deposition. After 200 years, the  lower quartile pH values were projected to change by -0.8
 units under current deposition and -1 .7 units under increased deposition.

      Projected  median calcium plus magnesium concentrations increased from  115 to about 123 ^eq
 L"  during the first 50 years and then decreased to about 108 to 110 /j.eq L"1 by year 100, with a further
 decrease to 105 peq  L"1  at year 200 under both current and increased  deposition.

      There was a differential change projected among streams based on  their initial ANC concentrations.
 Streams with higher initial ANC (based on NSS  Pilot Survey data) were  projected to have a greater
 change in  ANC than  streams with lower  initial ANC.  This result is illustrated by the change in the
frequency intervals of streams in different ANC categories in the histograms (Figures 10-68 and 10-69)
and Table 10-15.  The projected changes for ANC concentrations in streams with initial ANC between
25 and  100 /ieq L"1 and between 100 and 400 ^eq L"1 were -14 ^teq L"1 versus -24 peq L"1 over a 40-
 year period under current deposition.  The projected changes for ANC concentrations in streams with
initial ANC between 25 and 100 /zeq L"1 and 100 to 400 peq L"1  were -21 versus -34 //eq L"1 over a 40-
year period under increased deposition.

10.11.2.2  Restricted Target Population Projections Using ILWAS and MAGIC
      An estimated 567 streams in the target population were  represented  in simulations using both
ILWAS and MAGIC.  These streams were considered  undisturbed based on the chloride concentrations
and  had  NSS Pilot Survey ANC concentrations less than 200 peq L"1. All the watersheds .had positive
sulfur retention.
                                              749

-------
                               SBRP Stream Reaches
                                 Priority Class A -E
                                   Model = Magic
                               Deposition = Constant
500
400
CO
CO
2 300
05
i—
CD
-a 200-
E

£
100-
o-

%
\
%
%
%
i





I
\
cf





I



7
X
g

\




c
83!
i
I





%
%
m

i
— K!
V
1
3
.
!


1-





LtoJnr^
            -40-15 10 35  60  85110135160210235285310335360410460510
                                                         D MAGIC Year 10
                                                         H MAGIC Year 50
                           SBRP Stream Reaches
                             Priority Class A - E
                               Model = Magic
                     Deposition = Ramped 20% Increase
       500
400
E
ns
2 300
CO
"o
^ 200-
E

100-
o-


















I
i









J=l-

















V
I
i


















I
I









fl









I


















-










I



















jn.n, ln „_
           -40-15 10 35  60 85110135160210235285310335360410460510
                                                        D MAGIC Year 10
                                                        ta Year 50 Ramped
Figure 10-68.  MAGIC ANC population distributions at year  10 and  year 50 for current and
increased deposition, SBRP streams, Priority Classes A -1.
                                        750

-------
                                SBRP Stream Reaches
                                   Priority Class A -E
                                    Model = Magic
                                Deposition = Constant
500
400
CO
E
CO
£ 300
55
•5
-S 200-
|
100-
o-*








— I

J



|









Irij







ir
i
[5
?
jx
i
i
^

\










1h. i
10 20 30 40 50 60 70 80 90 100110120130140
[SO 4"] (ueqL'1)












150160170180
D MAGIC Year 10
El MAGIC Year 50
                               SBRP Stream Reaches
                                 Priority Class A - E
                                   Model = Magic
                          Deposition = Ramped 20% Increase
            500
            400
£
(0
2 300-
co
"o

-------
 10.11.2.2.1  Deposition scenarios -

      ILWAS and MAGIC projected similar changes in ANC, sulfate, and  pH after 50 years.  Changes
 projected for  streams with lower initial  ANC concentrations using MAGIC were  greater than those
 projected using ILWAS (Figures 10-70 through 10-72).  The ILWAS model performed 50-year rather than
 200-year simulations,  because of time and computational restrictions, and  comparisons are therefore
 made only for this 50-year  period.

      Median ANC concentrations using the ILWAS model were projected to decrease from 87.4 to 72.4
 l*eq L"  (-15.0 /^eq L"  ) under current deposition and from 87.4 to 71.8 peq L'1  (-15.2 ^eq  L"1)  for
 increased deposition (Table 10-19).  Median  ANC using MAGIC was projected to decrease from 118.1
 to 85.5 (-32.6) /zeq L"  for current deposition and from 118.1  to 80.1  (-38.0) jueq L"1  for increased
 deposition for the 50-year  simulation  period.  Differences between the change projected by the two
 models were 17.6 Meq L" at current deposition and 22.8 »eq L'1 at increased deposition.

      Median sulfate concentrations using the ILWAS model were projected to increase from  25.0 to
58.9 (+33.9) /zeq L'  for current deposition and from 25.0 to 69.1  (+44.1) peq L'1 for increased deposition
 (Table 10-19).  The median sulfate increases projected using MAGIC were from 37.2 to 75.3 (+38.1)
 '1
                                                                           .       .       .
 L'  for current deposition and from 37.2 to 91.8 (+54.6) //eq L'1 for increased deposition.  Differences
 between the changes projected using the  two models were 4.2 Meq L'1 for current deposition and 10.5
 fj.eq L"1  for increased deposition.

      Median pH values using the ILWAS model were projected to decrease from 7.0 to 6.8 (-0.2) for
 current deposition and 7.0 to 6.8 (-0.2) for increased deposition (Table 10-19).  The median  pH values •
 projected  using MAGIC  decreased from  7.0  to 6.9  (-0.1) for  current deposition and from 7.0 to 6.8
 (-0.2) for increased deposition.

      Median calcium plus magnesium concentrations using the ILWAS model were  projected to increase
 from  82.3 to 95.7 (+13.4) Meq L'  for current deposition  and from 82.3  to 103.4 (+21.1) ^eq L"1 for
 increased deposition (Table 10-19).  The median calcium plus magnesium concentrations  using MAGIC
 were  projected to increase from 115 to 114.8 (-0.2) ^eq L'1 for current deposition and from 115 to  120.4
 (+5.4) neq L"  for increased deposition. Differences between the change projected using the two models
 were  13.2 //eq  L'1 for current deposition and 15.7 /*eq L'1 for increased deposition.

      Watersheds in the  SBRP had an estimated median sulfur retention of 46 percent for current
 deposition and 48.3 percent for increased deposition (Table 10-19) after 50 years using ILWAS  Median
 sulfur retention for SBRP watersheds using MAGIC was projected to vary from about 24.5 percent for
 current deposition to 25 percent for increased deposition.

      None of the streams in the SBRP was projected to become acidic within 50 years using the ILWAS
model for either current or increased deposition.  There were 129 (23 percent) streams that might become
acidic at current deposition levels within 50 years using MAGIC and an estimated 159 (28  percent) that
might become acidic for increased  deposition levels within  50 years.
                                              752

-------
          tOr
       O
       0.
       O
          0.6
       

                                               l°-4
                                               i
                                               o o-2
                                                          0.0
                                                           -wo
                                                         0     100   200    300
                                                           ANC  
-------
           tOr
        o
        £L
        S 0.6
        Q_
          0.4
        O
          0.0
                SBRP Stream  Reaches
                Priority Class = A  & B
                Deposition  = Constant
                      Year  =  0
    100
  [SO,*]
                              200
                              L-1)
                                       300
                                                 SBRP  Stream Reaches
                                                Priority Class =  A & B
                                            Deposition = Ramp 20% Increase
                                                       Year = 0
                                           to
                                                         J5 0.8
                                         £"0.6
                                         Q.
                                                         o o-2
                                                           0.0
  100       200
[SO,*-]  (jieq  L-i)
          10 r
          0.8
          0.6
        T= 0.4
        .2
        ;a

        o o-2
          0.0
                Deposition •=  Constant
                     Year -  20
                    100       200
                  [SO,1-] 
        £0.4
          0.2
          0.0
Deposition «= Constant
     Year -  50
            0        100       20O
                  [SO,*-] <(ieq L-i)
                                       300
                                                             Deposition  •= Ramp 20 Increase
                                                                      Year -  50
                                                           to
                                         5 0.8
                                         o
                                         t-
                                         0>
                                         I0-4
                                         i
                                         o
                                                           0.0
                                                     100       200
                                                   ISO,*-] (jieq L-1)
                                                                                        300
Figure 10-71.  Comparison of ILWAS and MAGIC projections for sulfate concentration at years 0,
20, and 50 for SBRP streams, Priority Classes A and B, under current and increased deposition.
                                                    754

-------
            o.o
                  SBRP Stream  Reaches
                  Priority Class » A & B
                  Deposition  = Constant
                        Year  =  0
            tOr
         _§ 0.8

         O
         a.
         £ 0.6
         a.

         

                             I"
                             3


                             O
                                                              0.0
                                 4.0 4.5  SJO 5.5  6.0  6.5  7.0  75 8.0
                                               PH
                  Deposition «=  Constant
                       Year -  50
            to
         O O.B
         " 0.6
         Q_
         7= 0.4
         £
         3


         O
           0.0
            4.0  4.5  5.0  55  &0 6.5  7.0  75 8.0
                           PH
                                 Deposition  = Ramp 20% Increase
                                           Year  - 50
                                to
                                                              0.8
                             o

                             2 0.6
                             Q.

                             O
                             3=
                             £
                             a

                             O
                                                              0.0
                                 4.0  45  5.0 65  6.0  6.5  7.0  75 8.0
                                               PH
Figure 10-72.   Comparison of  ILWAS and MAGIC projections for pH at years 0, 20,  and 50  for
SBRP streams, Priority Classes A and B,  under current and increased deposition.
                                                      755

-------
Table.  10-19.   Descriptive Statistics of Projected  ANC,  Sulfate,  Percent  Sulfur
Retention, and .Calcium Plus Magnesium for SBRP Streams in Priority Classes
A and B Using  ILWAS and MAGIC for Both Current and Increased Deposition
Model Mean

ANC
Model Year 0
ILWAS 98
MAGIC 109
Model Year 20
ILWAS 91
MAGIC 100
Model Year 50
ILWAS 79
MAGIC 87
SO.2'
Model Year 0
ILWAS 31
MAGIC 49
Model Year 20
ILWAS 47
MAGIC 61
Model Year 50
ILWAS 71
MAGIC 79
EH
Model Year 0
ILWAS 6.82
MAGIC 6.82
Model Year 20
ILWAS 6.77
MAGIC 6.62
Model Year 50
ILWAS 6.64
MAGIC 6.05
% S Retention
Model Year 0
ILWAS 74
MAGIC 58
Model Year 20
ILWAS 60
MAGIC 47
Model Year 50
ILWAS 40
MAGIC 33
Std.
Dev.



37
46

33
48

31
49


16
26

18
28

23
31


0.23
0.23

0.23
0.34

0.28
0.58


11
21

11
22

14
22
Min.
Current


22
20

21
10

19
-3


12
12

29
13

42
16


6.32
6.23

6.27
5.92

6.10
5.20


37
24

29
18

20
5
P_25
Deposition


83
70

78
62

53
52


21
30

35
43

55
69


6.71
6.76

6.63
6.71

6.44
6.63


64
35

55
28

22
21
Median



87
118

85
99

72
85


25
37

40
54

59
75


6.96
6.99

6.91
6.91

6.84
6.85


80
65

65
48

46
24
P_75



118
152

104
142

99
124


40
68

54
76

88
95


7.09
7.09

7.05
7.06

6.95
7.00


80
77

66
67

49
30
Max.



159
208

145
210

126
215


73
99

93
118

118
144


7.27
7.23

7.27
7.24

7.23
7.25


89
89

82
88

74
85
                                                                continued
                            756

-------
Table 10-19 (Continued)

Model
Ca + Ma
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC

ANC
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
SO,2'
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
Model Year
ILWAS
MAGIC
BtL
Model Year
ILWAS
MAGIC

Mean

0
88
107
20
90
111
50
94
114


0
98
109
20
90
97
50
79
79
0
31
49
20
48
65
50
82
95

0
6.82
6.82
Std.
Dev.


27
36

26
35

26
33
20%


37
46

33
48

32
49

16
26

18
30

29
38


0.23
0.23

Min.


39
50

41
53

44
57
Increase


22
20

21
8

19
-10

12
12

30
15

46
19


6.32
6.23

P_25


64
85

62
86

73
88

Median


82
115

98
121

96
115

P_75


104
127

105
128

113
134

Max.


128
191

132
186

143
181
in Deposition


83
70

78
61

51
46

21
30

35
46

62
83


6.71
6.76


87
118

85
95

72
80

25
37

41
57

69
92


6.96
6.99
Model Year 20
ILWAS
6.77
MAGIC 6.55
Model Year 50
ILWAS
MAGIC
6.60
5.72
0.23
0.38
0.30
0.69
6.27
5.83
6.06
4.86
6.63
6.70
6.41
6.57
6.91
6.89
6.83
6.82


118
152

104
140

100
114

40
68

54
81

100
112


7.09
7.09

7.05
7.05
6.96
6.96


159
208

144
209

126
216

73
99

94
128

136
175


7.27
7.23

7.27
7.23
7.23
7.25
                                                              continued
                            757

-------
Table 10-19  (Continued)
Model Mean
% S Retention
Model Year 0
ILWAS 74
MAGIC 58
Model Year 20
ILWAS 64
MAGIC 51
Model Year 50
ILWAS 42
MAGIC 32
Ca + Ma
Model Year 0
ILWAS 88
MAGIC 107
Model Year 20
ILWAS 90
MAGIC 113
Model Year 50
ILWAS 101
MAGIC 120
Std.
Dev.


11
21

10
20

15
23


27
36
26
36
27
35
Min.


37
24

36
25

18
4


39
50
41
53
52
60
P_25


64
35

60
32

26
22


64
85
61
86
79
88
Median


80
65

68
51

48
25


82
115
99
122
103
120
P_75


80
77

70
69

50
29


104
127
105
130
122
147
Max.


89
89

84
88

77
86


128
191
133
189
149
187
                            758

-------
                                                           3idQuatila-f
                                                           (1.5 x Interquartile Range)"
                                                           3
-------
                                                           3rd Quart* +
                                                           (1.5 x Inteiquutile Range)"
                                                           SiriQuartile

                                                           Moan

                                                           Median

                                                           IttOuartils

                                                           IslOuartile
                                                           (1 £ x Intetquartifa Range)"


                                                           "Not to exceed extreme value
                                              Constant
                          300-

                          250-


                          200—

                       Ij 150-

                       .i 100-
                       O
                       ^  50-

                           0—


                          -50—
o
SE
CO     T
CC     OC
>•     >-
S
OC
8
CM
OC
                                              Ramped

300-
250-
200-
100-
50-
o-
-50-
-100—



o











o
DC











8
SE







§ § S
E
3E
3E
]

o
i






F
3

Figure  10-74.   Box and whisker plots for ANC  distributions in 10-year  intervals projected  using
MAGIC for SBRP streams, Priority Classes A and B, for current and increased deposition.
                                                  760

-------
                                120-

                                100—

                            T  80—
                            "ir
                            A  60-


                                20-

                                 O
                                                o
                                                cc
                                                   Constant
s
                                                                3idQuartile +
                                                                (1JS x Interquartile; Range)**
                                                                SrdQuaitila
                                                                Mean
                                                                Median
                                                                IrtQuartile
                                                                1st Quartile
                                                                (1 £ x Interquartile Range)**

                                                                **Not to exceed extreme value
                               200—1
                               150—
                            ?
                            -a-  100-
                                50—
                                                   Ramped
                                                        CC
                                                        >-
        cc
        >-
                                                                                 s
Figure  10-75.  Box and whisker plots for sulfate distributions in 10-year intervals projected using
ILWAS  for SBRP streams, Priority Classes A and B, for current and  increased deposition.
                                                  761

-------
                                                          SrdQuartHe*
                                                          (1JS x Interquartile Range)"

                                                          3rdQuartile


                                                          Mean


                                                          Median


                                                          IstQuartile


                                                          IstQuartile
                                                          (1 £ X Irteiquartile Range}"



                                                          **Not to exceed extreme value
                                             Constant
200-

150-
JJ
1 100-






]
o
i



•











s
oc









o
CT
EC






3
0
DC









O
u>
OC


r


~i
J
o
o
i

"i






o
s
EC

E
L


]
J

                                              Ramped
                                 o
                                 oc
•8
oc
                                                   O
                                                   to
                                                        CC
o
10
oc
                                                                     oc
1
oc
                          200-]



                       ~ 150-
                       L


                       I 100



                       «  50-



                           o-
Figure 10-76.  Box and whisker plots for sulfate distributions in 10-year intervals projected using

MAGIC for SBRP streams, Priority Classes A and B, for current and increased deposition.
                                                  762

-------
                                                            3idQuartile +
                                                            (1JS x Inteiquartile Range)**
                                                            SidQuartite

                                                            Mean

                                                            Median

                                                            IctOuaitile

                                                            1«tQuartHe
                                                            (1.5 x Irteiquartile Range)"


                                                            **Not to exceed extreme value
                                               Constant
                           8-1
                           7—
                           6-
                           5—
                                 cc
° 8
EC OC
>- >
— — _
	 _



o o o
CO ^ U>
OC DC DC
>
J
>
J |~
1 — ~~1 —
_| — —

                           8—1
                           5—
                                                Ramped
                                          cc.
                                          >-
s
>-
                                                            §
s
§=
Figure 10-77.  Box and whisker plots for pH distributions in 10-year intervals projected using ILWAS
for SBRP streams, Priority Classes A and B, for current and increased deposition.
                                                  763

-------
                                                            3rd Quartife +
                                                            (1.5 x Inteiquartile Range)"
                                                            3rd Quartile

                                                            Mean

                                                            Median

                                                            IstQuartile

                                                            1s) Quart ile
                                                            (1 £ x Interquartito Range)"


                                                            **Not to exceed extreme value
                                               Constant
                           7—
                           5—
                                        £E
                                                         EC
10    ^
fif    fir*
>-    >-
                                                Ramped
                           7—
                                       o:
                                                        C3
                                                        •««-
                                             oc
                                                               10
                                                               oc
Figure  10-78.  Box and whisker plots for  pH  distributions in  10-year intervals projected  using
MAGIC for SBRP streams, Priority Classes A and B, for current and increased deposition.
                                                  764

-------
  10.11.2.2.2  Rate of change of ANC, suifate, and pH over 50 years -

       The  change in median ANC and suifate concentrations and  in pH for streams in the SBRP are
  shown in box and whisker plots (Figures  10-73 through 10-78).   Median ANC was projected to change
  by about -15 ^eq L'  over the 50-year period using the ILWAS model for both current deposition and an
  increase in deposition.   MAGIC projected median changes in ANC of about -33 ^eq  L'1 for current
  deposition  and about -38 ^eq  L'1 for increased deposition (Table 10-19).   The change  in ANC was
  relatively small for the first 10 to 20 years and then decreased relatively linearly for the next 30 years.

       Median suifate concentrations, estimated from the ILWAS model,  were projected to increase  by
  about 34 neq L~   over the 50-year period for current deposition and about 44 peq L'1 for increased
  deposition  over the  50  years (Table  10-19).  Using  MAGIC, the median suifate concentrations were
  projected to  increase by about 38 ^eq L'1 for current deposition and about 55 /zeq L'1 for increased
  deposition.  There was a relatively linear  increase in suifate concentrations over the 50-year period for
  both models.

       Median and lower quartile pH values were projected to change less than 0.2 units for both ILWAS
  MAGIC over 50 years for either deposition scenario.

      There was an indication that the changes in ANC and suifate  were functions of the initial (NSS -
 Pilot Survey ) ANC using the ILWAS model (Table  10-19). A larger increase in suifate concentrations
 and a larger  decrease in ANC in  the lower ANC groups  (i.e., 25 < ANC <  100 peq L-1) than in the
 higher ANC groups (i.e., 100 < ANC < 400 Meq L'1 ) were projected with the ILWAS model.  Relatively
 similar changes in ANC  and suifate among ANC  groups were projected, however, with  MAGIC.  This
 result is  indicated  in the number of streams that change frequency intervals for distributions  of ANC and
 suifate concentration  over the 40-year period  (Figures 10-79 through 10-82).

 10.11.3  Regional Comparisons

      This section  focuses on regional comparisons among the aquatic systems in the NE and the
 SBRP. Although the representative northeastern systems are lakes and the SBRP systems are streams,
 it is  watershed processes  that control  projected  changes  in  ANC and  suifate.   Comparisons  of
 relationships  between ANC and  suifate in these systems, and of  changes  in pH and calcium plus
 magnesium with changes in suifate, can reveal similarities and differences in these processes between
 the regions.

 10.11.3.1 Northeastern Projections of Suifate Steady State

      All  three models projected that northeastern  lakes would be at suifate steady state within 50 years
 at current levels of  deposition (Figure 10-83). To examine suifate steady state in the NE, projected suifate
 concentrations are  compared with steady-state suifate concentrations computed using current deposition
 and mass balance.   A  1:1  line indicates perfect agreement between the two values.  These suifate steady-
 state projections are consistent with the percent sulfur retention of northeastern watersheds presented
 in Tables  10-14, 10-16 and 10-17.  With a  30  percent reduction,  the  projected suifate values fall below
the 1:1 line,  indicating a reduction in lake suifate concentrations within a 50-year period compared to the

                                             765

-------
                             SBRP Stream Reaches
                               Priority Class A -B
                                 Model = ILWAS
                             Deposition = Constant
      
-------
                            SBRP Stream Reaches
                              Priority Class A -B
                                Model = Magic
                            Deposition = Constant
      CO

      CO
      £
     55
      CD
     .a
      |
        150
        100-
         50-
                       ,-i
 I
       I
I

             -40  -15   10
35   60    85
   ANC&ieqL-1)
         110
160  185  210
                                                         n  MAGIC Year 10
                                                         E3  MAGIC Year 50
                           SBRP Stream Reaches
                             Priority Class A - B
                               Model = Magic
                      Deposition = Ramped 20% Increase
        200
        150
     CO
     co
     CD
     35
     o  100-j
     =1
        50-

                   1
                              I

                                              • r-B
      a   prl

            -40   -15   10   35    60   85   110  135  160  185   210
                                                        D MAGIC Year 10
                                                        H Year 50 Ramped
Figure 10-80.  MAGIC ANC population distributions at year 10 and year  50  for current and
increased deposition, SBRP Priority Class A and B streams.                     current ana
                                         767

-------
                          SBRP Stream Reaches
                            Priority Class A -B
                              Model = ILWAS
                          Deposition = Constant
ouu
250-
4~](

ieql_-n








I
90









1
100 110








1771



120 130 140
D ILWAS Year 10
B Year 50 Ramped
Figure 10-81.  ILWAS sulfate population distributions at year 10 and year 50 for current and
increased deposition, SBRP Priority Class A and B streams.
                                         768

-------
                          SBRP Stream Reaches
                             Priority Class A -B
                              Mode! = Magic
                          Deposition = Constant
300
250
w
1 200-
2
55
*o 150"
.O
1 100-
50-
o-






1
nil
1 Vn W

10 20 30








r— |











M~

\
1


1
m

40 50 60 70 80 90 100 110 120 130 140


[SOl"] OieqL-1) n MAGIC Year 10
H MAGIC Year 50
      300-
      250-
                          SBRP Stream Reaches
                            Priority Class A - B
                              Model = Magic
                    Deposition = Ramped 20% Increase

      200-
•Q 150-
XI
1 100-
50-
n-
1





1



1

I

!

S
m m
          10  20  30   40  50  60  70  80   90  100  110  120 130 140
                                                       O MAGIC Year 10
                                                       B Year 50 Ramped
Figure 10-82.  MAGIC sulfate population distributions at year 10 and year 50 for current and
increased deposition, SBRP Priority Class A and B streams.
                                         769

-------
                          NE  Lakes
                    Priority Class = A & B
                       Model  =  MAGIC
                    Deposition - Constant
               0        100       200       300
               Steady State [SO.,*-] (p.eq L-1)
                                                              o
                                                              10
                                                              _ 200
                                                              CD
                                                              tr
                                                              §.100
                                                              O
                                                              co
                NE Lakes
         Priority  Class  = A  & B
              Model =  ETD
          Deposition =  Constant
     0        100       200       300
     Steady State [SO4*-]  (neq  L-i)
             300r
                   Priority Class =  A &  B
                       Model = ILWAS
                    Deposition - Constant
               o        100       200
               Steady  State  [SO^l (^eq
                                           300
                                                                300r
                                                              ta
                                                              *~
                                                              o
                                                              to
                                                              M 200
                                                             O
                                                             CO
                                                                100
         Priority Class  = A - E
              Model = ETD
          Deposition « Constant
     o        100       200        soo
     Steady  State  [SO4*-] (neq  L-i)
            300
          o
          in
          „ 200
          01
          I 100
          O
          CO
                   Priority Class «  A - E
                      Model  -  MAGIC
                   Deposition • Constant
                                                                300r
o
10
^ 200
CO



IT
3 too
C3
CO
              o        wo        200       soo
               Steady State [SO4*-] (jieq L-i)
         Priority  Class  - A  - I
            Model  «= MAGIC
         Deposition - Constant
0        100       200
Steady  State  ISO«*]  (jxeq
                                300
PTnf1i S?o3'   ^SS!^ of projected  sulfate versus sulfate  steady-state concentrations using
ETD, ILWAS, and MAGIC for NE lakes.
                                                      770

-------
sulfate concentrations projected for current levels of deposition (Figure 10-84).  The watershed sulfur
retention values calculated on the basis of sulfur input/output indicate the watersheds are near zero sulfur
retention (i.e., near sulfate steady state), after 50 years with a 30 percent deposition decrease.  The
estimated time to sulfate steady state in the NE  is less than 50 years for both current and decreased
deposition.

      Comparisons of  projected sulfate concentrations among models indicates excellent agreement
among all models for both current  and decreased deposition (Figure 10-85).  Fewer data for ILWAS
comparisons than for MAGIC and  ETD are shown, because only 28 lakes were  simulated.   The 1:1
relationship among models, however, is evident.

10.11.3.2 Southern Blue Ridge Province Projections of Sulfate Steady State

      Projections of sulfate steady state using MAGIC indicate sulfate steady state might be reached in
the SBRP within 200 years under current and  increased deposition (Figure 10-86).  The 1:1 line on the
figure indicates agreement between the projected and steady-state sulfate  concentrations under current
deposition, consistent with the projections  of watershed sulfur retention presented in  Table 10-19.  The
relationship between projected sulfate concentrations assuming a 20 percent increase in sulfate deposition
indicates these sulfate concentrations  lie  above the 1:1 line for current deposition,  because of the
increased sulfate loading and greater sulfate steady-state concentrations.  The estimated time to sulfate
steady state in the SBRP is about 200 years, compared to less than  50 years in the  NE.

10.11.3.3 ANC and Base Cation Dynamics -

      All three models projected changes in ANC, sulfate, and pH.  Only ILWAS and MAGIC,  however,
projected changes in base cations.  Relationships between changes  in ANC and sulfate concentrations
and between changes in pH, calcium plus magnesium,  and sulfate concentrations are examined in the
following sections.

10.11.3.3.1  Northeast -

      Comparisons of projected ANC concentrations among models for northeastern watersheds  after
50 years are shown in Figure 10-87.  The 1:1 line indicates excellent agreement among  model projections.
The comparisons for ILWAS contain  only about 25 data  points so the relationships are not  as apparent.

      The changes in ANC concentrations as functions of change in sulfate concentrations are shown
in Figure 10-88 for all three models.   For current deposition, the relationships are not  apparent  because
the changes in ANC and sulfate concentrations were projected to be  quite small. A negative trend with
decreased deposition is apparent for MAGIC and ILWAS because of greater changes in ANC and sulfate
concentrations.  Given the uncertainty in the projections, however, the indicated trend is not significant.

      The pH  - ANC relationship for each  of the models is compared in Figure 10-89. There is good
agreement between the  ETD and MAGIC relationships but greater scatter in the ILWAS pH -  ANC
relationship. The ILWAS ANC - pH  relationship is modified by seasonal changes in the pCO2 function
                                              771

-------
                          NE Lakes
                   Priority Class = A & B
                       Model = MAGIC
               Deposition = Ramp 30% Decrease
               0        100       200       300
               Steady State  [SO4*1 (jieq L-I)
                                                                NE Lakes
                                                          Priority Class = A  & B
                                                               Model  = ETD
                                                     Deposition = Ramp 30% Decrease
                                                    300r
                                                            s.

                                                            8
                                                            „ aoo
                                                            to
                                                            cr
                                                            §.100
                                                            o
                                                            CO
                                                                      8? f
                                                      0        100       200       300
                                                      Steady State [SO,*] (neq L-i)
                   Priority Class = A  & B
                      Model  = ILWAS
              Deposition  - Ramp 30% Decrease
             300
           O
           in
           .. 200
           to
           O"
           §.100
          O
          a)
               0       100        200       300
               Steady State [SO,*-] (jxeq L-I)
                                                          Priority Class = A  - E
                                                              Model = ETD
                                                     Deposition  - Ramp 30%  Decrease
                                                              300r
                                                      0        100       200       300
                                                      Steady State [SO,*]  (jieq  L-1)
                   Priority Class  -= A  - E
                      Model - MAGIC
              Deposition  - Ramp  30%  Decrease
             300r
o
IO
^ 200
ta
             100
          o
          CO
    0        100       200
    Steady  State  [SO4»-]
                                          300
                                        L-I)
                                                          Priority Class - A -  I
                                                             Model - MAGIC
                                                     Deposition  -  Ramp  30% Decrease
                                                   300r
                                                            o
                                                            10
                                                              200
                                                            cr
                                                            o
                                                              wo
0        100       200       300
Steady State [SO4«-1 (jieq L-1)
Figure 10-84. Comparison of projected sulfate concentrations under decreased deposition with the
current sulfate steady-state concentrations using ETD, ILWAS, and  MAGIC for NE lakes.
                                                    772

-------
             NE  Lakes
       Priority Class = A - E
       Deposition = Constant
  Calculated [SO/-] at 50  Years
300r
           T 250
            §200
           O
           CO
           CJ
           <50
               0    50   100   150   200  250  300
                    ETD  [SCV-J  (u,eq L-I)
                                                                             NE  Lakes
                                                                       Priority Class = A - E
                                                                 Deposition  •= Ramp 30% Decrease
                                                                  Calculated [SCX,*] at 50  Years
                                                                300r
                                                     0   50   100   150  200  250   300
                                                         ETD [S04*-] (fteq L-I)
                   Priority  Class  - A  & B
                    Deposition =  Constant
              Calculated lSO4*-J  at 50 Years
             300r
           Y250
           CT
           §200
           O
           co
             100
           <50
               0   SO   100   150  200  250   300
                  ILWAS ISO,"] (jieq L-1)
                                                         Priority Class - A &  B
                                                    Deposition  = Ramp 30% Decrease
                                                    Calculated  [SO^] at 50  Years
                                                                300r
                                                              V250
                                                 cr
                                                 §200
                                                 ",150
                                                 O
                                                 CO
                                                 o100
                                                 o
                                                 < so
                                                     0    50   100   150   200   250  300
                                                        ILWAS  [SOy*-]  (|ieq L-I)
                   Priority Class -  A &  B
                   Deposition  - Constant
              Calculated  ISO.,*-] at  50 Years
            300r
          -^250
           g-200
          a
          uj so
                                                        Priority  Class  - A & B
                                                   Deposition  «=  Ramp 30% Decrease
                                                    Calculated [SO,2-] at 50 Years
                                                             ,-.250
                                                -T-150

                                                O
                                                CO.


                                                Ill 50
              0    SO   100  150   200   250  300
                  ILWAS  (SCVH|ieq L-I)
                                                    0    SO   100   150   200  250  300
                                                        ILWAS  ISO,,*]  dieq-L-i)
Figure 10-85.  Comparison of projected  sulfate concentrations between models for NE lakes after
50 years under current and decreased deposition.
                                                      773

-------
       300r
     W
              SBRP  Stream  Reaches
                 Model =  MAGIC
              Deposition = Constant
                   Year = 100
         0        100        200       300
         Steady  State [SO,2'] (jxeq L"0
          SBRP  Stream  Reaches
             Model =  MAGIC
     Deposition = Ramp  20% Increase
               Year = 100
                                                      300r
     0        100        200       300
     Steady State [SO42-] (jj.eq L-I)
      300r
    O
    o
    CM
      200
      100
    o
    CO
              Deposition = Constant
                  Year = 200
        0         100       200       300
         Steady State  lSO42-]  (p,eq L-1)
    Deposition  = Ramp 20%  Increase
               Year = 200
                                                      300
o
o
CM

re
                                                      200
                                                      100
O
CO
     0         100       200       300
     Steady  State  [SO42-]  (jaeq L-I)
Figure 10-86. Comparison of projected sulfate versus sulfate steady-state concentrations for SBRP
streams using MAGIC under both current and increased deposition.
                                            774

-------
            400r
         V  300
         .3- 200
         o
         o
         o
                        NE Lakes
                  Priority  Class =  A - E
                   Deposition = Constant
                    ANC  at  50  Years
           -100!
                   0    100    200   300
                   ETD ANC (fieq L-1)
                                         400
                                                                 NE  Lakes
                                                           Priority Class = A - E
                                                      Deposition = Ramp 30% Decrease
                                                             ANC at 50 Years
                                                             400
                                                   cr
                                                  .3. 200

                                                  O

                                                  < 100
                                                  g
                                                  o
                                                             -100
                                                      -100    0    100    200   300   400
                                                            ETD  ANC  (jieq  L-1)
           400
         3- 200

         o

         < 100
         o
         (3
                  Priority  Class  -= A & B
                  Deposition =  Constant
                    ANC at 50  Years
                   0    100   200   300    400
                  ILWAS ANC (jieq  L-1)
                                                          Priority  Class =  A &  B
                                                      Deposition  =  Ramp  30%  Decrease
                                                             ANC  at 50 Years
                                                  cr
                                                  3? 200

                                                  O
                                                  < 100
                                                  g
                                                  o
                                                  <   0
                                                      -100    0    100   200   300    400
                                                           ILWAS ANC (fieq  L-i)
        ,-, 300

Q
tu   o
           -100
                 Priority Class = A &  B
                  Deposition = Constant
                    ANC at  50 Years
            -100    0     100   200   300   400
                 ILWAS ANC (jieq L-i)
                                                          Priority Class -  A &  B
                                                     Deposition = Ramp 30% Decrease
                                                             ANC at  50  Years
                                                             400
                                                          ~ 300
                                                          2 200



                                                          2  100

                                                          Q
                                                          us   o
                                                            -wo
                                                     -100    0     100   200   300   400
                                                          ILWAS ANC (fieq L-i)
Figure 10-87.   Comparison of  projected ANC between models in NE  lakes after 50 years under
current and decreased deposition.
                                                      775

-------
                        NE  Lakes
                  Priority Class = A  - E
                      Model  =  MAGIC
                   Deposition = Constant
                        0  10  20  30  40  SO
                   A [SO,2-]  (neq L-1)
           NE Lakes
     Priority Class =  A -  E
        Model = MAGIC
Deposition  = Ramp 30% Decrease
20
15
.—. 10
3,

o- s
 o < -5
0
0 o -10
o
o
00 °0
° 0 » °0
0
0°0° °°° 0
0 °°°0 0*^ #0 ° °
o ^"oggo o o°
° 0 °0
o° «.«•

o
o o
o
-30 -20 -10 0 10 20 30 40 50 -80 -60 -40 -20 0
A [SO,2-] (neq L-1) A [SO,2-] (jieq L-1)
Priority Class = A & B Priority Class •= A & B
Model - ILWAS Model « ILWAS
Deposition - Constant Deposition - Ramp 30% Decrease
20
15

— 10
-^
IT S
§. „
-=^ 0
O
Z -5

< -10

-15
-20
30
25

— 20
0 -^15


-------
                                  NE  Lakes
                                  AH Models
                                 pH vs. ANC
8.0
7.5
7.0
6.5
£.6-0
5.5
5.0
4.5
4.0
-1(
-
«o"<*°'°
sf^"^
/• *
TO *
I.
J% . o MAGIC
JV v ETD
- v<| • ILWAS
	 1 	 | 	 | 	 | 	
)0 0 100 200 300 4
                               ANC (jieq
                           SBRP  Stream Reaches
                                 All  Models
                                pH vs.  ANC
8.0
7.5

7.0
6.5
£.6.0
5.5
5.0
4.5
4O
-
o
«t • o °
^0°°^°
O^ A
o *

o MAGIC
• ILWAS

-100    0    100    200
         ANC {|xeq  I_
                                             300
400
Figure 10-89. Comparison of pH - ANC relationship for each of the models.
                                     777

-------
and organic acid production/decomposition.  Comparisons of projected pH values between models are
shown in Figure 10-90.  There  is greater  scatter about the 1:1 line at the lower pH projections  for
comparisons between all three model pairs with greater convergence on the 1:1 line at higher pH values.

      Comparison of changes in calcium plus magnesium concentrations as a function of changes in
sulfate concentrations is illustrated in Figure 10-91 for MAGIC and ILWAS.  Minimal changes in calcium
plus magnesium and sulfate under current deposition  resulted  in  a grouping  of lakes in the upper
quadrant of the graph about the (0,0) point.  The relationship between projected changes in calcium plus
magnesium and  sulfate concentrations under decreased deposition, however, was relatively linear for both
MAGIC and ILWAS.

      The projected rate of change for ANC and calcium plus magnesium, although small for the NE, is
continuous and  does not appear to asymptotically approach steady-state concentrations. This result is
illustrated by a plot of the median ANC and calcium plus magnesium concentrations over time  for 100
years using the MAGIC results under both current and decreased deposition (Figure 10-92).  For current
deposition, median ANC remains relatively constant for the first 20 years and then decreases.  Median
calcium plus magnesium concentrations, however, decrease over the entire 100-year period, although the
rate of change slowly decreases.  This rate of change is not significant considering the uncertainty in the
projections, but  the trend is apparent.

      The projected change in ANC assuming decreased deposition approaches a peak in 50 years and
then declines  slightly over the next 50 years (Figure 10-92).  The  projected change  in calcium plus
magnesium decreases more  rapidly over the 100-year  period with decreased deposition  than under
current deposition.  The change  over the last 50 years was less than during the first  50 years.

      The median sulfate concentration decreases  from year 0 and asymptotically approaches a steady-
state value by year 50 under current  deposition.  Median  sulfate  concentrations were projected to
decrease linearly by 22 //eq L"1 by year 30  and asymptotically approach steady state by year 50 under
decreased deposition.

      Median pH is projected to change less than 0.1 unit over 100 years regardless of the deposition
scenario.  However, if the change in pH is compared with the original calibrated  pH at  year 0, all three
models indicate the greatest change in pH occurs in lakes with initial pH values between about 5.0 and6.5
(Figure 10-93). Under current deposition, those lakes with pH values between 5.0 and 6.0 were projected
using ILWAS  and MAGIC to have the greatest decrease in pH. The ETD model also  projected these
lakes might experience  the greatest change, but in both the positive and negative directions.  Under
decreased deposition, all three models projected lakes with initial pH values between  5.0 and 6.0 might
have a net increase in pH  of from 0.1 to 1.0 pH units.

10.11.3.3.2  Southern Blue Ridge Province -

      Comparisons of model-projected ANC, sulfate concentration, and pH for SBRP watersheds after 50
years are "shown  in Figure 10-94.  The  1:1  line indicates  an  apparent relationship  among  model
projections, but there are relatively few points for inter-model comparison as well as considerable scatter
about the 1:1 line.
                                              778

-------
                         NE Lakes
                   Deposition «=  Constant
                  Model  pH at 50  Years
            7.5


            7.0


            6.5


            '6.0


            S.5


            S.O


            4.5
            4.0'
             4.0  4.5  SJO  5.5  6.0  6.5  7.0  7JS
                         ETD PH
            NE Lakes
 Deposition >= Ramp 30% Decrease
      Model pH at  50 Years
 4.0  4JS  5.0  5.5  6.0  6.5  7.0  7.5
            ETD pH
                  Deposition -  Constant
           4.0
            4.0  4.5   SJO   6.5  «.0  0.5  7.0  7.5
                       ILWAS pH
                                                              Deposition -  Ramp  30%  Decrease
                                                             7JO
                                                           g
                                                             6.0
                                                             E.O


                                                             4.5
4.0  4.5  5.0   SJS   6.0   8.5   7.0   7.5
           ILWAS pH
                 Deposition  - Constant
                                                              Deposition - Ramp 30% Decrease
            4.0  4JS  SJO  5.5  8.0  tJS  7.0  TJS
                      ILWAS  pH
                                                             7JSr
                                                             7.0


                                                             &5
                                                           0.6.0
                                                            i6"6

                                                             SJO

                                                             4£
                                                             4J3
4.0  4.5  5.0  SJS  6J)  6.5  7.0  7.5
           ILWAS  pH
Figure 10-90.  Comparison  of projected  pH  values between models for  NE lakes after 50  years
under current and decreased deposition.
                                                       779

-------
        10
     if -10
     0)
       -30
    o
       -50
       -60
                    NE  Lakes
             Priority Class =  A  -  E
                 Model  = MAGIC
              Deposition  = Constant
-80    -60    -40    -20     0
       A  [SO,2']  (jieq  L-I)
                                      20
 of
                NE  Lakes
         Priority Class =  A - E
             Model  =  MAGIC
    Deposition = Ramp 30% Decrease
    10r
55-30
                                                       -50
                                                        -60
                                                         -80    -60    -40    -20     0
                                                               A [SO42-J (p.eq  L-I)
                                  20
       10
     of -10
    A
    o>
      -30
    O
      -50
      -60
             Priority Class  = A  & B
                Model = ILWAS
             Deposition =  Constant
-80    -60    -40    -20     0
       A [S
-------
              138r
 cr

3.

O130

<

 c
 «J
=5126
 ffi
                          NE  Lakes
                    Priority  Class = A -  I
                       Model  •=  MAGIC
                    Deposition = Constant
                0  10  20  30 40 SO  60 70 80 90 100
                       Simulation  Year
                                                              NE  Lakes
                                                         Priority  Class = A - I
                                                           Model  =  MAGIC
                                                   Deposition  = Ramp 30% Decrease
                                                  138r
                                                             O130
                                                             01
                                                             =6126
                                                              122
                                                    0 10 20 30 40 50 60  70 80 90 100
                                                           Simulation Year
            ffi 190
            £
            CD

            "180

            "
            CO
            O
                    Deposition - Constant
                  10 20 30 40  50 00 70 80  90 100
                       Simulation Year
                                                   Deposition
                                                  200r
                                                            of 190
                                                0>
Ramp 30%  Decrease
                                                    0  10  20  30 40 SO  60 70 80 90 100
                                                           Simulation  Year
                    Deposition  - Constant
             120r
            SJOO
            a
            •o
            S70
              60
   0 10 20  30 40 50 60  70 60 90 100
          Simulation Year
                                                   Deposition
                                                  120
                                                                           Ramp  30%  Decrease
                                                                0 10 20  30 40 SO 60  70 60 90 100
                                                                       Simulation Year
Figure 10-92.   Change in median  ANC,  calcium  and  magnesium, and  sulfate concentrations
projected for NE lakes  using MAGIC under current and  decreased deposition.
                                                     781

-------
                         NE Lakes
                      Model = MAGIC
                   Deposition = Constant
             0.4
           Q.

           <
              4.0  4.5  5.0  5.5  6.0   6.S  7.0  7£
                     Year 0 Model pH
           NE Lakes
         Model = MAGIC
Deposition = Ramp 30% Decrease
                                                            a.
                                                             -0.4
                                                             -0.8
 4.0  4.S   5.0  5.5  6.0  6.5  7.0  7.5
        Year 0 Model pH
             1.2r
             0.4
                        NE Lakes
                      'Model = ETD
                   Deposition - Constant
              4.0  4.5   SJO  5.5  6.0  6.5  7.0  7.5
                     Year 0 Model pH
           NE  Lakes
          Model  = ETD
Deposition •= Ramp 30%  Decrease
                                                            o.
                                                             0.4
 4.0  4.5  5.0  5.5  BJJ  6.5  7.0  7.5
       Year  0 Model pH
             I2r
           CL
            0.4
                        NE  Lakes
                      Model  - ILWAS
                   Deposition  - Constant
              4.0   4.5  5.0  5.5  6.0  6.5  7.0  7.5
                    Year  0 Model pH
           NE Lakes
        Model «=  ILWAS
Deposition  - Ramp 30%  Decrease
                                                             0.8
                                                             0.4
4.0  4.5  5.0  5.5  6.0  6.5  7.0  7.5
       Year  0 Model pH
Figure 10-93.  Comparison of the change in pH after 50 years as a function of the initial calibrated
pH for MAGIC, ETD and ILWAS  on northeastern lakes.
                                                     782

-------
             400r
T  300

cr
o>
.i 200

O

<  100
g
(3

I
             -100
                    SBRP  Stream  Reaches
                   Priority Class = A  & B
                      ANC at 50  Years
                    Deposition  = Constant
    -100    0    100   200    300   400
         ILWAS ANC ((ieq L-I)
                                                          SBRP  Stream Reaches
                                                         Priority Class «= A & B
                                                            ANC at  50 Years
                                                     Deposition = Ramp 20%  Increase
                                                   400r
                                                             - 200
                                                           O
                                                           < 100
                                                           o
                                                           o
                                                             -100
                                                               -100    0    100   200   300
                                                                    ILWAS  ANC {fieq L-I)
                                                                                          400
             300
                   [SO4*-] at  50 Years
                   Deposition «  Constant
               O   SO   WO   ISO  ZOO   250  300
                  ILWAS  ISCVI (neq L-1)
                                                         [SO,1-] at 50 Years
                                                     Deposition =  Ramp 20% Increase
                                                            r 2so •
                                                     0   50  100   150  200  250  300
                                                        ILWAS  [SO.*]  (neq L->)
             7.5


             7.0


           £6.5


           §"

           •§ 5.5

             5.0
             4.0
                     pH at 50 Years
                   Deposition - Constant
              4.0 4.5  5.0  5.5 6.0  6.5 7.0  7.5 8.0
                        ILWAS  pH
                                                            ph at 50 Years
                                                    Deposition - Ramp 20% Increase
                                                   8.0

                                                   7.5

                                                   7.0

                                                 1.6.5


                                                 8"
                                                 I 55

                                                   5.0

                                                   4.5
                                                             4.0
                                                    4.0  4.5  5.0 5.5  6.0 8.5  7.0  7.5 8.0
                                                              ILWAS pH
     MA       wc°'?Parisons of Projected ANC and sulfate concentrations and pH between ILWAS
and MAGIC after 50 years for SBRP  streams.
                                                     783

-------
      The changes in ANC as functions of change in sulfate concentrations are shown in  Figure 10-95
for both ILWAS and MAGIC.  A similar figure (Figure  10-96) illustrates the MAGIC projections for all 32
streams simulated in the SBRP, not just the comparable 14 ILWAS watersheds. The projected changes
in ANC concentrations were  negatively correlated with the projected changes  in sulfate concentrations
for both current and increased deposition. The relationships between the change in ANC and change
in sulfate, computed using a weighted regression for MAGIC were AANC = -2.8 - 0.372 ASO42"(r2 = 0.28)
for current deposition and AANC = -1.05 - 0.441  ASO42" (r2 = 0.42) for increased  deposition.  The
changes in calcium plus magnesium concentrations as functions of change in sulfate concentrations are
shown in Figure 10-97 for both the ILWAS and MAGIC results under current and  increased deposition.
Linear regression models assume no error in the independent variable with all the error assumed for the
dependent variable.  Therefore, a structural regression model  is required to compute the  slope of the
regression  line because a structural  regression model accounts for error in both  the independent and
dependent variables.  The structural regression model, however, requires additional analyses, which are
ongoing. Computing the slope of the relationship of calcium and magnesium versus sulfate using linear
regression to estimate an "F" factor (Henriksen, 1982), is not recommended.

      Median ANC concentrations were relatively constant for the first 20 years under current deposition,
and then were projected to decrease linearly over the remainder of  the 200-year period (Figure  10-98).
The rate of ANC decrease from year 10 to year 100 was greater under increased  deposition (i.e., -0.47
jjeq L"1 yr"1) than for current deposition (i.e.,  -0.28 /zeq L"1  yr"1).  From year 100 to year 200, however,
the rate of  change in ANC was similar for both deposition scenarios.

      Median  calcium plus magnesium concentrations were projected  to increase until about year 40
and then decrease for the rest of the simulation for both deposition scenarios (Figure 10-98).  The rates
of change in median calcium plus magnesium concentrations from year 50  to year 100 for current and
increased deposition were -0.22 and -0.28 //eq L"1 yr"1, respectively.  The rates of change in calcium plus
magnesium concentrations from  year 100 to  year 200 were -0.03  ^eq L"1 yr"1 for both  current and
increased deposition.

      Median sulfate concentrations were projected to increase at rates of 0.76 fj,eq L"1  yr"1 for the first
50 years, 0.43 /teq  L"1  yr"1 from year 50 to year 100,  and  0.14 jueq  L"1 yr"1 from year 100 to year 200
under current deposition (Figure 10-98).  Under increased deposition, the projected rates of change in
median sulfate concentrations were 1.1 /*eq L"1 yr"1 for the first 50 years, 0.66 jueq L"1 yr"1 from year 50
to year 100, and  0.09 /jeq L"1  yr"1 from year  100 to  year 200.  Both deposition  scenarios resulted in
median sulfate concentrations near sulfate steady state after 200 years.

     About 44 percent (669 streams) of the DDRP streams in the SBRP had pH values below 7.0 with
17 percent  (262 streams) having pH less than 6.75.  Comparing the  projected change in pH versus the
initial pH at year 0, however,  indicates that streams with initial pH values less than 6.75  might decrease
between -0.5 and -1.0 units within 50 years under current deposition and might have greater than -1.0
unit decrease under increased deposition  (Figure 10-99).   By year 200, streams with pH less than 7.0
might experience  pH decreases between -0.25 and -0.5 under increased deposition with some streams
projected to have greater than a"-2.0 unit decrease.
                                              784

-------
              SBRP  Stream Reaches
              Priority Class =  A & B
                 Model = MAGIC
              Deposition = Constant
10


0
^^
^-
-J -10

0"
Q)
-3--20
O
2
<-30
<3
-40
-50
-
o
o
o
0
o
-
o
°^°o
°° 0 0 °
o
o
0
o

	 	 1 	 1 	 1 	 1 	 1
0    20    40    60    80    100
     A [S
-J -10
O"
d)
3- -20
0
<-30

•^
-40

O
o
o
0 °
o

• •.< .
0 0°
°0
o

0
	 1 	 1 	 1 	 1 	 1
                                                         0    20    40    60    80    100
                                                              A tSO42-] (neq L-I)
        10r
             Priority  Class  = A  & B
                 Model =  ILWAS
              Deposition =  Constant
0

•^
-J -10
o-
O>
3-20
0
<-30
-40

-50
-
o
0 °° °
0 ° °0 °°
0 0 0 0

0


0
	 1 	 : 	 1 	 1 	 1 	 1
0    20    40    60    80
     A [SO,*] (jieq  L-I)
                                     100
                                                   Priority Class = A &  B
                                                       Model  = ILWAS
                                               Deposition =  Ramp  20% Increase
10
0
5-10
o-

-------
              SBRP Stream  Reaches
              Priority Class  =  A  -  E
                 Model  = MAGIC
              Deposition = Constant
10
0
r 	 t
-J -10

CD
-3--20

o
<-30
<]
-40
-fin
0
o °
o
°

o o « o o o O 0 o o 1 1 0 20 40 60 80 A [SO,2-] (fieq L-1) 100 SBRP Stream Reaches Priority Class = A - E Model = MAGIC Deposition = Ramp 20% Increase cr cu A* o <-30 -40 -50 •o# «> o 20 40 eo eo A [SCV-] (jieq L-I) 100 70 r 50 CD 30 10 a • O -50 Priority Class = A - E Model = MAGIC Deposition = Constant 0 O 0 Oo 8 o ° 20 40 60 80 A [S042-] (fieq L-I) 100 Priority Class = A - E Model = MAGIC Deposition = Ramp 20% Increase 70r 50 CD - 30 -10 ca O -30 -50 o o o o &' 00 O 00% 0 20 40 60 80 A [S042-] (neq L-I) 100 Figure 10-96. Comparison of projected AANC and Asulfate relationships and A(calcium and magnesium) and Asulfate relationships for SBRP Priority Class A - E streams using MAGIC. 786


-------
              SBRP  Stream Reaches
              Priority Class  =  A & B
                 Model = MAGIC
              Deposition =  Constant
      o-
      CD
      0>
70 r


50


30


10 o o
   o
        -10
      rt
     O
       •-30
       -50
             o
           O O
o  o
 o o   o

  o 8
    o o
               20     40    60    80
               A  [SO,*]  (jieq  I_-1)
                              100
                                                     SBRP  Stream Reaches
                                                     Priority  Class  =  A  & B
                                                         Model = MAGIC
                                                Deposition  = Ramp 20%  Increase
70
L 50
CD
3- 30

^"•^
ra 10
^5
"
T"10
«
o
"-30
-50
o
-

'
o
0 0
O o O O °
o
0 °
03ft
o °



	 1 	 1 	 1 	 1 	 1
                                                 0     20     40     60    80    100
                                                      A  [S042']  (^teq L-i)
       70 r
             Priority  Class  = A  & B
                 Model = ILWAS
              Deposition =  Constant
L 50
o-.'
CD
3- 30
..-j
TO 10
^
T|
F-io
CO
O
"-30
-50

o
o
0
o o
0 o o
o ^ o
o
0 ° O
OO o
	 1- 	 1 	 1 	 1 	 1
         0     20     40     60    80    100
              A  ISCV"]  (jieq L-1)
                                                    Priority  Class = A  & B
                                                        Model  =  ILWAS
                                                Deposition = Ramp  20% Increase
70
L 50
o-
CD
3- 30
. — .


-------
                      SBRP Stream Reaches
                      Priority Class = A  - E
                         Model  «= MAGIC
                      Deposition = Constant
                                                        SBRP Stream Reaches
                                                        Priority Class - A  - E
                                                           Model  •= MAGIC
                                                   Deposition =  Ramp 20% Increase
                                                  150r
                                                              J125
                                                              0100
                                                              
-------
             I
  tOr


  0.5


  0.0


 -0.5


  •10


 -ts


 -2.0
               -2.5
                      SBRP Stream Reaches
                         Model  « MAGIC
                      Deposition «= Constant
                            Year = 50
                 6.0     6.5     7.0     7.5     8.0
                      Simulation Year  0 pH
                                                          SBRP Stream  Reaches
                                                             Model  = MAGIC
                                                     Deposition ~  Ramp  20% Increase
                                                               Year =  50
    10


   0.5


   0.0


  . -0.5


  3 -tO


   -1.5


   -2.0
                                                                  -2.5
                                                     6.0     6JS     7.0     7.5     8.0
                                                          Simulation Year  0 pH
             CL
 to

 0.5

 0.0

-0.5




-ts


-2.0
               -2.5
                      Deposition =  Constant
                           Year -  100
               "6.0      6.5      7.0     7^     8J3
                      Simulation Year 0  pH
                                                     Deposition •= Ramp 20% Increase
                                                               Year - 100
                                                                  10


                                                                  0.5


                                                                  0.0
                                                                  -1.5


                                                                  -2.0
                                                            9JS     70)     7.5     8.0
                                                          Simulation Year 0  pH
             o.
             <
 to


 0.5


 0.0


-0.5


-10


-ts


-2.0
              -2.5
                      Deposition  •-- Constant
                          Year  - 200
                6.0     8.5     7.0      7.S      8J3
                     Simulation Year 0 pH
                                                     Deposition  = Ramp 20% Increase
                                                               Year - 200
3
 to


 0.5


 0.0


-0.5


-10


-ts


-2.0
                                                     8.0     «£      7.0     7.5     8.0
                                                          Simulation Year 0  pH
Figure 10-99.  Comparison of the change in pH after 200 years as a function of the initial calibrated
pH for MAGIC on SBRP streams,  Priority Classes A - E.
                                                       789

-------
 10.12  DISCUSSION

 10.12.1  Future Projections of Changes in Acid-Base Surface Water Chemistry

      The  Level 111  Analyses used typical year deposition scenarios to examine the potential  effects of
 alternative  deposition levels on future changes in surface water chemistry.  The typical year, as discussed
 in Section 5.6, represents the  average meteorology for a 30-year period of record and  the average
 deposition  for a 3- to 7-year period  of record adjusted for the average meteorological year.  Deposition
 was then estimated for each of the watersheds considered  in the Level III Analyses.  The typical year
 scenario enabled each modelling  group to use the same input and provided a common  basis for
 comparing changes in  surface water chemistry as functions of comparable  deposition among all the
 models.  The intent was not to forecast future meteorological or deposition conditions,  but rather to have
 a common basis for comparison among model results.  Comparable watershed morphometry, physical
 and chemical soils data, and surface water chemistry data also were provided to each of the modelling
 groups, enabling them to  assess and contrast  the different model formulations and projections. These
 models integrate much of our knowledge on how watershed processes control surface water acidification,
 and comparing the  output from these models,  in part, provides a test of how  well we understand these
 processes. There are different hypotheses on how these processes operate and different  philosophies on
 how to integrate this information in the  models (Eary et al., I989; Jenne et a!., 1989).  These results are
 not intended, and should not be interpreted, as forecasts of conditions that might be expected over the
 next 50 to  200  years.

 10.12.2  Rate of Future Change

    The Panel  on  Processes  of Lake  Acidification  raised questions on  the  extent  of surface water
 acidification, the processes that control changes in surface  water chemistry  (including surface water
 acidification and chemical  improvement), and the rate at which these processes occur.  The  extent of
 acidic  and  low ANC surface waters in the  United  States was  addressed through the NSWS.   The
 processes that control changes  in surface water chemistry were discussed in Section 3 and summarized
 in Galloway et al. (1983a), Church and Turner (1986), Reuss and Johnson  (1986), and Martin (1986).

     The DDRP was initiated because scientists did not concur on how watershed processes control the
 rate and magnitude of surface  water acidification and how to project such changes in surface water
 chemistry.  A primary area of disagreement among scientists on the Panel was whether future ANC
 decreases would be gradual over a  period of centuries or perceptible over years to decades,  i.e.,  they
disagreed about the rate at which acidification might occur.  The  rates at which changes  in sulfate
adsorption and base cation supply and surface water acidification and chemical improvement might occur
 in northeastern  lakes and SBRP streams are discussed below.

 10.12.2.1  Northeast

     Changes that  might  occur in the  NE over the  next 100 years (summarized in Figure  10-92) are
consistent with various conceptual models of surface water acidification (Galloway  et al., 1983a; NAS,
1984; Church and Turner, 1986; Cosby et al., 1985a,b,c; Reuss and Johnson,  1986).
                                             790

-------
       Sulfate deposition in the NE has declined since the 1970s concurrent with declining sulfur emissions
 in the NE (OTA, 1984;  Kulp, 1987).  The decline in sulfate concentrations at the start of the projections
 for the NE under current deposition (Figure 10-92) reflects this deposition decrease as the watersheds
 approach a sulfate steady-state concentration that is lower than it was in the 1970s.  The relatively
 constant ANC  concentrations under current deposition for the first 20 years of the projection occurred
 primarily because the decline in sulfate concentrations of about 8 /jeq L"1 was compensated by a decline
 of about 8 /zeq L"  in  calcium plus magnesium concentrations. Sulfate concentrations asymptotically
 approached steady state  after 20 years, changing by about 2 to 3 peq L'1 over the  next 80 years.  A
 continual depletion  of about 8 //eq L"1 in  base cation concentrations (calcium plus magnesium) was
 projected during this 80-year period as sulfate approached steady state, however, which  resulted  in the
 continual decline in ANC of about 4 /*eq L"1 over this same 80-year period.  These results are consistent
 with  observations made  in  Plastic Lake,  Ontario,  Canada where ANC  concentrations continued  to
 decrease following a reduction in sulfate deposition even though sulfate concentrations remained relatively
 constant in the lake (Dillon  et al., 1987). The ANC decrease in Plastic Lake was attributed  to depletion
 of the available pool of base cations in the watershed (Dillon et al., 1987), although no soil measurements
 were made. A depletion  of the pool of available soil base cations was projected for the northeastern
 watersheds using both  ILWAS and MAGIC and resulted in similar ANC decreases in the northeastern
 lakes.

      All  three  models projected that northeastern watersheds might be at or  near sulfate steady state
 within 50 years assuming either current or decreased deposition.  All three models projected decreased
 ANC concentrations over 50 years and  that additional lakes might become acidic,  because of the slow
 but continual decrease in base cation and ANC concentrations. The lakes currently not acidic that  might
 become acidic  over the  next 50 to 100 years represented about 3 percent of the 3227 lakes in the MAGIC
 target population. When compared with the ELS-I target population of 7157 (many of which have ANC
 concentrations  exceeding 400 peq  L'1), this additional percentage of  acidic lakes  represents less than
 1 percent of the population.   The ELS-I target population, however, included only lakes larger than 4 ha.
 Ongoing analyses of small lakes indicates that the ratio of smaller acidic lakes (< 4 ha) to  acidic lakes
 larger than 4 ha is about 2:1  (Sullivan et al., submitted). Considering these small lakes might increase the
 projected percentage of acidic lakes over the next 50 years to 2 percent. The models,  however, support
 the hypothesis that future ANC decreases in the NE will be gradual over a period of decades  to centuries
 rather than years to decades.

     Following the 30 percent decrease in sulfate deposition beginning in  year 10, there was a  rapid
 increase in  projected ANC over the next 40 years (Figure 10-92). This  11 fj.eq L"1  increase in ANC
 occurred  because the concurrent projected decrease in median sulfate concentrations of about 22 /*eq
 L" occurred with a projected decrease in median base cation concentrations (calcium  plus magnesium)
 of about 11  jueq L"1.  This rapid increase in ANC probably occurred because the watersheds were initially
 near sulfate steady state.  A  rapid increase in  ANC  might not be expected  if the systems are not at or
 near sulfate steady state (Cosby et al., 1985a,b,c). All three models projected this rapid increase in ANC
following the 30 percent decrease in sulfate deposition.  Even though the watersheds were  nearly at
sulfate steady state within 50 years under decreased deposition, there was a continued decrease in base
cations projected from year 50 to year  100, which resulted in a small but  continued decrease in ANC
concentrations.
                                              791

-------
      Although there was no apparent relationship between the rates of change in ANC and sulfate and
the initial ELS-I ANC concentration, the projected rate of change under current deposition in the NE was
small.  If the majority of the watersheds are near sulfate steady state, then most of these systems might
be expected to respond relatively quickly to changes in sulfate concentration regardless of the initial ANC.

      Projections for  all three  models indicated  that as  many  as 125  currently acidic  lakes  might
chemically improve (increase in ANC) in 50 years assuming a 30  percent deposition decrease.  This
estimate represents about 77 percent of the estimated 162 currently acidic DDRP target population lakes,
but only about 4  percent of the 3277 lakes in the MAGIC target  population.  The number of lakes
estimated to chemically improve was moderated by the continued decrease in ANC from year 50 to year
100:  after year 100, the estimated number had decreased  to 113 (70 percent).

      Differences among model  projections were more apparent for Priority Class A and B lakes for three
reasons. First, the sample size for this priority class is small and are available for  comparison.  Second,
this priority class includes low ANC systems, which  have the greatest variability  in terms both of ANC
measurements (Linthurst et  al.,  1986a) and model  calibration.  The ILWAS  and  MAGIC models are
calibrated on base cations and acid anions and ANC is a computed, not a calibrated value (i.e., ANC =
sum base cations - sum acid anions).  ELS-I field measurements for  many lakes indicate cation or anion
deficits that reflect the accepted sampling and measurement error in  the analysis.  The models, however,
require charge balance  so the calculated ANC concentrations  following calibration  might not equal the
measured ANC in the  lake or stream.  This difference between  calibrated and measured ANC values for
the models was generally greatest at the low ANC concentrations where the relative  measurement errors
also are greater. The differences between models, however, are well  within the uncertainty bounds about
the projections. Third, MAGIC performs hindcasts as  part of its calibration/projection exercise and, thus,
simulates the declining sulfur deposition levels over the past 10 years.  These declining sulfur deposition
levels continue to exhibit a cumulative effect over the first 10-20 years of the projections.  ILWAS and ETD
assume historically deposition values are the same as current  deposition values and calibrate to them,
which also contributes to the differences among models.

      The change in pH projected using MAGIC might be underestimated because  the initial or calibrated
ANC concentrations at year 0 were greater than the ELS-I ANC concentrations. Because of the pH - ANC
relationship, the unit change in  pH for each unit change in ANC decreases as  the ANC increases (i.e.,
at higher ANC concentrations,  pH  changes  are less).  To assess  this possible underestimate in pH
change, the change in ANC projected using MAGIC was added to the ELS-I ANC,  and a derived pH was
estimated  using the pH  - ANC relationship incorporated in  MAGIC (Figure 10-100).  The change in the
derived  pH is  similar to  that  in the modelled  pH for current deposition, although  the maximum change
is greater.  Under decreased deposition, the  estimated change  in  pH is greater with the derived rather
than modelled pH values, but only for a few lakes (Figure  10-100).  Because the changes in ANC are
both small and not influenced  by the initial ANC, the change in pH does not  appear to be greatly
underestimated.

10.12.2.2. Southern Blue Ridge Province

      Projected changes in surface water chemistry that might occur in the SBRP  were  shown in Figure
10-98.  ILWAS  and  MAGIC projected similar changes  in ANC, calcium plus magnesium,  and sulfate over
                                             792

-------
                     CO
2.0


1.6





0.8


0.4
                    o
                    in o.o
                      -0.4
                      -0.8
                                     NE Lakes
                                  Model =  MAGIC
                               Deposition = Constant
                                       o    Model pH
                                       A    Derived pH
                                  A A   A
                              1 - 1 - 1 - 1
                        4.0   4.5   5.0  5.5   6.0   6.5  7.0   7.5
                              Simulation  Year 0  pH
                                     NE  Lakes
                                 Model  = MAGIC
                        Deposition  = Ramp  30%  Decrease
                       2.0 r
                       1.6
                       0.8
                    o
                    10 o.o
                      -0.4
                      -0.8
                o    Model  pH
                A    Derived pH
                        4.0   4.5  5.0   5.5   6.0  6.5   7.0   7.5
                              Simulation  Year  0  pH
Figure 10-100.  Comparison of projected MAGIC change in pH versus derived pH after 50 years
for NE lakes.
                                         793

-------
 the 50-year period.   MAGIC projections for the SBRP, however, suggest that substantial changes also
 occur between 50 and 200 years.  This discussion, therefore, focuses on the MAGIC projections.

       For the first 50 years, both models projected a decrease in  ANC and an increase in base cation
 and sulfate concentrations for both deposition scenarios. The decrease in ANC concentrations over the
 first 20 years was slight, but a relatively constant linear decrease in ANC after 20 years was projected.
 Under current deposition, sulfate concentrations increased linearly for the first 50 years from about 37 to
 75 /teq L" while base cations  increased from about 110 to 123 jueq L"1 by year 30.  Increased  sulfate
 concentrations were compensated by increased base cation transport from the watershed and relatively
 little change  in ANC for the first 20 to 30 years. However,  when base cations began to decrease, ANC
 concentrations also decreased from about 122 to 100 peq L"1  from year 20 to year 100, respectively.
 Over the  interval from year 30  to year 100, sulfate concentrations increased by about 35 neq L"1, base
 cations declined  by about 15 /zeq L'1,  and ANC decreased by about 20 /*eq L'1, projections which  are
 consistent with charge. balance requirements and  current understanding of soil processes (Reuss and
 Johnson,  1986; see Sections 3  and 9).  Although the rates of sulfate increase and base cation decrease
 changed from year 100 to year 200 compared with year 50 to year 100,  the ratio or relationship between
 increased sulfate concentrations and decreased base cations remained  relatively constant because  the
 rate of change in  ANC concentrations  was relatively linear from year 30 to year 200.  The  SBRP
 watersheds were asymptotically approaching sulfate steady state by year 200, and median watershed
 sulfur retention had declined to about 5 percent.

      The two models differed  in the projected number of streams that might become acidic within 50
 years under current deposition.  The ILWAS model projected no acidic streams while MAGIC projected
 130 streams  that might become acidic in 50 years assuming current deposition.  The  estimate of 130
 acidic stream reaches, however, is derived from differences in the projections for one SBRP stream with
 a  relatively large  weight.  This  stream's ANC decreased from an initial  concentration of about 20
 "           "
    to 3 /zeq L"1 within 50 years. Given the uncertainty in the projections, 130 is probably the maximum
estimated number of streams that might become acidic within 50 years.  MAGIC projections also indicated
additional streams might become acidic over the next 200 years in the SBRP, with between 12 and 15
percent of the systems  potentially  becoming acidic by 100 years and  200 years, respectively,  under
current deposition.

    Changes  in surface water chemistry projected for the SBRP under increased deposition showed
similar patterns to those projected  under current deposition (Figure 10-98).   The rate at which sulfate
asymptotically approached the steady-state concentrations with increased deposition was greater than
that under current deposition because of the change in  sulfate loading during the first 100 years. The
rate of increase in stream sulfate concentrations during the initial phase of approaching steady state is
nearly  linear and becomes asymptotic as the soil solution sulfate concentration approaches  the steady-
state sulfate concentration.  Higher loadings with the 20 percent increased sulfur deposition scenario
resulted in the SBRP soils approaching the new sulfate steady state more quickly on the linear portion
of the  curve.  The rate of change in sulfate from year 100 to year 200 under increased deposition was
less than under current deposition because the increased loading over the first 100 years resulted  in the
watersheds being nearer to sulfate steady state.  This increased sulfate loading also resulted in greater
base cation depletion rates over the first 100 years. The rate of change in base cations from  year 100
to year 200 under increased deposition was slightly greater than under current deposition. Because the

                                              794

-------
 rate of change in sulfate under increased deposition was less and the rate of change in base cations was
 greater than  under current  deposition, there was a slight decrease in the rate of  change  in ANC
 concentrations from year  100 to year 200.

      The increased deposition and more rapid increase toward sulfate steady state resulted in  a larger
 number of streams that might become acidic by year 200. The estimated numbers of streams that might
 become acidic by year 100 and year 200 were 159 (11  percent)  and  337 (24 percent)  streams.

      The models also support the hypothesis that future ANC decreases in the SBRP will  be  gradual
 over the period of decades to centuries rather than occur over years to decades.  Streams in the SBRP
 might experience a slow but steady decline  in ANC over the next  200 years assuming constant or
 increased deposition.  The stream population in the SBRP typically had higher initial ANC concentration
 than streams  in other  geographic regions of the Southeast.  Thirty percent of the DDRP SBRP stream
 population had ANC concentrations between 25 and 100 /*eq L"1, and 70 percent of the stream reaches
 had ANC > 100 /*eq L'1 .  Extrapolating results from the SBRP to the  population of other streams in the
 Southeast, therefore, might not be appropriate because the proportion of streams with lower initial ANC
 concentrations in other southeastern  regions is greater  than in the SBRP  (Kaufmann et al., 1988).  In
 addition, the projected changes in pH in the SBRP stream accompanying these changes in ANC might
 range from -0.5 to -1.0  over 50 years and up to -2.0 pH unit changes over 200 years. Other southeastern
 streams with lower current ANC might exhibit even greater pH changes within 50 years than projected
 for SBRP streams.

 10.12.3  Uncertainties and  Implications for Future Changes in Surface Water Acid-Base Chemistry

      Uncertainty is defined as intrinsic variability plus error.  Intrinsic variability represents the natural
 variability or  noise  in  the  systems  that cannot be  reduced.   The  components of error  include
 measurement error, sampling error, model structural error, prediction error, and population estimation error
 (Beck, 1987).  The uncertainty analyses conducted for the Level III models quantitatively estimated many
 of these error  components (although the total error was  not partitioned into its respective components)
 and incorporated this error in the confidence bounds around the model projections.  Unknown or poorly
 understood  processes,  however,  are  more difficult to estimate quantitatively  but can be qualitatively
 discussed. The implications of these processes on estimates  of future change in ANC and pH are listed
 in Table 10-20.

 10.12.3.1  Deposition  Inputs

      Analyses were performed to determine the effect  of deposition input uncertainty on the model
 projections (Section 10.10.2).  These uncertainty estimates were used to establish confidence intervals
 about the model projections in Appendix C.  Analyses indicated the input uncertainty contributed about
 half of the total uncertainty in the projections with the  other half arising from parameter uncertainty.
 Uncertainty in  dry deposition, particularly of base cations, is certainly a major contributor to deposition
 input uncertainty.  The approaches used to estimate the deposition inputs, however, were reasonable,
 based on input from the  deposition  modellers, conversations with technical experts on dry and wet
deposition, analyses of existing data, and conventional theory.  In  part, underestimates or overestimates
                                              795

-------
Table 10-20.  Effects of Critical Assumptions on Projected Rates of Change
Assumptions Resulting in Under-
Estimates of ANC and pH Changes
1.    Mineral weathering overestimated
2.    Nitrate assimilation overestimated
3.    Total sulfur deposition underestimated
4.    Calibrated ANC greater than observed
5.    Watershed land use changed
6.    Episodic acidification of surface waters
7.    Biotic uptake/assimilation reducing
     available base cation pool
8.    Effects at distribution extremes over-
     smoothed through aggregation
Assumptions Resulting in Over-
Estimates of ANC and pH Change
1.    Mineral weathering underestimated
2.    Organic acids buffer surface water chemistry
3.    Total sulfur deposition overestimated
4.    Calibrated ANC less than observed
5.    Watershed land use changes
6.    Desorption is not the reverse of adsorption-
      hysteresis-related delays in change
7.    Weathering and sulfate adsorption increased by
      decreased soil pH
                                            796

-------
  in anion or cation  deposition inputs are compensated by increasing or decreasing mineral weathering
  rates, respectively, of the anion or cation species to match observed surface water chemistry  Watershed
  exchange pools are tightly coupled with deposition inputs.

       This tight coupling of declining  base cation  concentrations  to declining surface  water sulfate
  concentrations was recently reported for Hubbard Brook  (Driscoll et al., 1989b).  Two mechanisms were
  md.cated that can contribute to this coupling: (1) atmospheric deposition of base cations and (2) release
  of base cations from mineral weathering or watershed pools of exchangeable base cations (Driscoll et
  3l., 1989D).

       For the Level III projections the typical year deposition/precipitation scenario was repeated each
  year for 50 years,  so annual atmospheric deposition was  constant for the 50-year  period (with daily
  meteorological variations).  For the 30  percent  deposition  decrease, only sulfate  concentrations were
  reduced in deposition with charge balance maintained by adjusting hydrogen  ion concentration   Base
  cation concentrations were not decreased in either deposition scenario.  For these projections,  surface
  water base cation concentrations were tightly coupled with sulfate concentrations through the depletion
  of soil exchangeable base cations. Depletion of soil exchangeable base cations occurred because sulfate
  moved through the watersheds as a mobile anion.  Under decreased  deposition, the reduction in sulfate
  concentration was compensated by soil base cations and a subsequent increase in ANC.  These patterns
 were consistent with those observed at Hubbard Brook.

      While  the projected  changes in  surface water  sulfate concentrations  are  consistent with the
 depletion of watershed pools  of  base cations, these processes cannot be decoupled from  atmospheric
 processes in natural watersheds. Atmospheric deposition of base cations clearly is an important process
 that must be investigated in assessing the effects of sulfate deposition on surface water chemistry  The
 calibrated models used in the  Level 111 Analyses represent an  excellent opportunity for evaluating different
 hypotheses related to  atmospheric deposition and watershed  processes.  Simulation experimentation on
 different hypotheses represents one of the most important uses of watershed models.

      The deposition inputs, indeed, might be highly inaccurate.  The intent, however, was  not to forecast
 but rather to project the effects of alternative sulfur deposition scenarios on future changes  in surface
 water acid-base chemistry. Additional analyses are being proposed as  part of the 1990 NAPAP Integrated
 Assessment but it is likely that this issue will remain beyond 1990.

 10.12.3.2  Watershed Processes

      Each of the three models has different formulations and different data requirements.  If the three
 models provide similar projections for similar reasons,  however, greater confidence can be placed in the
 conclusions.   Questions remain, however, as to whether the models incorporate  the key watershed
 processes affecting surface water acidification and  how important  the model formulations,  operational
assumptions, and parameter selection are on the  long-term projections.

      The key  watershed processes incorporated in  each model  were listed  in Table  10-1  and are
discussed in detail in Eary et al. (1989) and Jenne et  al. (1989).  All three models focus on the effects
of sulfur deposition  on surface water acidification.   Each  model  considers total  deposition acidity,

                                              797

-------
 including  nitrate,  but the nitrogen dynamic formulations  included in each model, including ILWAS, are
 rudimentary.   Because most of eastern forested watersheds are  nitrogen-limited  (Likens et al.,  1977;
 Swank and Crossley, 1988), nitrogen inputs are effectively removed from the soil  complex. Deposition
 inputs of  nitrate  are about twice the ammonium inputs for the  eastern  United  States  (Kulp, 1987).
 Although nitrification has an acidifying effect (Lee and Schnoor, 1988), nitrate assimilation has an alkalizing
 effect (Lee and Schnoor, 1988).  Nitrate concentrations are low in receiving lakes and streams, indicating
 nitrate is not moving as a mobile anion.  Median nitrate concentrations measured during  the ELS-I for
 northeastern lakes were less than 1 ^eq L'1.  Median nitrate concentrations for SBRP streams were about
 10 peq L"  . This does  not preclude soil acidification, however, because biotic processes might influence
 surface water  chemistry. The assumption that nitrogen is not a primary contributor to chronic surface
 water acidification and,  therefore, that nitrogen dynamics do not have to be explicitly modelled represents
 a limitation of the models,  rather than a short-coming in the DDRP design.  Nitrate also  might be an
 important  component  of episodic acidification.   The  DDRP, however, is not addressing  episodic
 acidification.

       Changes in  soil pH might influence mineral weathering rates and sulfate adsorption capacities.  Plot
 experiments have indicated these processes can be affected by decreased soil solution pH.  Although
 these effects might  occur, median soil  solution pH were projected  to change less than 0.1 units in the
 NE and less than  0.2 units in the SBRP.

       One of the operational assumptions of the Level III Analyses was that the relationship of organic
 acids to other chemical species would  remain constant.   Krug and  Frink  (1983) hypothesized  that
 reversing surface  water acidification  by strong mineral acids could result in increased  dissociation of
 humic acids and mobility of organic acids and, therefore, return naturally acidic lakes to  their original
 state. The historical acidic status of the currently acidic lakes is unknown, so  the estimated chemical
 improvement of the 125 currently acidic lakes might be liberal.   Historical reconstruction of water
 chemistry for Adirondack Lakes should  be available in  the fall of 1989 and might be compared with the
 DDRP projections  of chemical improvement for the same lakes.

      Mineral weathering is critical for  all long-term projections, but is the process about which little
 information can be obtained. The mineral weathering parameters are calibration parameters but are not
 completely unconstrained. The range over which these  parameters can vary while maintaining reasonable
 ranges for  other, better characterized parameters (e.g., selectivity coefficients) and still match  observed
 surface water chemistry constituent concentrations (e.g., silica, calcium, sodium, and other base cation
 concentrations) is bounded.  All three models yield similar long-term projections, even  though ETD and
 ILWAS use a fractional  order weathering  formulation based on hydrogen ion and MAGIC  uses  a  zero
 order weathering formulation.  Long-term projections,  however, are sensitive to  the mineral weathering
 parameters in all three models.  The sensitivity of the MAGIC and ETD models to changes in the mineral
weathering parameters  was identified in  Table  10-10.   Although mineral weathering  rates cannot be
unequivocally estimated, the model formulations and mass balance approaches used  in the models might
be analogous to the  mass balance approaches used to  estimate weathering in watershed studies (Velbel
 1986b; Paces 1973).

      Data aggregation  might  result in  underestimates of  change  in the  tails or extremes  of  the
distributions.  Soil horizon physical and  chemical attributes are averaged (weighted)  to Master horizons,

                                              798

-------
 Master horizons aggregated  to  sampling classes, and sampling class attributes aggregated to the
 watershed values, which are  used for model calibration.   This averaging or aggregation process will
 preserve the central tendency in  watershed attribute, and subsequent projected effects, but will reduce
 the variability or extremes in the  distribution of soil horizons through watershed attributes  While these
 extremes represent a small  proportion of the target population, the changes in these watersheds might
 be underestimated so the changes in ANC or pH might be greater than projected.

      Although data  are not  available  for model confirmations  of  long-term  projections  short-term
 calibration and confirmation studies on Woods Lake, Panther Lake, and Clear Pond  indicate the RMSEs
 among the models and the observed standard errors of the data were similar.   Identical data were
 provided to each of the modelling groups in performing the  projections; a consistent,  methodological
 approach was used for the sensitivity analyses and the long-term  projections; and uncertainty analyses
 were  performed for all three models.  The rates of change for different constituents were comparable
 among models  and  the processes controlling  changes  in  surface water  chemistry  under different
 deposition  scenarios and among regions were similar among and between models.  Even through there
 are differences in model structure, process formulations, and temporal  and spatial scales  the model
 projections were remarkably similar. Regardless, long-term  projections can be confirmed only with long-
term periods of record  (Simons and Lam,  1980), which do not exist.  Moreover, this study does not
establish the adequacy of the formulations representing important watershed processes, the procedures
for spatial aggregation of data, or the  calibration approaches for long-term acidification projections.

10.13  CONCLUSIONS  FROM LEVEL III ANALYSES

      Conclusions from the Level  III Analyses  follow:

           All three models produced comparable results for the northeastern watersheds. ILWAS and
           MAGIC produced comparable but more variable results for the SBRP.

           All three models projected minimal changes in ANC and sulfate concentrations and pH for
           lakes in the  NE over the next  50  years at current deposition rates.  The median changes
           in ANC,  sulfate, and pH over the  next 50 years were  -1 to -5^eq L'1, <0.1  PH units, -0.1
           to  -5 /*eq L-1, respectively, each of which is within the projection  error of  the respective
           analyses.

           ETD and MAGIC projected about 3 percent and  5 percent, respectively, of the lakes in
           Priority Classes A - E that  are currently not acidic might become acidic  within 50 years at
           current deposition and 2 and 3 percent, respectively, at decreased deposition. ETD estimated
           about  22 and 46 percent of  the  currently acidic lakes in  Priority Classes A -  E might
           chemically improve  (i.e., increase in ANC) in 50 years for current and decreased deposition
           respectively.  MAGIC estimated about 39 percent and 77 percent, respectively, of the currently
          acidic lakes might improve in 50 years for the entire target population.

          All three models projected reduced lake sulfate and increased ANC concentrations and pH
          with a  30 percent reduction in deposition.  The median changes in sulfate, ANC, and pH,

                                             799

-------
 respectively, were -23 to -28 /zeq L"1, +6 to +10 peq L'1, and 0 to +0.5 pH units over 50
 years.

 MAGIC and ILWAS projections of changes in ANC, sulfate concentrations, and pH for SBRP
 streams over 50 years were similar but there was considerable scatter in the comparisons
 because of the small sample size.

 For current deposition, MAGIC projections in the SBRP indicated the change in median sulfate
 after 50, 100, and  200  years was 38,  60, and 74 /^eq L"1, respectively.  The  changes in
 median ANC after 50, 100, and 200 years were -11, -23, and -46 peq L"1, respectively. The
 median percent sulfur retention at 0 years and after 50, 100, and 200 years was 65 percent
 and 27 percent, 15 percent, and 6 percent, respectively.  The changes in median pH after
 50, 100, and 200 years were -0.04, -0.09 and  -0.20, respectively.

 The percentage  of SBRP stream reaches that  might become acidic after 50, 100, and 200
 years was  < 9, 11, and 14 percent, respectively, for current deposition and 11, 11, and 24
 percent for increased deposition.

 With a  20  percent  increase in deposition, MAGIC projections  for the  SBRP indicated the
 changes in median sulfate concentrations after 50,  100, and 200 years, respectively, were 55,
 87, and 96 /*eq L"1. The changes in median ANC after 50, 100, and 200 years, respectively,
 were -19, -41, and -64 /teq L"1.  The changes in median pH after 50,  100, and 200 years,
 respectively, were -0.07, -0.12, and -0.32.

 Based on the Level III projections, lakes in the NE might  not change significantly over the
 next 50 years with current deposition.

 Acidic lakes in the NE might improve chemically with a 30 percent reduction in deposition
 assuming organic acid  relationships with other  chemical  constituents  remain constant,
 although some lakes might continue to acidify.

 Streams in the SBRP might experience a slow but  steady decline in ANC and a  linear
 increase in  sulfate  concentration  over  the next 50 years  assuming current or increased
 deposition.   About 10 percent of the SBRP streams might become acidic within 50 years.
The stream population in the SBRP typically had higher initial ANCs than  streams in  other
 geographic  regions of  the  Southeast.   Thirty  percent of  the population  had  ANC
 concentrations between 25 and 100 /*eq L"1, and 70 percent of the stream  reaches had ANC
 >  100  peq L"1.  Care  should be taken in extrapolating results from the SBRP to the
population of other streams in the Southeast.
                                  800

-------
                                            SECTION 11

                                      SUMMARY OF RESULTS

  11.1  RETENTION OF ATMOSPHERICALLY DEPOSITED SULFUR

  11.1.1  Current Retention

       On average, watersheds in the Northeast have sulfur budgets near steady state, with negligible net
  retention of atmospherically deposited sulfur (Section 7).  A small proportion of northeastern watersheds
  have significant net retention, which appears to be controlled by reduction in wetlands or within lakes.  In
  contrast, net retention in stream systems of the Southern Blue Ridge Province is high, averaging about  75
  percent. These observations are qualitatively consistent with theory (Galloway et al., 1983a; NAS, 1984) and
 with site-specific budgets summarized by Rochelle et al. (1987).

       The Mid-Appalachian Region is a zone of transition between the NE and SBRP in terms of observed
 current sulfur retention. Because of the similarities between soils in this region and the SBRP, it is likely that
 this region at one time retained as much of the elevated sulfur deposition as is now evident in the SBRP (i.e.,
 70 - 80 percent).  It is also likely that continued high sulfur deposition is bringing soils near steady state,
 leading to reduced sulfur retention, perhaps very dramatically in the westernmost area (Subregion 2Cn  of
 the National Stream Survey, which now has median percent sulfur retention of only 3 percent) (Plate 11-1),
 and has led to the low ANC and acidic stream reaches (excluding stream reaches affected  by  acid mine
 drainage) identified there by the National Stream Survey (Kaufmann  et al.,  1988).  The Mid-Appalachian
 Region is the subject of additional in-depth soil sampling and  analyses now underway within the DDRP.

      Results of the sulfur  input-output analyses are consistent with results of Level I regression analyses
 summarized in Section 8.  Regression analyses indicate that  in the NE,  sulfate  concentrations are more
 highly correlated with sulfur deposition than with any watershed characteristic, as would be expected for
 systems at  or near steady state (i.e.,  systems where sulfur input equals  output). Additionally in the NE,
 percent watershed sulfur retention is correlated with the extent of wetlands and wet soils  on watersheds
 (Section 8.5). This provides empirical support for the hypothesis that,  to the limited extent  sulfur retention
 is observed in NE watersheds, reduction in wetlands is the principal retention mechanism.

      In  the SBRP, sulfate concentrations are correlated principally  with  edaphic factors.  Sulfate
 concentrations are relatively high in watersheds with high proportions of  shallow soils and in catchments
 having soils with low adsorption capacity.  Similarly, percent sulfur retention increases with soil depth and
with sulfate  adsorption capacity of soils.  In both the NE and SBRP, watershed disturbance  (e.g., mining
activity) is associated with elevated surface water sulfate concentrations.

11.1.2 Projected Retention

      Using deposition scenarios described in Section 5.6, projections were made of future sulfur retention
in the NE and SBRP using both a single factor (Level  II) adsorption  model (Section 9.2)  and the three
integrated models discussed in Section 10. For the sake of consistency, projections presented graphically

                                              801

-------
Plate 11-1. Sulfur retention and wet sulfate deposition for National Surface Water Survey subregions
in the eastern United States.
                                             802

-------
                         NSWS  SUBREGIONS
                 MEDIAN  %   SULFUR  RETENTION
                 AND  WET  SULFATE  DEPOSITION
                                                           a.as
MEDIAN  PERCENT
SULFUR  RETENTION





H  20  - 40

    40  - 60

    60  - 80

    80  - 100
   2-50

  2.25-.
2.00''
Average Annual

Wet Sulfaie

Deposition (g nfa yr~')
                                                       Eastern Lake Survey
                                              2.25
                                                                 Median
                                                       Subregion  X Retention
                                                         1A
                                                         IB
                                                         1C
                                                         ID
                                                         IE
                                              -H
                                               8

                                              -9
                                              -12
                             X2.00
                                                       Notional Stream Survey
                                                       Subregion  % Retention
                                                         2Cn
                                                         2Bn
                                                         3B
                                                         n
                                                         2As
                                                         3A
                                               3
                                              40
                                              34
                                              50
                                              75
                                              Deposition for 1980 - 1984
                                              (A- Olsenr Personal Communication)

-------
  in this section are from the Level ill MAGIC model. Because different target populations were modelled by
  the four models (i.e., Level II and three Level III models) and because the projected results vary somewhat
  among those populations, comparisons will be discussed qualitatively.

       In the NE, median lake sulfate concentrations are already very close to steady state. For the scenario
  of constant deposition, all of the models thus projected only small changes in median sulfate concentration,
  and projected those changes t© occur relatively rapidly (10-20 year lags). Among the  Level fll models,
  MAGIC and ETD project small decreases in median sulfate concentration during the next 20 to 50 years,
  whereas ILWAS projects very small increases.  Slight (3 - 5 percent) positive sulfur retention is projected by
  all three models by year 50, with in-lake reduction as the principal retention mechanism.  The  differences
  in the direction of changes for sulfate concentration result  from differences in target lake populations, in
  process representation by the models, and in calibration procedures; absolute differences among projections
 are minor and are relatively unimportant.  For the scenario of decreased sulfur deposition, the models
 consistently project substantial decreases in median lake sulfate concentration by year 50.  MAGIC and ETD
 project decreases in median sulfate of about 40 /*eq L1 in  50 years; ILWAS projects a somewhat slower
 decrease and a smaller, but still significant decrease of 21 Ateq L1 in median lake sulfate.
       Changes projected by the Level II sulfate model are very similar to those projected by MAGIC and
 ETD. The Level II model projects only a small median decrease (7 peq L1 ) in sulfate concentration by year
 20 for the constant deposition scenario, and a decrease in median sulfate of 40 ^eq L1 by year 50 for the
 decreased sulfur deposition scenario.  The principal difference in projections between the Level II and III
 models is that the Level II model projects all watersheds to eventually reach exactly steady state, rather than
 the small positive sulfur retention projected by Level  111  models.  This results from differences in the
 processes considered by the models; the Level II model considers only sulfate sorption by soils, whereas
 the Level III models include in-lake reduction, which accounts for the slightly positive retention at long time
 intervals.

      Projected changes in sulfate concentrations for SBRP surface waters occur much more slowly than
 in the NE, and are much larger in magnitude. Median sulfate retention in SBRP watersheds is currently about
 75 percent, but retention is projected to decrease sharply over the  next several decades  (Plate 11-2).
 Results were available for the Level II model (Section 9) and two of the Level III models (MAGIC and ILWAS)
 (Section 10); all three models projected generally similar changes for sulfate in the SBRP.  For the constant
 deposition scenario, the two integrated models project increases in median stream sulfate of roughly 15 Meq
 L1 in the next 20 years and about 40 neq L1 in 50 years;  median percent retention  is projected to decrease
 by about 40 percent over the 50-year period. For the increased deposition scenario, slightly larger increases
 in median sulfate concentration, of slightly over 50 peq  L'1, are  projected by year 50. The Level II model
 projects  somewhat faster increases for sulfate,  with  increases of 31  and 56 ^eq L1 in median sulfate
 concentration at 20 and 50 years, respectively. The Level  II model and MAGIC both project that rates of
 increase in sulfate  concentration will decrease by year  100 as  SBRP watersheds approach steady state
 (ILWAS projections were not made beyond 50 years) (Section 10). The cumulative increases projected for
 median sulfate at 100 and 200 years are 60 and 74 /*eq L1 for MAGIC and 66 and 81 peq L1 for Level II.
The differences among the models at 20 and 50 years are attributable to differences in hydrologic routing
 in the models and to assumptions about the chemistry of deep subsoils.  The 20-  and 50-year projections
                                              803

-------
Plate 11-2.  Changes in sulfur retention in the Southern Blue Ridge Province as projected by MAGIC
for constant sulfur deposition.
                                           804

-------
                  I   SULFUR   RETENTION
                        Model  =  MAGIC
                  Deposition  =  Constant
'YEAR 0 = NSS Sample
                                                   3rd Quariile +
                                                   (1.5 x Interquartile Range)
                                                   3rd Quoftile

                                                   Uean
                                                   tsl Ouarliie
                                                I   1st Quortile -
                                                   (1.5 x Interquartile Range)'
' Not to exceed extreme value.

-------
 occur during the period when the models project the most rapid changes in suifate concentration, and can
 be regarded as a measure of uncertainty in the projections. In terms of the most important aspects of sulfur
 dynamics, the three models are consistent.  All project that under the deposition scenarios simulated, the
 delayed response phase of SBRP watersheds would end for suifate, and that there would be substantial
 increases in suifate concentration in the next 20 to 50 years.  Such changes would be accompanied by
 decreases in surface water ANC to a degree dependent upon the relative leaching of acids and base cations
 from watershed soils.

      The results of the various suifate analyses are all internally consistent. Level II projections of base year
 suifate in watersheds of the NE and SBRP are consistent with, and provide a mechanistic explanation for,
 analyses by  Rochelle and  Church  (1987), summarized  in Section 7.3,  showing  watersheds in  the
 northeastern United States to be at or near steady state for sulfur and watersheds in the SBRP to have high
 net sulfur retention. The very short suifate response times projected for the NE are also consistent with
 results of regression analyses in Sections 7 and 8, which indicate that deposition is the principal control on
 surface water suifate in the NE, and that significant sulfur retention (where observed), is probably attributable
 to suifate reduction in lakes and/or wetlands rather than to sorption.  Similarly, the long response times
 predicted by dynamic models for the  SBRP are consistent with results of the Level I regression analyses,
 which found suifate concentration and percent sulfur retention to be correlated with soil variables directly
 affecting adsorption capacity of soils (i.e., soil thickness and isotherm parameters).

 11.2   BASE CATION SUPPLY

 11.2.1  Current Control

      Base cations are supplied from watersheds to surface waters by two processes acting in concert. The
 initial source is mineral weathering, which is a slow process that supplies base cations to the soil exchange
 complex. Equilibrium between the exchange complex and soil water (and thus waters delivered to lakes
 and  streams)  is reached quickly.   It  is generally accepted that weathering rates are likely to  change
 negligibly or increase only slightly due to the effects of acidic deposition since only slight decreases in soil
 pH are  likely.  If weathering supplies base cations to surface waters at rates equal to or greater than rates
 of acid  anion deposition,  then systems would be relatively  "protected". If weathering rates are low and
 cation exchange dominates base cation supply rates, then the rate of depletion of base cations from the
 exchange complex becomes an important determinant of rates of surface water acidification. Our analyses
 indicate that surface water ANCs > 100 jueq L1 cannot be explained by the cation exchange model of Reuss
 and Johnson (1986); thus, ANC generation appears to be dominated by weathering in these systems and
 they, presumably, are relatively protected against loss of ANC (Section 9). Surface waters with ANCs  < 100
 neq L1  are likely controlled by a mix of weathering and cation exchange. The exact proportion of the mix
 is difficult to determine.

 11.2.2  Future  Effects

      Single factor base cation analyses, using the models of Reuss and Johnson (1986) and of Bloom and
 Grigal (1985), were developed as a "worst-case" analysis by (1) considering only processes occurring in the
top 1.5-2 meters of the  regolith and (2) setting mineral weathering  rates to zero (i.e., assuming that the

                                              805

-------
 supply of base cations was totally controlled by cation exchange).  This analysis indicated that depletion
 of base cations from the exchange complex would occur under the sulfur deposition scenarios simulated.
 The effect on surface water ANCs was initially slight but was not negligible.  The magnitude of soil base
 cation depletion was projected to accelerate in the future. At current levels of deposition, about  15 percent
 of the lakes in the ELS target population are potentially susceptible to significant depletion of exchangeable
 cations and, thus, depletion of associated surface water ANCs.  The greatest portion of such changes is
 projected to occur on a time scale of about 50 years. In the SBRP, a greater percentage of systems are
 projected to be susceptible to adverse  effects, but at longer time scales (i.e., about 100 years) than
 northeastern systems.

      Any effects of base cation depletion would be  superimposed upon effects resulting from changes in
 sulfate mobility in soils. The combined effects were simulated using the Level III watershed models and are
 summarized in the next section.

 11.3 INTEGRATED EFFECTS ON SURFACE WATER ANC

      The three Level III watershed models (Section 1.3.4) were used to project the integrated  watershed
 and surface water responses to the sulfur deposition scenarios. Results among the models were remarkably
 comparable.  For example, within modelling Priority Classes A and  B in the NE (Section 10) and for the
 decreased sulfur deposition scenario, the MAGIC,  ETD, and ILWAS models project changes (at 50 years)
 in the median target population ANC for ANC groups <0 and 0 - 25 peq L1 within 2 /zeq L1 (5 - 7 peq L"1)
 and 3 peq L'1 (10 -13 ^eq L"1), respectively. For the ANC group 25 - 100 ^eq L"1 the ILWAS and MAGIC
 models project increases in median ANC within 1 /*eq L'1 (5.4 - 6.3 /*eq L'1). Increases in the median ANC
 of this group (25 -100 ^eq L"1) under these conditions  projected by the ETD model are quite a  bit greater
 (i.e., ~14^eq L"1 vs. ~6 ^eq L"1).

      The greatest disagreement among the model projections (at 50 years) is for  the increased sulfur
 deposition scenario in the SBRP.  For modelling Priority Classes A and B and ANC group 100 - 400 /*eq L1,
 the ILWAS model projects a decrease in median ANC of 7 peq L'\ whereas the MAGIC model  projects a
 decrease  of 24 jueq L1.   Otherwise, comparative results  among the models are  remarkably uniform,
 especially among the lower ANC groups of systems  that are of the greatest concern.

      Results from MAGIC are presented here because this model was successfully calibrated to the largest
 number of watershed systems in the two regions (i.e., 123 of the 145 DDRP sample watersheds, representing
 a target population of 3,227 systems in the  NE; and  30 of the 35 DDRP sample watersheds, representing
 a target population of 1,323 stream reaches in the SBRP).

      As discussed  in Section 10, the  watershed modelling analyses make use of watershed soil
 representations as aggregated from the DDRP Soil Survey. Because of the focus of the DDRP on regional
 characteristics and responses, soils data were gathered and aggregated so as to capture the most important
 central tendencies of the study systems. As a result,  extremes of individual watershed responses probably
are not fully captured in the analyses presented here (see discussion in Sections 8 and 10). Those systems
that are projected to respond to the greatest extent or most quickly to current or altered levels of sulfur
deposition might, in fact, be expected to respond even more extensively or more quickly than indicated here.
This should be kept in mind when reviewing the simulation results presented in this Section.

                                              806

-------
  11.3.1 Northeast Lakes

       Results of the projections for both deposition scenarios are presented in a couple of ways. Plate 11 -3
 and Table 11-1  illustrate the projected change in the median ANC at 50 years for lakes classified into four
 ANC groups (i.e., <0 peq L'1, 0 - 25 /*eq L1, 25 -100 /*eq L'1, and  100 - 400 /teq L1). These projections
 indicate a general, very slight decline in ANC over the 50-year period under the current deposition scenario
 and an increase of roughly 5 - 15 ^eq  L'1  in ANC for all groups under the decreased  sulfur deposition
 scenario. Plates 11-4 and 11-5  illustrate the overall projected ANCs for the target population at 20, 50 and
 100 years for the constant and decreased deposition scenarios, respectively.

       Table 11-2 presents the population estimates (with 95 percent confidence intervals) of northeastern
 lakes  having values of ANC <0 ^eq L1 and <50 /zeq L1 at 20 and 50 years as projected by the MAGIC
 model for the two deposition scenarios.  The ANC = 0 /*eq L1 value is used to define acidic systems, and
 the ANC value of 50 /zeq L1 (for index values as sampled in the NSWS,  see Section 5.3) has recently been
 suggested as useful in approximating the  level at or below which systems are susceptible to severe episodic
 acidification (i.e., brief  periods of ANC depression to very low or negative values) (Eshleman, 1988) with
 consequent adverse effects on biota. It is extremely important to keep in mind that these values only serve
 as indices in an otherwise smooth continuum of surface water chemistry conditions and responses to acidic
 deposition.  It is also important to remember that  adverse biological effects occur at higher  ANCs (i.e.,
 greater than 50 ^eq L1) in systems that previously (i.e., prior to the advent of acidic deposition) were
 adapted to more circumneutral conditions (Schindler, 1988).

      As indicated in Table 11 -2, under the constant deposition scenario, the number of lakes with ANC <0
 »eq L1 increases at 50 years whereas the  number of lakes with ANC <50 peq L'1 remains essentially
 constant.  For the scenario of decreased  sulfur deposition, a marked decrease is projected in the number
 of systems with ANC <0 and ANC  <50 ^eq L1. Plate 11-6 shows the changes in pH for  northeastern lakes
 at 50 years as projected by MAGIC.  MAGIC projects the greatest change for the lowest ANC group.  For
 this group the change projected  by ILWAS is virtually identical to that projected by MAGIC. Projections by
 ILWAS and ETD for the higher ANC groups are somewhat greater than projections by MAGIC (see Section
 10).

      Model projections indicate a mixed  response of northeastern lake systems at current levels of sulfur
 deposition. Slight decreases in median ANCs are projected for all ANC groups, along with a slight increase
 in the number of systems with ANC < 0 ^eq L1.  The number of systems having ANC  <  50 /*eq L'1 (and
thus potentially susceptible to episodic acidification), however, is not projected to change appreciably.
 Projected  responses to decreased  sulfur deposition  show a clearer pattern; MAGIC projects surface water
ANCs to increase and the number of lakes with ANC <0 /jeq L1 and  ANC <50 peq L1 to decrease.  Such
a response would be consistent qualitatively with reported changes in the chemistry of lakes near Sudbury,
Ontario, following reductions of  sulfur dioxide emissions from the Sudbury smelter (Dillon et al.,  1986;
Hutchinson and Havas,  1986; Keller and Pitbaido, 1986).
                                             807

-------
Plate 11-3.  Change in median ANC of northeastern lakes at 50 years as projected by MAGIC (see
Section  1.3.4 for definition of the deposition scenarios used).
                                           808

-------
CHANGE  IN  MEDIAN  ANC
  Year  10  to  Year 50
    Model  = MAGIC

-------
Table 11-1. Weighted  Median  Projected  Change in  ANC at  50  Years for
Northeastern DDRP Lakes
                        <0
                ANC Group (neq L1)

             0-25        25-100        100-400
Target
Population
162
Change in Median faeq L"1)  -2
(deposition = constant)3

Change in Median (/j.eq L'1)   5
(deposition = decreased)
398
              -2
              10
                         1054
               -1
               10
                           1612
                                         15
 See Section 1.3.4 for definition of the deposition scenarios used.
                                    809

-------
Plate  11-4. ANCs of northeastern lakes versus time, as projected by MAGIC for constant  sulfur
deposition.
                                           810

-------
                          ANC  vs.   TIME
         Mode I  -  MAGI C;   Depos i t i on
                       ANC  Group(s)  =  A
=  Constant
      Maximum
      3rd Quartile i
       (1.5 x Interquartile Range)"
      3rd Quartile
      Mean
      Median

      1st Quortile
      1st Quartile -
       (1.5 x Interquartile Range)"
      Minimum
Not to exceed extreme value.
                                                'YEAR 0 = Phast I NSWS Sample

-------
Plate 11-5.  ANCs of northeastern lakes versus time, as projected by MAGIC for decreased sulfur
deposition.
                                           811

-------
                 ANC  vs,  TIME
Mode I   =  MAG 1C;   Deposition
             ANC  Group(s)  =  Al
                                             =  Decreased
      Maximum
      3rd Quartile -f
       (1.5 x Interquartile Range)"
      3rd Quortile
      Mean
      Median
      1st Quartile
      1st Quartile -
       (1.5 x Interquartile Range)"
      Minimum
Not to exceed extreme value.
                                                           NSWS Sample

-------
 Table 11-2.   Lakes in the NE Projected to Have ANC Values <0 and  <50 ueq L'1
 for Constant and Decreased Sulfur Deposition8-1*
Time from
Present (yr)
°NSWS *
°calibrated *
20 #
50 #
Constant
ANC <0
162d
5
161e
5
161 (245)
5(8)
186 (251)
6(8)
Deposition
ANC <50
880d
27
648e
20
648 (319)
20 (10)
648 (329)
20 (10)
Decreased
ANC <0
162d
5
161e
5
136 (230)
4(7)
87 (237)
3(7)
Deposition
ANC <50
880d
27
648e
20
621 (313)
19 (10)
586 (331)
18 (10)
  Projections are based on 123 lake/watersheds successfully calibrated by MAGIC.  Projections at 20
  and 50 years are based on the MAGIC calibrated values at year 0.  The calibrated values at year 0
  can vary from the values observed by the NSWS (see footnote "e" this table and also Rgure 10-42).
  If modelled changes in ANC are combined with observed NSWS ANC values at year 0 (rather than
  with model-calibrated ANC at year 0), resulting projections of ANC in years  20 and 50 are obtained
  that sometimes differ from the values given here (for example, 248 lakes [rather than 186] would be
  projected to be acidic at year 50 under current levels of deposition).  Projections presented in
  this table, therefore, are best used to indicate the direction and relative magnitude of potential
b changes rather than absolute numbers of systems with ANC values less than 0 or 50 ^eq L"1.
  See Section 1.3.4 for definition of the deposition scenarios used.
  # is the number of lakes; % is percent of the target population of 3,227 lakes;  () indicate 95
  percent confidence estimates relative to NSWS estimates at year 0.
  Indicates estimate from NSWS Phase I sample for the same 123 lakes;  target population  = 3 227
  lakes.
  # is the number of lakes and % is the percent of target population of 3,227 lakes as estimated
  from the MAGIC calibration to the NSWS Phase I sample.
                                          812

-------
Plate 11-6. Changes in median pH of northeastern lakes at 50 years as projected by MAGIC (see
Section 1.3.4 for definition of the deposition scenarios used).
                                           813

-------
CHANGE  IN  MEDIAN  pH
 Year 10 to Year 50
   Model =  MAGIC

-------
       Because of the highly organic nature of some soils in the NE, the exact nature of chemical "recovery1
 of northeastern lakes is uncertain. To our knowledge, there are no field studies in that region that carefully
 document such a situation over a sufficient time period to cast much light upon this subject. As discussed
 in Section 1, it has been hypothesized that leaching of organic acids could be controlled by changes in soil
 water pH (e.g., as caused by acidic deposition) and that this, in turn, could have  important effects on surface
 water pH values (Krug and Frink, 1983; Krug, 1989). In this hypothesis, a decrease in precipitation acidity
 would result in an increase in leaching of organic acids to surface waters, partially offsetting (i.e., toward
 lower pH) pH increases associated with the "improved"  chemical quality of the atmospheric deposition.
 Recently, Wright et al. (1988) noted such an effect in a stream catchment in Norway where acidic deposition
 was excluded and reconstituted,  more circumneutral waters  were substituted as  "rain".  The catchment
 studied by Wright et al. (1988) has extremely thin, organic soils and, thus, is a  site almost ideally suited to
 the observation of such an effect. Wright et al. (1988) noted that in other areas of Norway having soils of
 a more mineral nature (and probably much more similar to soils of the type found on DDRP northeastern
 study sites) the potential for enhanced mobilization of organic anions would likely be much suppressed and
 minor relative to the  effects of decreasing sulfur deposition.

      Even if there  was an appreciable  increase  in  organic acid leaching as a response to reduced
 deposition acidity, the net effect would be beneficial to aquatic biota inasmuch as  it would most likely be
 accompanied by  reductions in surface water concentrations of inorganic monomeric aluminum, which is
 highly toxic to fish (Baker and Schofield, 1982).

      Thus, although the exact chemical response  of  the  northeastern DDRP  systems is  unknown,
 projections indicate an improvement in surface water quality as a consequence of reduced sulfur deposition
 in the region.

 11-3.2 Southern Blue Ridoe Province

      Plate 11-7 and Table 11-3 illustrate the projected changes (MAGIC model)  in median ANC at 50 years
for stream reaches in the SBRP.  The MAGIC model used in this analysis was successfully calibrated to 32
of the 35 DDRP SBRP stream  reach watersheds. Two stream reaches had ANC >  1000 neq L"1 and were
dropped from this presentation. The remaining 30 stream  reaches had ANC >  25 /zeq L"1 and <  400 fieq
L1 and represent a target population of 1,323 stream reaches in the SBRP. The projected changes in median
ANCs have been computed for the same ANC groups (25 -100 /jeq L1 and 100 - 400 //eq L1) as for the
NE (Plate  11-3).  Plates 11-8 and 11-9 illustrate the overall projected ANCs for the target population at 20,
50, 100, and 200 years for the current and increased deposition scenarios, respectively.

      Table 11-4 presents the  population estimates (with 95 percent confidence intervals) of SBRP stream
reaches having ANC  <0 Meq L"1 and <50 peq L"1 at 20 and 50 years as projected by the MAGIC model for
the two deposition scenarios.  The 95 percent confidence intervals about  these projections are broad but
understandable, given the low number of systems available for simulation (30) and the inherent uncertainties
involved in such a complex simulation of environmental response.
                                              814

-------
Plate 11-7.  Change in median ANC of Southern Blue Ridge Province stream reaches at 50 years as
projected by MAGIC (see Section 1.3.4 for definition of the deposition used).
                                           815

-------
CHANGE IN  MEDIAN  ANC
  Year  10  to  Year 50
    Model  = MAGIC
                      ncreased
                    5  Deposition

-------
 Table 11-3.  Weighted Median Projected Change in ANC
 at 50 Years for DDRP SBRP Stream Reaches
                                   ANC Group Cueq L'1)
(deposition = constant)
                                 25-100
               100-400
Target
Population
Median Change (p


eq L1)
. .\ 9

407
-14

916
-24
Median Change (/jeq L1)
(deposition = increased)
-20
                  -34
aSee Section 1.3.4 for definition of the deposition scenarios used.
                              816

-------
Plate 11-8.  ANCs of Southern Blue Ridge Province stream reaches versus time, as projected by
MAGIC for constant sulfur deposition (see Section 1.3.4 for definition of the deposition scenarios
used).
                                           817

-------
                       ANC  vs,  TIME
       Model  =  MAGIC;  Deposition  =  Constant
              ANC  Group(s)  =  <400  ueq  I/1
                                                  3rd Quariile +
                                                  (1.5 x Interquartile Range) '
                                                  3rd Ouorlile

                                                  Uean

                                                  Median

                                                  Is! Quortile

                                                  1st Quortile -
                                                  (1.5 x Interquartile Ronge) '
'YEAR 0 = NSS SdmpU
                                            ' Not to exceed extreme value.

-------
Plate 11-9.  ANCs of Southern Blue Ridge Province stream reaches versus time, as projected by
MAGIC for increased sulfur deposition (see Section 1.3.4 for definition of the deposition scenarios
used).
                                           818

-------
r
                                  ANC  vs,   TIME

                  Model  =  MAGIC;  Deposition  =  Increased
                         ANC  Group(s)  =  <400  ueq  L'
  -i
                                                             3rd Quarlile
                                                             (1.5 x Interquartile Range
                                                             1st Quortile

                                                             1st Quortile - .
                                                             (1.5 x Interquartile Range) *"
             'YEAR 0 = NSS Sample
1 Not to exceed extreme value.

-------
 Table  11-4.  SBRP Stream Reaches Projected to Have ANC Values  <0 and
  <50 //eq L'1 for Constant and Increased Sulfur Deposition8^
Time from
Present (yr)
°NSWS f
0 *e
calibrated £"
20 #
50 #
Constant
ANC <0
Od
0
oe
0
0
0
129 (295)
10 (22)
Deposition
ANC <50
58d
4
187e
14
187 (310)
14 (23)
203 (333)
15 (25)
Increased
ANC <0
Od
0
Oe
0
0
0
159 (291)
12(22)
Deposition
ANC <50
58d
4
187e
14
187 (314)
14 (24)
340 (359)
26 (27)
   Projections are based on 30 stream/watersheds successfully calibrated by MAGIC.  Projections at 20
   and 50 years are based on the MAGIC calibrated values at year 0.  The calibrated values at year 0
   can vary from the values observed by the NSWS (see footnote "e" this table and also Figure 10-70).
   If modelled changes in ANC are combined with observed NSWS values at year 0 (rather than with
   model-calibrated ANC at year 0), resulting projections of ANC  in years 20 and 50 are obtained that
   sometimes differ from the values given here (for example, zero stream reaches [rather than 129]
   would be projected to become acidic by year 50 under current levels of deposition; also, although
   projections from the ILWAS model for  median regional decreases in ANC over 50 years are comparable
   to those  projected by MAGIC for the same watersheds [see  Table 10-15], ILWAS does not project any
   SBRP watersheds to become acidic by year 50).  Projections presented in this table, therefore, are
   best used to indicate the direction and relative magnitude of potential changes rather than absolute
   numbers of systems with ANC values less than 0 or 50 peq  L"1.
   See Section 1.3.4 for definition of the deposition scenarios used.
  # is the number of stream reaches; % is percent of the target population of 1,323 stream
d reaches;  () indicate 95 percent confidence estimates relative to NSWS estimates at year 0.
  Indicates estimate from NSWS Pilot Stream Survey sample for the same 30 stream reaches;
  target population = 1,323  stream reaches.
  # is the number of stream reaches and % is the percent of  the target population of 1,323 stream
  reaches as estimated from the MAGIC  calibrations to the NSWS Pilot Stream Survey sample.
                                          819

-------
       Plates 1 1 -1 0 and 1 1 -1 1 show decreases in pH of SBRP stream reaches as projected by MAGiC and
  ILWAS, respectively, for the increased sulfur deposition scenario.  Changes projected by the two models
  are highly comparable.

       Model projections for the SBRP stream reaches indicate decreased surface water quality under
  scenarios of either current or increasing sulfur deposition.  As noted in Sections 9 and 10, responses to
  changes ,n sulfur deposition levels in the SBRP are  projected to be slower than those in the NE- i e
  there is a considerable lag in the response of the systems due to the storage of sulfur in the soils ' The
  result is that there is a delay not only in the acidification of surface waters in the region, but also in any
  potential recovery if sulfur deposition were to be decreased. Due to the fact that soils in this region are
  much less organic in nature than those in the NE (e.g., wetlands in the SBRP are virtually non-existent-
  mean stream DOC  at lower nodes was <1 mg L~\ these model projections are uncomplicated  by
  potential effects of organic acid leaching.

       Projections of stream water quality response for the DDRP SBRP target population clearly indicate
 future adverse effects of sulfur deposition at increased or current levels.

  11.4  SUMMARY DISCUSSION

      The NE is currently at sulfur steady state and sulfate  concentrations in surface waters would
 respond relatively rapidly to decreases in sulfur deposition. Associated with these changes would be
 increases in surface water ANC.  Continued sulfur deposition at current levels is gradually depleting the
 cat,on exchange pool  in northeastern soils with  consequent decreases in surface water ANC   Such
 changes are relatively slow and minor, however, relative to direct effects of increased anion mobility  in
 watersheds on surface water chemistry.
                   H *"        ^ CUITently retai"nin9  "^ ^ee^^^ of the atmospherically
                  the average but soils are projected as becoming more saturated with regard to sulfur

?, nr? TT Ttl0nS ^ Pr°JeCted t0 bS inCreaSI'ng iP the SUrfaCe Waters of the re9jon- ™« response
•s projected to be marked over the next 50 years at either current or increased levels of sulfur deposition
as are decreases in stream water ANC.  Superimposed upon this effect is a relatively minor acidification
effect of base cation depletion.
fGaHow         o         °\DDRP ana'yses are 
-------
Plate 11-10.  Changes in pH of SBRP stream reaches as projected by MAGIC (see Section 1.3.4 for
definition of the deposition scenarios used).
                                           821

-------
                        pH  vs,  TIME
      Model   =  MAGIC;  Deposition  =   Increased
             ANC  Group(s)  =  <400  ueq  I/1
                                                  3rd Quorlile +
                                                  (1.5 x Interquartile Range)'

                                                  3rd Ouartile

                                                  Uedion


                                                  1st Quortile
                                                  1st Quortile -
                                                  (1.5 x Interquartile Ronge)"
'YEAR 0 = Uodel Year 0
                                             1 Not to exceed extreme voltie.

-------
Plate 11-11.  Changes in pH of SBRP stream reaches as projected by ILWAS (see Section 1.3.4 for
definition of the deposition scenarios used).
                                           822

-------
                       pH  vs,  TIME
      Model  =  ILWAS;  Deposition  =  Increased
             ANC  Group(s)  =  <400  ueq  I/1
'YEAR 0 = yodel Year 0
                                                3rd Quartile
                                                 1.5 x Interquartile Range}'
                                                1st Quartile
                                                1s,i Quartile -
                                                 1.5 x Interquartile Range)'
                                           1 Not to exceed extreme value.

-------
                                           SECTION 12



                                          REFERENCES
 Abrahamsen, G. 1980. Acid precipitation, plant nutrients, and forest growth, pp. 58-63.  In- D Drabl^s and

                                                                    a an lntemational Conference'
 Akaike, H. 1969. Fitting autoregfessive models for prediction. Ann. Institute Statist. Math.  21:243-247.



 Aimer B  W Dickspn, C  Ekstrom, and E. Hornstrom. 1978. Sulfur pollution and the aquatic ecosystem, pp.


      fnc  NewYor?'NYagU'         '" ^ Environment' Part ll: Ecological Impacts. John Wiley & Sons,
       c!c        n  RAJ Linthur|tD 1?84' The Acidic Deposition  Phenomenon and Its Effects:  Critical
      Assessment Review Papers. EPA/600/8-83/0 16bf. U.S. Environmental Protection Agency, Washington,




Appleby, P.G., and F Oldfield. 1978. The calculation of lead-210 dates assuming a constant rate of supply

      of unsupported Pb-21 0 to the sediment. Catena 5: 1 -8.



April>
                     oo™' lnfluence of geology on lake acidification in the ILWAS watersheds. Water,
                   .  26.3/3~386.
Arnold R.W. 1977. Clean brush approach achieves better concepts in soil survey, pp. 61-92 In- Qualitv
      Conference Proceedings, Bergams, New York, NY.                                 '   ' v*u Aluminum toxicity to fish in acidic waters. Water,  Air, Soil Pollut.
     ,nfvtPrwt   A       D          'W-  Ogburn' '"• 1985'  Sediment acid  neutralization in
     softwater lakes. Water, Air, Soil Pollut. 25:215-230.
                                             823

-------
 Baker, L.A., P.L. Brezonik, and E.S. Edgerton. 1986a. Sources and sinks of ions in a softwater acidic lake
       in Florida. Water Resour. Res. 22:715-722.


 Baker, LA., P.L Brezonik, and C.D. Pollman. 1986b. Model of internal alkalinity generation: Sulfate retention
       component. Water, Air, Soil Pollut. 31:89-94.

 Baker, LA., C.D. Pollman, and J.M. Eilers. 1988. Alkalinity regulation in softwater Florida lakes Water
       Resour. Res. 24:1069-1082.

 Barrow, N.J. 1967. Studies on the adsorption of sulfate by soils. Soil Sci.  104:342-349.

 Barrow, N.J.,  and T.C.  Shaw. 1977. The  slow reactions between soil and anions: Effect of time and
       temperature of contact between an adsorbing soil and sulfate. Soil  Sci. 124:347-354.

 Barrow, N.J., K. Spencer, and W.M. McArthur.  1969. Effects of rainfall and parent material on the ability of
       soils to adsorb sulfate. Soil Sci. 108:120-126.

 Bartz, J.K., S.K Drouse, K.A. Cappo, M.L. Papp, G.A. Raab, LJ. Blume, M.A. Stapanian, F.C. Garner, and
       u.b.  Goffey. 1987. Direct/Delayed Response Project:  Quality Assurance Plan for Soil Samplinq
       Preparation and Analysis. EPA/600/8-87/021. U.S. Environmental Protection Agency, Washington, DC!
       HIy  pfj*


 Bayley S.E., R.S. Behr, and C.A. Kelly. 1986. Retention and release of S from a freshwater wetland Water
       Air, Soil  Pollut.  31:101-114.


 Bayley S.E., D.H. Vitt, R.W. Newbury, K.G.  Beaty, R. Behr, and C. Miller. 1987. Experimental acidification
       of a Spftagvwm-dominated peatland: First year results. Can. J. Fish. Aquat. Sci. 44(Suppl. 1):194-205.

 Beamish R.J.,  and H.H. Harvey. 1972. Acidification of the La Cloche Mountain lakes, Ontario, and resulting
      fish mortalities.  J. Fish. Res. Board Can.  29:1131-1143.

 Beck,  M-B. 1987.  Water quality modeling:  A review of the analysis of uncertainty.  Water Resour. Res.
      23.1393-1442.


 Belsley, D.A., E. Kuh,  and R.E. Welsch. 1980. Regression Diagnostics. Wiley-lnterscience, New York, NY.

 Best, M.D., L.W. Creelman, S.K. Drouse, and D.J. Chaloud. 1986. National Surface Water Survey Eastern
      Lake Survey (Phase I), Analytical Methods Manual. EPA-600/4-86/011. U.S. Environmental Protection
      Agency,  Las Vegas, NV.


 Bettany J.R., J.W.B. Stewart, .and E.H. Halstead.  1373. Sulfur fractions and carbon, nitrogen and sulfur
      relationships in grasslands, forest and associated transitional soils. Soil Sci. Soc. Am. J. 37:915-918.

 Beven, K.J. 1986. Runoff production and flood frequency in catchments of order: An alternative approach
      pp. 107-132. In:  V.K. Gupta,  I. Rodriguez-lturbe, and E.F. Woods, eds. Scale Problems in Hydroloqy
      D. Reidel Publ. Co., Dordrecht, The Netherlands.

 Beven K J., and M.J. Kirkby. 1979. A physically based, variable constanf contributing area model of basin
      hydrology. Hydrol.  Sci. Bull. 24:43-69.


Billings, MP 1980. Geologic Map of New Hampshire. U.S. Geological Survey, Washington, DC. (scale
      1.250,000).


Birks,  H J.B.  1985.  Recent  and  future mathematical  developments in  quantitative palaeoecoloqy
      Palaeogeogr., Palaeoclimat.,  Palaeoecol.  50:107-147.
                                              824

-------
 Bloom, P.R., and D.F. Grigal. 1985. Modeling soil response to acidic deposition in nonsulfate adsorbing
      soils. J. Environ. Quai. 14:489-495.

 Bloom, P.R., M.B. McBride, and R.M. Weaver. 1979a. Aluminum organic matter in acid soils:  Buffering and
      solution aluminum activity. Soil Sci. Soc. Amer. J. 43:810-813.

 Bloom, P.R., M.B. McBride, and R.M. Weaver. 1979b. Aluminum organic matter in acid soils: Salt-extractable
      aluminum. Soil Sci. Soc. Amer. J. 43:813-815.

 Bogucki,  D.J., G.K. Gruendling,  and E.B. Allen.  1987.  Final Report:  Evaluating Accuracy of Photo
      Interpretation for Land Use, Wetland, Beaver, and Stream Data Collected for Direct/Delayed Response
      Project  (DDRP) Watersheds. Rep.  Center for Earth and Environmental Science, State University of
      New York, Plattsburgh, NY.  (Unpub.). 191  pp.

 Bogucki, D.J., G.K. Gruendling, E.B. Allen, K.B. Adams, and M.M. Remillard.  1986. Photointerpretation of
      historical (1948-1985)  beaver activity  in the Adirondacks. pp.  299-308. In: Proceedings of the
      ASPRS-ACSM Fall Convention. September 28-October 3, Anchorage, AK.

 Bolan, N.S., D.R. Scotter, J.K. Syers, and R.W. Tillman. 1986. The effect of adsorption on sulfate leaching.
      Soil Sci. Soc. Am. J. 50:1419-1424.

 Bornemisza, E., and R. Llanos. 1967. Sulfate movement, adsorption, and desorption in three Costa Rican
      soils.  Soil Sci. Soc. Am. Proc. 31:356-360.

 Brakke, D.F., D.H. Landers, and J.M. Eilers. 1988. Chemical and  physical characteristics of lakes in the
      northeastern United  States. Environ. Sci. Technol. 22:155-163.

 Brezonik,  P.L, LA. Baker, and I.E.  Perry. 1987.  Mechanisms of alkalinity generation  in acid-sensitive
      softwater lakes, pp.  229-262. In: R. Hites and S.J. Eisenreich, eds. Sources and Fates of Aquatic
      Pollutants. Advances in Chemistry, Series 216. American Chemical  Society, Washington, DC.

 Bricker, O.P., A.E. Godfrey, and E.T. Cleaves. 1968. Mineral-water interaction during the chemical weathering
      of silicates,  pp.  128-142.  In: Advances in Chemistry,  Series 73. American Chemical Society,
      Washington, DC.

 Brown, P.M. 1985. Geologic Map of North Carolina. Dept. Natural Resources and Community Development,
      Div. Land Resources, Raleigh, NC.  (scale 1:500,000).

 Buol, S.W.,  F.D. Hole, and R.J. McCracken. 1980. Soil Genesis and Classification. Iowa State  University
      Press, Ames, IA.

 Buso, D.C., S.W. Bailey, S.F. Baird, J.W. Hornbeck, and C.W. Martin. 1985.  Watershed Interactions Affecting
      Pond  Acidification. Research Report No. 62. New Hampshire Water  Resource Research Center,
      Durham, NH.

 Byers, G.E., R.D. Van Remortel, J.E. Teberg, M.J. Miah, C.J. Palmer, M.L Papp, W.H. Cole, A.D. Tansey,
      D.L Cassell, and P.W. Shaffer. 1989. Direct/Delayed Response Project: Quality Assurance Report for
      Physical and Chemical Analyses of Soils from the Northeastern United  States. EPA/600/X-88/136.
      U.S. Environmental Protection Agency, Environ. Monitoring Systems Lab., Las Vegas,  NV. 216 pp.

Calles, C.M. 1983. Dissolved inorganic substances. Hydrobiologia 101:13-27.

Camburn,  K.E., and J.C.  Kingston. 1986. The genus Melosira from soft-water lakes with special reference
      to northern Michigan, Wisconsin, and Minnesota, pp. 17-34. In: J.P. Smol, R.W. Battarbee, R.B. Davis,
      and J. Merilainen,  eds.  Diatoms and Lake Acidity. Dr. W. Junk, Dordrecht, The Netherlands.
                                              825

-------
 Campbell, W.G., and M.R. Church. 1989. EPA uses GIS to study lake and stream acidification. Federal Digital
      Cartography Newsletter 9:1-2.

 Campbell,  W.G., M.R.  Church,  G.D. Bishop,  D.C. Mortenson, and S.M. Pierson. 1989. The role for a
      Geographic Information System in a large environmental project. Internat. J. GIS. 3:349-362.

 Cappo, K.A., L.J. Biume, G.A. Raab, U.K. Bartz, and J.L Engels. 1987. Analytical Methods Manual for the
      Direct/Delayed Response  Project Soil Survey. EPA/600/8-87/020. U.S. Environmental Protection
      Agency, Environ. Monitoring Systems Lab., Las Vegas, NV.

 Cardwell, D.H.,  R.B. Erwin,  and H.P. Woodward.  1968.  Geologic Map of West Virginia. West Virginia
      Geological and Economic Survey.

 Carlston, C.W. 1963. Drainage Density and Streamflow. U.S. Geological Survey Professional Papers 422-C,
      Washington, DC. 8 pp.


 Chan, W.H., D.B. Orr, and R.J. Vet. 1982a. The Acid Precipitation in Ontario Study (APIOS). An Overview:
      The Cumulative Wet/Dry  Deposition Network.  Report #ARB-15-82-ARSP, Ontario Ministry  of the
      Environment.

 Chan, W.H., D.B. Orr, and  R.J. Vet.  I982b. The Acid Precipitation in Ontario Study (APIOS). An Overview:
      The  Event Wet/Dry Deposition  Network. Report  #ARB-11-82-ARSP, Ontario  Ministry  of the
      Environment.


 Chang,  M., J.D. McCullough, and A.B. Granillo. 1983. Effects of land use and topography on some water
      quality variables in forested east Texas. Water Resour. Bull. 19:191-196.

 Chang, J.S., R.A. Brost, I.S.A. Isaksen, S. Modronich, P. Middleton, W.R. Stockwell, and C.J. Walcek. 1987.
      A three-dimensional Euierian acid deposition model: Physical concepts and formulation  J Geophys
      Res. 92:14681-14700.

 Chao, T.T., M.E. Harward, and S.C. Fang. 1962a. Movement of 35S tagged sulfate through soil columns. Soil
      Sci. Soc. Am. Proc. 26:27-32.

 Chao, T.T.,  M.E. Harward, and S.C. Fang. 1962b. Adsorption and desorption phenomena of sulfate ions in
      soils. Soil Sci. Soc. Am. Proc. 26:234-237.

 Chao, T.T., M.E. Harward, and S.C. Fang. I964a. Anionic effects on sulfate adsorption by soils. Soil Sci Soc
      Am. Proc. 28:581-583.

 Chao, T.T., M.E. Harward, and S.C. Fang. 1964b; Iron or aluminum coatings in relation to sulfate adsorption
      characteristics of soils. Soil Sci. Soc. Am. Proc. 28:632-635.

Chen, C.W., J.D. Dean, S.A. Gherini, and R.A. Goldstein. 1982. Acid Rain Model: Hydrologic Module Jour
      Env. Engr. ASCE  108:302-318.

Chen, C.W., S.A. Gherini, R.J.M. Hudson, and J.D. Dean. 1984. The Integrated Lake-Watershed Acidification
      Study, Volume 2: Hydrologic Analysis. C.W. Chen, ed. Rep. No. EA-3221. Electric Power Research
      Institute, Palo  Alto, CA.

Chen, C.W., S. Gherini, R.J.M. Hudson, and J.D. Dean.  1983a. The Integrated Lake/Watershed Acidification
      Study, Volume 1:  Model and Procedures. Rep. No. EA-3221. Electric Power Research Institute  Palo
      Alto, CA.
                                             826

-------
       svn,,,-0^- Hudson' and R'A- G°'dstein. 1983b. Modeling of Precipitation
       Series, Volume 9. Ann Arbor Sciences, Butterworth Publishers, Boston, MA. 175 pp.
              -                Steady-state kinetics and dissolution mechanisms of albite. Am. J. Sci.




 Christophersen N. and R F. Wright. 1981 Sulfate flux and a model forsulfate concentrations in streamwater

       at Birkenes, a small forested catchment in southernmost Norway. Water Resour. Res.  17:377-389

                                                           «*
      T'he •Sirecf/D^vpH^9 "* T i™10?"1 6ffeCtS °f acidic deP°sition on surface *** chemistry:-
      The Direct/Delayed Response Project. Eos, Transactions, American Geophysical Union 70:801-813.
 Church^ M.R  and R.S. Turner. 1986. Factors Affecting the Long-term

                                                     - NTIS
     > n0Ln'^Rv ' D!nniS> an? S^K- Se"kop- l989' Re-examination of Interim Estimates of Annual Sulfur Dry
      Depos.t.on Across the Eastern  United States. EPA/600/4-89/026. U.S. Environmental  Protection
      Agency, Atmospheric Research and Exposure Assessment Laboratory, Research T^e pSTSS



Clayton J.L 1986 An estimate of plagioclase weathering rate in the Idaho batholith based upon chemical


                                                     - Dethier' «* Rites
                                                            b*nce
Cochran, W.G. 1977. Sampling Techniques, pp. 259-261. John Wiley & Sons, Inc., New York, NY.
     Rnnnto pLe?' ^J^ ^'^ Remortel, M.L Papp, and G.R. Holdren.  1987a. Direct Delayed
     Response Project: F.eld I Operations and Quality Assurance Report for Soil Sampling in the Southern

     Blue R dge Province of the United States, Volume I: Sampling. EPA/600/4-87/041 US Environmental
     Protection Agency, Environ.  Monitoring Systems Lab., Las Vegas, NV. 205 pp.      tnvironme™al
                        M       ' M'G- Johnson- and G-R- Holdren. 1987b. Direct/Delayed Response

                                                                  ^ ^ironmenyta, Protection





                                 .                          « *»
                                            827

-------
  Cook, FLB., M.L Jones, D.R. Marmorek, J.W. Elwood, J.L Malanchuk, R.S. Turner, and J.P. Smol 1988 The
       Effects of Acidic Deposition on Aquatic Resources in Canada: An Analysis of Past, Present, and Future
       Effects. Oak Ridge National Laboratory, Environ. Sci. Div., Oak Ridge, TN.

  Cooper J.R  J.W. Gillam and T.C. Jacobs. 1986. Riparian areas as a control of nonpoint pollutants, pp
       li  1-  *   £r      e> ed'  Watersned  Research Perspectives.  Smithsonian Institution  Press
       Washington, DC.

  Cosby  B.J.,  G.M  Hornberger,  J.N. Galloway, and R.F. Wright.  1985a. Modeling the  effects of acid
       deposition: Assessment of a lumped parameter model of soil water and streamwater chemistry. Water
       Kesour. Res. 21:51-63.

                   AH°mbeIge,r' J-N. Galloway,  and  R.F. Wright. 1985b.  Time  scales of  catchment
                 :0   quantitative model for estimating freshwater acidification. Environ. Sci. Technol.
 Cosby B.J., G.M. Wright, G.M. Hornberger, and J.N. Galloway. 1985c. Modeling the effects of acid
       deposition: Estimation of long-term water quality responses in a small forested catchment  Water
       Kesour. Res. 21:1591-1601.
      wtnvI; EB-*astetter' J'N- Galloway- and R-F. Wright. 1986a. Estimating catchment
      water quality response to acid deposition using mathematical models of soil ion exchange processes.
      \36Ou6rnri3 38. 77-95.

 Cosby  B.J  G.M.  Hornberger, R.F.  Wright, and  J.N. Galloway.  1986b.  Modeling the effects of acid
                :o5°        long-term sulfate dynamics by soil sulfate adsorption. Water Resour. Res.
 Cosby, B J  P.G. Whitehead, and R. Neale. 1986c. A preliminary model of long-term change in stream
      acidity in southwest Scotland. J. Hydrol. 84:381-401.                                   *>"«*«"

         ^'l G-M;.1Hornberger, P-F.  Ryan,  and D.M. Wolock. 1989. MAGIC/DDRP Final Report  Vol  1-
                                                     QA/Qa  Internal  Report- EPA Environmen tai
                 Hornbe,rgpeKr> and R-F Wright ln press. A regional model of surface water acidification in
                      ?"brat.on and val|dation using survey data. In: J. Kamari, ed. Proceedings of the
      D RedPubCo   CS     m9 °" Environmental lmPact Models to Assess Regional Acidification.
                                                                          n alfis°' °f the tropics-

Cowardln  LM  V  Carter  F.C. Golet, and E.T.  URoe.  1979. Classification of Wetlands and Deepwater
      Sh ?nnw H%  f d Statu/ FWS/OBS-79-31 • Bio'ogical Services Program, U.S. Dept. of the Interior,
      Fish and Wildlife Service, Washington, DC. 131 pp.

Cozzarelli  I.M. J.S  Herman, and R.A. Parnell, Jr. 1987. The mobilization  of aluminum in a natural soil
      system: Effects of hydrologic pathways. Water Resour. Res. 23:859-874.

        i; A'C-.Edwards'. J'M- Reid- D-A- MacLeod, and M.S. Cresser. 1986. The use of catchment studies
         Sf I8'"! 2 ^  wealte™3 rates in two contrasting upland areas in northeast Scotland, pp.
     Academic P     CH^d ^FL         i6r' ^^ ^^ °f Chemical Weathe"ng of Rocks and Minerals.
                                             828

-------
  Cronan  C_S  and G.R. Aiken. 1985. Chemistry and transport of humic substances in forested watersheds

        of the Adirondack Park, New York, NY. Geochim. Cosmochim. Acta 49:1697-1705.



  Cronan  C.S, W.A  Reiners, B.C. Reynolds, Jr., and G.E. Lang. 1978. Forest floor leaching: Contributions

        from mineral, organic and carbonic acids in New Hampshire subalpine forests. Science 200:309-31 1 .
  Cr°n£Tn Jh'p No'S £r'ah aAdHSMSkibinokl' 1987' F°rest vea-etati°n in ^lation to surface water chemistry
       in the North Branch of the Moose River, Adirondack Park, NY. Biogeochemistry 3:121-128.
                               °f
                                      deposition to surfaces in deciduous and pine canopies. Environ.
                                   Sulfur constituents and cycling in waters, seston, and sediments of an
                   lake. Limnol. Oceanogr. 30:1196-1207.




      '«nHB;hM J' !^-Che!!' and J'.P- Nakas" 1982' Organic and inorganic sulfur constituents of a forest soil
       and their relationship to microbial activity. Soil Sci. Soc. Am. J. 46:847-852.
                        ' 3nd. S'C- Schindler- 1984- Dynamics of organic and inorganic sulfur constituents


                               ft^f**- '"= EL  St°ne'  ed F°reSt S°"S and Treatment lmPacts'
                  H      ,lMh N°rth  American Forest Soils  Conference, June 1983.  Dept.  Forestry
            ies, and Wildlife, University of Tennessee, Knoxville.

                                                               the oxide/water
       BMnn,     --';       '  C'W- Chen' R'A- Goldstein- P-W-H. Chan, and R.K. Munson. 1986.
       Big Moose Basin: Simulation of response to acidic deposition. Biogeochemistry 3:141-161.


 De Grosbois E., P.J. Dillon  H.M. Seip,  and R. Seip. 1986. Modelling hydrology and sulfate concentrations
       in small catchments in central Ontario. Water, Air, Soil Pollut. 31:45-57


 Dethier D.F. 1 986. Weathering rates and the chemical flux from catchments in the Pacific Northwest, U.S.A


                                           Dethier' eds- Rates of Chemical Weathering of Rocks and
      amnhPrioHn     WA Sch,eider' 1 ^ The USe °f Calibrated Iakes and watersheds for estimating
      atmospheric deposition near a large point source. Water, Air, Soil Pollut. 18:241-258.



Dillon  PJ.R.A. Reid and R. Girard.  1986. Changes in the chemistry of lakes near Sudbury  Ontario

      following reductions of SO2 emissions. Water, Air, Soil Pollut. 31 :59-65.




                                      1987> The rate of acidification of aquatic ecosystems in Ontario,





                 R   'eof loo ^al°r influence °n the hydrol°ay of New Hampshire and Vermont, USA.
               i. bull. 26:399-413.




    > «0n'ov98t4'tA sy|i°Ptic.suryey of headwater streams in Shenandoah National Park, VA, to evaluate

     sensitivity to acidification by acid deposition. M.S. Thesis. University of Virginia, Charlottesville, VA.
     1 oo pp>




                     J'B' Th°mpSOn' Jr" and MP- Billings- 1961' Centennial Map of Vermont,  (scale
D°"'
                                              829

-------
         HQ         TnJ.977' Carbonate- nalide, sulfate, and sulfide minerals, pp. 75-98. In: J.B. Dixon
       and S.B. Weed, eds. Minerals in Soil Environments. Soil Science Society of America, Madison, WI.


         mn 3t"d '• Sevaldrud- J.9?0-  "-ate acidification, fish damage, and utilization of outfields- A
       compare survey of six h.ghland areas, southeastern Norway, pp. 354-355. In: D. Drabk&s and A


                                                                  °f an lnternatlonal conference:
                                                                              <*


 Draper, N.R., and H. Smith. 1981. Applied Regression Analysis. John Wiley & Sons, Inc., New York, NY.


 Drever, J.I. 1982. The Geochemistry of Natural Waters. Prentice-Hall, Inc., Englewood Cliffs, NJ. 388 pp.
         ninnNlQvnu        °f SOme dilute acidified lakes and streams in the Adirondack
      Region of New York State. Ph.D. Dissertation. Cornell University, Ithaca, NY.



                                           characteristics of Adirondack lakes. Environ. Sci. Technol.
on fish in
        ?he 'NCorth^SnknhaonfdthF JMUnango-- 1 ^ Londitudinal and temP°^' trends in the water chemistry
        me North Branch of the Moose River. Biogeochemistry 3:37-61
                                                     1980' Effect °f aluminum
          L?Ji Wy.sl;?wski cf- Consentlni, and M.E. Smith. 1987b. Processes regulating temporal and

                                         ' a low-°rder woodland —" ln the "
                  JO'    E  Likens'  and M'C- Feller 198a Effei:ts <* acldte deposition on the
                                                                   Bro* New Hampshire- and
Driscoll, C.T., G.E. Likens, LO. Hedin, J.S. Eaton, and F.H. Bormann. 1989b. Chanqes in the chemistrv of

           et      "                        Br°°k ExPerim^ta. Forest, N
                                                                '" ™-** ^-'on analyses of




                                     '" Environmental Plannin9- W-H- Freeman and Company, New





                                                                       **
     aij^  RfL-^De"nis- ln. Revision. On the development  of an inference technique used in the
     estimat,on o  dry deposition of Mg+, Ca+, Na+, and K+. Atmospheric Research and Exposure

     Assessment Laboratdry, U.S. Environmental Protection Agency, Research Trilng e Park NC
                                            830

-------
                           i' ClnriSand C-E Bollin9er- 1Q86. RELMAP: A Regional Lagrangian Model of
              tion User's Guide. EPA/600/8-86/013. NTIS PB86 171 394/AS. U.S. Environmental Protection
       Agency, Research Triangle Park, NC.


                ' GoSS> ^En-Ve ASter' a£d JA Rogalla- 1983' Hydrologic control of lake susceptibility to
                  . Can. J. Fish. Aquat. Sci. 40:1896-1904.

  Eilers, J.M., P. Kanciruk, R.A. McCord, W.S.  Overton, L Hook,  D.J. Blick, D.F. Brakke, P.E. Kellar MS
       DeHaan M E. Silverstem, and D.H. Landers. 1987. Characteristics of Lakes in the Western United
                           ..^31!  9omPendi|Jm  for  Sele<*ed  Physical   and  Chemical  Variables.
                         .  U.S. Environmental Protection Agency, Washington, DC.

  Eilers, J.M  D.H. Landers, and D.F. Brakke. 1988. Chemical and physical characteristics of lakes in the
       southeastern United States. Environ. Sci. Technol. 22:164-172.

  Elassal, A.A., and V.M. Caruso. 1983. Digital elevation models. U.S. Geological Survey Circular 895-B. 40 pp.

  Engstrorn, D.R 1987. Influence of vegetation and hydrology on the humus budgets of Labrador lakes  Can
       J. Fish. Aquat. Sci. 44:1306-1314.

 Environmental Systems Research Institute. 1986. ARC/INFO User's Manual, Version 3.2. Redlands, CA.

 Eralp, A.E., and M.B. Tomson. 1978. pH averaging. J. Water Pollut. Control Fed. 389-392.

                               regional  episodic acidification of surface waters using empirical models.
 Eshleman, KN and H.F. Hemond. 1985. The role of organic acids in the acid-base status of surface waters
      at Bickford watershed, Massachusetts. Water Resour. Res. 21:1503-1510.

 Eshleman, K.N., and P.R Kaufmann. 1988. Assessing the regional effects of sulfur deposition on surface
      water chemistry: The Southern Blue Ridge. Environ. Sci. Technol. 22:685-690.

 Evangelou, V.P., and R.E. Phillips. 1987. Sensitivity analysis on the comparison between the Gapon and
      Vanselow exchange coefficients. Soil Sci. Soc.  Am. J. 51:1473-1479.
     > wH'J98?' Fnr£St C°Ver types °f the United States and Canada. Society of American Foresters,
Falkengren-Grerup  U., N. Linnermark, and G. Tyler. 1987. Changes in acidity and cation pools of south
      Swedish soils between 1949 and 1985. Chemosphere 16:2239-2248.


Fay' Ju"J?' ?o°m?' K3,nd S' Kumar'  1986'  Modelina of the 1900-1980 trend of precipitation acidity at
      Hubbard Brook, New Hampshire. Atmos. Environ. 20:1825-1828.

Fernandez, I J. I985a. Acid deposition and forest soils: Potential impacts and sensitivity, pp 223-239 In-
      D.D Adams and W. Page, eds. Acid Deposition - Environmental, Economic, and Policy Issues Plenum
      Publ. Corp., New York, NY.

Fernandez, I.J.  I985b.  Potential effects  of  atmospheric  deposition on forest  soils  pp  237-250 In-
      Symposium on the Effects of Air Pollutants on Forest Ecosystems. The Acid Rain Foundation, St. Paul,


Fernandez, I.J., and  P.A. Kosian. 1987. Soil air carbon dioxide concentrations in a New England spruce-fir
      forest. Soil Sci. Soc. Am. J. 51:261-263.
                                             831

-------
  Fisher, R.A. 1985. The logic of inductive inference (with discussion). J. Royal Stat. Soc. 98:39-54.

  Fitzgerald  J W., and D.W. Johnson. 1982. Transformations of sulphate in forested and agricultural lands
       pp. 411-426. In: A.I. Moore, ed. Sulphur, Volume 1. British Sulphur Corp., London.

  Fitzgerald, J.W., and M.E. Watwood. 1987. Sulfur retention in hardwood forest litter and soil: Mechanisms
       and environmental parameters, pp. 119-126. In: NAPAP Aquatic Effects Task Group VI Peer Review
       Project Summaries, Volume I, May 17-23. New Orleans, LA.

  Fitzgerald, J.W., T.Q Strickland, and W.T. Swank. 1982. Metabolic fate of inorganic sulphate in soil samples
       from undisturbed and managed forest ecosystems. Soil Biol. Biochem. 14:529-536.

  Fletcher  R  and M.J.D.  Powell. 1963. A rapid convergent descent method for minimization. Computer J
       b: 163-1 68.
R'G'
                                       '' 1985' lndications of tne caPacity for rapid reversibility of lake
 Foss, W.M. 1953a. After the storm. NY State Conserv. 7:2-3.


 Foss, W.M. 1953b. Salvage Report (12). NY State Conserv.  7:4-5.


 Freney J-R. 1961. Some observations of the nature of organic sulphur compounds in soil. Aust. J. Agric.
      liWo.  I tL. , ^r^_Ti~T'vJi(£ .
 Freund, R.J., and R.C. Littell. 1986. SAS System for Regression. SAS Institute, Inc., Gary, NC. 167 pp.

 Fry, B 1986 Stable sulfur isotopic distributions and sulfate reduction in lake sediments of the Adirondack
      Mountains, New York. Biogeochemistry 2:329-343.

 Fuller, W.A. 1987. Measurement Error Models. John Wiley & Sons, Inc., New York, NY. 440 pp.

 Fuller R.D., MB. David and C.T.  Driscoll. 1985. Sulfate adsorption relationships in forested Spodosols of
      the northeastern USA. Soil Sci. Soc. Am. J. 49:1034-1040.

 Fuller R.D  C.T.  Driscoll,  S.C.  Schindler, and  M.J. Mitchell. I986a. A simulation model  of  sulfur
      transformations in forested Spodosols. Biogeochemistry 2:313-328.

 Fuller, R.D.  M.J  Mitchell, H.R. Krouse, B.J. Wyskowski, and C.T. Driscoll. 1986b. Stable sulfur isotope ratios
      as a tool for interpreting ecosystem sulfur dynamics. Water, Air, Soil Pollut. 28:163-171.

 Fuller RD  C.T. Driscoll G.B  Lawrence, and S.C. Nodvin. 1987. Processes regulating sulphate flux after
      whole-tree harvesting. Nature 325:707-710.

      >, GM iWesta"' and. p- Sollins-  1989. The study of soil chemistry through quasi-steady-state models-
      I. Mathematical definition of model. Geochim. Cosmochim. Acta. 53:595-601.

Gaines Jr., G.J., .and H.C. Thomas.  1953. Adsorption studies on clay minerals. II. A formulation of the
      tnermodynamics of exchange adsorption. J. Chem. Phys. 21:714-718.

Galloway,  J.N  C.L Schofield,  G.R. Hendrey, N.E. Peters, and A.H. Johannes. 1980. Sources  of acidity in
      three  lakes acidified during  snowmelt, pp. 264-265. In: D. Drabl^s and A. Tollan, eds  Ecological
      Scfc D  • Acid,freciPitation. Proceedings of an International Conference, Sandefjord, March 11 14
      SNSF Project, Oslo-As, Norway.
                                              832

-------
 Galloway J.NG.E. Likens, W.C. Keene, and J.M. Miller. 1982. The composition of precipitation in remote
       areas of the world. J. Geophys. Res. 87:8771-8787.

 Galloway, J.N., S A Norton, and M.R. Church. I983a. Freshwater acidification from atmospheric deposition
       of sulfunc acid: A conceptual model. Environ. Sci. Techno!. 17:541-545.

 Galloway J.N., C.L Schofield  N.E. Peters, G.R. Hendrey, and E.R. Altwicker. 1983b. Effect of atmospheric
       sulfur on the composition of three Adirondack lakes. Can. J. Fish. Aquat. Sci. 40:799-806.    P


                                             1984' ACW PreCipitati°": Natural versus anthropogenic
 Galloway J.N., G.R. Hendrey, C.L Schofield, N.E. Peters, and A.H. Johannes. 1987. Processes and causes

      J  Fish Aquat  Sc"              "™™* '" ^ West'central Adirondack Mountains, New York. Can.
 Gapon, Y.N. 1933. On the theory of exchange adsorption in soils. J. Gen. Chem. USSR 3:144-160.

 Garrels, R.M  and FT.  Mackenzie. 1967. Origin of the chemical compositions of some springs and lakes.
      pp. 222-242. In. Advances in Chemistry, Series 73. American Chemical Society, Washington, DC.


                    Application of the ILWAS model to the Northern Great Lakes States. Lake  Reserv.
    >  A Mrf'SwSS'Sl^ GJ' ST18?' o986- The r°'e °f a'kaline materials in Palpitation chemistry:
      A brief review of the issues. Water, Air, Soil Pollut. 30:245-251 .
      *«m     D Afal verifLcati.on  of Polygonal resource maps: A low-cost approach to accuracy
      assessment. Photogram. Engin. Remote Sensing 52:839-846.

Georgakakos, KP., G M. Valle-Filho, N.P. Nikolaidis, and J.L Schnoor. In Press. Lake acidification studies-
      The role of input uncertainty in long-term predictions. Water Resour. Res.

Gherir£ S.A. L  Mok, R.J. Hudson, G.F. Davis, C.W. Chen, and R.A. Goldstein. 1985. The ILWAS model-
      Formulation and application. Water, Air, Soil Pollut. 26:425-459.
                                c,lR- April' °- Chen' N' Clesceri' C- Cronan' C- Dri^oll, RJ. Johannes
       nm       , M      - !,nd C' Schofield 1989- Reaional Integrated Lake-Watershed Study (RILWAS):
     Summary of Major F.ndmgs. EPRI RP-2174-1. Electric Power Research Institute, Palo Alto, CA.
                                  ' R'K"
                                                               1988' B'Ue Lake Acidification Study.



                                             833

-------
1985'
                                                        lake-watershed acidification study: Summary.
  Goldstein R.A  S.A. Gherini, C.T. Driscoll, R. April, C.L Schofield, and C.W. Chen. 1987. Lake-watershed
       acidification in the North Branch of the Moose River: Introduction. Biogeochemistry 3:5-20.

  Gorham, E. 1955. On the acidity and salinity of rain. Geochim. Cosmochim. Acta 7:231-239.


  G°rham> E'C ar!u-N-E 9etenbeck- 1986- Sulphate in bog waters: A comparison of ion chromatography with
       Mackereths cation-exchange technique and  a revision of earlier views on  cause of bog  acidity.
       CCOlOyy /4.O99-903.


  Gorham, E. P.M. Vitousek, and W.A. Reiners.  1979. The regulation of chemical budgets over the course of
       terrestrial ecosystem succession. Ann. Rev.  Ecology Systematics 10:53-84.


                                                          correlations in the eastern United
 Gorham, E S J. Eisenreich J. Ford, and M.V. Santelmann. 1985. The chemistry of bog waters, pp. 339-363
       in: w. Stumm, ed. Chemical Processes in Lakes. Wiley-lnterscience, New York,  NY.

 Gorham E., J K Underwood, F.B. Martin, and J.G. Ogden, III. 1986. Natural and anthropogenic causes of
       lake acidification in Nova Scotia. Nature 324:451-453.                                 ^u=>e& ui
        ndthth,    -5- Ko9' and GJ- All°rd 1988' Maps of Runoff in the Northeastern Region
      and the Southern Blue Ridge Province of the United States During Selected Periods in 1983-85 U S
      Geological Survey Open-File Report 87-106, Washington,  DC. 2 plates. 8 pp.                '  '  '

 Greb, S R.K. Munson, S.A. Gherini, LE. Gomez, C.W. Chen, P.J.  Garrison, and D.R. Knauer. 1987  RILWAS-

                                          ** Dep°sition on Crystal' Vandercook, and Little Rock Lakes.
 Grennfelt, P. 1987. Deposition processes for acidifying compounds. Environ. Technol. Letters 8:515-527.

 Groterud, O. 1984. A conceptual model of lake acidification. Verh. Internal. Verein. Limnol. 22:686-691.
             >  £" nC; 5s0chwandtner' and K- E|dridge.  1985. Historic emissions of sulfur and nitrogen
                                    Jj0- EPA/600/7-85/^a- U.8- ^ironmenta, Protection
HainesT_ A. 1981 Acidic precipitation and its consequences for aquatic ecosystems: A review. Trans. Am.
      risn. oOC. 110:669-707.
      inpwF   JHAvk/ie,laSZei-1983- A.7?lonal survey of the chemistry of headwater lakes and streams
      in New England: Vulnerability to acidification. In: Air Pollution and Acid Rain. FWS/OBS-80 #15 U S
      Dept. Interior, Fish and Wildlife Service, Washington, DC. 141 pp.
      HMhhaHR-.f' and G'R- HendreV- 1980-  Experimental acidification of a stream in
      Hubbard Brook Experimental Forest, New Hampshire. Ecology 61 :976-989.

Hammond, E.H  1964. Classes of land surface form  in the forty-eight  states,  Annals  of the Assoc. of
      American Geographers, 54(1) map supplement No. 4.


                              ' AdSOIpti°n and desorPtion of sulfate in some soils of the West Indies.
                                             834

-------
                                 ***** °f Tennessee' State of Tennessee Dept. Conservation, Division of


 Haren M.F., and R.D. Van Remortel. 1987. Direct/Delayed Response Project: Field Operations and Quality
       SSS? sft  ^ f°r ^"Sampling and  Preparation in the Southern  Blue Ridge Provlr.ee oMhe
       FnS  M   ;  ° T h PreParat'0n- EPA/6°0/4-87/041. U.S. Environmental Protection Agency,
       Environ. Monitoring Systems Lab., Us Vegas, NV. 69 pp.


 Harvey H.W. 1969. The Chemistry and Fertility of Sea Waters. Cambridge University Press  NY 240 pp

       ha                                    S°IUbility
 Hayden, N J. 1987. Sulfate Retention and Release in Six Eastern Soils. M.S. Thesis. Dept. Civil and Environ
       Engineering, Michigan State University, East Lansing, Ml. 165 pp.                        environ.






 Helgeson KQ, W.M. Murphy, and P. Aagaard. 1984. Thermodynamic and kinetic constraints on reaction

       SSL?, mon9 ™nerals and Aqueous solutions: II. Rate constants, effective surface area and the
       hydrolysis of feldspar. Geochim. Cosmochim. Acta 48:2405-2432.
    -K DUnkle' ^f6- lnPut-°utPut Begets of Selected Nutrients on an Experimental

raomall, paAS°7npp         ^ US°A F°reSt ^'^  Northeastern Forest  Experiment


                             's Bog, Concord, Massachusetts. Ecological Monographs
                  ' Biogeochemistry
         th    M- E,fhleman-A;1984- Neutralization of acid deposition by nitrate retention at Bickford
      Watershed, Massachusetts. Water Resour. Res. 20:1718-1724.


      anHHvHr-h Ga"7ay' S'A: N°rt°n' C'L Schofield' P-W. Shaffer, and D.A. Burns. 1980. Geological
      and Hydrochermcal Sens.t.vity of the Eastern United States to Acid Precipitation. EPA/600/3-80/024
      U.S. Environmental Protection Agency, Washington, DC. 100 pp.                        °u/u<^.


Henriksen AJ979. A simple approach for identifying and  measuring acidification of freshwater. Nature
      ^/o.O4^i-o4.
Henrlksen, A. 1980 Acidification of freshwater - a large scale tltration, pp. 68-74. In: D. Drabtos and A

                                                                  of an lntemational' Conference:
           . Li™o|C22n692S698baSe °ati°n oonoentrations due to freshwater acidification. Varh. Internal.
                                            -"'*"«<=- « "«™asn to the acldfty o, surface waters
                               precipltatlon cn a sma" add lake in
                                                                                       Norway
                                                        iC Carb°" -"-ntrations in acid,c lakes ,n
                                             835

-------
 Herczeg, A.L, and R.H. Hesslein. 1 984. Determination of hydrogen ion concentration in softwater lakes using
       carbon dioxide equilibria. Geochim. Cosmochim. Acta 48:837-845.

 Herlihy, AT., A.L. Mills, G.M. Hornberger, and A.E. Bruckner. 1987. The importance of sediment sulfate
       reduction to the sulfate budget of an impoundment receiving acid mine drainage. Water Resour. Res.
       23:287-292.

 Hewlett, J.D. and A.R. Hibbert. 1967. Factors affecting the response of small watersheds to precipitation
       m humid areas,  pp. 275-289. In: W.E. Sopper and H.W. Lull, eds. Symposium on Forest Hydraloqv
       Pergammon Press, New York, NY.                                                  yu,u,uyy.

 Hicks B.B  R.P. Hosker, Jr., and J.D. Womack. 1986. Comparisons of wet and dry deposition derived from
       the first year of trial dry deposition monitoring, pp. 486-490. In: Symposium on Acid Rain: I Sources
       and Atmospheric Processes. Presented before the Division of Petroleum Chemistry, Inc. American
       Chemical Society, April 13-18, New York, NY.

 Hillman  D.C.J  J.F. Potter, and S.J. Simon. 1986. National Surface Water Survey: Eastern  Lake Survey
       (Phase I), Analytical Methods Manual. EPA/600/4-86/009. U.S. Environmental Protection Agency, Las


 HingstonF^LRj^ Atkinson, A.M. Posner, and J.P. Quirk. 1967. Specific adsorption of anions. Nature
       c. i o. I T'Oy" i ^to i .
 Hingston, F.J., A.M. Posner, and J.P. Quirk. 1972. Anion adsorption by geothite and gibbsite: I. The role of
      tne proton in determining adsorption envelopes. Soil Sci. 23:177-192.

 Holdren, G.R., and C.I. Liff.  In Preparation. An expanded version of Reuss' model for cation exchange
      equilibria: A user's guide. U.S. Environmental Protection Agency, Corvallis, OR.

 Holdren  G.R  and P  M. Speyer. 1985. PH dependent changes in the rates and stoichiometry of dissolution
      of an alkali feldspar at room temperature. Am. J.  Sci. 285:994-1026.

 Holdren, G.R T.M. Brunelle, G. Matisoff, and M. Wahlen. 1984. Timing the increase in atmospheric sulphur
      deposition in the Adirondack Mountains. Nature 31 1 :245-248.

 Holdren  G.R., C.I. Liff, and  D.L Cassell.  In Press. Cation exchange models and the prediction of soil
      buffering capacity. In: R. Bassett arid D. Melchior, eds. Chemical Modeling in Aqueous Systems II
      ACS Symposium Series, Washington, DC.

            'S;; and CA  Federer-  1985- Estimating  the buffer capacity of  forest soils.  J.  Forestry
            ~
Hornberger, G.M.  K.J. Beven, B.J. Cosby,  and D.E.  Sappington. 1985. Shenandoah watershed study-
      Calibration of a topography-based, variable contributing area hydrological model of a small forested
      catchment. Water Resour. Res. 21:1841-1850.

Hornberger, G.M B.J. Cosby, and J.N. Galloway. 1986.  Modelling the effects of acid deposition: Uncertainty
      and spatial variability in estimation of  long-term  sulfate dynamics in a region. Water Resour. Res.
      ££'.\ 293-1302.
                                                                             •-x

Hornberger, G.M  B.J. Cosby, and R.F. Wright. 1987a. Analysis of historical surface water acidification in
      southern Norway using a regionalized  conceptual model  (MAGIC), pp. 127-132 In- M B Beck ed
      Systems Analysis in Water Quality Management.  Pergammon Press, New York NY
                                             836

-------

                                         me;ieorol°9ical »* Chemical filterpack monitoring system for







                                              ' SU"ate from sulfate
                             ^

                               - S°"
"8Cid ra'"" Wtth
                                                                         n«ra.e..o-su,fa,e ratios.

Hurlbu^Jr., C.S, and C. Klein. 1977. Manual of Mineralogy. John Wiley & Sons, Inc., New York, NY. 532



Husar, R.B. 1985. Manmade SOX and NOX emissions and trends for eastern North America. 48 pp.
                             M                   °Xid6S and trends for eastern North America, pp.
                                                                         Oeposition, ed. Acfd
Stream and its valley- Verh-
                                                                                    . Verein.
     NY  (sca     so 000)
                                 Se0'°9iC MTo°f N6W Y°rk' Adi™dack, Hudson-Mohawk and Lower

                                       m ^ SdenCe SeiViCe' Map and Chart Series No' 15'
                                            837

-------
 Jeffries, D.S., R.G. Semkin, R. Neuteuther, and M. Seymour. 1986. Influence of atmospheric deposition on
       lake mass balances in the Turkey  Lakes watershed, central  Ontario. Water, Air,  Soil Pollut.
       30; 1033-1044.


 Jeffries, D S., J.R.M. Kelso, and I.K. Morrison. 1988. Physical, chemical, and biological characteristics of the
       Turkey Lakes watershed, central Ontario, Canada. Can. Spec. Publ. Fish. Aquat. Sci. 45:3-13.

 Jenne, E.A  LE. Eary, LW. Vail, D.C. Gin/in, A.M. Liebetrau, LF. Hibler, T.B. Miley, and M.J. Monsour 1989
       ??„,««    °n and Ana|ysis of Tnree  Dynamic Watershed Acidification Codes (MAGIC ETD and
       ILWAS). PNL-6687/UC-11.  Richland, WA.                                               '


 Johannes, A.H.  and  E.R. Altwicker. 1980. Atmospheric inputs to three  Adirondack lake watersheds  pp
       256-257. In: D. Drabl^s and A. Tollan, eds. Ecological Impact of Acid Precipitation: Proceedings of
       an International  Conference, Sandefjord, March 11-14. SNSF Project, Oslo-As, Norway.

 Johannes A.H., J.N. Galloway, and D.E. Troutman.  1980. Snow pack storage and ion release pp 260-261
       in: D.  Drabl^s and A. Tollan, eds. Ecological Conference Impact of Acid Precipitation: Proceedings
       of an International Conference, Sandefjord, March 11-14. SNSF Project, Oslo-As, Norway.

 Johannes, AH.  E.R.  Altwicker, and  N.L Clesceri. 1981. Characterization of acid precipitation in  the
       Adirondack Region. EPRI EA-1826.  Electric Power Research Institute, Palo Alto, CA.

 Johannes, A.H., E.R. Altwicker, and N.L Clesceri. 1984. Atmospheric Inputs to the ILWAS Watersheds in the
       Integrated Lake-Watershed Acidification Study, Volume 4: Summary of Major Results EPRI EA-3221
       Tetra Tech, Inc., Palo Alto, CA.                                                '


 Johannes, A.H  E.R Altwicker, and N.L Clesceri. 1985. The integrated lake-watershed acidification study-
       Atmospheric inputs. Water, Air, Soil Pollut. 26:339-353.

 Johnson, D.W. 1984. Sulfur cycling in forests. Biogeochemistry 1:29-43.

 Johnson, D.W., and D.W. Cole. 1980. Anion  mobility in  soils: Relevance to nutrient transport from forest
       ecosystems. Environ. Internat. 3:79-90.

 Johnson D.W., and G.S. Henderson. 1979. Sulfate adsorption and sulfur fractions in a highly weathered soil
       under a mixed deciduous forest. Soil Sci. 128:34-40.

 Johnson  D.W. and J.O.  Reuss. 1984. Soil-mediated effects of atmospherically deposited sulphur and
      nitrogen. Phil. Trans. R. Soc. Lond. 3056:383-392.

 Johnson D.W., and D.E. Todd. 1983. Relationships among iron, aluminum, carbon, and sulfate in a variety
      of forest soils. Soil Sci. Soc. Am. J. 47:792-800.

Johnson D.W., and D.E. Todd. 1987. Nutrient export by leaching and whole-tree harvesting in a loblolly pine
      and mixed oak forest. Plant and Soil 102:99-109.

Johnson D.W., J.W. Hornbeck, J.M. Kelly, W.T. Swank, and  D.E. Todd. 1980. Regional patterns of soil
      sultate  accumulation: Relevance to ecosystem sulfur budgets, pp. 507-519 In- DS  Shriner  CR
      Richmond and S.E. Lindberg, eds. Atmospheric Sulfur Deposition: Environmental Impact and Health
      Effects. Ann Arbor Science, Ann Arbor, Ml.


Johnson, D.W.  J. Turner, and J.M. Kelly. 1982a. The effects of acid rain on forest nutrient status Water
      Resour. Res. 18:449-461.
                                              838

-------
        urnriQhr  r'         '     Lindberg' D'D' Richter' D"S' Shriner' D'E T°dd, and J.
       Turner. 1982b. Cycling of organic and inorganic sulfur in a chestnut oak forest. Oecologia 54:141-148.

  Johnson, D.W., G.S. Henderson and D.E. Todd. 1988a. Changes in nutrient distribution in forests and soils
       or walker Branch watershed, Tennessee, over an eleven-year period.  Biogeochemistry 5:275-293.

  Johnson, M.G  1986. Clay mineralogy and chemistry of selected Adirondack Spodosols.  Ph.D  Dissertation
       Cornell University, Ithaca, NY. (Diss. Abstr. 86-7348) 152 pp.                        u'fcsenauon.


  Johnson, M.G   P.W. Shaffer, D.L. Stevens, K.W. Thornton, and R.S. Turner. 1988b. Predicting and
       Forecasting  Surface Water Acidifibation:  A Plan  to Assess  Data Aggregation  Effec s  US
       Environmental Protection Agency, Environ. Res.  Lab., Corvallis, OR. 93 pp

  Johnson, N.M., C.T. Driscoll, J.S.  Eaton, G.E. Likens,  and W.H. McDowell! 1981. "Acid rain" dissolved
       aluminum and  chemical weathering at the Hubbard Brook Experimental Forest, New Hampshire
       Geochim. Cosmochim. Acta 45:1421-1437.
Johnson^
                     Likens, M.C. Feller, and C.T.  Driscoll.  1984. Acid rain and soil chemistry. Science
                                               Multivariate Statistical Analvsis-
                                                                                       Hall,  Inc.,
          imt'H--  Bubenzer' G'B- Lee' RW-  Madison- ^d J.R. McHenry. 1984. Nutrient trapping by
      sediment deposition in a seasonally flooded lakeside wetland. J. Environ. Qual. 13(2):283-290.

 Johnston C A., N.E. Detenbeck, J.P. Bonde, and G.J. Niemi. 1988. Geographic information systems for
      cumulative impact assessment. Photogram. Engin. Remote Sensing 54:1609-1615.


            G'K MOS' Qoi My^' BT' Rhyne> and JA Watts- 198a User's Guide for tne Soil Description

                                                                  °*k ™*  Nati°nal ^°^>
      rLninn m°f,nf', a"d J'^ ^]amS- ^ Chronoloay. magnitude and paleolimnological record of
      changing metal fluxes related to atmospheric deposition of metals in New England   pp 23-35 In-
      u.P. Bncker, ed. Geologic aspects of acid deposition. Ann Arbor Science, Ann Arbor, Ml.
                 thrt           oH- LanderS' D-R Brakke' and R'A- Linthurst- 1986a- Characteristics
       h      ,w •  t,   S^T,Unitfd States' Volume lll: Data  Compendium of Site Characteristics and
      Chemical Variables. EPA/600/4-86/007c. U.S. Environmental Protection Agency, Washington, DC. 439
      |jp.


Kanciruk, P., M. Gentry, R. McCord, L Hook, J.M. Eilers, and M.D. Best. 1986b. National Surface Water

      nRM?y/VMaf£rnJ ^e,So-^ey l^386 !)> Data Base DictionarV- Environ. Services Div. Pub. No. 2778.
      ORNL/TM-10307. Oak Ridge National Laboratory, Oak Ridge, TN. 77 pp.

Kaufmann, PR  AT^Herlihy, J.W& Elwood, M.E. Mitch, W.S. Overton, M.J. Sale,  J.J. Messer, K.A. Cougar,
      OH   •  k'r^   Reckn.ow- AJ- Kinney- S.J- Christie, D.D. Brown, C.A. Hagley, and H.I. Jager. 1988
      Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern United States  Volume I-

                    ^^
                                             839

-------
 Keller, W., and J.R. Pitbaldo. 1986. Water quality changes in Sudbury area lakes: A comparison of synoptic
      surveys in 1974-1976 and 1981-1983. Water, Air and Soil Pollut. 29:285-296.

 Kelly, C.A., and J.W.M. Rudd. 1984. Epilimnetic sulfate reduction and its relationship to lake acidification.
      Biogeochemistry 1:63-77.

 Kelly, C.A.,  J.W.M. Rudd, R.B. Cook, and D.W. Schindler. 1982.  The potential importance of bacterial
      processes in regulating rate of lake acidification. Limnol. Oceanogr. 27:868-882.

 Kelly, C.A., J.W.M. Rudd, R.H. Hesslein, D.W. Schindler, P.J. Dillon, C.T. Driscoll, S.A. Gherini, and R.E.
      Hecky. 1987. Prediction of biological acid neutralization in acid-sensitive lakes. Biogeochemistry
      3:129-140.

 Kerekes, J., S. Beauchamp, R. Tordon, C.  Tremblay, and T. Pollock. 1986. Organic versus anthropogenic
      acidity in tributaries of the  Kejimkujik watersheds in western Nova Scotia. Water,  Air  Soil Pollut
      31:165-173.

 Khanna, P.K., J. Prenzel, K.J. Meiwes, B. Ulrich, and E. Matzner. 1987. Dynamics of sulfate retention by acid
      forest soils in an acidic deposition environment. Soil Sci. Soc. Am. J. 51:446-452.

 King, P.B., R.B. Neuman, and J.B. Hadley. 1968. Geology of The Great Smoky Mountains National Park,
      Tennessee and  North Carolina. U.S.  Geological Survey, Professional Paper 587.

 Knox, C.E., and T.J. Nordenson. 1955. Average annual runoff and  precipitation in the  New England-New
      York area. U.S.  Geological Survey Hydrologic Investigations Atlas HA-7. 6 pp.

 Kramer, J.R. 1986. Re-evaluation of trends in surface water alkalinity in New York and New Hampshire. Dept.
      Geology, McMaster University,  Hamilton,  Ontario, Canada.

 Kramer, J.R., and S.S.  Davies. 1988. Estimation of non-carbonate protolytes for selected lakes in the Eastern
      Lake Survey. Environ. Sci. Technol. 22:182-184.

 Kramer, J.R., and A. Tessier.  1978. Acidification of aquatic systems: A critique of chemical approaches.
      Environ. Sci. Technol. 16:606-615.

 Kramer, J.R., A.W. Andren, R.A. Smith, A.H. Johnson, R.B. Alexander, and G. Oehlert. 1986. Streams and
      lakes, pp. 231-299. In: Committee on Monitoring and Assessment of Trends in Acid  Deposition, ed.
      Acid Deposition: Long-Term Trends.  National Academy Press, Washington, DC.

 Krug, E.G. 1987. Acid deposition/watershed interactions: The importance  of humic substances and
      biological processes. AAAS  Annual Meeting,  Manuscript.

Krug, E.G. 1989. Assessment of the theory and hypotheses of the acidification of watersheds. Illinois State
      Water Survey Division, SWS Contract Report 457. 252 pp.

Krug, E.G., and C.R. Frink. 1983. Acid rain on acid soil: A new perspective. Science 221:520-525.

Krug, E.G.,  P.J.  Isaacson,  and  C.R. Frink. 1985. Appraisal  of some current hypotheses  describing
      acidification of watersheds. J. Air Pollut. Contr. Assoc. 35:109-114.

Krug, W.R., W.A. Gebert, D.J. Graczyk, D.L Stevens, B.P. Rochelle, and M.R. Church.  In Press. Runoff Maps
      for the Northeastern,  Southeastern, and Mid-Atlantic United States for 1951-1980. U.S. Geological
      Survey Water Resour. Invest. Report  No. 88-4094.
                                              840

-------
  Kulp, L  1987. Interim Assessment, The Causes and  Effects of Acidic  Deposition Volume I- Executive

       Summary. National Acid Precipitation Assessment Program, Washington! DC




       w^'tor'oh'^ ^°bba' D-'S- Jeffries> and D" Craig' 198a Modelling stream chemistry for the Turkey Lakes
       Watershed: Comparison with 1981-84 data. Can. J. Fish, and Aquatic Sci. 45:72 80
       Ooerato'nSand->AJJ- Lee' ^-L Stevens, W.G. Campbell, and M.G. Johnson.  1987a. Field

       Operations and Quality Assurance/Quality Control for Direct/Delayed Response Project Soil Maooinn

       Act,v,t,es  ,n the Southern Blue Ridge Region. EPA/600/3-88/016 NTIS  PB88 ?95 722/AS  U S

       Environmental Protection Agency, Corvallis, OR. 122 pp.                          722/A5.  U.S.
            0ea«ons         .W^ SteV^S' R'S' Tumer' W'G' CamPbe11' and M'G' Johnson. 1987b.
            Operations and Quality Assurance/Quality Control for Direct/Delayed Resoonse ProiPrt Snil
       Mapping Activities in  the Northeast Region. EPA/600/3-87/017 ^ NTIS PB88 17° 301 /AS  US
       Environmental Protection Agency, Corvallis, OR.  127 pp.                         -301/AS.  U.S.



                             ^^
                         w                            f °r9anio and inor9anic sulfur constituents In
                         water. Intern. J. Environ. Anal. Chem. 14:245-256.
                                                   'E Ke"ar' M'E Silverstein- R'D- Schonbrod, R E
                                n                  ^ EP' Meier- 1987- Characteristics of Lakes in




                                                      198a Eastern uke survey: Regionai estimates
      mHainn    .                 ^^tion: A general model for metal adsorption onto geological
      materials, pp. 1-18. In: P.M. Tewan, ed. Adsorption from Aqueous Solutions. Plenum Press, New York,




Lasaga, A.C. 1984a. Chemical kinetics of water-rock interactions. J. Geophys. Res. 89:4009-4025.



                                              - pp- 269-274-  ln:  Fourth
                         in9' 1985' The determlnallon of soil sensitivity to acid deposition. Water, Air, Soil

     effect of
                             on surface water chemistry. Environ. Manage. 13:95-108.
                                             841

-------
  Lee, J.J., D.A. Lammers, D.L Stevens, K.W. Thornton, and K.A. Wheeler. 1989b. Classifying soils for acidic
       deposition aquatic effects: A scheme for the northeastern U.S. Soil Sci. Soc. Am. J. 53:1153-1163.


  Lee, J.J., D.R. Marmorek, K.W. Thornton, D.L. Stevens, and D.A. Lammers. 1989c. Direct/Delayed Response

       £?^«9efinlt!on °f  Soil SamP|in9 Classes and  Selection of Sampling Sites for the Northeast.
       EPA/600/3-89/041.  U.S. Environmental Protection Agency,  Environmental Research Laboratory
       Corvallis, OR.                                                                              '


  Lee, S. 1987. Uncertainty Analysis for Long-Term Acidification of Lakes in Northeastern USA Ph D Thesis
       University of Iowa, Iowa  City.                                                   ....


  Lee, S  and J.L  Schnoor. 1988. Reactions that modify chemistry in lakes of the NSWS Environ  Sci
       Tecnnol. 22:190-197.


  Lee, S.J   K.P.  Georgakakos, and J.L. Schnoor. In  Review.  An uncertainty analysis of long-term lake
       alkalinity predictions. Water Resour. Res.


 Lepisto, A., P.G.  Whitehead,  C. Neal,  and  B.J. Cosby.  1988. Modeling  effects  of acid  deposition-
       Estimation of long-term water quality responses in forested catchments in Finland. Nordic Hydrology
       1 9i99~1 20.


 Levine E.R., and EJ. Ciolkosz. 1986. A computer simulation model for soil genesis applications  Soil Sci
       Soc. Am. J. 50:661-667.


 Levine E.R., and EJ. Ciolkosz.  1988. Computer simulation of soil sensitivity to acid rain. Soil Sci. Soc. Am.
       j. o^.iiuy~^ 1 5.
 Lewis, Jr  W.M., and M.C. Grant. 1979. Change in the output of ions from a watershed as a result of the
      acidification of precipitation. Ecology 60:1093-1097.


 Liegel L.H., MR. Church, D.L. Cassell, W.G. Campbell, D.J. Bogucki, G.K. Gruendling, and E B Allen  In
      Review  Using color infrared aerial photography in regional acid precipitation research. Photogram
      Engm. Remote Sensing.


 Liesko I., K. Zotter, and F. Szakal. 1987. Clarification of surface waters affected by acid rain pp 391-398
      In: R. Perry,  R.M.  Harrison, J.N.B. Bell and J.N. Lester, eds. Acid Rain: Scientific and Technical
      Advances. Selper, Ltd., Ealing, London.


 Likens G.E., F.H. Bormann, R.S. Pierce, J.S. Eaton, and N.M. Johnson. 1977. Biogeochemistry of a Forested
      Ecosystem. Springer-Verlag, Inc., New York, NY. 146 pp.


 Un>  J'fl 2-677 694  Schn°°r' 1986- Acid PreciPitation model for seepage lakes. J. Environ. Engineer.



 Lin, J.C.,  J.L Schnoor, and G.E. Glass. 1987. Ion budgets in a seepage lake. pp. 209-227. In: R A Heits and
      b.J. Eisenreich, eds. Sources and Fates of Aquatic Pollutants. Advances in Chemistry Series 216
      American Chemical Society, Washington, DC.                                          «^,D,


 Lindberg  SE.   and C.T.  Garten,  Jr. 1988. Sources of sulphur in forest  canopy  throughfall. Nature
      ooo: i4o- 151 .


Lindberg, S.E., G.M  Lovett, D.D. Richter, and D.W. Johnson. 1986.  Atmospheric deposition and canopy
      interactions of major ions in a forest. Science 231:141-145.


Lins, H.F. 1986. Recent patterns of sulfate variability in pristine streams. Atmos. Environ. 20:367-375.


                                               842

-------
                      fI' J'M' E"ers' D'F- Brakke' W-S- Overton, E.P. Meier, and R.E. Crowe. 1986a
                     of Lakes  in the Eastern United States, Volume I: Population  Descriptions and
                                      EPA/600/4-86/007, U.S.  Environment,, ProtS?
 Linthurst, R A  D.H Landers, J.M. Eilers, P.E. Kellar, D.F. Brakke, W.S. Overton R E  Crowe E P  Meier  P
       Kanc,ruk, and D.S. Jeffries. 1986b, Regional chemical characteristics of lakes in X America  Part
       II - Eastern United States. Water, Air, Soil Pollut. 31:577-591.                       America. Part


 Loucks O.L  G.E. Glass, J.A.  Sorensen, B.W.  Liukkonen, J. Allert, and  G.  Rapp  Jr  1986  Role  of

       Sol ^Srai0:^17 V6rSUS ^ WaterSh6d Pr°Pertl'eS h WiSC°nSin lake acSification Water AJr,
LU"'
                'E S°PPer" 1966' FaCt°rS that inf'Uence streamflow ^ the Northeast. Water Resour. Res.
                  urwR Sensitivitv , of Stream ^sins in  Shenandoah National Park to Acid
                .  USGS  Water  Resources  Investigation  Report  85-4115  US  Geoloaical  SUIVPV
      Washmgton, DC, and Dept.  Environ. Sci, University of Virginia, Richmond VA. 61 pp           V>

 Madansky, A. 1988. Prescriptions for Working Statisticians. Springer-Verlag, Inc., New York, NY. 295 pp.

 Mairs, D.F. 1967. Surface chloride distribution in Maine lakes. Water Resour. Res. 3:1090-1092.

                       ^^^
Matalas KO, and  B. Jacobs.  1964. A Correlation  Procedure for Augmenting Hydrology  Data US
      Geological Survey Professional Paper 434, Washington, DC. 7 pp.
     tadro.n         'L JaCkS°n- 197a Glbbslte solubilit>' and thermodynamlc properties of
     hydroxy-alummum ions in aqueous solutions at 25 C. Geochim. Cosmochlm. Acta 43:
                                                         n- ^600/3-30/013. US. Environment

     ^25o!ooo). Ge°'0giC Map of Pennsy'vania. Dept.  Environmental Resources,  Harrisburg, PA. (scale
                                            843

-------
D1erS'
                               B"?dowskt G'B- Lawrence, and M.B. David. 1984. Organic and inorganic
                                                New York lakes: Effect of ••* sediment depth
                                                              soils
                                                                                            °f
           'irtf™™°±f1ter' iSrf°K-8?' and SA Ncrton' 1988' Comparison of sulfur concentrations within
           sediment profiles. Hydrobiologia 157:219-229.
 Mohnen, V.A. 1988. The challenge of acid rain. Scientific American 259:30-38.
                   Rf '  Gt°°u I9?8' Stream  chemistrV in the New Jersey pinelands: The influence of
                  and watershed disturbance. Water Resour. Res. 24:1091-1100.

               eds- Add Rain:  ScientifL and Technical
Murdoch PS  N E Peters^ and R.M^ Newton. 1984. The integrated lake-watershed acidification study. In:
      fn*titut2  Pa\oA\lo', CA     '             * Hydrologic Ana|ysis' EA-3321' Electric Power Research

                  - WhtLtehrad> BJ' COSby' G'M' Hornberger- a^ R.F. Wright. 1987. Regional modeling
                      P6   MWa/ re9'onrin.southwest Scotland. In: J. Kamari, ed. Proceedings of the
      D Reidel Publ Co             O" Environmental lmPact Models l° Assess Regional Acidification.
                                       alterati°ns
                                                                 boreal *«- -«m. by beaver
                                                         Alkalinity

National Academy of Sciences. 1984. Acid deposition: Processes of lake acidification.  Summary of a
                           f^frHh  CRoun?DCommi^ion  on Physical  Sciences, MathematicZ and
           Washng on   : 1 1      "     '       O" Processes of Lake Acidification. National Academy
                                            844

-------
                                    - AddM D^Position: Long-Term Trends. National  Research Council

                                                          Resources- Envi™




                                    £
                                      °f SU'PhUr deposition on '^e-water chemistry in Ontario, Canada.
                                                                                           - Soil
       aonvinMRh^;           relationshiP between surface water chemistry and
       geology in the North Branch of the Moose River. Biogeochemistry 3:21-35.



          hN" 1987' Modeling the direct versus delayed response of surface waters to acid deposition in

       northeastern United States. Ph.D. Thesis. University of Iowa, Iowa City. 288 pp.       aeP°SIIIon In




                                                                 1«* A generated soft water

 NBSS°2rw:Ff ?'G" f^' 3nd  J'D- Miller' 1982' Forest 9rowth as a P°ssible cause  of soil and  water
      acidification: An examination of the concepts. Oikos. 39:40-49.




                                                  8fect °f PH °n SUlfete adsorPtion ^ a forest soil.




             K; If2' Tht6 6ffAe,C^ °f SUlfate on aluminum concentrations in natural waters: Some stability
              in the system AI2O3 - SO3 - H2O at 298 K. Geochim. Cosmochim. Acta 46:681-692.
                         ATHA- Haines' KJ- Stromborg, and J.R. Longcore. 1982.  Distribution of Surface

                                                         Atias-  Nationai Atm°spheric
Nyborg  M L 1978. Sulfur pollution and soil, pp. 359-390. In: J.O. Nriagu, ed. Sulfur in the Environment, Part

      II. Ecological Impacts. Wiley-lnterscience, New York, NY.
   >  PT'iharnmm u"*^' ^ ^ ^ Under Shifting Cultivat|on. Commonwealth Bureau of Soils
     Tech. Comm. No. 51. Commonwealth Agricultural Bureaux, Farnham Royal, Bucks.
        llfv0 OTA° RlnSTon!"!', I98-4' Add Rainnand TransP°rted Air Pollutants: Implications for Public
        licy. OTA Report 0-204. U.S. Government Printing Office, Washington, DC. 323 pp.
     n
-------
           p 1984' Add DeP°sition  Ann"al Data Summaries: 1980,  1981,  1982
       U.S. Environmental Protection Agency, Washington, DC.       •    ° >  ***>*•



                 ' W'H- Chan' T'L Clark' M'A- Lusis- P-K. Misra, and R J Vet In Press

                                                                 ^
New Yo'rkTeg^n
INSW TorK negion.
                       U         Surface, Wate- * Map of the New Engiand and
                      . U.S. Environmental Protection Agency, Corvallis, OR.
Growers. 1983. Tota. alkalinity of surface waters - a national map. Ann. Assoc. Am.

                                                                      water
  H" ""' '"• 1965' Ge°IOgiC Map °f the Crystalline Ro^s of South Carolina, (scale




       ^                                       -a, Technical Report No. 115,












                   eqUilibriUm between Around water and granitic rock. Geochim.
        1 9?a The mecha"ism °f ^"Ifte adsorption on iron oxides. Soil Sol. Soc. Am.
         '987' Pr°PertleS and Ma"a9eme"t <* *>">« Soils. John Wiley & Sons, Inc.,
                                               Survey BulL
                         846

-------
S    9Im
                               ina' 'igandS diSP'aced' and Chan9es in ««rface charge.





               deS°rpti°n °f sulfate and char9e relationships in allophanic clays. Soil



           ^
          "**
                                                 Water
                                                      effect °f aoid V"*****-"

                   198T- Ch9ml0a' Processes governing soil and water acidification.
       ^^^








PJ' D°"in9' 1985' Th9 ro'e °' °rganlc matter in sc" acidification. Aust. J. Soil Res.
                                                               °f the

                          847

-------
                                        a'e":
                                                                                      *»
                                                                     "— • R—
                                                                    ca,,on in ,he Upp, Great




Romesburg, H.C. 1984. Cluster analysis for researchers. Lifetime Learning Publications, Belmont, CA.
                                                                                        344
 RosenbrocMiR, 960. An automatic method for finding the greatest or least value of a function. Comput.
                                                                       *«*•

                                                 ' 1988'
                                                  for streanwater
                                                                     *»•« cycle in ac,d«ed
      tPA/600/3-88/021b. U.S. Environmental Protection Agency, Washington, DC. 597 pp
                              1975' AdSOrpti°n °f SU'Phate bV - sandy ,oam soi. (caicic cambiso.).




SAS In^ute, Inc. 1985. SAS User's Guide: Statistics, Version 5 Edition. SAS Institute, Inc., Gary, NC. 956






                                 lde f°r PerS°nal C°mPUters' Version 6 Edition. SAS Institute, Inc.,




SAS Institute, Inc. 1988. SAS/STAT User's Guide, Release to 6.03 Edition. SAS Institute, Inc., Gary, NC.




                       criteria for discrimi^ting among alternative regression models. Econometrica
                                           848

-------
      nfHn     ,-T- DriSC°IL 1987a" ComParison of terrestrial and hypolimnetic sediment generation
      of acid neutralizing capacity for an acidic Adirondack lake. Environ Sci. Technol. 21:988-993










Schindler, D.W. 1986. The significance of in-lake production of alkalinity. Water, Air, Soil Pollut. 30:931-946.


Schindler, D.W. 1988. The effects of acid rain on freshwater ecosystems. Science 239:149-157.


Schindler, D W  M.A. Turner, M.P. Stainton, and G.A. Linsey.  1986b. Natural sources of acid neutralism
      capacity in low-alkalinity lakes of the Precambrian Shield. Science 232 :844847         neutralizing



      -'C- ' M"!i MitChe"' TJ> Sc0tt> R'D- Fuller' and CT- Drisco11- 1986a. Incorporation of 35S-sulfate into
      inorganic and organic constituents of two forest soils. Soil Sci. Soc. Am. J 40:457-462
          Ridge TN
                       hS' A'E °sborne-Lee- J-A- Watts, J.C. Goyert, and R.S. Turner. In Review User's

                             ^  ^^ °ak Ridge Nati°na' Laboratory. Computing and Telecomm Div.
                                                                          chemical and hydroloqic
                                                                            ^search
                       --               in a sma" stream as affected

                                                 D'L


                                         chemical str^ture and reaction of humic substances, pp.

                                                 ' NAT° Conference Series 1= ^co.ogy, Voiume'S.
      tnm                  - 1985' Acidificati°n °f aquatic and terrestrial systems, pp 311-338  In- W
     Stumm, ed. Chemical Processes in Lakes. Wiley-lnterscience,  New York, NY.
                                                 Chemical Weatherin9
                                                                      *e neutralization of acidic
                                                                          ^position in the eastern



                                                                           acidic deposition in the








                                                                             'LWAS basins- Water'


Schott J and RjA Bemer. 1 984_ Dissolution mechanisms of pyroxenes and olivines during weathering, pp.
      35-54. in. J.I. Drever, ed. The Chemistry of Weathering. D. Reidel Publ. Co., Boston, MA.
                                                          at risk <°



                  ^^                                             -
                                            849

-------
  Searle, S.R. 1987. Linear Models for Unbalanced Data. John Wiley & Sons. Inc., New York, NY.
                                                     patterns and trends ln

Shreider, Y.A.  1988. The Monte Carlo Method. Pergammon Press, New York, NY.
      . Sroaouaf         "' ^ *"""' diS"ibU"°n and °*ai"  '"
                                                                                   watershed.




                                   **' di
                                                   model
                                                                         » «**-oa Water
Smith, R.A. ,972. Air and Ra,n: The Beginnings of a Cnemical Climatology. Longmans, Green, London

Sm

                        - 198a C°rrelatlons bet^" °"«n, sulpt,a,e and regional SOa emissions.

                           d M'a W°'man- 1987' Wa.er.qUa,*y trends In the nation's rk,ers. Science
                                           850

-------
 Snedecor, G.W., and W.G. Cochran. 1967. Statistica. Methods. Iowa State University Press, Ames 593 pp





 M8
 S°"
S°"
Soil
                    : fa"
                             **
                                      l' USDA S°" Co"*™*-" Service, Agricultural Handbook No.





                                                    " C°nselv a"°" S«**. *<*»*- Handbook
                                                                     °f Analysis for Soil
Sokal, R.R., and F.J. Rohlf. 1969. Biometry. W.H. Freeman and Co., San Francisco, CA.

Sposito, G. 1977. The Gapon and Vanselow selectivity coefficients. Soil Sci. Soc. Am. J. 41:1205-1206.
                       ^^
                                       CllemiCal foundation of the
                                                                              " ««<>. Soil Sol.
                                             "
                                                    »*
        Me^rS So. £££3?
           S
                                                                *n by a random s*c,,cn o,





                                            interPret*" - •" PH trend in the United State, Bull.





                                         iS' ComP°sltl0"' Reaai°"^ Wlley-lnterscience, New York.





                                      F0rma"°n 3"d ml™»*ation of organic sulfur in fores, soils.











                                                  S0il's sensi™* to acid PrecipBation. Acta Agric.











                     J'L Schn00r' 1987' Aqua"° dlemlstr>' of acw ^position. Environ. Sci. Technol.
                                          851

-------
  Sullivan  T.J., J.M Eilers, M.R. Church, D.J. Blick, K.N. Eshleman, D.H  Landers and M S
       Atmospheric wet sulphate deposition and lakewater chemistry. N^ure 331 ^07 609
                                                           volume 66: Forest
       Ecplog,cal Studies, VoIUme66:Fores,HydroloayandEcologya.!Zfe,fspri4SeS!gyNewY«k;
                        ' a"d JX Ash' 1 ^ Mioroblal 'ransformations of sulfats in forest soils. Science

                                                                      zs

                                                                        ytelds
                                                                  « "- —"v
Tisdale, S.L, and W.L Nelson. 1975. Soil Fertility and Fertilizers. Macmillan Publ. Co., New York, NY.
Tomovic, R. 1963. Sensitivity Analysis of Dynamic Systems. McGraw-Hill Co., New York, NY.
                                         852

-------

                                                                  soil
                                                           of lakes ln the
                                                 tarvesting ™trie«
                                                                                 nutnen, resets.
                                                                    in
                                 ' Anthr°W"ic  a"d natural  acidification in terrestrial  ecosystems.

Da'a' W9'nla' Water
    ' Wginia' Water
                                                                         ' U'S. Geologica, Survey

                                                                                         Survey
                                                           Addfflcation
                              CT' DriSC°"- 1983' A«ta*n and alkalinization of soils. Plant and Soil
VanLoon, G.W.  1984. Acid rain and soil. Can. J. Physiol. Pharmacol. 62:991-997.
van MH                                             n soii acidmcation
                                             853

-------

                                   •—**»«• —a" - •>"*>«•« parses, so» colloids, and


                                               ' BaS'OS' a"d
                                                                                 Date Analysis.


                                                              °' SUlfate '" Sll
                                                                                         soil.


                                                                      , S«udy Han^oo, Dep,.
                                                                       fram
                                                                                      , Wes,
Wetzel, R.G. 1975. Limnology. W.B. Saundere Co., Philadelphia, PA. 743 pp.
                                            854

-------
 Whitehead, P.G., S. Bird, M. Hornung, B.J. Cosby, C. Neal, and P. Paricos. In Press-a. Stream acidification
      trends in the Welsh uplands: A modeling study of the Llynbrianne catchments. J. Hydrology.

 Whitehead, P.G., B. Reynolds, M. Hornung, B.J. Cosby, C. Neal, and P. Paricos. In Press-b. Modeling long-
      term stream acidification trends in the upland Wales at Plynlimon. Process. Hydrology.

 Wollast, R. 1967. Kinetics of the alteration of K-feldspar in buffered solutions at low temperature. Geochim.
      Cosmochim. Acta 31:635-648.

 Wolock, D.M. 1988. Topographic and soil hydraulic control of flow paths and soil contact time: Effects on
      surface water acidification. Ph.D. Thesis. University of Virginia, Charlottesville, VA.

 Woiock,  D.M., G.M.  Hornberger,  K.J. Beven, and W.G.  Campbell.  1989.  The relationship of catchment
      topography and soil hydraulic characteristics to lake alkalinity in the northeastern United States. Water
      Resour. Res. 25:829-837.

 Woodruff, J.F., and J.D. Hewlett. 1970. Predicting and mapping the average hydrologic response for the
      eastern United States. Water Resour. Res. 6:1312-1326.

 Wright, R.F. 1982. A model for streamwater chemistry at Birkenes, Norway. Water Resour. Res. 18:977-996.

 Wright, R.F. 1988. Acidification of lakes in the eastern United States and southern Norway: A comparison.
      Environ. Sci. Technol. 22:178-182.

 Wright, R.F., and B.J. Cosby. 1987. Use of a process oriented model to predict acidification of a manipulated
      catchment in Norway. Atmos. Environ. 21:727-730.

 Wright, R.F., and A. Henriksen. 1978. Chemistry of small Norwegian lakes, with special reference to acid
      precipitation. Limnol. Oceanogr. 23:487-498.

 Wright, R.F., and A. Henriksen. 1983. Restoration of Norwegian lakes by reduction in sulphur deposition
      Nature 305:422-424.

 Wright, R.F., B.J. Cosby, G.M. Hornberger, and J.N. Galloway. 1986. Comparison of paleolimnological with
      MAGIC model reconstructions of water acidification. Water, Air, Soil Pollut. 30:367-380.

 Wright, R.F.,  E.  Lotse,  and A. Semb.  1988. Reversibility of acidification shown  by  whole-catchment
      experiments. Nature 334:670-675.

Yuretich, R.F., and G.L Batchelder. 1988. Hydrochemical cycling and chemical denudation in the Fort River
      watershed, central  Massachusetts: An appraisal of mass-balance studies.  Water  Resour  Res
      24:105-114.

Zen, E-an, ed. 1983. Bedrock Geologic Map of Massachusetts. U.S.  Geological Survey, Washington DC
      (scale 1:250,000).
                                              855

-------
                                         SECTION 13

                                         GLOSSARY
 13.1 ABBREVIATIONS AND SYMBOLS

 13.1.1 Abbreviations
 ADS
 AERP
 ANC
 AREAL-RTP
 CDF
 Cl
 CIR
 Acid Deposition System
 Aquatic Effects Research Program
 Acid neutralizing capacity
 USEPA Atmospheric Research  and Exposure Assessment Laboratory - Research
 Triangle Park

 Cumulative distribution function
 Confidence interval
 Color infrared photography
 DDRP
 DEM
 DOC
 DQO

 ELS-I
 EMSL-LV
 EPA
 EPRI
 ERL-C
 ERP
 ETD

 GIS

 IBM PC
 ILWAS
 IQR

LAI
LTA
 Direct/Delayed Response Project
 Digital elevation models
 Dissolved organic carbon
 Data quality objective

 Eastern Lake Survey-Phase I
 USEPA Environmental Monitoring and Systems Laboratory - Las Vegas
 U.S. Environmental Protection Agency
 Electric Power Research  Institute
 USEPA Environmental Research Laboratory - Corvallis
 Episodic Response Project
 Enhanced Trickle Down Model

 Geographic Information System

 International Business Machines Corporation - personal computer
 Integrated Lake/Watershed Acidification Study
 Interquartile range

Leaf area index
Long-term annual average deposition
                                            856

-------
  MAGIC
  MLRA

  NAS
  NADP/NTN
  NAPAP
  NCDC
  NE
  NHAP
  NOAA
  NRC
  NSS-I
  NSWS

 ORNL
 OTA

 PCA
 PNL

 QA
 QC

 RADM
 RELMAP
 RCC
 RILWAS
 RMSE
 RSD

 SAB
 SAS
 SBR
 SBRP
 SCS
 SE
 SOBC
 SOEBC
 SUNY-P

TMY
  Model for Acidification of Groundwater in Catchments
  Major land resource areas

  National Academy of Sciences
  National Acid Deposition Program/National Trends Network
  National Acid Precipitation Assessment Program
  National Climatic Data Center
  Northeast Region
  National High Altitude Photography
  National Oceanographic and Atmospheric Administration
  National Research Council
  National Stream Survey-Phase I
 National Surface Water Survey

 Oak Ridge National Laboratory, Tennessee
 Office of Technology Assessment

 Principal component analysis
 Battelle-Pacific Northwest Laboratories

 Quality assurance
 Quality control

 Regional Acid Deposition Model
 Regional Lagrangian Model of Air Pollution
 Regional Coordinator/Correlator
 Regional Integrated Lake/Watershed Acidification Study
 Root mean square error
 Relative standard deviation

 Science Advisory Board
 Statistical Analysis System
 Southern Blue Ridge
 Southern Blue Ridge Province
 Soil Conservation Service
 Standard error
 Sum of base cations
 Sum of exchangeable base cations
 State University of New York, Pittsburgh

Typical meteorological year
                                           857

-------
   UMW
   UDDC
   USDA
   USDOI
   USFS
   USGS
   UTM

   WA
   WBA
   Upper Midwest
   Unified Deposition Database Committee
   U.S. Department of Agriculture
   U.S. Department of Interior
   U.S. Forest Service
   U.S. Geological Survey
   Universal Transverse Mercator

   Watershed  area
   Watershed  Based Aggregation
  13.1.2 Symbols

  2As
  2Bn
  2Cn
  2X
  3A
  3B

  A
  AC_BaCI
  AH
  AL
  AI_AO
 AI_CD
 AI_PYP
 AI3+
 ALPOT
 ANN_AVG
 AVG_EL
 AW

 B_CENT
 B_LEN

B_PERIM

B SHAPE
  Southern Blue Ridge subregion (NSS Pilot Survey)
  Valley and Ridge subregion (NSS Pilot Survey)
  Northern Appalachians subregion (NSS Pilot Survey)
  Southern Appalachians subregion (NSS Pilot Survey)
  Piedmont subregion (NSS Pilot Survey)
  Mid-Atlantic Plain subregion (NSS Pilot Survey)

  acid that is leached out of the soil
  barium chloride triethanolamine exchangeable acidity
 area of all open water bodies in drainage basin, in kilometers squared
 area of primary lake, in kilometers squared
 aluminum, acid oxalate extractable
 aluminum, citrate dithionite extractable
 aluminum, pyrophosphate extractable
 aluminum ion
 aluminum potential (pH - 1/spAI)
 flow-weighted annual average sulfate concentration
 average elevation; (MAX_ELEV +  MIN_ELEV)/2, in meters
 total watershed area, in kilometers squared

 drainage basin centroid expressed as an X,Y coordinate
 length of drainage basin:  air-line distance from basin outlet to farthest upper point
 basin, in kilometers
the length of the  line which defines the  surface divide of the drainage basin,  in
kilometers
                   basin shape ratio; B_LEN2/WS_AREA
                                             858

-------
  B_WIDTH
  BS  Cl
  C
  C_TOT
  Ca + Mg-DRY
  Ca+Mg-WET
  Ca_CI
  Ca2+
  CaCI2
  CEC_CI
  cr
  C02
  COMPACT

 DDENSITY

 ELONG
 FRAG
 LJ +
   total
 H20_WS
 H2O
 H2S04
 H5up
 ha
 HC03-
 H-DRY
 H-WET
I
IND_AVG

INT

K
  average basin width; WS_AREA/B_LEN, in kilometers
  base  saturation  calculated from  unbuffered  1N  ammonium  chloride  CELod
  exchangeable bases

  correction factor for the decrease in acidity due to the protonation of bicarbonate
  carbon total
  the annual loading of Ca plus Mg in dry deposition
  the annual loading of Ca plus Mg in wet deposition
  exchangeable calcium in unbuffered 1N ammonium chloride
  calcium ion
  calcium chloride
  unbuffered  1N ammonium chloride cation exchange capacity
  chloride ion
  carbon dioxide
  compactness ratio; ratio of perimeter of basin to the perimeter of a circle with equal
  area; (PERIM)/(2 * (* AJ5)
 drainage density; TOTSTRM/WS_AREA

 elongation ratio; (4 * WS_AREA)/L_BEN
 fragments > 2 mm diameter

 hydrogen ion
 total effective acidity (H+ + NH4+ - NO3")
 ratio of open water bodies area to total watershed area; H20_AREA/ws_area
 water
 sulfuric  acid
 the percent of a watershed covered by bedrock with sensitivity codes of 5 and 6
 hectare  (2.47 acres or ten thousand square meters)
 bicarbonate  ion
 annual hydrogen ion loading  in dry deposition
 annual hydrogen ion loading  in wet deposition

 amount  of effective acidity in  deposition
flow-weighted average sulfate  concentration for the index sample time frame (spring
or fall)
total length of intermittent streams as  defined from USGS topographic  maps  of
aerial photos, in kilometers
hydraulic conductivity.
                                            859

-------
  K+
  K_CI
  Keq ha"1
  kg
  km
  kso4

  L_CENT
  L_PERIM
  LIMEPOT
  ln(a/TanB)
  ln(a/KbTanB)
  LTA-rbc
  LTA-zbc

 M_PATH
 M04
 MAX_EL
 MAX_REL
      "1
 mg
 Mg_CI
 Mg2+
 MINEL
Na+
Na_CI
NECMPON

NECMPOS
NEIDLGD
NH/
  potassium ion
  exchangeable potassium in unbuffered 1N ammonium chloride
  Kiloequivalent per hectare
  kilogram
  kilometer
  sulfate mass transfer coefficient (m yr"1)

  primary lake centroid expressed as X,Y coordinates
  perimeter of primary basin lake, in kilometers
  lime potential (pH -  i/apCa)
  an index of flowpath partitioning used in the TOPMODEL hydrologic model
  an index of flowpath partitioning used in the TOPMODEL hydrologic model
  long-term  annual average, reduced dry base cation
  long-term  annual average, zero dry base cation

  estimate of mean flowpath, in meters
  miscellaneous land area mapped as quarry pits
 elevation at approximately highest  point, in meters
 maximum  relief;  MAX_ELEV - MIN_ELEV, in meters
 microequivalents per liter, unit of concentration
 milligram
 exchangeable magnesium in unbuffered 1N ammonium chloride
 magnesium ion
 elevation of primary lake,  in meters

 sodium ion
 exchangeable sodium in unbuffered 1N ammonium chloride
 data file with soil and miscellaneous area components of  map units for the DDRP
 Northeast region
 map unit composition data file for the DDRP Northeast region
identification legend data file for the DDRP Northeast region
ammonium  ion
nitrate ion
                  hydroxide ion
                                           860

-------
   PC02
   PER_DD
   PERIMRAT
   PERIN

   PH_01M
   PH H2O
  REL_RAT
  ROTUND
  RTR

  S
  SBC_CI
  Sd  ~
  SE_MP_CM
  SE_MP_UN
  SECMPNT

 SEDBMNT
 Si02
 SO4_B2
 SO4_EMX
 SO4_H2O
 SO4_PO4
 SO4_SLP
 SO4_XIN
 S042'
 SO4-DRY
 SO4-WET
 [S042-]ss
SOILDEN
   partial pressure of carbon dioxide
   drainage density calculated from perennial streams only; PERIN/WS_AREA
   ratio of the lake perimeter to the watershed perimeter;  Lake Perimeters/B_PERIM
   total perennial stream length as defined from USGS topographic maps and aerial
   photos, in kilometers
   pH (0.01 M CaCI2)
   pH (deionized water)

   runoff estimate (length time"1)
   Average annual runoff;  interpolated to each site from Krug  et al. (in press) runoff
   map, in centimeters
   correlation coefficient
   coefficient of determination, the proportion of variability explained by a regression
   model
  relief ratio;  (MAX_ELEV-MIN_ELEV)/B_LEN
  rotundity ratio;  (B_LEN)2/(4 * WS_AREA)
  lake retention time, in years

  sum of base cations
  sum of base cations as measured in unbuffered 1N  ammonium chloride
  dry sulfur deposition (mass length'2 time-1)
  map unit composition data file for the DDRP Southern Blue Ridge region
  map unit identification legend data file for the DDRP Southern Blue Ridge region
  data file with soil and miscellaneous area components of map units for the DDRP
  Southern Blue Ridge region
  Southern Blue Ridge Mapping  Database Management System
 silicon  dioxide
 half saturation constant
 adsorption asymptote
 sulfate, water extractable
 sulfate, phosphate extractable
 slope of sulfate adsorption isotherm at zero net adsorption
 zero net adsorption concentration for sulfate, determined from adsorption isotherms
 sulfate
 annual loading of sulfate in dry deposition
 annual loading of sulfate in wet deposition
 steady state sulfate concentration
soil bulk density
                                            861

-------
  STRMORDER
  SUB_BAS(n)
  THKA
  TOT_DD

  TOTSTRM
  V6

  WA:LA
  WM_PATH
  WS_AREA
  WS LA
 surface water sulfur (mass length"3)
 maximum stream order (Horton) of streams in the watershed
 area of each subcatchment in the drainage basin, in kilometers squared
 wet sulfur deposition (mass length'2 time'1)

 soil thickness, adjusted for FRAG
 estimated drainage density based on crenulations
 identified on topographic map
 total stream length; combination of perennial and intermittent, in kilometers
 hydraulic residence time, in years
 hydrologic retention time, in years

 volume of primary lake, 106m3

watershed area to lake area ratio
estimate of weighted mean flowpath, in meters
total watershed  area, in kilometers squared
ratio of the total watershed area to the area of the primary lake
 13.2 DEFINITIONS
 texa   soiutiono                 ** appr°ximate solution obtai^ using a numerical mode, and
 the exact so ution of the govern.ng equations (or a known standard concentration), divided by the exact
 solution  (or known  standard concentration).                                             Y

 ACID ANION  - negatively charged ion that combines with hydrogen ion to form an acid.
                                "**" '
                                                produce
                                                                                      ions-
Sx.DA?ION  fnE " Tff WJth hjgh C0ncentration of meta|s. sulfate, and acidity resulting from the
OXIDATION of sulf.de mmerals that have been exposed to air and water (usually from mining activities).
                                             862

-------

                                                                                       an
       to
     ° <
                           " B"
                                            W'" a" AC'D NSJTRAUZIN8 CAPACITY less than or

                                                                          ACID
               crY.         h exp9rienced  any temporary  or  permanent •»•
               CAPACITY or a so.l that has experienced a reduction In  BASE SATURATION.
                                                    USed to transfom <=oncen,ra«on data to sal,
                                          anaiysis' modffied
 AFFORESTATION - the natural process through which non-forested lands become forested
 AGGRADING FORESTS - forests in which there Is a net annual accumulation of biomass.

                                              a set of data to a sinaie °aicuiated °

                                             SO"S W*h an
                                                                               35 percent
The
may
                       a spunous periodic solution or mask a real periodic phenomenon.
                                          863

-------

                               ^^^
                            ^^^
                           "
                                           *"»— •*"* commonly .hough, ,o be a pre.
                                          in
  ALUMINUM BUFFER RANGE - pH 4.2 - 2.8
                                           either an acid °r
 ANAEROBIC - without free oxygen (e.g., hypolimnetic lake waters, sediments, or poorly drained soils).
 ANALYTE - a chemical species that is measured in a water soil, or tissue sample.

                                • physicai and chemicai ™s °- — «*
 ANALYTICAL DUPLICATE - a QUALITY CONTROL sample made by splitting a sample.
 ANION - a negatively charged ion.

 ANION CATION BALANCE - a method of assessing whether all CATIONS and ANIONS have been
                             ^
                                                        ^  *• - in whlc, ANIONS are
ANTHROPOGENIC - of, relating to, derived from, or caused by human activities

                                       -imate form °f a
                                                                  or actions.
AQUEOUS SPECIES - any dissolved ionic or nonionic chemical entity.
                                        864

-------
  AQUIC - a moisture regime of soils in which a water table and reducing  conditions occur near the
  SUfTaCG.



  AQUIFERS - below-ground stratum capable of producing water as from wells or springs.


  AQUO LIGAND - a water molecule held to Fe or Al in a clay edge or hydrous oxide by ligand exchange.


  ARC -represents line features and borders of area features. One line feature may be made up of many
  arcs. The arc is the line between two nodes.


  ARC/INFO - a commercial geographic information system (GIS) software used to automate, manipulate
  analyze, and display geographic data in digital form.



                      ' CharaCteriStiCS °r Other Pr°Perties associated with a specific feature, area on a
 AVAILABLE TRANSECT - a transect identified to represent a map unit and listed for random selection.


 BASE CATION - a nonprotolytic CATION that does not affect ACID NEUTRALIZING CAPACITY- consists
 principally of calcium, magnesium, sodium, and potassium.


 BASE CATION  EXCHANGE - the process  by which  BASE CATIONS (Ca2+  Mg2+  Na+   K+)  are

 adsorbed or released from negatively charged sites on soil particles from or to, respectively, soil solutions
 buch exchange processes  are instrumental in determining pH of soil solutions.
    h* CAT'°NhS"PPLY - 0).«he POO. of BASE CAT.ONS (Ca2+, Mg2+, Na+, K+) in a soil availab.e for
 exchange with hydrogen ,ons (H+). The base cation pool is determined by the CATION EXCHANGE
 CAPACITY of the soil and the percentage of exchange sites occupied by BASE CATIONS.


 BASE SATURATION - the percentage of total soil CATION EXCHANGE CAPACITY that is occupied by
 exchangeable cations other than  hydrogen and aluminum, i.e., the base cations Ca2+, Mo2-, Na+, and




 BEDROCK - solid rock exposed  at the surface of the earth or overlain by unconsolidated material.


 BEDROCK GEOLOGY - the physical and chemical nature and composition of solid rock at or near the
 earth s surface.


 BEDROCK LITHOLOGY - see LITHOLOGY.


 BEDROCK SENSITIVITY SCORES - a six point sca.e, developed for DDRP, designed to distinguish the
relative reactivities of different lithologies.


BEDROCK UNITS - the smallest homogenous entity depicted on a bedrock map.
                                           865

-------
  B.AS - a systematic error in a method caused by artifacts or idiosyncracy of the measurement system.




  B10MASS - the quantity of paniculate organic matter in units of weight or mass.

                                                              e°n
                        func,,ons
                                     from
                                                 ""
                                                                      samp,e with

                                 °' a U°'Ume °' «* ""**• "« "-». «• solutions, vote,
CALCITE - a mineral with the formula CaCO3.  A carbonate mineral
^^
                                                C°NTBOL
                                                                    contains on,y the
                                   °f 3 SyStem defined as a function °f the quantity or size of
                                       ag" SU'fate ads°rption or cation e-hange) for which
                                       866

-------
   CARBON-BONDED SULFUR - a reduced form of organic sulfur, charactered by C-S bonds.
  reactions gene n

  CATCHMENT - see WATERSHED.
  CATION - a positively charged ion.
                                                       "' ra*°"+"°** '" — '• Association
                                                           °' "  * ^"^ "
                                                                ' in which ACIDI° OATIONS
  CATiON EXCHANGE CAPACITY - the sum total of exchangeable cations tna, a soil can absorb.
                                                                - •* '
 CHRONIC ACIDIFICATION - see LONG-TERM ACIDIFICATION.
 C.RCUMNEUTRAL - close to neutrality with respect to PH (pH = 7); in natural waters, pH 6 - 8.
                                 ^
                      "
                                   *" *- "*
                                                             '" • - °' temperature
CLOSED LAKES - a lake with a surface water inlet but
                                               no surface water outlet.
COLLINEAR - see MULTICOLLINEARITY.
                                        867

-------
                                       °r more dissimiiar
  COMPONENTS - see MAJOR COMPONENTS, MINOR COMPONENTS, and MAP UN.T COMPOS.TION.

  CONSOCIATION - a map unit dominated by a singie soil taxon (or miscellaneous area) and similar soils.

  CONTOUR LINE - a line connecting the points on the land surface that have the same elevation




  COOK'S D - a regression statistic designed to indicate LEVERAGE POINTS.

  COVERAGE - a digital analog of a single map sheet; forms the basic unit of data storage in ARC/INFO


                            =-
 DATABASE FILE - a collection of records that share the same format.


                                           t0 the 6arth>S Surface * -y of a number of chemical
 DEPTH TO BEDROCK - depth to solid, fixed, unweathered rock underlying soils.
                                                      LAYER  - depth to a layer in soils or
                                                  !.g., bedrock, dense till or fragipan).

                            SAMPLE - a QUALITY CONTROL sample that contains the ANALYTE
                            :he contract required detection limit.

DIAZO - a photocopy whose production involves the use of a coating of a diazo compound.

DIGITIZATION - the process of entering lines or points into a GEOGRAPHIC INFORMATION SYSTEM.
                                         868

-------
            COORDINATES - lines or points that have been entered into a GEOGRAPHIC INFORMATION
                  REDUCTION - a
'" «*"* ^ oxidized chemical species (e.g., SO, - S)  i
  in the absence <* free
                                                                                          is
  DISSOCIATION - separation of an acid into free H* and the conjugate base of that acid (e.g.. H,CO  -


  NH4OH ->  rS/ " OrT"0" * ' **** "° " "" ^'^  "" *"' ^'"^^ ^ °' the ""« 
-------
                                        iakes sampied
                                                       by
            " '" S°"
                                  °RDER °' mi"6ra' S°"S Wlth "° <* ™y P°°"/ developed genetic
                                                                  CAPAC.TV dunng s,orm ,,cws
                  " "
                                          " S°" °r9anl°
                                                               characterized by 0-0-SO3 or N-
                                         CATI°NS '" **
EXTENSIVE PAPAMETERS - variables that depend on the size (extent) of the
                                                  that
                                                                                in  exchange
                                                                      system.
FELDSPARS - a group of tectosilicate minerais that are the most abundant group in the earth's crust



of boundaries, and map detail in relation to survey objectives.                    Placement
                                           870

-------
                                               °< tores< ta" "—i on present occupancy of an
                                                                                ers;

    roughly 3.5.                               p   =  1'0) fractlon of an alkali-soluble soil extract; pK




 GAINES THOMAS FORMULATION - a formu.ation used  to describe exchange processes.


 GAPON - a formulation used to describe exchange processes.



 GENERIC BEDROCK TYPE - see GENERIC ROCK TYPE.



 GENERIC ROCK TYPE - a general classification of different BEDROCK i INIITQ int
 the primary LITHOLOGY.                              BCUHUOK UNITS into groups according to



 GEOGRAPHIC INFORMATION 
-------
  GREAT GROUP - in Soil Taxonomy, the level of classification just below SUBORDER, e.g., Haplorthods.
  GROUNDWATER - water in the part of the ground that is completely saturated.
  HETEROSCEDASTIC - referring to a statistical situation in which variances are not all equal



                                       condition as
 HISTIC SOILS - organic-rich soils.
 HISTOSOLS - in Soil Taxonomy, the ORDER of soils formed from organic PARENT MATERIAL
 HOMOSCEDASTIC - referring to a statistical situation in which the variances may all be considered

                                          physioai and/or


HYDROLOGIC RETENTION TIME - see HYDRAULIC RESIDENCE TIME.
                              csr r,
                                       872

-------
 o™* '™
                                         <* a"
                                                                       (-Pec,a»y S or N) to
                                              nstructlon - a dam: ais°
  INCLUSIONS - see MINOR COMPONENTS.
average annual runofl.
                                                 poten"ai °f °°ntact
                                               "*"* "°W rate 
-------
   shorthand ,o desc^lZn     ""   ° C°mP°UndS °" S* b"-1" th<* «"«• *
   KAOLINITE - a two-layer day mineral with the chemical formula ALSLO (OH)
                                                            2  2  5*   '4"
  KRIGING - a technique for spatial interpolation.


  LABEL - represents point features or is used to assign identification numbers to POLYGONS.



                                                                 °f ^' — s, and dams as
LAND COVER - see FOREST COVER TYPE.


LAND USE - the dominant use of an area of land (e.g., crop land).



                               '" °' «" to' -*™ « * -** related ,o Escape
 pTcesles™
 LANGMUIR ISOTHERM - hyperbolic adsorption isotherm (used in this proiec, for sulfate) o, the form






 LARGE-SCALE MAPS - 1:24,000, 1:25,000, or 1:62,500 sca,e U.S. Geological Survey topographica, map,


 LEACHING - the transport of a solute from the soil in the soil solution.


 LEVERAGE POINT - a data point that strongly influences the parameter estimates in a regression








LIMESTONE - a rock type consisting primarily of CALCITE.


LINEAR BUFFER - land area within a set distance of a lake or stream.
                                          874

-------
rock or
                                                         BEDROCK
              • *""
                                  USUa"y
                                                            etc., ,ha, compose ,he bu,k of ,he
                                               landform "»' is *•'—• ^ tocalfced landscape

  LOWER NODE - the downstream NODE of a STREAM REACH.
                                             area characterized by
                                      °r miSCe"aneOUS
                                                             are identified in the name of a
 MALLOWS' CP - a criterion for selecting one of a sequence of regression models.




 MAP COMPILATION - the process of checking and measuring soil map unit data.




 MAPP.NG PROTOCOLS - instructions that guide the fie,d mapping and provide for quaiity control.




 MAP SYMBOL - a symbol used on a map to identify map units.




 MAPPING TASK LEADER - the person responsible for field mapping activities.



 MAP UNIT - see SOIL MAP UNIT.
                                   P~
                                                                of a,, sol, components and
                              a DATABASE FILE that contains a" components and
                                         (c°mp0nents are identified by an assigned code, i.e.,
MAP UNIT CORRELATION - see SOIL CORRELATION.
                                         875

-------
  ™nh                           °" 3 maP Unique'y identified with a sVmbo1- A delineation of a soil
  map has the same major components as identified and named in the map unit.
                                                                                           as
th

       C°EfT'CIEIIT8 '
           ^ t
                                                                           —ation-of-mass
                                               r "*' COnStant USed in models of "^ alkalinity
                                                rem°Val rate of a reactant from «lullon.  Specific
 andiffusion                                    S6diment by a" pr°CeSSeS' includi"9 sedimentation
 for sSracrossr" T   H-'   6 ^ tranSfer C°effiCient f°r SU'fUr iS eSSentia"y a diffusion ~
 for sulfate across the water-sed.ment interface; for nitrate a  biological uptake/sedimentation rate.

 MASTER HORIZONS - the most coarsely based delineations within a pedon.  Usually  A/E horizons

                                "' B horbons ar
                                            i"9 known
         ,                                                                   ^ ANAL/TE to a
        3iiquot.


 MAX - the maximum sensitivity code observed on a WATERSHED.


 MEAN - the weighted average of sensitivity codes for a WATERSHED.


 MEDIAN (M) - the value of x such that the cumulative distribution function F(x)  = 0.5; the 50th percentile.


 METASEDIMENTARY - rocks or bedrock formed from metamorphosis of sedimentary rocks.


 MICAS  - a group of primary phylosilicate minerals, frequently including biotite and muscovite.


                   REG'°N " ^ °f *" *"* ^oac "*** considered by the DDRP, consisting
MINERAL WEATHERING - dissolution of rocks and minerals by erosive forces.

                                            876

-------
      ™puvs,             mlSC*'ne°us — «« « « ktaflM in ,he name of
      map unit.  Many areas of these components are too small to be delineated separately.
  MISCELLANEOUS LAND AREAS - see MISCELLANEOUS AREA
                                     STOCHAST'C
                                                        or selection of random numbers to
  MOTTLING - spots or blotches of different color in a soil, including gray to black blotches in ooorlv
  drarned soils due to presence of reduced iron and other metals.                           P   V
                 srs^r1*11 processes throu9h  wwch **" •*—  '- »
 NODE - the points identifying either an upstream or downstream end  of a REACH.
                                                  dO88 nOt make the dassical Distributional
 NON-SILICATE IRON  AND  ALUMINUM - soil iron and/or  aluminum occurring in the soil as an
 amorphous or a (hydrous) oxide phase rather than as an ion incorporated within a sHicate minS lattice
                     DESCR1PTION ' a record <* ^ Definitions of a soil series and other relevant
in to        h         eneS' TheSe definiti°nS are the framework within whi^ ™* of the detailed
inftxmabon about so.ls of the United States is identified with soils at specific places  These demons
also provide the principal medium through which detaHed information about the soi and te ^hlvoTa
one place is projected to similar soils at other places.                                oenavior at

ORDER - in Soil Taxonomy, the highest level of classification, e.g., SPODOSOLS.

ORGANIC ACID - organic compound possessing an acidic functional group; includes fulvic and humic
                                        877

-------
  ORGANIC ANION - an organic molecule with a negative net ionic charge.
                      r^^
                         'dentfflable M» h°ri2°" »«** '" —- - ^0 percent organic matter
 OUTLIER - observation not typical of the population from which the sample is drawn.
                                                   lns ' from a lower to a hisher
 PARENT MATERIAL - the material from which soils were formed.
                          S°"
POLYGON - represents area features.
                                              8ample that  fe
                                                                        ""e  particular
                                                             «• Mu-y about , ^ a
                           whioh 9ases' "quids' or plant roots penetrate

                                              to provwe reproducibie
                                                          with
                                       878

-------
   e^na newMcoordENtTS " T^T *"" C°mbinations " the «Wha. data,  which geometrically
  represent a new coordinate system with axes in the directions of maximum variability.


  PROBABILITY SAMPLE - a sample in which each unit has a known probability of  being selected.


  QC CHECK SAMPLE  - a  QUALITY CONTROL sample that contains  the ANALYTE of interest at a
  concentration in the mid-calibration range.
                                                   StUdy iS adequate'y planned and implemented to
                                                  to ensure that data  quaiity meets  the
                                         °3)  that divide a population into four equal classes- each
 QUARTZ - a crystalline form of silicon dioxide (SiO2).


 QUARTZITES - a metamorphic rock-type composed of primarily QUARTZ.
   crernHn          ***** ^ ' °2 ' °3 ' ^ ^ dMde 3 P°pulation into five e^' classes,
 each  representing  20 percent of the  population;  used to provide additional  values  to compare
 characteristics among populations of lakes and  streams.                                  compare
      nnt.K               '"        Weatherin9) for wh^h the long-term ability to supply
     ion products (e.g., base cations) is constrained by reaction or transport kinetics.


 RCC TRANSECTS - transects conducted by the Regional Coordinator/Correlator (RCC).
 Sum                              rePresented as blue lines on 1 :250,000-scale U.S. Geological
 Survey  maps.   Each reach  (segment)  is defined  as the length  of stream between two blue line
 confluences.  In the NSS-I, stream reaches were the sampling  unit.
          SIht  H^ rf"°"Ship between the rate of « <*emical reaction and the concentration of
          substrate, defined by the value of the exponent of, that substrate.


REACTIVITY SCALE  - any  of  a number of relative  scales designed to  categorize the qeneral
"weatherability" of different LITHOLOGIES.                                    eyunze ine general


REACTIVITY SCORE -  see REACTIVITY SCALE.
          ,                                    that C°ntains a" the rea9ents used ^d in the same
         used in preparing a soil sample for analysis.
                                            879

-------
  REDUCTION/OXIDATION - chemical reaction in which substances gain or lose electrons.



             .S^ri^ the ?Tt6rminOUS United States where a substantial number of streams with
             less than 400 /zeq L  can be found.




                         " PhySI'°graphiC 3reaS  that reflect a maJ°r 'and-shaping  process over a long





                             " a correlated  and controlled legend for an  entire  region (see SOIL



             S°UrCe-reCept0r model desi9ned to estim*e dry deposition of sulfur; not used directly in





             elative to GIS activities> a format designed by the user for printing out information containin9



 RESERVOIR - a body of water collected and stored for future use in a natural or artificial lake.



                                                                       Variable a"d *• value
 REUSS MODEL - a numerical model used to describe exchange processes in a soil environment.



 RIPARIAN - a zone bounding and directly influenced by SURFACE WATERS.



 ROBUST - a statistical procedure that is insensitive to the effect of OUTLIERS.



 ROUTINE TRANSECTS - transects conducted by field soil scientists responsible for the mapping.
          ™«                                      displaced for the soil exchanae complex by
BASE CATIONS (from neutral salts). The result is a short-term increase in the acidity of associated water



SAMPLING CLASS - see SOIL SAMPLING CLASS.




SAMPLING CLASS CODE - a three-character code assigned to each SOIL SAMPLING CLASS.



SAMPLING CLASS COMPOSITION - the  relative proportion of sampling classes in a map unit.
   nnt   7T T^" ^ "* ^ mm '" diameter; also a soil texture class Gaining at least
   percent sand, and  whose percentage of silt, plus  1.5 times the percent clay, does not exceed 15.
                                           880

-------
                                 -
                                       (S°
                                                 Pr0dUCt (°f diSSOlved ions> to the s°^ility Product
                                                dS 1'°' the SO'Uti0n iS s"P^aturated with respect to
      eandu81'0" S6qUenCe  f0"OWing imPlementation  of a  controi  or mitigation
  strategy and the subsequent effects associated with this deposition sequence.

                                        minerai phase formed
  SEEPAGE LAKE - a lake with no permanent SURFACE WATER inlets
                                                                 or outlets.

                                                         *«*•
                                                                                  spec.es ,„ an
  SENSITIVITY CODES - see BEDROCK SENSITIVITY SCORES.
 SILICA - the dissolved form of silicon dioxide (SiO2).


 SILT - a soil separate consisting of particles between 0.05 and 0.002 mm in equivalent diameter- also a
 so,l texture class containing at least 80 percent silt and < 12 percent clay.


 SILVICULTURAL PRACTICES -  forest management practices to increase wood yields- thinning prunina
 fertilization, spraying with herbicides/insecticides, and irrigating.                                  9>

 SIMULATION - replication of the prototype using a model.

 SKELETAL SOILS - soils with at least 35 percent rock fragments in the control section.

 SLOPE PHASE - the slope gradient of a soil map unit or taxonomic unit expressed in percent.
c^vex, plane?
                                    ""*"*
                                                                °f the landscaPe ^ concave-
SLOPE SHAPE DOWN - shape of the land surface at right angles to the contours of the landscape.

SMALL-SCALE MAP - 1 :250,000-scale U.S. Geological Survey map.
                                            881

-------
  SMECTITES - a family of three-layer day minerals.
                                                  'hat Sen/es as a natural ""*"" *»• «*>
,o°h?sys,emR'NG
                            ' *"
                                                                     fr°m
                                         " S°" "
                                                             '"
                                                                         «««on o, acids
                           ' 3"

                                          °' 3 S°" ^ »«* » 'hs Absence o, exchange
                                     ln soil Taxonomy in w*h ciasses
                                                                        SOIL
 SOIL LEGEND - see SOIL IDENTIFICATION LEGEND.


 mLirousaLc8--!*^0!^?,^^.^ "amed * te™ - •* ~  components or
 is uniquely identified c

                                                                     states-

SOIL SOLUTIONS - those aqueous solutions in contact with soils.
                                                °f
                                         882

-------
 SOIL TAXONOMIC  UNIT - a kind of soil described  in terms of ranges in soil properties of the
 polypedons referenced by the taxonomic unit in the survey area.

 SOIL TEXTURE - the relative proportion by weight, of the several soil particle size classes finer than 2
 mm in equivalent diameter (e.g., sandy loam).

 SOIL TEXTURE MODIFIER - suitable adjectives added to soil texture classes when rock fragments
 exceed  about 15 percent by volume, for example, gravelly loam.  The terms "very" and "extremely" are
 used when rock fragments exceed about 35 and 60 percent by volume, respectively.

 SOIL TRANSECT - a distance on the surface of the earth represented by a line on a map.  Transects
 can be  straight, dogleg, or zigzag.

 SOLID PHASE  EXCHANGERS - those components of soils, primarily organic matter, clay minerals and
 mineral  oxides,  that serve as the sites for exchange reactions.

 SOLUM - soil layers that are affected by soil formation.

 SPECIATION MODEL - a numerical model used to describe the distribution of aqueous species among
 various  possible complexes and ion pairs; usually for the purpose of estimating single ion activities.

 SPECIFIC ADSORPTION - adsorption of sulfate by ligand exchange,  often involving exchange of two
 ligands and formation of a bridged (M-O-SO2-O-M) structure.

 SPODIC HORIZONS  - a soil  horizon in which iron oxides, aluminum oxides, and organic matter have
 accumulated from higher horizons.

 SPODOSOLS -  in Soil Taxonomy, the ORDER of mineral soils with well-developed SPODIC HORIZONS.

 SPRING BASEFLOW INDEX PERIOD - a period  of the year when streams are expected  to exhibit
 chemical characteristics most closely linked to ACIDIC DEPOSITION.  The time period between snowmelt
 and leafout (March  15 to May 15 in the NSS-I) when NSS-I stream reaches were visited coinciding with
 expected periods of highest geochemical and assessment interest (i.e., low seasonal pH and
 sensitive life stages of biota).

STABILITY (NUMERICAL OR COMPUTATIONAL) - ability of  a scheme to control the propagation or
growth of small perturbations introduced in the calculations. A scheme is unstable if it allows the growth
of error  so that  it eventually obliterates the solution.

STANDARD DEVIATION - the square root of the variance of a given statistic.

STEADY-STATE - the condition that occurs when a property  (e.g.,  mass, volume, concentration) of a
system does not change with time. This condition requires that sources and sinks of the property are in
balance  (e.g., inputs equal outputs; production equals consumption).
                                            883

-------
                                                ,                processss b
                               , or prevented from reaching receiving SURFACE WATERS
                                                 or a compound containing
                                                          losss eiectrons
   SULFITIC - containing sulfide minerals, usually pyrite.
SURFACE WATER - streams and lakes.




SURFACE WATER RUNOFF - preclpfction ,ha, flows
                                                 overiand ,o reach SURFACE WATERS.
  SYNOPTPC - re.a,ing ,o or displaying oondMons as they ex,s, a, a poin, in ,ime over a broad area.
                                                                                 commoniy
                                                            by
 TICS - registration or geographic control points for a COVERAGE.




 TILL - unstratified material deposited by glaciers.
TOPMODEL - topographically based, variable
                                        source area hydrologic model.




                                           885

-------
  TOPOGRAPHIC MAP - a map showing contours of surface elevation.

  TRANSECT - see SOIL TRANSECT.
TRANS^C ™
                                                  °f
                                         a'°ng
                                                                                  line (see
  TRANSECT POINTS - locations along a TRANSECT where data are collected.

  TRANSECT SEGMENT UNION - all transect stops in the same map unit on a WATERSHED.

  TRANSECT STOPS - see TRANSECT POINTS.

  TRANSFORMATION  ERROR - ca.culates the residual mean square error of the digitized TIC locations
  ana the existing TICs.
              " *
^ *** i
                                       °bservation at unc°ntrolled representative locations in the
 TYPICAL YEAR (TY) DEPOSITION DATA - a dataset of atmospheric deposition developed within the
 UURP for specific use with the integrated watershed models.

 UNCERTAINTY ANALYSIS - the process of partitioning modelling error or uncertainty to four sources-
 intrinsic natural variability,  prior assumptions/knowledge, model identification, and prediction error.

 UNIVERSAL TRANSVERSE MERCATOR (UTM) PROJECTION - a standard map projection used by the
 U.S. Geological  Survey.

 UPPER NODE - the upstream NODE of a STREAM REACH.

 UPSTREAM REACH NODE - see UPPER NODE.
                           °r P°intS 3S  rePresented  in a UNIVERSAL TRANSVERSE MERCATOR
VALIDATION  - comparison of model results with  a set of prototype data  not used for  verification
Comparison includes the following: (1) using a dataset very similar to the verification data to determine
 he vahdity of  the model under conditions for which it was designed; (2) using a dataset  quite different
from the verification data to determine the validity of the model under conditions for which it was not
designed but could possibly be used; and (3) using post-construction  prototype data to determine the
validity of the  predictions based on model results.

VANSELOW EXCHANGE FORMULATION - a formulation used to describe soil exchange reactions.
                                          886

-------
VARIABLE - a quantity that may assume any one of a set of values during the analysis.

VARIABLE SOURCE AREA - A topographically convergent,  low transmissivity area within a watershed
that tends to produce saturation excess overland flow during storm runoff periods.

VEGETATION - see FOREST COVER TYPE.

VERIFICATION - check of the behavior of an adjusted model against a set of prototype conditions.

WATERSHED - the geographic area from which SURFACE WATER drains into a particular lake or point
along a stream.

WATERSHED STEADY STATE - a condition in which inputs of a constituent to a WATERSHED equal
outputs.

WATERSHED SULFUR RETENTION - retention of sulfur by any of a number of mechanisms within a
WATERSHED.

WEATHERED BEDROCK - soft or partly consolidated BEDROCK that can be dug with a spade.

WEATHERING - physical and chemical changes produced in rocks at or near the earth's  surface by
atmospheric agents with essentially no transport of the altered materials.

WEIGHT - the inverse of a sample's inclusion probability; each sample site represents this number of sites
in the TARGET POPULATION.

WET DEPOSITION - for the purposes of the DDRP, atmospheric deposition of materials via rain or snow.

WETLAND - an area, generally with hydric soils, that is saturated, flooded, or ponded long enough during
the growing season to develop anaerobic conditions  in the upper soil horizons and that is  capable of
supporting the growth of hydrophitic vegetation.

ZERO-ORDER  REACTION -  a chemical   reaction,  the  rate of which  is  independent of  reactant
concentration.
                                           887

-------