United States
             Environmental Protection
             Agency
vvEPA
             Great Lakes National
             Program Office
             536 South Clark Street
             Chicago, Illinois 60605
GLNPO # 87-8
EPA-905/2-87-003
June 1987
Tri-State
Tillage Project

Modeling Component
Applying the Answers
Model to Assess the
Impacts of Conservation
Tillage on Sediment and
Phosphorus Yields to
Lake Erie
                    Do not WEED. This document
                    should be retained in the EPA
                    Region 5 Library Collection.

-------
                                 FOREWORD

The U.S. Environmental Protection Agency (USEPA) was created because of
increasing public and governmental  concern about the dangers of pollution
to the health and welfare of the American people.  Noxious air, foul water,
and spoiled land are tragic testimony to the deterioration of our natural
environment.

The Great Lakes National  Program Office (GLNPO) of the USEPA was established
in Region V, Chicago, Illinois to provide specific focus on the water
quality concerns of the Great Lakes.  The Section 108(a) Demonstration
Grant Program of the Clean Water Act (PL 92-500) is specific to the Great
Lakes drainage basin and  thus is administered by the Great Lakes National
Program Office.

Several  sediment erosion-control projects within the Great Lakes drainage
basin have been funded as a result  of Section 108(a).  This report describes
one such project supported by this  Office to carry out our responsibility
to improve water quality  in the Great Lakes.

We hope  the information and data contained herein will help planners and
managers of pollution control  agencies to make better decisions in carrying
forward  their pollution control  responsibilities.
                                 Valdas V. Adamkus
                                 Administrator, Region V
                                 National  Program Manager for the Great Lakes

-------
                                                           EPA-905/2-87-003
                                                           June 1987
                                                           GLNPO #87-8
                  TRI-STATE TILLAGE PROJECT
"Modeling Component Applying The Answers Model  to Assess The Impacts
   of Conservation Tillage on Sediment and Yields to Lake Erie"
                         Final  Report
                         Prepared by:

                David B. Beasley, Ph.D., P.E.
                     Associate Professor
              Agricultural  Engineering Department
                      Purdue University
                   West Lafayette, IN  47907
                Great Lakes Grant No.  R005717
                             For
                         GLNPO# 87-8
             U.S.  Environmental  Protection Agency
             Great Lakes National  Program Office
                  230 South Dearborn Street
                   Chicago, Illinois  60604
                         September 1984

-------
                           Disclaimer

This report has been reviewed by the Great Lakes National  Program
Office, U.S. Environmental Protection Agency and approved  for pub-
lication. Approval  does not signify that the contents necessarily
reflect the views and policy of the USEPA, nor does mention of
trade names or commercial  products constitute endorsement  or rec-
ommendation for use.

-------
                     TABLE OF CONTENTS








Section                                                Page




INTRODUCTION  	   1




PROJECT AREA  	   2




STUDY METHODOLOGY  	   6




PROJECT RESULTS  	  13




SUMMARY AND CONCLUSIONS  	  17




REFERENCES  	  23
                            m

-------
                       LIST OF TABLES
Table                                                  Page

  1.  Major Soil Group Descriptions and Spatial
      Extent  	   4

  2.  Land Use and Management Descriptions  	   9

  3.  Cropping Distribution for the Major Soil
      Groups (from CTIC)  	  10

  4.  Scenarios Used in the Simulation Study  	  10

  5.  General Watershed Information (First
      0.5 Percent Sample)   	  14

  6.  General Watershed Information (Second
      0.5 Percent Sample)   	  15

  7.  Predicted Sediment and Phosphorus Yields for
      Each Scenario on Each Watershed - First 0.5%
      Sample  	  16

  8.  Predicted Sediment and Phosphorus Yields for
      Each Scenario on Each Watershed - Second 0.5%
      Sample  	  17

  9.  Predicted Sediment and Phosphorus Yields for
      Eight Scenarios on Each Soil Group  — Western
      Basin Totals	  18

 10.  Unit Area Loading Information  	  21
                             IV

-------
                      LIST OF FIGURES








Figure                                                 Page




  1.  Study Area Drainage and Site Location Map  	   3




  2.  Study Area Soil Groups Map  	   5

-------
                             Final Report of the

               Modeling Component — Tri-State Tillage Project
      "Applying the ANSWERS Model to Assess the Impacts of Conservation
           Tillage on Sediment and Phosphorus Yields to Lake Erie"
                        David B. Beasley, Ph.D., P.E.
                             Associate Professor
                     Agricultural Engineering Department
                              Purdue University
                          West Lafayette, IN  47907
INTRODUCTION

     In late 1981, a modeling study was proposed as a part  of  the  Tri-State

Tillage  project.   This  study had two major objectives.  The primary goal of

the program was to provide basin scale projections of sediment and  phosphorus

yield  reductions  in  the  Western Basin of Lake Erie attainable from various

degrees (i.e., spatial extent) and types of tillage management.   Secondarily,

a study of the size of the sample area needed to make accurate projections was

to be undertaken.


     Two stratified 0.5% samples (based on soil  association  groupings)  were

simulated  for  a  number of tillage management scenarios.  Basin-wide projec-

tions of sediment and phosphorus yield were compared to observed  river  mouth

data  for  the  study  area.   Finally,  the projections were refined based on

actual conservation tillage implementation rates to provide data for  tracking

purposes.   Data  for  1982 and 1983 land use patterns in the project area (on

each major Soil Group) were obtained from the Conservation Tillage Information


                                     1

-------
Center   (CTIC) and used to better describe the actual patterns and percentages




of cover and management in the simulations.  The  "potential"  scenarios  were




also modified to correspond to the CTIC information.






     The yields predicted from each Group, as well as  from  the  Basin  as  a




whole,  are  consistent with long-term river mouth and tributary monitoring of




sediment yields in the project area.  Predicted unit area loads of  phosphorus




indicate  that  major  reductions in diffuse phosphorus loading will be rather




difficult to obtain.  This is due to the fact that the  amount  of  phosphorus




reduction  required  by Annex III of the International Joint Agreement is of a




similar magnitude to the unit area agricultural loading.






     Descriptions of the project area, study objectives, and developed  metho-




dology  are  included.   The results of the final simulations are described in



detail and emphasized in the Summary and Conclusions section.






PROJECT AREA






     The Tri-State Tillage Project  encompasses  approximately  24,200  square




kilometers in Northwest Ohio, Northeast Indiana and Southeast Michigan (Figure




1).  The Maumee, Portage, and Sandusky River basins make up  the  majority  of




the  study area.  Some "near lake" drainage area is also included.  Thirty-one




counties in the three states are participating in the project.  The  locations




of all of the watersheds simulated in the study are shown in Figure 1.






     The methodology developed in this project is based on  the  premise  that




"representative"  samples of the basin could be identified and modeled using a




"stratified" sampling technique.  The stratification would be based  on  simi-

-------

-------
larities  in  the  drainage  characteristics  of the  numerous soil  associations

within the basin.  The distribution  of  the  major   soil   associations   (Soil

Groups)  in  the  study  area is depicted in  Figure  2 and  a description of the

characteristics of each Group appears in Table  1.  Soils   in   the   study  area

vary from sands to loams to clays to organic  mucks.



          Table 1.  Major Soil Group Descriptions and Spatial  Extent
Group
1
2
3
4
5
6
7
8
Drainage
Class
SPD-VPD
WD-SPD
VPD
MWD-VPD
SPD-VPD
WD-MWD
WD-VPD
WD-VPD
Surface USDA-SCS
Soil Hydrologic
2
Texture Group
SiCL-C
L-SiCL
Muck
L-SiL
SiL-SiCL
L-SiL
FS-SiL
SL-SiL
D
C-D
A-D
B-D
B-D
B-C
A-B
B-D
Parent Percent of
Material Basin Area
Lacustrine-Till
Till
Organic
Outwash
Lacustr ine
Ou twa sh- Al 1 uv ia 1
Outwa sh-Al 1 uv ial
Till
28.6
50.5
0.2
4.6
9.3
0.5
4.5
1.8
 1 Drainage class abbreviations are:   VPD—very poorly drained,  SPD—somewhat
   poorly drained, MWD—moderately well  drained, WD—well drained.

 2 Some groups (particularly 3 and 4)  contained soils which were listed as
   drained/undrained,  thus the large range.

 3 Groups 4 and 6 were later combined.
     Cropping patterns are also variable.   However,  the  predominant   agricul-

tural crops are corn,  soybeans and small grains.  Much of  the area  is  composed

of soils with less than adequate internal   drainage   and   very  little  slope.

Thus, both surface and subsurface drainage systems are numerous.

-------
in
c:
-5
 rv
 3=
 -5
 fD
 CT5
 -s
 o
MAUMEE
  PORTAGE
   SANDUSKY

    BASIN
     SOIL
      GROUPS

-------
STUDY METHODOLOGY






     The ANSWERS  (Areal Nonpoint Source Watershed Environment Response Simula-




tion)  model,  developed at Purdue University (Beasley, 1977; Beasley and Hug-




gins, 1982), was chosen as the predictive tool for the study.   Since  ANSWERS




is  an  event-oriented,  distributed parameter model, simulation of the entire




study area was deemed both unreasonable and unnecessary.  Instead,  a  strati-




fied  sample  of the area was described, simulated and the results extended to



the much larger study area.






     The methodology required that a number of "representative"  subwatersheds



be selected and simulated.  Ideally, these subwatersheds would have an area of




1,000 to 1,800 hectares so that sediment and nutrient yields at their  outlets




could  be  eventually  expected  to  reach  the larger receiving body of water



(stream, river or lake).  This size of watershed normally would be expected to




support  a  second-  or  even third-order stream system.  It would, therefore,




drain to a well defined, continuously flowing channel system that should  pro-



vide  for  ultimate delivery to the receiving body.  Originally, a 0.5 percent




sample of the area was chosen.  This corresponded to 9 watersheds of the  size




mentioned  above.  Later work in the project included selecting and describing




an additional 0.5 percent sample and comparing simulation results  with  those




from  the first sample.  In doing so, a determination was made of the adequacy




of the different size samples (0.5 and 1.0 percent)  to  describe  the  entire




basin response of the different Soil Groups as indicated by monitored informa-




tion.

-------
     The watersheds, mentioned above, were chosen based on a  stratified  sam-



pling criteria.  The stratification involved grouping soils  (actually associa-




tions) with similar water movement characteristics  (internal drainage, surface




texture, hydrologic group, and parent material) in order to simplify and "nor-




malize" data collection and expand applicability of simulation  results.   The




application  of  this  classification  and stratification scheme resulted in 8




major Soil Groups.  The Soil Groups (shown in Figure 2) were  planimetered  to



determine  the  areas in each grouping.  The Soil Group descriptions and their




areal extent are shown in Table 1.  These major Soil Groups included more than




75  soil types.  Since neither Group 3 nor 8 accounted for more than 4 percent




of the area (both less than 2 percent), they were eliminated from further con-




sideration.   Since Groups 4 and 6 were similar, they were combined to form an




area large enough to require one watershed.






     Watersheds were then selected from areas mapped as belonging to a partic-




ular Soil Group.  The maps used in this original site selection were the state




soil association maps for Ohio, Indiana, and  Michigan.   Care  was  taken  to




select  drainage  areas in counties that were participating in the project and




that had modern soil surveys.  The topography and detailed  soils  information




from  the  selected  areas were then collected and assembled, using USGS topo-




graphic and SCS Soil Survey  maps,  into  detailed  watershed  data  files  as




described  in  the ANSWERS Users Manual (Beasley and Huggins, 1982).  Based on




the watersheds selected to represent the various groups, 9 were  required  for




both of the 0,5 percent samples.






     The cropping data, including management, was selected using a statistical




process  which  "populated" the watershed in a manner consistent with cropping

-------
patterns on the group being represented.  Once the percentages of the  various



land  uses  were  known,  a randctn number generator was used to "populate" the




watersheds with land uses so that the divisions of land uses in the watersheds




and  the entire county were the same.  The cropping and management assumptions




are detailed in Table 2.  Information derived from  the  Conservation  Tillage




Information  Center  (CTIC)  data  base  was used to describe the land use and




managenent patterns on the various Soil Groups throughout the study  area  for




the  final simulations.  Those percentages, upon which the conclusions of this




study are based, are shown in Table 3.  The scenarios  chosen  for  simulation




are  presented  in  Table  4.   There  were  two  actual years (1982 and 1983)




included along with the 6 potential  or  hypothetical  scenarios  (which  were




based  on modifications to the 1982 percentages).  The 1982 data is considered




the baseline, since that year has been so designated by USEPA.  'The 1983  data




is the first complete year of CTIC data.






     For the simulations, an element size of 1 hectare (100 meters on a  side)



was  chosen.  This led to 1,000 to 1,800 overland flow elements plus somewhere




between 125 and 300 channel flow elements for a chosen  watershed.   The  time




step  chosen  for the simulation was one minute.  The rainfall event simulated



corresponded to an 8-year return period,  1.5-hour  duration  storm  with  the




antecedent  soil  moisture  at  field capacity.  This storm typically produced




sediment and total P yields  that  corresponded  well  with  observed  average




annual  yields on a number of watersheds within the basin  (Black Creek area of




Allen Co., IN).  The definitive work on this "design storm" concept  was  per-




formed  by  Conrad  Heatwole (1980) and was utilized in both the Indiana Model



Implementation Project  (MIP) and Allen County Special ACP Projects.

-------
                Table 2.   Land Use and Management Descriptions
Land
Use

Row Crop


Other Crop

Hay-Pasture
Woods
Built-up
Tillage Area in
Management Vegetation

Conventional
Chisel Plow
Mini Till
No-till
Conventional
Chisel Plow
No- till
	
	
	
- % -
25
30
60
80
60
60
85
90
85
60
Max. Roughness Residue
Height Cover
- on -
5.1
6.4
6.4
7.6
4.6
5.1
6.4
3.8
7.6
3.8

< 30%
30% to 60%
40% to 70%
> 70%
< 30%
30% to 60%
> 60%
	
	
	
USLE
"C"4

0.50
0.10
0.05
0.02
0.10
0.08
0.02
0.04
0.01
0.20
  Row crop values depict corn, other crop values describe a combination
  between small grains and truck crops.
o
  Conventional consists of fall turn plow and spring disk, chisel  plow
  is fall tillage with spring harrow or similar, mini-till is either
  ridge plant after field cultivation or standard till-plant technique.
  For other crop categories, these definitions really only designate
  differences in residue cover.

  Residue covers are based on high predictivity.  Thus, lesser yields
  would probably produce more erosion and sediment yield.

  The values are reported for crop stage 1 - the period after planting
  until the plants reach 1 month in age.


     The 75-plus soil types were sorted for similarities based  on  hydrologic

group,  textural class, parent material, internal drainage, soil erodibility -

"K", and permeability.  The result was 13 different  soils  response  classes.

When  conservation  tillage  was  applied on a specific soil, the infiltration

capacity of that soil was assumed to increase.  The more conservative the til-

lage systems and the more residue that remained on the surface, the higher the

infiltration rates (Beasley and Huggins, 1982).

-------
   Table 3.  Cropping Distribution for the Major Soil Groups (from CTIC).
Group
1
2
4+6
5
7
Row Crops
63
58
63
50
55
Crop
Other Crops
19
18
16
15
16
or Cover*
Pasture
5
10
8
15
10
Forest
8
9
8
15
14
Built-up
5
5
5
5
5
* Row crops include:  corn, soybeans, sorghum,  etc.   Other crops include:
  small grains, beets, berries, etc.  Pasture  includes hay crops.   Forest
  includes: brushland and swamp.  Built-up includes: homesteads, farmsteads,
  and other semi-pervious, smooth areas.
               Table 4.  Scenarios Used in the Simulation Study

   Scenario	                      Description

      1       Baseline.  1982 basin averages for conventional,  chisel
              plow, reduced till, and no-till on row and other  cropland.

      2       1983 basin averages for conventional,  chisel plow,  reduced
              till, and no-till on row and other cropland.

      3       1/4 of the conventional tillage on both row and other crops
              replaced with fall chisel plowing.

      4       1/4 of the conventional tillage on row crops only replaced
              by till-plant minimun tillage techniques.

      5       1/4 of the conventional tillage on both row and other crops
              replaced with no-till techniques.

      6       1/2 of the conventional tillage on both row and other crops
              replaced with fall chisel plowing.

      7       1/2 of the conventional tillage on both row and other crops
              replaced with no-till techniques.

      8       All conventionally tilled areas switched to no-till techniques
              (most conservative scenario).


                                    10

-------
     The phosphorus projections used in this study are the result of an  equa-



tion which relates particulate phosphorus yield (85% - 90% of Total  P)  to sed-




iment yield.  This equation was the result of an extensive  data  analysis  of




the  water  quality  information gathered in the Black Creek Project (Lake and




Morrison, 1977) by Dr. Darrell W. Nelson, formerly of the Agronomy  Department




at Purdue.  The equation used was:








                             P = ,00058*SED**1.12








      where:   P = particulate phosphorus yield in kg/ha,






             SED = predicted sediment yield in kg/ha.








The predicted particulate P yield was then divided by 0.85 to produce an esti-




mate  of Total P yield.






     When the representative watersheds had been described and crop and  soils



information  supplied  to  the  model,  a  number of scenarios were simulated.




These scenarios described the "as is" or baseline conditions in the watersheds




as  well as potential conditions involving the changing of management to forms



of tillage with greater conservation potential.  Thus, a planner  could  study




the  impact of various levels of conservation tillage application (all the way




to 100 percent no-till) and compare them to existing levels.
                                     11

-------
PROJECT RESULTS






     The primary reason for conducting  this   study  was   to   provide   accurate




estimates of the efficacy of a large-scale application of conservation tillage



on sediment and phosphorus yields in the Western Basin of Lake  Erie.   However,




other questions also needed to be answered.   The question of  how large an area




to simulate to arrive at consistent  results  is quite   important.    >From   a




planner's  standpoint, the question of  how much time must be  allotted  to apply




the methodology described herein is also very important.






     The work conducted in the first two  years  of the   study  involved  the




selection  and  simulation  of  two  0.5  percent  samples of the total basin.




Numerically, this involved 9 watersheds in each year.  While  project personnel




spent  large  amounts  of  time  defining the basin, detailing  how Soil Groups




would be described,  selecting  appropriate   watersheds   for  simulation,  and




entering data, we believe that a routine application of this  methodology could




be done much more efficiently.  Selection of watersheds,  data preparation, and




simulation  should  be possible on a 3.5 to  5.0 days per  1,000  elements basis.



Thus, it should be possible to complete work on a 1,500 element  watershed   in




less  than  two  weeks.  Although cost  information  and data on  individual com-



puter requirements is generally not applicable to other systems,  the   simula-




tions  presented  herein  typically  cost $8  to $14 on  Purdue's CDC-6600 and




required 200 to 350 CPU seconds.  Simulations  were also run   on a   Digital




Equipment  Corporation  VAX  11/780  and  typically required  500 to 800 CPU




seconds.
                                     12

-------
     The Group 1 and Group 2 soils were represented  by  multiple  watersheds.




Two  catchments  were  used for Group 1 and four for Group 2.  The output from




the  individual watersheds were combined to give a Group average.  Each of  the




watersheds  that made up the two 0.5 percent samples are described in Tables 5



and  6.






     In an effort to more closely detail the cropping and  management  systems




presently  in use, the CTIC Data Base was utilized.  Table 3 displays the land




use  percentages for the baseline (1982) scenario for each of  the  major  soil




groups.






     The simulation results for the 8 scenarios for both of  the  0.5  percent




samples are shown in Tables 7 and 8.  Watersheds composed primarily of Group 2




soils stand out as sources of sediment and phosphorus in both samples.   Since




the  soils in this Group are typically more erodible, the topography more vari-




able, and the infiltration rates fairly low, it is quite  logical  that  these




soils would have the highest erosion and phosphorus yields.  Each of the other




Groups had at least one factor which lessened  its  impact  when  compared  to




Group  2.  All other Groups had lower average values of slope, and several had




much higher infiltration capacities.  Also, other Groups contained soils which



were more resistant to erosion than soils classified in Group 2.






     When the information in Table 1 is combined with that in Tables 7 and  8,




the  problem  with  the  Group 2 soils becomes even more evident.  Since these




soils make up over 50 percent of the basin, any large-scale reduction  in  the




phosphorus  yields  to the Western Basin of Lake Erie will have to be achieved




by reducing Group 2 phosphorus yields.  Table 9 highlights this information by
                                     13

-------
          Table 5.  General Watershed Information (First 0.5 Percent Sample)
No.
Name
Location
Description
      Edgerton-Carson and
      Miller Ditches

      Unnamed Tributary to
      Auglaize River

      Bottern, Kurtz, and
      Roth Ditches

      East Branch
19    Aurand Run
20    Kyle Prairie Creek
 7    Unnamed Tributary to
      Maumee River

 3    Cartwright Run
      Harris Ditch
               Allen Co., IN


               Allen Co., OH


               Allen Co., IN


               Seneca Co., OH


               Hancock Co., OH


               Mercer Co., OH


               Henry Co., OH


               Putnam Co., OH
               Lucas, Henry, and
               Fulton Co., OH
                Group 1 soils, 1,164 hectares,
                0.21 percent average slope.

                Group 1 soils, 1,118 hectares,
                0.44 percent average slope.

                Group 2 soils, 1,366 hectares,
                1.16 percent average slope.

                Group 2 soils, 1,192 hectares,
                1.74 percent average slope.

                Group 2 soils, 1,573 hectares,
                0.66 percent average slope.

                Group 2 soils, 1,372 hectares,
                0.68 percent average slope.

                Groups 4 and 6 soils, 1,827  hectares,
                0.53 percent average slope.

                Group 5 soils, 1,234 hectares,
                0.62 percent average slope.

                Group 7 soils, 1,606 hectares,
                0.66 percent average slope.
                                        14

-------
        Table 6.  General Watershed Information (Second 0.5 Percent Sample)
No.
Name
Location
Description
10    Unnamed Tributary to
      Auglaize River

11    Town Creek
21    Beaver Run
22    Unnamed Tributary to
      Black Creek

23    Harrison Creek
24    Lick Creek
16    Coon and Lick
      Creeks

17    Owl Creek
18    Unnamed Tributary to
      Beaver Creek
               Putnam Co., OH


               Van Wert Co., OH


               Allen Co., OH


               Mercer Co., OH


               Seneca Co., OH


               Williams Co., OH


               Henry Co., OH


               Williams Co., OH


               Wood Co., OH
               Group 1 soils, 1,628 hectares,
               0.23 percent average slope.

               Group 1 soils, 1,263 hectares,
               0.31 percent average slope.

               Group 2 soils, 1,535 hectares,
               0.69 percent average slope.

               Group 2 soils, 1,057 hectares,
               0.76 percent average slope.

               Group 2 soils, 1,560 hectares,
               0.72 percent average slope.

               Group 2 soils, 1,284 hectares,
               1.59 percent average slope.

               Groups 4 and 6 soils, 930 hectares,
               0.56 percent average slope.

               Group 5 soils, 1,661 hectares,
               0.38 percent average slope.

               Group 7 soils, 1,311 hectares,
               0.37 percent average slope.
* Land use was broken into five categories: row crops, other crops, hay or pasture,
  woodlands, and built-up areas.  For this table, cropland includes the first two
  categories.
                                     15

-------
Table 7.  Predicted Sediment and Phosphorus Yields for Each Scenario on Each Watershed - First 0.5% Sample.


#1
#2
#3
#6
#19
#20
#7
#8
#9
Soil
Watershed Group

- Edger ton-Car son and 1
Miller Ditches
- Unnamed Tributary to 1
Auglaize River
- Bottern, Kurtz, and 2
Roth Ditches
- East Branch 2
- Aurand Run 2
- Kyle Prairie Creek 2
- Unnamed Tributary to 4+6
Maumee River
- Cartwright Run 5
- Harris Ditch 7
Parameter

Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
1

210
0.27
600
0.88
1,330
2.14
2,210
3.80
680
1.02
840
1.28
400
0.55
530
0.77
130
0.17
2

210
0.27
600
0.87
1,250
2.01
2,060
3.51
640
0.95
790
1.20
370
0.52
530
0.76
130
0.15
3

210
0.24
520
0.75
1,140
1.81
1,880
3.17
600
0.88
730
1.10
350
0.48
450
0.65
120
0.15
Seen;
4
U-n/l
Kg/ 1
170
0.22
500
0.72
1,100
1.74
1,820
3.06
570
0.84
700
1.05
340
0.46
420
0.60
120
0.14
ario
5

160
0.21
480
0.68
1,060
1.66
1,760
2.94
540
0.78
670
0.99
320
0.44
410
0.58
110
0.14
6

170
0.21
440
0.63
940
1.45
1,540
2.54
510
0.74
620
0.91
290
0.40
380
0.53
100
0.12
7

120
0.15
350
0.48
770
1.16
1,280
2.06
400
0.56
490
0.70
250
0.33
300
0.41
80
0.10
8

50
0.06
110
0.13
220
0.29
350
0.48
150
0.19
160
0.21
90
0.11
90
0.11
30
0.04

-------
Table 8.  Predicted Sediment ard Phosphorus Yields for Each Scenario on Each Watershed - Second 0.5% Sample.


#10
#11
#21
#22
#23
#24
#16
#17
#18
Soil
Watershed Group

- Unnamed Tributary to 1
Auglaize River
- Town Creek 1
- Beaver Run 2
- Unnamed Tributary to 2
Black Creek
- Harrison Creek 2
- Lick Creek 2
- Coon and Lick 4+6
Creeks
- Owl Creek 5
- Unnamed Tributary to 7
Beaver Creek
Parameter

Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
1

210
0.27
170
0.21
690
1.04
860
1.31
940
1.45
1,750
2.92
550
0.80
380
0.54
140
0.17
2

200
0.26
170
0.21
660
0.97
810
1.23
880
1.36
1,620
2.68
510
0.74
380
0.53
130
0.16
3

190
0.24
150
0.19
620
0.91
740
1.11
810
1.24
1,470
2.41
470
0.67
340
0.47
130
0.16
Scena
4
kn/V
Kg/r
180
0.23
140
0.18
590
0.86
710
1.06
780
1.19
1,420
2.32
450
0.64
320
0.43
130
0.15
irio
5

170
0.21
140
0.17
550
0.81
680
1.02
750
1.12
1,360
2.21
430
0.61
310
0.42
120
0.15
6

180
0.23
130
0.17
520
0.75
640
0.94
710
1.07
1,240
1.99
400
0.56
290
0.39
110
0.14
7

130
0.17
100
0.12
410
0.58
520
0.75
580
0.85
1,030
1.61
320
0.44
230
0.30
90
0.11
8

60
0.06
40
0.04
140
0.18
160
0.21
200
0.26
280
0.38
130
0.16
90
0.10
50
0.06

-------
Table 9.  Predicted Sediment and Phosphorus Yields for Eight Scenarios on Each Soil Group — Western Basin Totals.
1 o
Group Sample

1 A
B
2 A
B
4+6 A
CO
B
5 A
B
7 A
B
Parameter

Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
Sed
P
1

280,300
400
131,500
165
1,546,000
2,520
1,295,400
2,055
49,400
65
67,900
100
119,300
175
85,500
120
14,200
20
15,200
20
2

280,300
395
128,000
165
1,448,200
2,345
1,212,900
1,905
45,700
65
62,900
90
119,300
170
85,500
120
14,200
15
14,200
15
3

252,600
345
117,700
150
1,329,000
2,125
1,112,100
1,730
43,200
60
58,000
85
101,300
145
76,500
105
13,000
15
14,200
15
Scenario
4

231,900
325
110,700
140
1,280,100 1,
2,045
1,069,300 1,
1,660
42,000
55
55,500
80
94,500
135
72,000
95
13,000
15
14,200
15
5

221,500
310
107,300
130
231,300
1,945
020,500
1,575
39,500
55
53,100
75
92,300
130
69,800
95
12,000
15
13,100
15
6

211,100
290
107,300
140
1,102,900
1,725
950,200
1,450
35,800
50
49,400
70
85,500
120
65,300
90
10,900
15
12,000
15
7

162,600
220
79,600
100
898,200
1,370
776,000
1,160
30,900
40
39,500
55
67,500
90
51,800
70
8,700
10
9,800
10
8

55,400
65
34,600
35
268,800
355
238,300
315
11,100
15
16,000
20
20,300
25
20,300
25
3,300
5
5,400
5
1 Groups 1 and 2 utilized multiple watershed simulations.  Figures reported are averages.




2 A and B samples correspond to the first and second 0.5 percent samples.

-------
showing  Basin-wide  tonnages  for  each  major  Soil  Group.   Several of the




scenarios indicate reasonable reduction levels can be achieved without extreme




changes  in management.  Obviously, levels of reduction approaching 50 percent




or more  (unless all cropland goes to no-till) will only  be  achieved  with  a



mixture of tillage and structural (terraces, sediment basins, waterways, etc.)




BMPs.






     A notable result is that the baseline  (1982 scenario) yield for  the  0.5




and 1.0 percent samples were within 85 kg/ha  (10 percent) of each other.  This




indicates that the information gained from  the 0.5 percent sample  would  have




been essentially the same as that gained  from the full 1.0 percent sample.






     An uncertainty analysis was conducted  on the 1.0  percent  sample   (since




each Soil Group had at least 2 observations) using the 90% Confidence Interval



as the test criteria.  The analysis looked  at the uncertainties  (based on sam-




ple  standard  deviation  and Student's t)  both within Soil Groups and for the




entire, basin-wide estimation.  Both sediment and phosphorus yields were sub-




jected to the analysis.  The results indicated that  for  the  1.0 percent  sample




 (18 watersheds) the uncertainty  associated with  sampling  ranged   from  26%




 (Scenario  8)  to  28%   (Scenario  1)  for   sediment estimations and  from 25%




 (Scenario 8) to 32%  (Scenario 1) for phosphorus estimations.   This   translated




to 745 _+ 207 kg/h
-------
     Sediment yields in the basin are low.   The  predicted  baseline  average




 (1.0  percent  sample)   of 745 kg/ha compares very well with the area weighted




averages of the Maumee (average of water years 1975-1978), Portage (average of




water years 1975-1978), and Sandusky (average of water years 1975-1979) basins




of 680 kg/ha/yr (USAGE, 1982).






     The data from both 0.5 percent samples were used to produce the  informa-




tion  shown in Table 10.  Long-term monitoring information (USAGE, 1979) indi-




cates that the mean annual  unit-area  P  loads  (both  agriculture  and  non-




agriculture)  for  the  basin  average  approximately 1.07 to 1.21 kg/ha.  The




modeling program indicates that  the  predominantly  agricultural  study  area




ranges  from 1.16 to 1.31 kg/ha (with individual Soil Groups ranging from 0.17




to 2.06 kg/ha).  Hence, the observed basin average and the  simulated  average




are  essentially  the same.  While the Group 2 soils have unit area loads of P




approaching 2 kg/ha, the entire basin is closer to 1 kg/ha.  Even though these




yields  are  relatively  high  when compared to some other agricultural situa-




tions, they are certainly not out of line with highly fertile and highly agri-




cultural  areas  in other parts of this country.  Tables 7 and 8 indicate that




substantial reductions in sediment and phosphorus yields will only occur  with



large  changeovers  to  more conservative tillage types (or addition of struc-




tural measures) .






SUMMARY AND CONCLUSIONS






     A model ing study was undertaken to help in assessing the effectiveness of




 increasing  amounts  of conservation tillage for reducing phosphorus and sedi-




ment yields to Lake Erie.  The study area included the  Maumee,  Portage,  and
                                      20

-------
                   Table 10.  Unit Area Loading Information
                                                  kg-P/ha
                   Mean Annual Monitoring Data
                   (from LEWMS reports)
                       1975 Basin Average           1.21
                       1976 Basin Average           1.10
                       1977 Basin Average           1.07

                   ANSWERS Predictions

                       0.5 percent sample           1.31
                       1.0 percent sample           1.17
Sandusky River basins and near-lake areas.  The  ANSWERS  model  was  combined

with  a stratified sampling procedure to produce estimates, based on represen-

tative watersheds, for major Soil Groups.   These  estimates  were  then  area

weighted to yield basin-wide predictions.


     Predictions for two 0.5 percent samples of the study area were  produced.

The second sampling was used to make an overall 1.0 percent sample.


     Statistical cropping and management data from county  Soil  Surveys  were

originally  used  in the watershed data files.  However, CTIC data was used to

update the  tillage  management  scenarios  and  produce  more  representative

descriptions for the final simulations presented herein.


     There are a number of important results and conclusions that can be drawn

from this project:


  1.  Since both the 0.5 and 1.0 percent sample results compared closely  with

      monitored  information,  a 0.5 percent sample appears to be adequate for

      the region under consideration.
                                     21

-------
2.  The Group 2 soils predominate in both  areal   extent   and   sediment   and



    phosphorus  yields.   These soils must be the  primary  target in any major




    reduction program.






3.  The "representative watershed"  concept  produced  results  that ware  con-




    sistent  with  both  USGS  and   LEWMS  information.   Thus,  the projected




    impacts of tillage management changes should  be  quite reasonable.






4.  The cost of producing these simulations is  only  a  very small fraction of




    what  a continuous monitoring program would cost.  Watershed description




    is the major cost and is a "one-time" expense.  In addition, results  are



    available much more quickly than monitoring data and  can be obtained  for




    hypothetical as well as actual  conditions.






5.  The amount of reduction achievable from  conservation  tillage   only   is




    probably  not  adequate  for  the  Annex  III goals.   The  1,000  + metric



    tonnes required of the study area could only be  achieved if  almost   50%



    of  the  basin  changes over to no-till.  If  only  the Group 2 soils were




    treated, approximately 60% of all Group 2 soils  in the row crop  or other




    crop categories would have to be switched over to  no-till  to achieve  the




    sought after reduction.






6.  A realistic, short-term expectation might be a 50  percent  switchover   to




    chisel  plowing   (scenario  6).   This  would  indicate a  reduction of  837




    tonne over 1982 yields (against an approximate 1,100  tonne goal), assum-




    ing all cropland  in the basin was treated equally.
                                   22

-------
  7.   The model  used  for  phosphorus prediction produces an estimate  of  total




      phosphorus.   Monitoring  data  suggests that as much as half of the bio-




      logically  available P is  soluble.  While the estimation technique  gives




      credit to  some  soluble P,  it may not be accurate for small, low sediment




      yield events.   Hence, the weight given to  soluble P  may  not  be  great




      enough.






  8.   ANSWERS produces estimates  of total  sediment  yield.   Most  monitoring




      programs  are   reporting  results based on  suspended sediment monitoring.




      While bedload  is not a significant problem in  some parts of  this  coun-



      try,  it  should certainly be  considered  in the upper Midwest!  Many of




      the particles moving through saltation are aggregates  that  are   almost




      entirely made  up of clay  and silt  particles.   These particles carry P in




      almost exactly the  same concentrations as  the  primary particles.   Their




      large weight,  when  compared to  dispersed,  suspended particles,  indicates




      the importance of adequately accounting  for their presence  in the  sedi-




      ment and P loading  sampling programs.






REFERENCES






  1.   Beasley, D.B.   1977.  ANSWERS:  A mathematical  model  for   simulating  the




      effects of land use and management on water quality.   Ph.D.  Thesis, Pur-




      due University, West Lafayette, IN.   266 pp.






  2.   Beasley, D.B.  and L.F. Huggins.   1982.    ANSWERS   Users  Manual.   EPA-




      905/9-82-001.    U.S. Environmental Protection  Agency,  Region V,  Chicago,




      IL.  54 pp.






                                      23

-------
3.  Heatwole, C.   1980.   Simulation of watershed   response   to  determine   a




    design  storm.   Project Report for CE 642 (Fall  1979)  for  Dr.  R.A.  Rao.




    Purdue University.  West Lafayette, IN.   42 pp.






4.  Lake, J. and  J. Morrison.  1977.   Environmental  impact  of  land  use  on




    water  quality,  final  report  of  the Black Creek project —  technical




    volume.   EPA-905/9-77-007-B.    U.S.  Environmental  Protection  Agency,




    Region V.  Chicago,  IL.   274 pp.






5.  U.S. Army Corps  of  Engineers.   1979.    Lake  Erie management  study.




    Methodology Report.   Buffalo,  NY.   146 pp.






6.  U.S. Army Corps of Engineers.   1982.  Lake Erie  wastewater  management




    study.  Final Report. Buffalo, NY.  241 pp.

-------
                                    TECHNICAL REPORT DATA
                             (Please readInilructions on the reiersL before completing!
 1  REPORT NO
   EPA-905/2-87-003
                                                            3 RECIPIENT'S ACCESSION-NO
 4. TITLE ANDSU8TITLE
   Modeling Component-Tri-State  Tillage Project
    "Applying the ANSWERS Model  to  Assess the Impacts of
     Conservation Tillage on Sediment  and Phosphorus Yields
                                    5 REPORT DATE
                                     May  1987
                                    6. PERFORMING ORGANIZATION CODE

                                    " 5GL
 7. AUTHOR(S)
  David  B.  Beasley, Ph.D., P.E.
                                    8. PERFORMING ORGANIZATION REPORT NO

                                      GLNPO  #  87-08
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Agricultural  Engineering Department
  Purdue University
  West  Lafayette, Indiana 47907
                                                             10. PROGRAM ELEMENT NO.
                                    11. CONTRACT/GRANT NO.

                                       Grant  No.  R005717-01
 12. SPONSORING AGENCY NAME AND ADDRESS
   U.S.  Environmental Protection  Agency
   Great Lakes National Program Office
   230  South Dearborn Street
   Chicago,  Illinois 60604	
                                    13. TYPE OF REPORT AND PERIOD COVERED
                                       Final 1982-1984
                                    14. SPONSORING AGENCY CODE
                                       Great  Lakes  National  Program
                                       Office,  USEPA,  Region V
 15. SUPPLEMENTARY NOTES
   Ralph  G.  Christensen, Project  Officer
   Great  Lakes Program
 16. ABSTRACT                                                                   ~~~	
   This  modeling study was undertaken  to  help in the assessing  the  effectiveness of
   increasing amounts of conservation  tillage for reducing phosphorus  and  sediment yields
   to  Lake Erie. The study area  included  the Maumee, Portage, and Sandusky River basins
   and the near-lake areas. The  ANSWERS model was combined with a stratified sampling
   procedure to produce estimates,  based  on representative watersheds,  for major Soil
   Groups. These estimates were  then area weighted to yield  basin-wide  predictions.
   Predictions for two 0.5 percent  samples of the study area were produced. The second
   sampling was used to make an  overall  l.O percent sample.  Simulations of the data
   collected are reported in the report.  Eight different scenarios  were done with
   varying residue cover and tillage methods with the different soil  types.
 7.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                               b.IDENTIFIERS/OPEN ENDED TERMS
                                                                          c.  COSATi Field/Group
  Modeling
  Sediment
  Phosphorus
  Tillage
  No-till
  Conservation tillage
  Conventional tillage
  Uater  quality
Runoff
Yields
Unit area
loads
 8. DISTRIBUTION STATEMENT
   Document  is  available to the  public  throug
   National  Technical  Information  Service
   (NTIS), Springfield, VA 22161
                       19 SECURITY CLASS (This Report)
                                       21  NO. OF PAGES

                                          32
                      20 SECURITY CLASS (This page)
                                                  22. PRICE
EPA Form 2220-1 (3-73)
                                            25

-------