Clean Lake Study
         • .  . / . • .   J~
of Dillon Reservoir

                        Summit County,
                        -      -  •      •
                               • *~* T  '• "_7
                                Colorado^
                   ' Y
                    1 - . •
                   :: :'• -*1 •---».
          % »> • .
          ' ,*
           -
        .
        -
                                     * '

                                    .

-------
                     DILLON  CLEAN  LAKES  STUDY
                          FINAL REPORT

                          (Auausc,  1932)
               Western Environmental  Analysts,  Inc.

         Principal Investigators:  William M. Lewis, Jr.,  Ph.D.
                                   James F. Saunders III,  Ph.D.
                                   David W. Crumpacker ,  Ph.D.
                                   Charles Brendecxe, Ph.D.
            Northwest Colorado Council of Governments
                          P. 0. Box 739
                     Frisco, Colorado  8C443
                          (303)  668-5445
This study was supported in part by a Clean Lakes Grant from
the U.S. Environmental Protection Agency under Section 314 of
the Clean Water Act.  The following entities also contributed
staff and financial support to this study:
AMAX                           Breckenridge Sanitation District
Denver Water Board             Copper Mt. Sanitation District
Frisco Sanitation District     Dillon-Silverthorne Joint Sewer
Keystone                         Authority
Summit County                  Town of 3rec:
-------
                            DISCLAIMER
This report has been reviwed by the sponsoring agencies a.-i
approved for publication.  Approval dees not signify that the
contents reflect the views or policies of those agencies.
Mention of trade names or products does nor constitute
endorsement.

-------
                           Table of Contents




                                                                   Page




List of Tables	  .    ii




List of Figures	    vi




Summary	•	     1




Acknowledgments  .........  	  .  	     9




Introduction	    H




Design of the Study  .	    17




Methods	    28




Physical Variables and Major Ion Chemistry of  the  Lake   	    50




Phosphorus and Nitrogen in Lake Water and  Sediments  	    89




Particulates and Phytoplankton Biomass   	  ...   108




Photosynthesis and Oxygen Consumption in the Water Column  ....   140




Nutrient Enrichment  Studies  	  ,  	   161




Horizontal Spatial Variation in the Lake	175




Overview of Limnology and Trophic  Status   .	   191




Chemistry of Nutrient Sources as They Enter  the  Lake	200




Total Nutrient Loading of the Lake	232




Nutrient Export  in Relation to Land Use	244




Separation of Nutrient Sources within the Watershed  	   269




The Dillon Clean Lakes Model  	   296




Predictions for Development Scenarios	.  .   318




Literature Cited	347




Public Participation Summary  	   357




Appendix	359

-------
                             List  of  Tables

Table 1.   Summary  of  analytical coverage  for  the  different  kinds  cf
           samples.

Table 2.   Summary  of  sampling  dates.

Table 3.   Comparison  of  extraction by  acetone  and seChanel  for Lake
           Dillon samples  on  1  March  1982.

Table 4.   Comparison  of  discharge measurements  made  by  U.S.G.S.  and by
           current  nieter.

Table 5.   Summary  of  statistics on duplicate  analyses.

Table 6.   Summary  of  information  on  spikes  for  quality  assurance.

Table 7.   Data  on  EPA standards.

Table 8.   Morphotnetric statistics for  Lake  Dillon.

Table 9.   Morphcmetric statistics on Lake Dillon, assuming  water  at
           spillway  level.

Table 10.  Estimated partitioning  of  outflow over  the  period of spiilvav
           loss.

Table 11.  Summary  of  water flow into and  out  of the  lake.

Table 12.  Seasons  of  layering  and mixing  in Lake  Dillon.

Table 13.  Major ion composition of Lake Dillon.

Table 14.  Means for conductance,  pH, and  alkalinity  at  three  depths in
           the water column.

Table 15.  Mean  concentrations  of  total phosphorus and phosphorus
           fractions at various depths.

Table 16.  Results  of  a two-way ANO^'A.

Table 17.  Comparison  of  total  ? values from various  studies of Dillon.

Table 13.  Average  total nitrogen  and nitrogen  fractions  at  different
           depth strata.

Table 19.  Results  of  a two-way ANCVA.

Table 20.  Summary of  average N and ? amounts  for  tne  whole  water
           column.

-------
Table  21.   Particulate  P  and  interstitial  ?  from  surface  sediments.

Table  22.   Mean  total participates  of  Lake Dillon in different  strata
            of  the  water column.

Table  23.   Information  on chlorophyll  in  1981  and 1982.

Table  24.   Summary of the phytoplankton  composition of  Lake Dillon.

Table  25.   Sizes of important Lake  Dillon  phytoplankton.

Table  26.   Composition  of organic matter  in  the  top 5 a  of  Dillon.

Table  27.   Comparison of  primary production  between Dillon  and  other
            Colorado lakes.

Table  28.   Percent saturation corresponding  to various  temperatures  and
            oxygen  concentrations at  the  elevation of Lake Dillon.

Table  29.   Information  required  for  prediction of AHOD.

Table  30.   Observed areal hypolianetic oxygen  deficits  compared with
            those predicted from  three  equations  by Cornett  and  Rigler-

Table  31.   Chlorophyll  concentrations  for  the  ten enrichment studies.

Table  32.   Enrichment treatments separated into  statistically coherent
            groups.

Table  33.   Particulate  P  concentrations  in the enrichment treatments.

Table  34.   Means for variables at the  index  station and  four main
            stations.

Table  35.   Means for variables at the  index  station and  four cain
            stations.

Table  36.   Summary of variation  among  the  five main stations.

Table  37.   Results of 14-station heterogeneity study.

Tabla  28.   Summary of trophic indicators  for Dillon.

Table  39.   Means and standard deviation  for  the  two 24-hour sampling
            series .

Table  40.   Discharge-weighted means  and  standard  errors  of  means for
            vater chemistry  variables in  the  three rivers.

-------
 Table  41.   Discharge-weighted means for the two small streams entering
            the lake.

 Table  42.   Discharge-weighted means for the two effluents discharging
            near the mouths of Miner's and Soda Creeks.

 Table  43.   Loading rates for bulk precipitation at the main collecting
            station.

 Table  44.   Summary of comparison for phosphorus loading by bulk
            precipitation.

 Table  45.   Summary of water flow into the lake.

 Table  46.   Summary of phosphorus loading of the lake.

 Table  47.   Summary of nitrogen loading of the lake.

 Table  48.   Summary of N and P loading.

 Table  49.   Yield of water, total P, and total N from two watersheds
            representing background  conditions and  frcm two watersheds
            with roads but  otherwise undeveloped.
 Table  50.   Summary of  statistical information on empirical
            establishment  of
            background  yield.
establishment of the relation Yn = aY^ for
Table  51.   Yield  of  water  and  nutrients  from a watershed supcorting
            residential  area  on sewer  and  urban development on sewer-

Table  52.   Yield  of  water, nitrogen,  and  phosphorus  for watersheds
            containing residential  septic  systems.

Table  53.   Yield  of  water  and  nutrients  from watersheds supporting ski
            slopes.

Table  54.   Yield  of  water, N,  and  P from  watersheds  with interstate
            highways  but  little else.

Table  55.   Yield  of  water, N,  and  P for  the  mining area on upper
            Tenmile Creek.

Table  56.   Yield  of  ? and  N  frcm  three WWT?  plants on  a per capita
            basis.

Table  57.   Average point source yields on a  per capita basis.

Table  53.   Comparison of predicted  and observed X  and  ? loads for  the
            two vears of  stucv.

-------
Table 59.  Breakdown of nutrient contributions  to  Dillon.

Table 60.  Percentage contributions  to loading  of  the  laka  aggregated
           in  three different ways.

Table 61.  Comparison of predictions  from  the Dillon Clean  Lakes  Model
           with observations for 1981 and  1982.

Table 62.  Results of the application of the Dillon Model  to  scenario
           data.

Table 63.  Results of the application of the Dillon Model  to  scenario
           data.

-------
                             List  of  Figures
Figure  1.   Orientation map  of  Lake Billon and  its  watershed.

Figure  2.   Map  of Lake Dillon  showing  the locations of sampling
            stations.

Figure  3.   Locations  of the stream sampling  sices.

Figure  4.   Design of  precipitation chemistry sampler.

Figure  5.   Bathymetric map  of  Lake Dillon.

Figure  6.   Depth-volume curve  for  Lake Dillon.

Figure  7.   Hypsographic curve  for  Lake Dillon.

Figure  8.   Calculated inflow as  obtained  by  the  Denver Water  Department
            for  Lake Dillon.

Figure  9.   Calculated inflow since 1963 for  Lake Dillon.

Figure  10.   Outflow from Dillon.

Figure  11.   Change of lake  level with  time over  the study period.

Figure  12.   Volume of monthly  inflow as a percentage of actual  lake
             volume.

Figure  13.   Time-depth diagram for temperature  at  the  index station,
             1981  (°C).

Figure  14.   Time-depth diagram for temperature  at  the  index station,
             1982  (°C).

Figure  15.   Thickness of the mixed layer  during  the ice-free  seasons  of
             1981  and  1982.

Figure  16.   Secchi depth and depth of  1%  light  in  Lake Dillon.

Figure  17-   Cumulative percent light extinction  due to particulates,
             chlorophyll,  and soluble material  fractions in Dillon.

Figure  18.   Conductance in  Lake Dillon, 1981-1982.

Figure  19.   Calculated inflow  depth of the Snake River, Blue  River,  and
             Tenmile Creek during stratification  in 1981 and 1982.

Figure  20.   pH  in Lake Dillon,  1981-1982.

Figure  21.   Alkalinity in Dillon,  expressed  as r.g/1 CO?.

-------
Figure  22a .   Total  phosphorus  and phosphorus  fractions in the top 5 m
              of  Lake  Dillon.

Figure  22b.   Total  phosphorus  and phosphorus  fractions in deep water
              (40-45 m)  of  Lake Dillon.

Figure  23.   Total nitrogen and nitrogen fractions  in the top 5 m of
             Lake Dillon.

Figure  24.   Time-depth  diagram for nitrate  in Lake Dillon,  1981 (ug/1).

Figure  25.   Time-depth  diagram for nitrate  in Lake Dillon,  1982 (ug/1).

Figure  26.   Time-depth  diagram for total particulates  in Lake Dillon,
             1981 (mg/1).

Figure  27.   Time-depth  diagram for total particulates  in Lake Dillon, 1982 (ir.g/1)

Figure  28.   Time-depth  diagram for chlorophyll a  in Lake Dillon, 1981 (ug/1).

Figure  29.   Time-depth  diagram for chlorophyll a_ in Lake Dillon, 1932 (ug/1).

Figure  30.   Scale drawings of  Dillon  phy to plankton:  Lyngbya (A),
             Oocystis  (B, G) ,  Synechococcus  (C, L) , microf lagellates
             (D,  H) , Synedra  (E),  Rhizosolenia (F) , Asterionella (I).
             Monoraphidium  ( J ) ,  Rhodomonas ( K) .

Figure  31.   Simplified  version of the chlorophyll  time-depth diagram with the  cai;
             contributions  of algal taxa indicated.

Figure  32.   Algal abundance  summarized  as in  Figure 31.

Figure  33.   Chlorophyll a  as a percentage of  total particulate dry
             weight .

Figure  34.   Production  per unit volume  versus depth for  selected times
             of 1982 as measured by C-14 uptake.

Figure  35.   Primary production per unit area  and  column  efficiency of
             photosynthesis.

Figure  36.   Depth of maximum photosynthesis and maximum  photosynthesis
             per  unit chlorophyll.

Figure  37a.  Time-depth diagram for oxygen  in Lake Dillon,  19S1 (-g.'l) .

Figure  37b.  Time-depth diagram for oxygen  in Lake Dillon,  1982
Figure 38.  Nutrient limitations as  shown  by  nhe  enrichment  experiments
            superimposed on the nitrate concentrations  near  the  surface .

-------
Figure  39.   Means  on various  dates  for chlorophyll for the 5-station
             heterogeneity  study.

Figure  40.   Summary of  seasonal  events,  generalized as a composite of
             1981 and 1982.

Figure  41.   Data that showed  statistical evidence of significant diel
             variation during  the  study of January 1981.

Figure  42.   Orthophosphate  data  for the  Blue  River at its point of
             entry  into  Lake Dillon.

Figure  43.   Values  of selected variables for  the Snake River as it
             enters  Lake Dillon.

Figure  44a.   Values of  selected  variables for the Snake  River as it
              enters Lake Dillon.

Figure  44b.   Values of  selected  variables for the Snake  River as it
              enters Lake Dillon.

Figure  45.   Nitrate values  for Tenmile Creek  at  its point of entry t:
             Lake Dillon.

Figure  46.   Selected variables for  bulk  precipitation near the Snake
             River WWT?.

Figure  47.   Itemization of  sources  of  ?  and N loading for Lake Dillon
             at the  points of  entry  to  the lake.

Figure  48.   Position of Lake  Dillon  on the original Vollenweider
             diagram.

Figure  49.   Representative  watersheds.

Figure  50.   Nutrient yield  from  various  sources  expressed on a per
             capita  basis, assuming  300 mm runoff.

Figure  51.   Nutrient  yield  from various  sources  expressed on an areal
             basis,  assuming runoff  of  300 mm.

Figure  52.   Observed  and predicted  yields fron segments  of the Snake
             River watershed.

Figure  53.   Observed  and predicted  yields from segments  of the Blue
             River watershed.

Figure  54.   Observed  and predicted  yields from segments  of the Tenmile
             Creek watershed.

-------
Figure 55.  Linear plot of  the inventory change  functions  for  nitrogen
            and phosphorus.

Figure 56.  Percentage of loading due  to various  sources  in  1932.

Figure 57.  Structure of the Dillon Clean Lakes  Model.

Figure 58.  The 19 segments used in modelling.

Figure 59.  Log-log  plot of total P and chlorophyll a.

Figure 60.  Graphical summary of Dillon Clean Lakes Modal  outputs  for
            four key lake characteristics.

Figure 61.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 62.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 63.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 64.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 65.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 66.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 67.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 68.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four key lake characteristics.

Figure 69.  Graphical summary of Dillon Clean Lakes Model  outputs  for
            four kev lake characteristics.

-------
                                 Summary




     I-  The  purposes  of  the  Dillon  Clean Lakes Project were as




follows:   (a)  to  provide  comprehensive  limnological information on the




lake, including its  seasonal  cycles  and its  basic  physical,  chemical,




and biological features with  special attention to  trophic status;  (b)




to provide information on the present nutrient sources  of the lake and




their relation to land use; and  (c)  to  construct a model that would be




capable of predicting  the trophic  status  of  the lake  given any likely




combination of future  development  patterns for the Lake Dillon




catchment.




     2.  The  limnological study  showed  that  the events  in Lake Dillon




can be divided according  to the  following seasons:  (a) ice  cover  frorr,




January through April, (b) spring  mixing  from the  beginning  to the end




of May, (c) summer stratification  from  June  through October, and




(d) fall mixing for  most  of November and  all of December.  During




summer stratification, the mixed layer  is stable at a thickness of




5-10 m until  September, at which time it  begins to thicken.




     3.  Transparency  is  lowest during  the first half of the




stratification period.  There is a sudden decrease in transparency in




June caused by the entry  of large  amounts of particulate materials from




the watershed.  Depending on  the amount of runoff, this reduces the




transparency at the  surface from about  4  m to less than 1 ni.  A large




proportion of  the  entering particulates settle out quickly,  but rapid




growth of algae causes transparency  to  stay  lew.  Mir.iraum secchi  derchs




in July, at the tice of the chlorophyll naxinuc, are  between 1.5 and




2.5m.   Transparency increases steadily after the  chlorophyll zaxirzur:

-------
 in  July because  of algal  nutrient  depletion.   Fall transparencies are




 high  (secchi,  3-5 m).




      4.   Runoff  from  the  three  rivers  enters  the lake at different




 depths  depending on the wetness  of the  year  (amount  of runoff) and the




 time  of year.  During  both  1981  and 1982,  river vater entered the lake




 at  progressively greater  depths  during  the stratification season.




 Entry level was  uniformly higher in 1982,  when there was much runoff,




 than  in 1981.  In both years  runoff entered  the upper water column in




 the early period of stratification and  the middle water column during




 the middle and late portions  of  stratification.




      5.   Annual  average total phosphorus  over  all depths was 6.6 ug/1




 in  1981  and 7.7  ug/1 in 1982.  Peak values at  the surface during runoff




 were  considerably higher  than this in both years, however (12-17 ug/1).




 Soluble  inorganic phosphorus  was consistently  present in very small




 amounts.   The upper water column showed a  steady loss of total




 phosphorus from  the beginning of stratification  until the increase in




 thickness  of the  mixed layer  in  September.  Soluble  organic phosphorus




 was almost totally  depleted by the  end of  August.




      6.   Total nitrogen in  the entire water column averaged 294 ug/1




 in  1981 and 447 ug/1 in 1982.  Lake  Dillon is  unusual in its very high




 ratio of  total nitrogen to  total phosphorus.   Thi:; is partly due to




 tertiary treatment, which removes a  lot of the  phosphorus but not




nitrogen, and to   the presence of mining, which  releases much nitrogen




into upper Tennila Creek.   Soluble inorganic nitrogen is  steadily




depleted from the upper water column after the  onset  of stratification.




Depletion is  essentially complete by the middle  of July.

-------
      7.   The  amount  of phosphorus in the sediments of Lake Dillon does




not differ  from the  amounts  typical of natural oligotrophic lakes.




      8.   Total  particulates  in the tipper water column (C-5 a) averaged




2.3 mg/1  in 1981 and 2.85 tag/1 in 1982.  In both years there were tvo




peaks  of  total  particulates  in the upper water colunn:  one at the cine




of runoff caused by  inorganic  particulates, and a second in July caused




by phytoplankton biomass.




      9.   The  maximum chlorophyll  a_ at the surface occurred during July




of both years and fell between 11 and 13 ug/1.  The  average for the




post-runoff stratification season was 6.7 ug/1 in 1981 and 7.3 ug/1 in




1982.  The  highest chlorophyll values in both years  were between 17 and




18 ug/1.




      10.  Repeated elements  of the annual composition cycle for




phytoplankton were Asterionella,  a diatom that is abundant under the




ice and during  early spring  in Lake Dillon, and Synedra, a diatom that




accounts  for  a  major portion of the July peak of chlorophyll in Lake




Dillon.   Irregular but very  large populations were observed of the




blue-green  alga Synechccoccus  and the diatom Rhizosolania.  Xo




nitrogen-fixing blue-greens  were  found.




     11.  Low ratios of  phosphorus,  nitrogen, and chlorophyll a_ to




organic matter  during stratification suggest significant nutrient




stress on the phytoplankton  during stratification.  The  ratio of




photosynthesis  to  chlorophyll  a_ was  highest during periods of ceap




mixing and  lowest  under  ice  cover and during the first half of




stratification,  indicating that the  greatest nutrient stress occurs

-------
 under ice and between the middle of June and the end of August.
                                                           ij
 Maximum daily photosynthesis was between 500 and 900 mgC/m /day.

      12.  Minimum oxygen 5 m above the lake bottom occurred in October

 of both years.  In 1981 the miniinun was 4.4 mg/1 and in 1982 it was

 4.6 mg/1.  This represents about 50% loss of oxygen from the saturation

 values.  The areal hypolimnetic oxygen deficit was well above the level

 expected for oligotrophic lakes.

      13.  Nutrient enrichment studies  showed that phosphorus limitation

 prevails between 1 January and 15 July.  After 15 July, the

 phytoplankton community is strongly nitrogen limited until September,

 when thickening of the epiliranion replenishes  nitrogen supplies.  At

 the beginning of October,  strong phosphorus limitation is

 reestablished.  Nutrient  limitation is  minimal during the month of May

 and during the last three  weeks  of  September.   Nitrogen and phosphorus

 limitations  are rather closely balanced,  as shown by the switch from

 one to  the other over  the  seasonal  cycle.

      14.   Studies  of horizontal  spatial variation over the lake show no

 statistical  evidence of sustained differences  between different parts

 of  the  lake.   Randomly changing  patterns  of variation between stations

 are  detectable  but  small for  both biological and chemical variables.

 The  lake can be  treated as a  functional unit.

     15.  A variety of  trophic indicators  indicate that Lake Dillon is

ziesotrophic at the present time.

     16.  Volume of water entering  the  lake by way of the three major

rivers was almost double in 1982 what  it  was in 1981.  The  time-

weighted concentration of both nitrogen and phosphorus  was  somewhat

-------
higher in  1982  than  in  1981.   The  time-weighted  averages  for  total




phosphorus concentration  in inflowing  rivers  varied  between 6 and




20 ug/1.   Annual  tine-weighted nitrogen  concentrations  varied between




200 and 600  ug/1  in  the three  rivers.  The  average  total  phosphorus




concentrations  in tertiary wastewater  treatment  plant  effluents  were




substantially boosted for all  such plants by  brief  or  extended




shutdowns  and malfunctions.




     17.   The total  amount cf  phosphorus  entering  the  lake  was 63CO




pounds in  1981  and 10,700 pounds in  1982.   The comparable figures  for




nitrogen were 189,000 pounds in 1981 and  333,800 pounds  in  19S2.




Expressed  per unit lake area,  the  phosphorus  loading corresponded  to




0.29 g/m2/year  in 1981  and 0.42 g/m2/year in  1982.




     18.   The total  nutrient loading was  divided according  to sources




on the basis of a detailed study of  small representative  watersheds and




watershed  segments.  Equations  for the following sources  were developed




which relate the  amount of nutrient  yield from a given  land surface to




the amount of runoff:  background  yield  (undisturbed),  nonpoint  yield




from residential  area on sewer,  nonpoint  yield fron  urban area on




sewer, yield from area on septic systems, yield  from ski  slopes, and




yield from interstate highway.   Equations were also  developed for  the




yield from the following sources that  did not show dependence on the




amount of  runoff:  Climax Molybdenum,  secondary  treatment plants




(package plants), tertiary treatment plants.




     19.   It was  shown from an  analysis of  the stream  segments that




river valley bottoms incorporating  substantial amounts  of wetland,




standing watar,  and  gravel beds  accumulate  significant  fractions of the

-------
 nutrient  yield  in dry years  and  export accumulated yield in vet years-




 An  equation  was developed  to account  for  this  storage and purging




 effect.



      20.   The yield  equations and  the storage  equations  were tested on




 the total runoff nutrient  yield  observed  in  1981  and  1982.  The




 predicted runoff yield  to  the lake  was 2300  kg and the observed was




 2900 kg in 1981.  In 1982, the predicted  was 4600 and the observed was




 48CO.  Nitrogen predictions  were similarly good.




      21.   By use of  the yield equations and  other information on




 nonrunoff nutrient yield,  a  complete  breakdown was made  of the




 phosphorus and  nitrogen sources  for Dillon in  1981 and 1982.  In 1982,




 which is  most typical of a run of years under  present trophic




 conditions,  tertiary plants  accounted  for 15%  of  phosphorus loading,




 secondary plants for 2.1%, Climax Molybdenum mining for  2.2%,




 background runoff for 45.6%,  precipitation for 13.1%, groundwater for




 1.9/o,  dispersed  nonpoint sources for  14.2%,  and a major  construction




 project in the  Snake  River bottom for  6.4%.




      22.   At the present time, total  loading of the lake can be divided




 approximately into four quarters.  The  first two  of these quarters are




 taken up  by natural  sources.   The third quarter is taken up by sources




 that can  be traced back to human waste  (septic, wastewater treatment




 plants),   and  the fourth quarter  is accounted for  by human activities




 that cannot be traced back to  waste disposal.




     23.   At  most 25% of the  present  loading or the lake could be




stopped by the most  elaborate  controls  short of diverting wastewater




elsewhere.  A comprehensive list is made  of  possible  additional

-------
controls.   The  most  important  items  or.  the  list include major reduction




in the  shutdowns  and  malfunctions  of  tertiary treatment plants,




conversion  of septic  systems  to  tertiary sever systems, containment




procedures  for  construction  in valley bottoms, and containment of the




present yields  of interstate  highway.




     24.  A model was developed  with  the following components:




(a) a land  use  component  that  accepts information on land use and




amount  of runoff, (b) a  trophic  status  component  that accepts output




from the land use component,  and (c)  an effects component that accepts




output  from the trophic  status component.   This model,  which is  called




the Dillon  Clean  Lakes Model,  predicts  the  nutrient yields by source




and by  watershed  segments  given  any  reasonable combination of future




land uses.  It  also  predicts  the total  loading and trophic status cf




the lake in terms of  total phosphorus and chlorophyll a,  and translates




this trophic status  prediction into  economically  significant measures




such as transparency  and hypoiimnetic oxygen  deficit.  The model




performs well on  1981 and  1982 data  and is  used to make predictions




about 10 different development scenarios described at the end of the




report.




     The modelling indicates  that  Lake  Dillon will move into the




eutrophic category if diversion  water rich  in phosphorus  is added in




quantity to the lake  or if high  growth  occurs without the adoption of




ncnpoint source controls or other  measures  net now in practice to




reduce  the  phosphorus loading  of  the  lake.  Under low growth or  higp.




growth with additional controls, the  condition of the lake cou_c be




held within the nesotrophic range, hut  wculd  suffer sccie  trc?r._c

-------
degradation or remain nore or less the same, depending on the exact




conditions.

-------
                       Credits  and  Acknowledgments




     This work was overseen  by the Dillon Clean Lakes  Steering




Committee, whose members  and their constituencies have been




exceptionally supportive  and helpful.   We are especially indebted to




the Committee's  Secretary, Mr. Tom Elmore,  who has answered innumerable




queries and solved many problems  for  us,  and  to the Chairman of the




Committee, Mr. Bruce  Bauiagartner,  for  his decisive leadership.




     Our own technical staff deserves  much  credit for  the smooth




completion of a  complex and  sometimes  difficult field  and laboratory




program.  Chemists for the project included Ms. Katherine Ochsr.er,  Mr.




Stephen Hamilton, Ms.  Margaret Robbins,  and Mr. Lewis  Dennis,  all of




whom were dedicated  to their work  well beyond what we  had a right to




expect.  The field crew,  which performed  its  duties well nany  times




under the most arduous conditions, included at various times Dr -  Robert




Epp, Mr. Steven  Murray, Mr.  Donald Morris,  Jr., Mr. George Kling, and




occasionally our chemists and  Mr.  Terry  Carter of the  Colorado




Department of Health,  who was  also helpful  in many other ways.  Ms.




Mary Marcotte did a  skillful job  of preparing the manuscript and Mr.




George Kling did the  same for  the  drafting.




     We thank the Denver Water Department,  Jerry Vest  and other members




of the Summit County  Planning  Department, the Summit County Engineering




Department, Barry Sheakley of  the  U.S.  Forest Service's Dillon Ranger




District, and Mr. Wesley  Nelson of the Colorado Division of Wildlife




for supplying us with various  kinds of  data.   We are grateful  to the




wastewater treatment  plant operators  and  especiallv to Mr. 3uc/. Wenger




of the Snake River Wastewater  Treatment  Plant,  where our field crsvs

-------
                                                                      10





did a considerable amount of sample  processing.  We  also  appreciate




occasional volunteer contributions from Dr. Thomas Frost,  Dr.  Suzanne




Levine, and Lie. Claudia Cressa, all of the C.U. Boulder  Limnology




Laboratory.




     Among ourselves we have divided the labor roughly  as  follows.




J. Saunders was field leader, ran the quality control and




troubleshooting for all aspects of the project, and  built  and




maintained the STORE! and other data files.  D.W. Crumpackar collected




all of the land use data from a combination of primary  and  secondary




sources and reduced the data to standard forms for modelling.




C. Brendecke supervised the collection of discharge  data, did  the  water




balance calculations,  and obtained  the U.S.G.S. records.  W.   Lewis  did




the data analysis,  writing,  and modelling.

-------
                                                                      11




                              Introduction




     Lake  Dillon  of  Siomit  County,  Colorado,  is an impoundment of the




Blue Fiver  just below its  confluence with the Snake River and Tennile




Creek..  The watershed drains  elevations  between lake level at about




2750 m and  the mountainous  headwaters of the  three ir.flowing rivers




at elevations as  high as  4300 n.   For at least the first decade after




its creation in 1963, Lake  Dillon was considered unequivocally




oligotrophic, as  shown by  its high  transparency.  Because of its




location,  Lake Dillon would be .axpectec  to remain oligotrophic




indefinitely if the  watershed were  uninhabited or very sparsely




inhabited.  Watersheds at  high elevations in  the Central Rockies are




seldom rich in phosphorus  because the parent  material, which lies




relatively close  to  the surface,  consists mostly of hard crystalline




rock that  is resistant to  weathering and poor in phosphorus.  In




addition,  the natural vegetative  cover effectively holds the




particulate phosphorus inventory,  as shown by the natural clarity of




streams in  undisturbed areas, even  at tines of peak runoff.




     The water of  Lake Dillon is  under the direct control of the Denver




Water Department,  which uses  Lake Dillon as the main storage facility




for the City of Denver.  Recreational activities on the lake are




managed by  the U.S.  Forest  Service.   The lake is a popular sport




fishery for rainbow  trout,  brown  trout,  and kokanee salmon,  and,




because of  a significant spawning run of brown trout up the  Blue River,




is of use  to the  State as a source  of brown trout eggs.  Numerous




permanent homes and  vacation  hones  lie within sight of the Lake and




aost of the owners of  these properties would  probably regard the blue

-------
 color  and  high  transparency  of  Lake  Dillon as  an aesthetically or




 economically valuable  amenity.   Furthermore, the appearance of the lake




 is  especially prominent  to the  general  public  because of the proximity




 of  the  lake to  Interstate 70.   Thus  it  is  clear  that  major changes in




 the algal  standing  crop  or transparency of Lake  Dillon could be




 undesirable from diverse viewpoints,  ranging from increased water




 treatment  costs for  the  City of  Denver  and lower property values for




 lakeside residents  to  loss of aesthetic  appeal for watershed residents




 and the general public.




     The uses of the watershed  are especially  diverse.   The U.S. Forest




 Service controls over  half the land of  the Lake  Dillon watershed; its




 holdings are principally on  the  steeper  slopes and at the higher




 elevations and are essentially undeveloped.  At  lower elevations, the




 watershed contains three major municipalities  (Frisco,  Breckenridge,




 and Dillon:  Figure 1).  There are many  other  smaller housing




 developments, some of  which  are  served  by  small  package plants,  some  by




 sewer through the four major treatment plants, and others by septic




 systems.  Four important ski areas are also located in the watershed




 (Breckenridge, Copper Mountain, Keystone and Arapahoe Basin).   Climax




Molybdenum, a major mining enterprise,  is  located  at  the headwaters of




Tenmile Creek.  Finally, the watershed serves a  large number of




nonresident visitors, who come in greatest  numbers for winter  sports




and in somewhat  smaller numbers for summer  recreation.   In 1981-1982,




the  time-weighted  average population of  the watershed was 19,000




persons and the  seasonal peak (winter) was  about  84,000 persons.




According  to  the Summit County Planning  Department, Summit  County,  half

-------
                                                                                  tISiWHOwER TUNNEL
 l-J
 I T

 1 '•

 >' J

 rt
[J
ru
I '
n>
                                                 SILVERTHORNE


                                                             ^

                                                             DILLON
ROBERTS

  1UNNEL
                                                                                                 Tailings pondt


                                                                                           •';,-J;/  Dr«dg«  lolling*


                                                                                          —(§)-- Stols  highway


                                                                                                           70
                        (|ipGLIWAX  MINE

-------
of which  is comprised of  the Lake Dillon  catchment,  was  the fastest




growing county in the United States in 1981.




     The  motivation of  the Dillon Clean Lakes  Project  is  twofold.




First, certain notable  changes, specifically a decrease  in




transparency, have occurred in Lake Dillon  in  recent years.  This,




coupled with the known  increase in population  and  intensity of land use




in the watershed, suggests that a comprehensive  evaluation of the




trophic status of the lake and of the present  nutrient sources would be




timely if not essential.  Secondly, additional development in the




watershed is expected,  while the incentives for  high water quality




remain the same or possibly even increase.  The  consequences of varying




degrees of additional nutrient loading of the  lake are not obvious and




must be anticipated by  some kind of predictive model capable of




approximating the trophic status of the laka under various




contingencies of watershed development.  Thus  the  twin purposes of the




Dillon Clean Lakes Study are evaluation of the present state of the




lake and construction of a model that will predict the future trophic




status of the lake given various assumptions about the direction of




development.




     A thorough study of Lake Dillon is well justified by  some of its




unusual features,  which may cause it to respond  to nutrient loading




somewhat differently than lakes that have been studied elsewhere.  It




is important to keep some of these points in mind  throughout the




evaluation of data and modelling.   First, the  background total




phosphorus concentrations for the Dillon watershed in complete absence




of human activity  would  be about 5 ug/1 P.  Such low background

-------
phosphorus concentrations  can maintain  lakes  cf  very low algal  standing




crop and thus of high  transparency  and  beautifully  blue  water.   In




contrast to the many lakes  that  occupy  watersheds whose  background




total phosphorus concentrations  are  several  times higher than  chis,




relatively little human  presence or  human  activity  is  required  co




double or triple the total  phosphorus concentrations.   The  vast




majority of lakes that have  been studied with  respect  to eutrophication




have higher background phosphorus concentrations than  Lake  Dillon,  so




good analogies are not always available.   We  will draw on information




from the lakes of the  Canadian  Shield and  parts  of  Scandinavia,  since




these lakes typically  have  low  phosphorus  concentrations,  ccol  water,




and short growing seasons  reminiscent of Lake  Dillon.   Reservoirs  ar.c




natural lakes in the Midwest and southeastern  United States, while




thoroughly studied, are  less similar to Dillon in these  important




respects.  Lake Tahoe  comes  to  mind  as  a good  comparison for Dillon  and




will occasionally be useful, but Tahoe  is  considerably more




oligotrophic than Lake Dillon,  owing largely  to  its  great  mean  depth




and small ratio of watershed area to lake  volume.




     The following report,  which is  based  on  field  work  extending  from




1 January 1981 to 31 December 1982,  deals  first  with the study  design




and methods of data collection.   A  comprehensive limnclogical analysis




follows ; the emphasis  of this is on  trophic  status  and nutrient




chemistry.  Next is an analysis  of  watershed  nutrient  yield, including




quantification of total  nutrient  loading and  separation  of  nutrient




sources based on land-use  analysis.  The last  portion  of the resort  is




devoted to description of  the Dillon Clean Lakes Model and

-------
                                                                      16




Its application to development scenarios provided by the Steering




Committee.

-------
                           Design  of  the  Study




     Lakes and watersheds  show  two major kinds  of  variability:




temporal and  spatial.   Sometimes  one category of  variation can be




studied much  more carefully  at  the expense  of the  other,  but a thorough




assessment and modelling  effort such as  that  required  for the Dillon




Clean Lakes Study would not  be  sound without  inclusion of extensive




information on both  spatial  and temporal variation.  At  the  same  tine,




very frequent sampling  over  an  extensive network  of  stations for  a




large suite of variables  is  seldom financially  feasible.   Fortunately,




such an approach is  seldom required.  For the Dillon Clean Lakes  Study,




the need for  information  on  temporal and spatial  variation was




fulfilled by  a dual  sampling program,  one part  of  which  was  designed




primarily to  provide detailed information for a few  sites on many dates




and the second to provide  a  detailed picture  of spatial  variation on a




smaller number of dates.   These two  approaches  were  supplemented  by




special data  collection programs  of  more limited  scope whose purpose




was to provide information that would  not necessarily  be  forthcoming




from the routine studies of  temporal and spatial variation.   The  data




base for the  Lake Dillon  Study  thus  consists  of three  components:  (1)




time series data set,  (2)  spatial survey data set, and (3) special




studies data  set.









Tine Series Data Set
     The tine series data  set  consists  of  three  parts:   routine  lake




data, routine stream data, and precipitation  chemistry  data.

-------
      (1)   Routine Lake Data.   A key station was selected near the




 center  of the lake in deep water.   This station is referred to as the




 index station;  its location is shown in Figure 2.   On each routine lake




 sampling  date,  the water column at  the  index station was sampled in 5-m




 increments  from top to bottom.  Variables  quantified at the index




 station are listed in Table 1.  In  addition,  one station was selected




 at  the  mouth of each of the four major  arms of the lake.  The locations




 of  these  four stations, which  are referred to as the main stations, are




 also  shown  in Figure 2.  The main stations were always sampled as part




 of  the  routine  lake sampling program.   However, samples were taken only




 at  the  top  and  bottom of the water  column  rather than over the entire




 vertical  profila  as at the  index station.   Analytical coverage is




 summarized  in Table 1.




      The  schedule  of  sampling  for the routine lake data collection




 series  is shown in  Table 2.  Samples were  taken on 32 different  dates.




 For most  of  the year,  the collections were biweekly.  During the period




 of ice  cover, collections were  less  frequent.




      (2) Routine Stream  Data.   All  overland flow was sampled as  close




 to the  lake as possible.  This  required collections  at eight stations;




 the resulting data make  up  the  routine  stream data set.  The eight




 stations,  with abbreviations as  shown in Figure 3,  were as follows:




 Snake  River Mouth (SRI), Blue River  Mouth  (BR1), Tenmile Creek Mouth




 (TCI), Miner's Creek Mouth  (MC2), Soda  Creek  Mouth (SC2),  Blue River




Outlet  (3RO), Frisco Effluent,   and  Snake River Effluenc.   Cther




etfluents  were represented  in the river mouth samplings  because  they  do




not  enter  the lake directlv.

-------
2.   Map of Lake Dillon showing the locations of sampling
    stations, including the index station (A?, the four rain
    stations (3-1), and che serve" stations (LS1-L59).

-------
 H-
 09
 m

 LO
 O
 Co
 rt
 H-
 O
 3
 cn
 01
 rt
 N
 ro
ro
fu
0
ti
oq
ro
in
                                                                                                     STREAM   SITE


                                                                                                     PRECIPITATION


                                                                                                     WWTP


                                                                                                      GAGING   STATION
                                                                                                  GALE
                                                                                                                                     r-o
                                                                                                                                     o

-------
   Variable
                  	Time Series	     Spatial Survey
                  Lake Index  Lake Main  Routine  Precipi-     Lake   Stream
                   Station     Stations   Strean   tation      Survey  Survey
Temperature

Conductance

Transparency

Discharge

Dissolved 07

PH

UCCL
PO^-P
NO-N
                                  +
                                                     +
Total Soluble P       +

Total Soluble N       +

Particulates          +

Partic. C             +

Partic. P             +

Partic. N             +

Chlorophyll a_         T

Priaary Production    +

Fhytoplankton         +
                                  +
                                  +
                                                     +
                                                     +
                                                                +
                                                                        +
Table 1.  Sunmary of analytical coverage for the different kinds  of  samples

-------

Routine
Lake/ Stream
23 Feb
30 March
13 Apr
11 May
26 May
08 June
22 June
13 July
27 July
10 Aug
31 Aug
14 Sept
23 Sept
12 Oct
26 Oct
09 Nov
23 Nov
21 Dec*

Routine
Lake/ Stream
01 Mar
15 Mar
05 Apr
19 Apr*
03 May*
17 May*
07 June
21 June
06 July
19 July
02 Aug
16 Aug
07 Sept
20 Sept
04 Oct
18 Oct
01 Nov
15 Nov
06 Dec*
1981

Stream Survey Lake Survey
26 Jan 09 Feb
27 Apr 18 May
05 May 29 June
18 May 20 July
01 June 24 Aug
08 June
15 June
06 July
03 Aug
05 Oct








1982

Stream Survey Lake Survey
25 Jan 08 Feb
10 May 24 May
31 May 14 June
07 June 26 July
14 June 23 Aug
21 June
28 June
12 July
09 Aug
11 Oct











Diel Enrichment
14 Jan 20 July
16 Nov 24 Aug
23 Sept

















Diel Enrichment
24 May
14 June
10 July
26 July
23 Aug
07 Sept
18 Oct












  Streams only.   Lake work judged unsafe.

Table 2.   Summary of sampling dates.  Precipitation chemistry samples
          were collected on the same schedule as routine lake and stream
          samples.

-------
     The schedule of  sampling  for  the  routine  stream data  set  was  the




same as for routine lake data  collection,  as  shown  in Table 2.   Table 1




shows the list of variables  analyzed for  each  sample.  In  addition to




the chemical determinations, discharge was  estimated at  each sampling




station on each date.




     (3)  Precipitation Chemistry.  Bulk,  precipitation (the combination




of dryfall and wet precipitation)  was  sampled  continuously near  the




Snake River Wastewater Treatment Plant (Figure 3).   The  sample was




collected whenever samples of  any  other kind vere collected on the lake




or in the watershed.  The analytical coverage  is  summarized in




Table 1.









Spatial Survey Series




     The spatial survey series  consists of  two parts: lake survey and




stream survey.




     (1)  Lake Survey.  The  lake survey stations  include the index




station, the four main stations, and nine  additional stations  scattered




over the lake (Figure 2) .  Analysis for each  station was based on  an




integrated sample taken from the top 5 meters.  A bottom sample  was




also taken at the index station and four nain  stations.  The analytical




coverage is summarised in Table 1.  The survey was  repeated 10 times




over the course of the two-year study interval, thus allowing  for




seasonal changes in spatial heterogeneity over the  surface of  the




lake.




     (2)  Stream Survey.  The stream survey series  is based on samples




taken at 33 sites in the watershed as shown in  Figure 3.   As with  the

-------
 lake  survey,  all  stations were  sampled  on  the  sane  day  so  that




 comparisons could  be made between  stations  for  a  given  date.   The 33




 stations  include  the stream  survey stations  already mentioned  above and




 the 11  stations of the  representative watershed sampling program (to be




 discussed below),  plus  14 other  stations distributed up and  down the




 nain  stems of  the  three  rivers  flowing  into  Lake  Dillon.   The  sampling




 dates are shown in Table 2.  A  few of the  stations  were not  sampled




 over  the  entire two years; these will be identified in  the course of




 data  analysis.









 Special Studies




      Special  studies include the following:  (1)  enrichment  experiments




 to determine  the identity of the limiting nutrient, (2) diel studies to




 determine the  degree of  24-hour variation in loading, (3)  sediment




 chemistry to  determine  the size of  the  nutrient inventory  in the lake




 sediments, (4) a representative watershed sample  series designed to




 provide export coefficients for major land use  types in the  Lake Dillon




 watershed, (5) supplemental rain chemistry  to provide additional




 information on atmospheric loading, (6) a groundwater study  to  show




 influence of groundwater on loading, and (7) a  land use survey.




      (1) Enrichment Studies.  On 10 different dates, water samples from




 the upper water column were enclosed in flexible  containers  and




 suspended within the lighted zone of the lake.  Inorganic  phosphorus




was added to certain of  these containers, inorganic nitrogen to  others,




and a combination of nitrogen and  phosphorus to still others.  Another




set of containers was left unaltered as a control.   Growth responses to




the various treatments  were determined  by changes in the concentration




of chlorophyll a within  the containers.  The objective  of  this work was

-------
to determine whether  the  phytoplankton  conznunity ai a given tine was




limited by phosphorus,  by nitrogen,  or  by neither of these elements.




The enrichment  experiments vere  done more frequently in 1932 than in




1981 after preliminary  evidence  indicated a late-summer switch between




phosphorus and  nitrogen limitation.   The  schedule of enrichment




experiments is  given  in Table  2.




     (2) Diel Studies of  Nutrient  Chemistry.   On cvo occasions the




nutrient transport  to the mouth  of  the  Blue River,  the Snake River,  and




Tenmile Creek was  studied over  a 2^-hour  period.  The objective was  to




determine whether  or  not  significant bias would  result if  loading of




the lake with phosphorus  and nitrogen were estimated on the basis of




samples taken at a  particular  time  of day or  exclusively during the




daylight hours.




     (3) Sediment  Chemistry.   Lake  sediment samples were taken at the




four main stations  and  at the  index  station on 26 July 1982.  The




interstitial waters and particulate  component of che sediments were




analyzed separately for all phosphorus  fractions.  The objective of




this work was to show whether  or not substantial reserves  of phosphorus




are present in  the  sediments of  Lake Dillon and, if so,  whether or not




significant amounts of  these reserves are present as soluble phosphorus




in the interstitial waters.




     (4) Representative Watershed Program.   Complete analysis of




present phosphorus  sources  and modelling  of future  phosphorus loading




both require sound  knowledge of  nutrient  export  coefficients associated




with different land uses  in the  Lake Dillon watershed.  Export




coefficients cannot be  derived from  data  on high-order screaks because

-------
 land uses  are typically mixed  in the watersheds of such streams.  For




 this reason,  a  set  of  representative watersheds was chosen,  each of




 which is drained  by a  low-order  stream and  whose land use is readily




 assignable to a particular  unmixed  category.   The land use categories




 chosen for representation are  as follows:   undisturbed forested




 watershed, residential areas served  by septic  systems, residential




 areas served  by sewer  systems, urban areas  served by sewers, ski




 slopes, roads,  mining,  and  interstate  highway-   Two watersheds were




 chosen for most types.   In  some  instances,  only one watershed for a




 particular type could  be found,  and  some of the sites were moved after




 the  first  year.  The discharges  and  concentrations of nutrients for




 each of the watersheds  were determined  for  each stream survey sampling




 date.   These  data were  used to compute  the  nutrient output per unit




 area for each of  the watersheds.




      (5) Supplemental Rain Chemistry.   After the 1981 data showed that




 precipitation would be  a significant contribution to  the  total




 phosphorus loading  of Lake Dillon, it  was decided that data  on




 atmospheric phosphorus  loading should  be taken  at several sites.  One




 additional station was  set up in  the summer of  1982 near  the Frisco




 Wastewater Treatment Plant, and  another was placed on the Denver Water




 Department's raft on the lake surface.  The purpose of the latter




 collector was to show whether the collection stations on  shore were




unduly influenced by terrestrial  dusts.




      (6) Groundwater Study.  Water and  nutrients enter Lake  Dillon




from  the air, as overland runoff, and  through groundwater-   In view of




 the geology of  the Lake Dillon watershed, groundwater  movement  was

-------
considered to be probably a minor source of nutrient  loading  for  the




lake, but a study was nevertheless nade of the grcundwarer  flow in




order to verify this impression.




     (7) Land Use Survey.  Partitioning of the present  nutrient loading




of the lake according to source requires a thorough quantitative




understanding of land use.  A land-use study was  thus carried  cut with




the objective of specifying the intensity, distribution, and  total




amount of various land uses in the Lake Dillon watershed.

-------
                                                                       28





                                 Methods









Lake Sample Collection




     Lakewater  samples were  always  taken  with  an integrating sampler of




the type described by Lewis  and  Saunders  (1979).  The sampler consists




of a PVC tube 5 m long with  closure  devices  at each end that can be




triggered by a  messenger.  When  the  sampler  is retrieved to the boat,




the water is released through  a  piece  of  surgical rubber tubing into an




integrating chamber, where the contents of the sampler are thoroughly




mixed.  Water is then drawn  from the integrating chamber into sample




bottles.  The integrating chamber incorporates a baffle that rises with




the water and thus prevents  contact  between  the air over the sample and




the water.  The integrator can thus  be used  on samples to be analyzed




for atmosphere-sensitive variables.




     The advantage of the integrating  sampler  is that it eliminates




sampling variance associated with vertical layering.  When successive




5-m increments  are taken from  the top  to  the botton of a water column




no layers will  have been missed.  At the  same  time, since the sampler




can be used on  successive 5-m  increments  in  the water column,




information on  gross vertical  structure of the water column can be




obtained with the sampler.




     Samples of Lake Dillon were  protected from heat and from direct




exposure to sunlight after they were bottled in the field.  Special




precautions were taken to protect the  chlorophyll samples from exposure




to light.  Sample bottles were always  rinsed copiously with deionized




water and dried at the end of each sampling  period.  A field log was

-------
maintained  of  sampling conditions and of problens that arose during




sampling.









Streamwater Sampling




      Streamwater samples were collected by insertion of a wide-south




plastic  bottle just  below the water surface in a reach of the stream




where  the flow was  fast across the entire cross section.  Care was




taken  not to stir up stream sediments in front of the sample.









Filtration  of  Water  Samples




     Filtration of  water samples was carried out at  the Snake River




Wastewater  Treatment Plant immediately after sample  collection.   Each




sample was  shaken prior to filtration and poured onto a filter tower




over  a Whatman GF/C  glass fiber paper (47 mm diameter).  The effective




pore  size of this paper is about 2 urn.  The purpose  of the immediate




filtration  was to reduce the  biological activity of  the sample to a




minimal  level  without the addition of any preservatives that might




affect sample  chemistry.  The filtered samples were  then returned to




Boulder  where  they were stored under refrigeration until analyzed.




Analysis of  soluble  constituents commenced the morning after sample




collection.  Analysis of labile constituents was completed the day




after sampling.









Analysis of  Particulates




     One filter  through which a known amount of water had been passed




was used for the  analysis  of  particulate phosphorus  (see below),  and,

-------
                                                                       30






 if chlorophyll was to be measured,  a second filter was used for this.




 The amount  of water that could be filtered by a single filter depended




 on the  amount of  particulates  in the sample,  but typically fell between




 100 and 1000 ml.




      Another filter for  each sample  was  dried to constant weight at




 60°C and weighed  with an analytical  balance.   A measured amount of




 water was poured  over this  filter, after which the filter was redried




 to constant weight at 60°C  in  an oven, cooled in a desiccator to room




 temperature and weighed  to  the nearest  .01 mg on an analytical balance.




 The difference between the  initial and final  weights was recorded as




 the amount  of particulate material on  the filter,  which was




 subsequently converted to mg/1 total particulates.  The filter was then




 saved for elemental  analysis with a  Carlo Erba Model 1102,  which gave




 the  amount  of  carbon  and nitrogen on the basis of  chromatography




 following automated  high-temperature combustion.   Because the amount of




 nitrogen  in  the particulate  fraction was a very small proportion of the




 total nitrogen, especially  for the streams,  the elemental analysis was




 performed only  on  the  samples  from the river  mouths, the lake index




 station,  and  selected  additional  samples from various locations.









 Particulate  ?




     One  filter from  each sample  was analyzed for  particulate




phosphorus following  the method  of Solorzano  and  Sharp (1980a).  This




relatively new method  is an especially sensitive  and reliable




alternative  to older methods,  which  often present  the problem of poor




or uncertain recovery.  The  Solorzano and Sharp method relies on

-------
decomposition of organic phosphorus  compounds  by  pyrolysis.




filters are dried with magnesium  sulfate,  heated  at  45G-50Q°C  for  one




to two hours, treated with hydrochloric  acid  to hydrclyze




polyphosphates, and then analyzed  for orthophosphate  by  the  standard




orthophosphate procedure (see below).  When used  in  combination  vith a




high-quality spectrophotometer, the  method allows  quantification of




very low amounts of particulate phosphorus, thus  eliminating problems




arising from the interpretation of data  that  include  many values below




detection limits.  Quality assurance information  en  this and other




methods is given at the end of this  section.









Soluble Reactive Phoschorus
     Soluble reactive phosphorus was measured by  the  ascorbic




acid-molybdate method (Murphy and Riley  1962).  This  is  the most widely




used method for low levels of soluble inorganic phosphorus.  When




combined with long-pathlength spectrophotometry,  soluble  reactive




phosphorus as low as 1 ug/1 can be measured.  Subtraction of a




turbidity blank at 885 nm is sometimes necessary  but  proved of




negligible importance on the filtered samples of  the  Dillon study.




Although the chemical species measured by  this  test will  be referred  to




as PO^-P, it is more properly characterized as  soluble reactive




phosphorus, since the test is sensitive  not only  to orthophosphate but




also to organic phosphorus compounds of  low molecular weight (Riglsr




1968, Levine and Schindler 1980).

-------
Total  Soluble  Phosphorus




     A pyrolysis method similar  to  that used  for  participate  phosphorus




was used  for total  soluble phosphorus  (Solorzano  and  Sharp  1980a) .   For




this method, a portion of the filtered sample is  evaporated to  dryness




with magnesium sulfate.  The sample  is then combusted at  450-5QO°C  and




the resulting  polyphosphates are hydrolyzed with  hydrochloric  acid.




This ±3 followed by determination of soluble  reactive phosphorus  by the




molybdate method.   Subtraction of the  soluble reactive  phosphorus from




total  soluble  phosphorus gives soluble organic phosphorus.   Addition of




total  soluble  phosphorus to particulate phosphorus gives  total




phosphorus.









Ammonium




     A modification of Solorzano's phenolhypochlorite method  (1969)  was




used for ammonium (see Grasshoff 1979).  This is  a -very sensitive




technique capable of detecting as little as 1 ug/1 NH^-N  with




appropriate spectrophotometry.  Hypochlorite  is added to  ammonium in




alkaline solution.  With the addition  of phenol and nitroprusside,




indophenol blue is  formed.  The indophenol blue concentration  is




determined spectrophotometrically.









Nitrite and Nitrate




     The widely used diazotization method originally  described  by




Bendschneider and Robinson (1952) was  used for nitrite.   The




combination of nitrate plus nitrite was determined by passage  of  a




portion of filtered sample through a reduction column containing  a

-------
cadmium-copper couple,  after  addition  of  a  buffer-   Nitrate  is  thus




converted to nitrite.   The amount  of nitrite  in  che  sample after




reduction is the sum of nitrate  and nitrite in  the  sample  before




reduction.  Subtraction of the nitrite  originally  in the  sample yields




the amount of nitrate.  The method is  extremely  sensitive; detection  is




below 1 ug/1 of N.




     After May 1982, all  nitrate measurements were  made  by ion




chromatography rather  than wet chemistry.   Ion  chrcmatography was




carried out with a Dionex Model  2110 chromatograph  using  a bicarbonate




eluent.  The automated  ion chromatographic  method  is highly  sensitive




and offers the advantage  of lower  variance  in measurements because of




reduced sample handling.









Total Soluble Nitrogen




     A new method outlined by Solorzano and Sharp  (1980b) was used for




total soluble nitrogen.   Analysis  of total  soluble  nitrogen  when




soluble organic nitrogen  is present in  small  amounts has  traditionally




been unsatisfactory because of low sensitivity  or  poor  recovery.  The




new method, which emphasizes  better control over pH  and  oxidizing




conditions, reduces these problems.  The  method  is  based  on  oxidation




of soluble organic nitrogen by potassium  persulfate  in an autoclave.




This converts all soluble nitrogen compounds  to  nitrate.  Nitrate is




subsequently analyzed by  the  reduction method described  abcve.  This




produces an estimate of total  soluble  nitrogen.  Soluble  organic




nitrogen can be obtained by subtraction of  the ammonium,  nitrate, anc




nitrite obtained in the other  analyses.   For  the low levels  of  soluble

-------
                                                                       34





 organic  nitrogen  present  in  certain  circumstances  for  the Dillon study,




 recrystallization of  reagents was necessary  to  improve  the precision of




 the  test.









 pH,  Alkalinity, and Oxygen




     pH  was  determined with  a Radiometer M29 pH meter  and combination




 electrode  on an unstirred sample.  Alkalinity was  determined  by




 titration  of a 100-ml aliquot to an  endpoint of pH 4.4  with N/44
      Dissolved  oxygen was determined by  the  azide modification of the




 Winkler method  (APHA 1975).









 Major Anions  and Cations




      Major ions not of direct nutritional significance  were  not




 determined routinely.  However, a few determinations  were  made,  and in




 these  instances the anions chloride and  sulfate were  determined  by ion




 chromatography  (Dionex 2110) and the cations were determined by  atomic




 absorption spectrophotometry (Varian AA6) after addition of  lanthanum




 oxide  and cesium chloride.









 Chlorophyll a




     Chlorophyll a_ was determined after  hot  methanol  extraction.  The




basic  procedure is that described by Tailing (1969) but has  been




modified some in accordance with recommendations presented in the




recent comprehensive methods evaluation  edited by Rai (1930). The




sample is filtered just after collection onto Whatman GF/C paper,

-------
                                                                      25




MgCC>3 suspension  (0.2 nl)  is  added  to  the  filter.   The filter is then




placed in a screwcap test  tute with 40 nl  of  90% methanol,  heated to




&5°C, and allowed  to boil  gently  for 30-60 seconds.   The tubes  are then




capped again and  stored  in the dark for 6  to  24 hours.  The absorbance




is measured at 665 and 750 nn relative to  90% methanol.   The absorbance




is then computed  as follows:  12.9  :: (-tJ65~E7r(p x (volume




solvent in ml/volume filtered in  liters) x 1/pathlength  in  cm.   The




coefficient of this equation  is lower  than Tailing's  original




coefficient (13.9).  The lower coefficient brings  the calculation into




line with recant  information  on the specific  absorbance  of  chlorophyll




in methanol as reported  in the summary by  Marker et al.   (1980).   The




data were also corrected for  a problem that has not been brought  out in




the literature but which was  discovered as a  result of a




seven-laboratory  aethods study coordinated by the  University of




Missouri in which  we participated.   Certain glass  filters release a




binder or fine colloidal fragnent  that does not centrifuge  readily.




These particles differentially increase the absorbance at 665 and




750 nm.  Tests on  our own  filters  showed  that the  fragments were




present and increased the  absorbance at 665 nn by  21.5%  more than the




absorbance at 750  nm.  All  of the  calculations were corrected for this




differential.  The resulting  difference in chlorophyll values was not




very great in percentage terns, however.




     The methanol  method was  selected  because recent  studies have shown




its extraction capabilities to be better than those of any  other method




(Riemann 1980).  We did a  comparison of methancl ar.d  acetone




specifically for Lake Dillon  on 1 March 1932.   Table  3 shows the

-------
                                                                       36
 results.   As  expectad  from the  literature,  methanol  is more efficient

 than  acetone  for  samples  that are  steeped but  not  ground in a

 hcmogenizer.   The differential  averages  about  15%.  As a rule of thumb,

 values  for Dillon should  be  divided  by 1.15 if they  are to be compared

 with  future or past  acetone-based  values for Dillon  derived from

 unground  samples. Ground acetone  samples would  be very close to the

 methanol  values,  as  shown by the recent  interlaboratory study in which

 we  participated.
Site
Tenraile arm
Index, 0-5 m
Index, 5-10 m
Index, 10-15 in
Chlorophyll
Methanol
6.6
5.9
3.5
3.1
a - ug/1
Acetone
5.6
5.4
3.35
2.4
Ratio M/A
1.18
1.10
1.05
1.28
Table 3.  Comparison of extraction by acetone  and methanol  for Lake
          Dillon samples on 1 March 1982  (samples were  steeped,  not
          ground).
Primary Production

     Primary production (photosynthetic carbon  fixation) was  determined

by the C-14 uptake method at the index station  on  each  routine  lake

sampling date.  The method was carried out essentially  as  described by

Lewis (1974).  Although the method is operationally  somewhat  complex,

the principle is very simple.  Water is removed from a  specific depth,

inoculated with C-14, and returned for a  few hours to the  depth from

which it came.  The sample is then retrieved, passed through  a  filter,

and the filter is washed to remove any soluble  C-14  label.  The

remaining C-14 label on the filter is that which has been  taken up

-------
photosynthetically  by  the  algae  during the period of incubation.  From




the total amount  of inorganic  carbon available in relation to the




amount of C-14  added,  one  can  calculate the total carbon fixation on




the basis of  the  C-14  fixation.




     Depths of  sampling  and  incubation for the Dillon work were as




follows:  0,  0.5, 1,  2,  5, and 10 m.  Greater depths are of little




interest from the viewpoint  of carbon fixation because there is




insufficient  light  to  support  much photosynthesis.




     At each  depth, two  125-ml transparent glass bottles were




suspended ("light bottles").   For each depth one bottle  was filled that




was completely  darkened  on the outside ("dark bottle").   Since




temperatures  were nearly uniform over the mixed layer, all of the dark




bottles were  incubated together  at a depth of 5 in.   The  dark bottles




served as controls  for nonphotosynthetic  inorganic  carbon uptake,




although this was typically  snail.  All three bottles were inoculated




with 1 ml of  C-14 in  the form  of sodium bicarbonate solution.  The




activity of the inoculum was 3 uCi/ml.  The samples were suspended in




situ for a timed  period  of 2.5-4 hours, typically between 10 a.m. and




2 p.m.  On the  day  of  the  incubation,  solar radiation reaching lake




surface was recorded continuously by a Belfort pyrheliometer so that




the amount of carbon fixation  could  be related to the amount of light.




     At the end of  the incubation, the bottles were removed from the




lake and placed in  a dark  box  so that  no  further photosynthesis could




occur.  The samples were then  transported to the Snake River Wastewater




Treatment Plant for  filtration.   As  each  bottle was removed frca the




dark box, it  was shaken  and a  100-ml subsample was  withdrawn for

-------
                                                                      38





 filtration.   This  subsample was  placed  onto  a  filter  tower over a




 Millipore  HA  filter  (effective pore  size  0.45  urn).  The  sample was




 filtered under gentle vacuvra and  the  filter  was  washed  two times with




 distilled  water.   The filter was  then placed in  a  liquid scintillation




 vial containing 12 ml of  scintillation  cocktail  (Aquasol II)  and




 capped.  The  filtered water, still containing  small amounts  of C-14,




 was returned  to Boulder for proper disposal  through a regulated




 radioactive waste  disposal system.   The scintillation vials  containing




 the labelled  algae were then counted  on a Beckman LS3133T scintillation




 counter.   Counting efficiency was determined by  addition of  standard




 amounts of C-14-labelled  toluene  to  selected samples.  Counting




 efficiencies  were  very stable in  the vicinity  of 80 to 90%.




     The total activity in each scintillation  vial was computed from




 the observed  activity and the machine efficiency after correction for




machine blank (typically  a very small correction in the  vicinity of  50




cpm).  The available inorganic carbon in the sample was  determined from




 the pH and alkalinity of  the sample and the  sample temperature




according  to  the tables of Saunders, Trama,  and  Bachmann (1962).




Carbon fixation in each bottle was computed  from the  following




information:   (1)   the activity of the sample,  (2)  the volume  of water




filtered,  (3)  the  available carbon in the sample,  (4)  the isotope




discrimination factor (1.06),  and (5) the amount cf C-14 added to the




sample.  The  dark  bottle  fixation for a given  depth was  then  subtracted




from the light bottle fixation for the  same  depth  to  yield  the observed




photosynthetic carbon fixation for the  incubation  period

-------
(:agC/m-V incubation  period) .   This  is  considered to be a good estimate


of gross photosynthesis  (Harris  1978).


     From the gross  photosynthesis at  each depth,  a vertical profile


could be made of  photosynthesis  at the  index station over the


incubation period.   A  profile of this  type was drawn on a piece of


graph paper for each one  of  the  incubation dates.   The amount of


fixation over all depths  was  obtained  by integration of the area under


this volume-specific fixation curve.   This yielded an estimate of gross


fixation per unit area (mg  carbon/m"/incubatlon period) .   Because the


incubation periods  were  of  slightly different lengths and the sunlight


conditions were different from one day  to the next, fixation per ur.it


area was converted  to  fixation per day  and to fixation per unit


sunlight as follows.  The amount of sunlight striking the surface of


the lake over the incubation  period was obtained from the pyrneliometer


sunlight trace by planimetric integration.  The fixation  per unit area


for the incubation  period was divided  by the amount of incident


sunlight during the  Incubation period  to yield the fixation rate/unit


area/unit of sunlight  (ing carbon/m-/langley).  The sunlight striking


the surface of the  lake  over  the entire day was then obtained by


planimetric integration  from  the pyrheliotneter charts.  The total


amount of sunlight  for the  entire  day was multiplied oy the fixation


per unit of light to obtain  the  total  estimated fixation  per unit area

                         i-\
for the day (mg carbon/m~,'day) .   This  last calculation assumes that,


on a given day, the  photosynthesis below a unit area is proportionate


to the amount of  sunlight striking the  surface.  This is  not -irecisel'-'


correct (Tailing  1971), but is  sufficiently true for present purpc~es .

-------
      Because photosynthesis has been measured by the C-14 method in a




 wide variety of lakes of different trophic status, the C-14




 measurements provide a method for comparing the photosynthetic activity




 of Lake Dillon with that of many other lakes.









 Temperature, Conductance, and Transparency




      Temperature profiles were measured with a YSI thermistor and




 submersible probe.  Conductance was determined in the laboratory with a




 Labline Model MC-1 meter standardized against KC1.  All conductances




 were temperature corrected to 25°C.




      Transparency of the lake was  determined by secchi disk, and




 sometimes also by submersible photometer (Whitney-Montedoro with




 submersible selenium photocell and opal filter).   The photometer




 readings were taken from the surface  to the limit of detection in




 increments  of 60 cm.   The secchi  disk is a standard white plate, 15 cm




 in diameter,  which is lowered to  the  limit of visibility.  In Lake




 Dillon,  the  secchi depth was  on  the average equal to the  depth of 19%




 surface  irradiance as measured by  light meter (standard deviation,




 5.7%).









 Discharge Measurements




     Discharge was  estimated  at  stream sampling stations  at the time of




 sampling.   In addition,  continuous  discharge measurements were obtained




 from four U.S.G.S.  gauging  stations.   The U.S.G.S.  gauging station




measurements  serve  two purposes:   (1)  they allow  us to check cur own




methods of measuring  discharge, and (2)  chey are  the main basis  for

-------
estimating total amount  of  water  entering the lake from each of the




three major rivers.   The  gauging  station locations are shown in




Figure 3.




     The Snake River  gauge  is  well  above the mouth of the Snake River.




In order to estimate  the  amount of  water entering the lake from the




Snake River, we added the values  from the U.S.G.S. gauges on the Snake




River and on Keystone Gulch and then  made a  small correction for the




remaining ungauged  portion  of  this  drainage.  Similar corrections were




made for the other  gauge  readings to  account for small areas below




gauges (Figure 3).  Also  for the  Blue River  we added  the sewage




treatment plant discharge to the  gauge reading.   The  sewage  treatment




plant discharge was obtained from plant operators on  sampling days.




     We made discharge measurements when we  took stream samples in




ungauged reaches of stream  for the  routine stream series and for the




stream survey series. Some of the  stream sampling sites were




sufficiently close  to culverts or bridges that the discharge could be




estimated by the use  of  standard  tables (Chow 1959,  BPR 1965).   This




was done for about  eight  sites in 1982 and slightly  more in  1981.  For




most of the stream  sites, standard  tables could  not  be used  because of




the irregularity of the  channel.   Consequently,  estimates were  nada by




the method of cross-sectional  current profiles.   The  depth profile of




the stream channel was mapped  at  a  location  where the bed was likely to




be stable.  The cross-sectional area  of the  stream as shown  by  this map




was divided into several  segments (three to  eight, depending en the




width and irregularity of the  stream).   Within each  segment, a  current




meter neasurement was  taken with  a Gurley Pygmy  current neter at a

-------
 depth  equal  to  60% of  the distance from the surface to the stream bed.

 This is  the  depth at which the velocity is considered to be about

 average  for  a segment.   The discharge of each segment was then computed

 on  the basis of its area and its  velocity of flow.  The discharges for

 the different segments  were added to  produce the estimate of total

 discharge  for the site.   The current  meter method is compared with the

 U.S.G.S. data in Table  4.   The deviations fall within the daily range

 of  fluctuations for a given site  and  are thus as close as could be

 expected.

Station
Tenmile Cr.
Blue River
Snake River
Number of
Determinations
20
21
21
U.S.G.S.
Mean, cfs
223
173
145
Current Meter
Mean, cfs
216
167
131
Diff .
%
-3
-4
-11
Table 4.  Comparison of discharge measurements  made  by U.S.G.S. and
          near  the  same location by  the current meter  method used in
          the field sampling program (1982).
     After a range of discharges had been  obtained  at  a  given site,  a

depth-discharge relationship was established  in  the form of a rating

curve.  After the rating curve was available,  discharge  estimates were

sometimes based on the curve rather than current  meter measurements.

     The Denver Water Department provided  daily  estimates of water

leaving the lake through the Roberts Tunnel and  by  the Blue River, and

daily estimates of water entering the lake.

-------
Precipitation Sampling




     Atmospheric  transport of nutrients  was  estimated  at  the  Snake




River Wastewater  Treatment Plant  site  by means  of  a  large bulk




precipitation chemistry sampler mounted  on a  tower 3 m above  the




ground.  The design of the sampler  is  given  in  detail  by  Lewis and




Grant (1978; Figure 4).  The opening of  the  sampler  is 0.20 m2,  which




is sufficiently large to supply measurable amounts of  phosphorus and




nitrogen over weekly or biweekly  collection  periods  at any tine  of  the




year, even when there is no wet precipitation.   The  collecting surface




consists of an inverted plexiglas pyramid which  funnels materials into




a narrow opening  that leads to an insulated  wooden box.   Wet




precipitation flows through the opening  into  a  plastic receiving




funnel.  The receiving funnel is  covered with a  fine nvlon screen so




that insects reaching the funnel  can crawl from  it.  The  funnel  is




connected to a surgical rubber hose that is  looped in  order to form a




vapor barrier, and the surgical rubber hose  is  connected  to a hard




glass receiving vessel with a collapsible plastic  overflow container.




The box housing the containers is heated by  means  of thermostatically




controlled light  bulbs.  The heat is sufficient  to melt snow  that




strikes the surface of the collector during  the  winter so that these




contributions are not blown from  the sampler.

-------
         CROSS SECTION
                                                      TOP  VIEW
                                                             TOP EO«C & PLYWOOD
                                -l S7YRCFQAM
                                 INSULATION
                               .' »~3/8"PLT'*000
 Figure  4.   Design of precipitation chemistry sampler  (from Lewis and
            Grant  1978).
      The  samples  were collected at weekly to biweekly  intervals

depending  on  the  schedule of other sampling programs.   On each

occasion,  the  collection vessels were removed from  the box and mixed in

a common container.   An. aliquot of 100 ml was removed  for determination

of pH and  alkalinity.   The surgical rubber hose was  then hooked to a

rinse container and  500 nil of deionized water was poured over the

collecting  surface,  thus rinsing the collecting surface, collecting

funnel, and tubing.   After the volume of the rinse  and the volume of

wet precipitation were  noted, the two were mixed  together.  The water

was then analyzed according to the protocol for lake samples.  The data

were corrected for the  amount of deionized water  added and for the

removal of the iOO-ntl aliquot for pH and alkalinity  determination.

-------
                                                                       -3





Concentrations were  converted  to  loading  rates on the basis of the area




of the collecting  surface  and  the elapsed tine since the last sample.









Special Studies




     The protocol  for  each of  the special studies is described along




with the results in  later  parts  of this  report .   All chemical analyses




and sampling were  carried  out  by  the  methods  described above unless




otherwise  indicated.









Quality Assurance




     The quality assurance program is intended to provide quantitative




information on the reliability and accuracy of analyses  for major




variables  in the Lake  Dillon  Study.   Three  types  of  information are




included:  duplicates,  spikes,  and EPA standards.




     (1)   Duplicates.   Approximately  10%  of the  analyses were




duplicated.  Table 5 summarizes  the resulting information.   The degree




of variability is  most  easily  evaluated  from  the  ratio of standard




deviation  to mean  (coefficient  of variation,  expressed as percent),




which is shown in  the  last column of  Table  5.  The degree of variation




for all variables  is acceptable,  and  for  most it  is  exceedingly low and




thus warrants no discussion.   Estimation  of tctal particulates stands




out as having the  highest  relative variation. We attribute this




largely to irregularities  in the  distribution of  particles  within the




stream (stream duplicates  are  based on two  samples taken saquer.tiallv)




rather than to variation in drying or weighing.   This  conclusion is




supported by a difference  in replicate variability for ^articulates in

-------
Number of
Variable Duplicates Mean
Oxygen
Conductance
pH
Alkalinity
Chlorophyll a
Total
Particulates
N02-N
N03-N
NH/.-N
Total Sol. N
?G4-P
Total Sol. P
Part. P
80
188
110
115
101
212
235
308
283
143
240
303
224
7
164
7
36
7
3
6
151
91
618
4
7
4
.59
.0
.43
.8
.29
.15
.76
.61
.63
.0
.56
.45
.95
mg/1
unho / eta

mg/1
ug/1
ng/1
ug/1
ug/1
ug/1
ug/1
ug/1
ug/1
ug/1
Standard
Deviation
0
4
0
0
0
0
0
5
5
73
0
0
0
.095
.4
.043
.76
.52
.56
.37
.91
.59
.0
.40
.56
.58
Coefricient of"
Variation, %
1.
2.
0.
2.
7.
17.
5.
3.
6.
12.
8.
7.
11.
3
7
6
1
1
8
5
9
1
6
8
5
6
  Ratio of standard deviation to mean, times 100.




Table 5.   Summary of statistics on duplicate analyses

-------
stream samples and lake  samples.   Lake  samples,  which come from the




same integrator, and which  are more  likely  to  have an evenly dispersed




particulate load of  fine  grain,  show lower  variability,  whereas stream




samples show higher  variability  than the  average.   Even  so,




determination of the particulate  load  of  streamj will not be affected




in any important way by  this  variability  because of  the  large number of




samples taken over time  at  each  site.   Replicate variability in




particulate load is  reflected to  some  degree  in  the  particulate




phosphorus analyses, as  expected.




     PO^-P shows slightly higher  relative variation  of duplicates




than most other constituents. This  is  explained by  the  extremely low




levels of PO/^-P characteristic of  the  entire  system  (note that  the




average PO^-P concentration for  all  duplicates  is  just slightly over




3 ug/1).  Since the  relative  variation  of a determination is to some




degree affected by the proximity  of  the analysis to  the  analytical




limit for the test,  the  slightly  higher relative variation for  FO^-?




is not unexpected.




     Total soluble nitrogen also  shows  slightly  higher variability than




most other variables.  We attribute  this  mostly  to the pK sensitivity




of the test, but do  not  consider  it  important  in the  interpretation of




results.




     (2) Spikes.  Table  6 summarizes  the  percentage  recovery of spikes




added to samples throughout  the course  of the  study.   The recovery is




typically quite close to  100%.  Only  the  total  soluble phosphorus and




nitrate recovery warrant  discussion.   For total  soluble  phosphorus, two




types of spikes were used:  dibasic  inorganic  chcsphorus  i, !-'.-> H.?0^)

-------
 and ATP.   The recovery of K2HPC>4 was  lower than for ATP, which was




 characteristically recovered near the 100% level.  Since K2HP04 is




 not expected in nature,  whereas  organic  phosphorus compounds more




 similar to ATP are expected  in nature, we believe that the percentage




 recovery  for ATP is more representative  of the  recovery of natural




 organic phosphorus by  the test.   In any  event,  the difference is small




 and thus  does not warrant correction.
Variable
P04-P
Total Sol. P
N02~N
N03-N
NH4-N
Total Sol. N
Number of Spikes
213
186
187
192
196
113
Mean % Recoverv
91.44
87.13
104.7
91.34
96.7
94.29
Table 6.   Summary of information on  spikes  for  quality assurance.









     For nitrate, the efficiency of  recovery  is about 90%.   Once again,




we feel that  this percentage  slightly underestimates  actual recovery of




nitrates in unspiked samples.  Addition of  the  spike  could  reduce the




efficiency of the reduction column by increasing the  load on it and in




this way leads to a small loss in recovery  that is  not likely to be




characteristic of real samples.  Thus we do not believe that the data




should be corrected.  'The effect of  any correction  would be very




minor, however.




     (3) EPA Standards.  Analysis of EPA standards  showed good




agreement of observed and expected values (Table 7).

-------

NH4-N,
N03-N,
P04-P,
pH

ug/1
ug/1
ug/1

EPA
19
31
3.1
7.4
Detected
19.3
33.4
3.0
7.7
Value




Fable  7.   Data on EPA standards (raw standards were diluted to
          appropriate range).

-------
                                                                       50





         Physical Variables  and Major  Ion  Chemistry of the Lake









Geology




     The geology of  the Lake Dillon region  need  be covered here only




briefly, since it has been  well documented  in  connection with the




construction of the  Roberts Tunnal and  the  exploitation of mineral




resources in the watershed.  For details, the  reader is referred to the




publication of WahlStrom and Hornback (1962),  which contains references




to earlier studies.




     Lake Dillon lies on a  bed of stream  and glacial deposits of




Quaternary age, especially  in the old riverbed,  and on sedimentary




rocks of Mesozoic age.  Except near the valley bottoms, the Snake River




and Straight Creek drainages and small adjoining portions of the Blue




River drainage are composed of igneous and  inetamorphic rock of




Precambrian age.  Nearer to the riverbed  at lower elevations the Snake




River drainage contains stream and glacial  deposits of Quaternary age.




In the upper Snake and Peru Creek there is  an  exposure of Tertiary




quartzes (nonzonite  and monzonite porphyry).




     There is a sharp break in the geology  along a line that runs north




and south to the east of the Blue River.  This line represents the




Williams Range Thrust Fault.  To the west of this fault,  we find




Mesozoic sedimentary rocks and roughly equal exposures of Tertiary




quartzes and feisite porphyry.  Along the valley bottoms  there are also




extensive deposits  of Quaternary age.

-------
Morphoaietry




     A bathymetric map  of  the  lake  is  shewn in Figure 5.  It is useful




to compare the central  portion of  the  lake to the lake arms because




differences between  the  aras and  the main body of the lake will be of




soiae interest throughout  the course of  this report.   Dillon Bay is




relatively deep and  opens  broadly  onto  the main body of the lake.  The




Snake River arm is narrowest but  has a  greater mean  depth than the




other arras.  The  Slue River arm is  shallow in its upstream half,  where




sediments are exposed at  times of  the  year when drawdown is most




severe.  The Blue River  arm opens  broadly onto the main body of the




lake, and this promotes  exchange  with  the center of  the lake.   The




Tenmile Creek arm has the  greatest  extent of  shallow water.  Shallow




water is found in the recessed area to  the west and  in the upstream end




of the arm.  Considerable  areas of  sediment surface  are exposed during




periods of extensive drawdown. Furthermore,  the circulation between




the main body of  the lake  and  the  Tenmile Creek arm  could be restricted




by the presence of islands near the mouth of  the arm.  Clearly the




greatest opportunity for biological or  chemical divergence from the




main body of the  lake is in this  arm.




     Figure 6 shows  the  depth/volume relationship, Figure 7 shows the




hypsographic curve,  and  Table  8 summarizes morphometric statistics




related to water  level.  Although  a small amount of  storage above




spillway level is possible, spillway level is for practical purposes




the maximum lake level,  so other  depths  are expressed in relation to




spillway level.  The Roberts Tunnel diverts water frcm the lake through




an intake pipe 52 m  below  the  spillway.   The  outlet  to the Blue River

-------
 Ti
 H-
 cro
 t:
 3"


"I
 ro
H-
o

a
(u
•d

o
MI
o
                                                                         Dillon Bay
Contour  interval,  40ft
                                     Tenmile

                                 Creek

                              Arm

-------
  OJ
  UJ
  Q
        0
      S80
         0
VOLUME  , meters^   x!0°

 75      150    225     300
                               0

                              10

                              20

                              30

                              40

                              50

                              60
             j	!	1
 60
120
ISO
240
             VOLUME ,  acre  feet   x 1000
                          CO
                          k.
                          CD

                          "a)
                          C
                          CL
                          UJ
                          Q
Figure 6.  Depth-volume curve for Lake Dillon.

-------
      0
<3J
X
S-
Q.
U
Q
    ISO -
       0
  AREA  , hectares  x 1000

3       6       9       12
    10          20          30

   AREA  ,  acres  x!00
                                                54
15
                                    0


                                    10


                                    20  I


                                    30  £


                                    40  h
                                         CL
                                         UJ
                                    50  Q


                                    60
 Figure 7. Hypsographic curve for Lake Dillon.

-------
is slightly below  this, at  a  depth  of  57 m  when  the  water  is  at

spillway level.  The original maximum  depth of the lake  was  68 m,  vhich

vould allow for  just over 1%  dead storage.   The  current  niaxinuzi  depth

is probably somewhat less than  this.   The greatest maximum depth

Treasured during  the course  of our study was 61 a but,  because of the

presence of the  old river channel,  some deeper spots are probably  still

present.
Sea Level Preference

Spillway
Maximum storage (flood)
Roberts Tunnel
Blue River outlet
Original maximum depth
feet
9017
9025
8846
8829
8795
meters
2748
2751
2696
2691
2681
Spillway
feet
0
+8
-171
-188
-222
Reference
rieters
0
+2 . 4
-52
-57
-68
Table 8.  Morphometric  statistics  for Lake Dillon.   Source:  engineering
          documents provided  by  the Denver Water Department.
     Table 9 summarizes the lake volume, area,  length  of  shoreline,  and

mean depth when the water is  at spillway leval.

Lake Volume
Lake Area
Length of shore
Mean depth
Watershed Area
English
262,000 acre-feet
3300 acres
24.5 miles
79.3 feet
212,900 acres
Metric
0.323 km3
13.35 km2
39.4 km
24.1 m
85,160 ha
Table 9.  Morphometric statistics on Lake Dillon,  assuming  water  at
          spillway level.  Source:  Denver Water Department.

-------
                                                                       56






 Hydrology




      We  summarize here only the  changes  in lake level and total flow




 into  and  out  of  the  lake.   A more  detailed consideration of the




 hydrology of  individual watersheds will  be taken up  later in connection




 with  the  analysis of nutrient  income  to  the lake.




      Figure 8 shows  the total  calculated inflow to the lake for the




 study period  as  estimated  by the Denver  Water  Department.  There are




 differences in the actual  inflow and  the calculated  inflow because the




 calculated inflow is estimated as  the  sura of the outflow (which is




 measured) and the change in water volume (which is estimated from the




 depth/volume  relationship).  Errors in equating calculated inflow with




 actual inflow are not  of concern here but  will be discussed in another




 section.   It  will also  be  shown  that groundwater inflow is at most




 5-10% of  the  total inflow.




      The  calculated  inflow to  the lake for 1982 was  about twice the




 amount for 1981.   In both  years  the peak inflow occurred in the month




 of June,  but  the  distribution  of inflow  around the peak was not quite




 the same  in the two  years.   In 1981 the  total  inflow was quite




 symmetrical around the  month of  June.  In  1982,  high flow extended well




 into  July.  In both  water  years  the months November  through March




 showed a  stable minimum  inflow of about  2,000  to 4,000 acre-feet per




month.




     Figure 9 shows  the  total  calculated inflow since 1963 (calendar




years).   The purpose of  this figure is to  put  the two years of the




present study into perspective hydrologically.   For  the period of




record, 1981 was  well below the  average  (only  1963 was  lower)  and 1982

-------
o 75 r
o
      50
   o>
      25
       0
                                                                  90
                                                                      a
                                                                       X
                                                               60
                                                                    O
                                                                    £
                                                                    \
                                                                   rO
                                                               30
           0    N   D
                      J    F   M    A
                               TIME
J    J   A    S
Figure  8.   Calculated inflow as obtained  by  the Denver Water Departner.:
           for Lake Dillon over the  study period or. a oioncnly basis
           (water years).

-------
  o
  o
  o
      300
  ©
  V.

  O

  O
   -  200
  u.

  5   100
  a
  r-
0
                                                      53
    I	I  I  l   l  j_ |   j  !   1  !  }   !  1
                                        t  I   >  !   !
      64  66  68  70  72  74  76  78   80   82


                         TIME
Figure 9.  Calculated inflow since 1963 for Lake Dillon (Denver Water

         Department).

-------
                                                                      59




was somewhat above the average  (13 of  the  19  other years are below




1982).




     Figure 10 shows  the  total  outflow per month for 1981 and 1962.




Water leaving the lake follows  some  combination  of three pathways:  (1)




the Roberts Tunnel, (2) the  Blue River outlst,  and (3)  the spillway.




Since the Roberts Tunnel  and  Blue River  subsurface outlet remove water




from points near the  bottom  of  the lake, they have not  been separated




in Figure 10, nor need they  be  treated separately  in any of the




analyses of this report.  Spillway losses  must  be  separated from other




losses, however.  In  1981 no  water left  the lake over  the spillway, but




in 1982 water left the lake  over the spillway between  9 July and i




November-  Since water going  over the  spillway  removes  water from the




epilimnion of the lake, it is important  to recognize the timing  and




extent of spillway flow.  The Denver Water Department  does not keep




records of the apportionment  of outflowing water between the spillway




and the outlet pipe,  although it is  known  that  spillway flow typically




dominates when the two occur  together.   Because  the surface and




deepwater nitrate and conductance differ considerably  at the time of




spillway flow, we can estimate  the relative contribution of spillway




outflow to total outflow  as  follows: C,  =  XC_ +  (1-X)  Cs where




Cc is the composite concentration of NO-^-N or conductance in the




Blue River below the  dam, Ct  is the  concentration  at the top and C^




at the bottom of the water column, and X is the  proportion of water




going over the spillway.  The values vary  some,  buc the estimate shows




that about SO" of the water reaching the Blue River between 9 July and




-4 November 1982 came  over the spillway.  Using  this percentage  as a

-------
O

§
   c
   O

   £
   0)
      25
   O
   -J
                                                                   60
                                                                 30
                                                                      O

                                                                      x
                                                              20
                                                                   c
                                                                   O

                                                                   £
                                                                   \
                                                                  KJ

                                                              10  2
                               _]	I	J	I	I	I	!
           ONDJFMAMJ   JAS

                                  TIME
Figure  10.  Outflow from Dillon  over the study period on a monthly

           basis.  (Denver  Water Department).

-------
                                                                      61




constant, we have computed  the  spillway  less  ca  a  weekly basis  over the




period of spillway flov (Table  10).




     Figure 11 shows the change in  lake  level  during  1981 and 1982.




The lake is typically drawn down  in  the  winter months  to neet demand




and legal requirements for  minimum  flow  in  the  Slue River daring the




period of low winter straamflow.  The  ~2ta  of  drawdown  in the two  years




was essentially the same, but  the eflecc  on lake  level  was mor~ drastic




in the water year 1982 because  the  snowpack of water  year 1931  was not




enough to fill the lake to  spillway  level.




     The combined effect of lake  volume  changes  and changes  in  total




inflow are illustrated in Figure  12  in terms  of  the percent  of  volume




added to the lake per month for each month  of  the  study.  Dilution is




most significant in June, less  so in May  and  July, and  minimal  in  other




months.




     Table 11 summarises the flow of water  into  and out of the  lake on




an annual basis during 1981 and  1982.  Both calendar  year and water




year figures are shown, but the  differenca  between these two




conventions is not very great.









Temperature




     Like most high temperate lakes  cf moderat2  to great depth, Lake




Dillon is dimictic.  Lakes  of this classification  show  two annual




episodes of complete mixing per year,  one of which is in the  fall  and




the other in the spring.  The year can thus  be divided  into  four




intervals:   (1) winter ice cover, (2)  spring mixing (spring  overturn),




(3) summer  stratification, and  (£) fall mixing (fall  overturn).

-------
$2
Hi
H-
2: 9020
o
>
!j 9000
UJ
LJ
o
2 8980
£T
CO
U
•\^ Q Q C Ol
62
. 3983
»*»«***««»»*
3aflsBM5B«aaai
C°oco 198!
°°ooco
C0000 Q__
°oco o^°°ooo0
«9 O0oooooooo°o00o» 00oCo00oo
"9
••..
* * • _ *
• •• •», *
* » *
1932 "•.........•
-

1 1 ! 1 i «!!!<- S
< nwn.iFMAM.i.iA<;
Figure 11.  Change of lake level with time over the study period.  Data
            points are midweek values (Source:  Denver Water
            Department).

-------
        30
     c
     o
     e
     *x
     
-------
                                                                       64
Acre feet per week
Week
9 Jul -
16 Jul
23 Jul
30 Jul
6 Aug -
13 Aug
20 Aug
27 Aug
3 Sept
10 Sept
17 Sept
24 Sept
1 Oct -
8 Oct -
15 Oct •
22 Oct •
29 Oct •
• 15 Jul
- 22 Jul
- 29 Jul
- 5 Aug
12 Aug
-19 Aug
- 26 Aug
- 2 Sept
- 9 Sept
- 16 Sept
- 23 Sept
- 30 Sept
7 Oct
14 Oct
- 21 Oct
- 28 Oct
- 4 Mov
Total
Outflow
4478
2458
5024
7228
3162
5396
5322
8942
3008
319*
2830
2476
2069
1858
1344
1774
1744
Outflow
from
Bottom*
3376
3951
3795
5546
679
1076
1064
1788
602
639
566
495
415
372
369
355
538
Outflow
over Spillway Outflow as
Spillway %/wk of 0-5 m layer
1102
2507
1229
1682
2483
4320
4258
7154
2406
2555
2264
1981
1654
1486
1475
1419
1206
2.5
5.6
2.8
3.8
5.6
9.7
9.5
16.0
5.4
5.7
5.1
4.4
3.7
3.5
3.3
3.2
2.7
Table 10.  Estimated partitioning of outflow over the period of spillway
           loss (1982).

-------

1981
Calendar Year
Water Year
1982
Calendar Year
Water Year
Calculated
Acre-feet

105,302
106,193

217,386
210,255
Inflow
te3

0.130
0.131

0.258
0.259
Measured
Acre-f ae

145,912
152,565

119.323
131,989
Outflow
•5
r j — '-'

0.180
0.138

0.143
0.163
Table 11.  Summary of water flew into and out of the lake, based  on
           daily statistics supplied by the Denver Water Depart-ent
           (cumulated monthly and daily statistics fron the  same  source
           differ slightly).

-------
                                                                       66





      In  both  1981  and  1982,  ice  cover  formed  on Lake Dillon in late




December  or early  January  (Figures  13,  14).   In both winters the




maximum  ice thickness  reached  about  0.5 in  in  midlake.  In 1981 the ice




was exceptionally  transparent  (i.e., without  bubbles or snow layers),




and,  because  of  the  small  amount  of  winter  snowfall, vas not covered




with  significant amounts of  snow  during the first  two months of




ice cover.  Because  an ice cover  of  this type transmits light almost as




well  as water  (Ragotzkie 1978), unusually large amounts of  light




reached the water  column during the  winter  of 1981-   By April of 1981




light penetration  was blocked more effectively because of some snow




accumulation  and thaw cycles at the  surface,  making  an opaque granular




layer at  the  top.  In 1982 there  was snow cover from the very




beginning, and it  lasted through  the entire winter.   Snow on the lake




surface did not accumulate to great  depth because  of wind exposure,  but




there were few open  areas  in the  winter of  1982.   Since snow has an




absorption coefficient of about 0.24 cm"1 (Sagotzkia 1978),  10 cm




of snow, which is  approximately the  amount  on the  lake in 1982,  would




remove about 90% of  the light.




     In both 1981  and 1982, the ice  cover broke up in early  May.




Temperatures under the ice in both years were between 1 and  3°C over




most of the water  column, as expected.  Temperatures closer  to the ice




were slightly lower  than near the bottom.   This  is a stable  condition




because the density  of water decreases  from 4°C to the freezing point.




     Warning of the  surface occurred slowly in both  years after the  ice




cover disappeared,  but especially so in 1981.   As  long as a  water




column is below 4°C,  warming at the  surface actually creates

-------
TI
f-*-
00
a.
IT)
13
fu
oo
OJ
a
rt
m
F3
10
ft
f.)
n
H-
O
            0 r-
                            ice
           10
     CD
     e
       **

     X
     r	

     Q-
     liJ
          30  -
          40  -
          50  -
                  J     F     M     A     M      J      J
                                                                                  0
N
f)
                                                      IS) 81

-------
OQ
C
M
0
n>
 I
a.
H-
tu
(a
B
o
f-l
(D

0
rt
C
rt)

H-
PJ
rt
H-
O
OD
10
 0)
^_
 
-------
instability because of the  increase  in  density of water as it is warded




from the freezing point to  4°C.  Even after  the water colunm warms to




5°C there is little resistance  to  mixing  because of  the very slight




change in density of water  in the  vicinity of 4°C.   Thus there is no




tendency to stable layering  until  the surface water  exceeds 5 to 5°C.




Until then, heat is efficiently  distributed  through  the entire water




column by the wind, and thus the water  warms only slowly at the




surface.




     In Lake Dillon some  beginning indications of layering could be




seen as the lake exceeded 5°C in the last half of May both years.




Although stratification was  very weak at  first, complete mixing of the




water column ended near the  last of  May in both years.   June 1 thus




marks the beginning of the  stratification season.




     By the last half of  June in both years  thermal  layering had become




very pronounced.  The upper  water  column  reached 12°C by the last half




of June in both years, and  the  thickness  of  the upper mixed layer at




this time was about 5 in (we  equate the  uprier nixed  layer with the




epilimnion, although the  epilinmion  is  sometimes defined somewhat




differently:  Hutchinson  1957).  The June thermocline was relatively




thick, extending from 5 m down  to  about 20 m.   la both  1931 and 1982




the upper water column continued to  take  up  he .at well after




stratification was first  established.   In 1981 the  lake reached naximun




heat content in the last week of July,  when  the temperature of the




epilimnion rose to just over 16°C.   In  1982  the warring occurred




slightly more slowly, producing maximum epili^netic  temperatura in the




first week of August, but the maximum epilimnetic temperature was

-------
                                                                       70


 essentially  the  same  as  in  1981.   In  both  years  the thickness of the

 mixed layer  remained  5-10 m for the first  half of  the stratification

 season,  i.e.,  until the  middle of  August,  but  the  average thickness was

 somewhat greater  in 1981 than in 1982 (Figure  15).  In both years

 cooling  of the epilimnion began in the last  half of August.  Ceding of

 the upper water  column reduced che stability of  stratification

 sufficiently to  allow gradual erosion of  the therEOdine by late August

 or early September in both  years.  In 1981 the thickness of the

 epilimnion had reached 15 in by the middle  of September and 25m by the

 middle of October.  In 1982, the epilimnion  had  thickened to 10 EL by

 the middle of September  and 25 m by the middle of  October.

      The stratification  season ended  with  complete mixing during the

 first week of November in both 1981 and 1982.  By  the tine complete

 mixing occurred,  the  lake had already cooled to  a  temperature of about

 6°C.  Isothermal  cooling accompanied  by complete mixing of the water

 column then  proceeded through November and most  of December until the

 establishment  of  ice  cover  at the  end of December.

      Table 12 gives a synopsis of  the layering and mixing events in

 Lake Dillon  in 1981 and  1982.  There  were  no significant differences in

 the timing of  major events  in the  two years.

	Season	Duration	

     Ice cover                        January; February, March, April
     Spring  mixing                     May
     Summer  stratification             June to  first week of November
        Fixed  epilimnetic thickness    June,  July,  August
        Increasing epilimnetic         September, October
           thickness
     Fall mixing                       Last three weeks of November,
	December	
Table 12.Seasons of layering and mixing  in Lake  Dillon.

-------
0.
LJ
  0


 10


20


30


40


50


60
                      r     i     i
                          198!
                                                                 I

          JFMAMJ    JASONO
Figure 15.  Thickness of the mixed layer during the ice-free seasons of
          1981 and 1982.

-------
Transparency




     Figure 16  shows  the  transparency  of  the  lake  over the two-year




study period both  in  terms of  the  secchi  depth  and  the depth of 1%




light.  The depth  of  1% light  was  approximated  from the secchi  depth as




follows:




                          nt - -ln(0.19)/zs




                          z.01 = -ln(0.01)/r,t




where nt is the extinction coefficient of photosynthetically active




radiation  (350-70Cnm, ra"1), zg is  secchi  depth  (m),  Z.QI is




the depth  at which 1% of  the surface photosynthetically active




radiation  (PAR) is found  (m).  The approximation assumes that the




secchi depth always falls at 19% of the surface irradiance-  The basis




of this assumption is a series of  comparisons between  light meter




readings and secchi depth readings as reported  in the  methods section.




The purpose of  the conversion  from secchi depth readings to depth  of 1%




light is to show the approximate depth at which photosynthesis  could




occur rapidly enough to match  the  respiration of algal cells (Tailing




1971).




     Figure 16 shows that significant photosynthesis in Lake Dillon




was always limited to the top  15m and was usually  limited to the  top




10 m of the water column.  The main features  of the  seasonal change in




transparency could be summarized as follows:  (1)  transparency  was




moderate to high under ice, (2) transparency  was consistently low  ever




the entire first half of  summer stratification, and  (3) transparency




was consistently high during spring and fall  mixing  and over the  last




nalr of summer stratification.  The readings  under  ice may be

-------
LsJ
a
     5 -
    10
    15
    20
r o
^ w
—r—
                          mi:
                                     stratified
                                                mix
           _i	L
                                           secchi  depth
                                            !%  light
JFM
                       AMJJA
                                 1981
                                                  0ND
5   5
1  10
&  '5
Q
   20
       L
               ice
                          mix
                            stratified
                                         mix
                                           secchi  aepth
                                            )%  liqht
             F    M    A   M
                                J    J
                                 1982
                                A    S   0    N    D
rigure  16.  Secchi depch and depth of 1% light ir. Lake L'illcn  -, ir.c-:-: 5-3--',--

-------
                                                                       74





underestimates, especially  in  1982 when  there  was  significant snow




cover.  Because the readings were taken  through a  hole  in the ice,




light could not reach the secchi disk as well  as it  would under




ice-free conditions.




     Four factors can affect the transparency  of the lakewater:  (1)




the extinction of light by  pure water, (2)  the extinction of  light by




organic materials dissolved in the water, (3)  the  extinction  of light




by chlorophyll, and (4) the extinction of light by organic and




inorganic particulate tnaterials.  The extinction coefficient  (Tt) can




thus be broken into components:




                   i± = nw  + nd + nc + np




where w represents pure water, d represents dissolved  substances,




c represents chlorophyll, and p represents  particles,   n^. is  the




exponential coefficient in  the equation that approximates the vertical




attenuation of light:




                        I,  = I0e-ntz




where z is depth (a), I is  irradiance, and  TV has  units a"-1-.




Values of n.^ and its components ara always  somewhat  approximate




because of the change in spectral composition of light  with depth




(Smith 1968).




     The effect of pure water is relatively constant through  the




seasons, although it does vary some for complex reasons having to do




with changes in spectral composition (nw is about  0.1 m"-1-).




Dissolved organic materials contribute to the  extinction of light in




the lake but are seasonally quite constant, as indicated by absorbances




of filtered samples from Dillon.  Dissolved organics, which are

-------
responsible for the extinction component ri^,  are  not  present  in




large amounts in Dillon.  From the absorbances  at 360nn and  the




equation of Lewis and Canfield (1977),  the  dissolved  organic  careen, is




estimated as 1.5-1.6 mg/1-  From  the absorbances  of  filtered  water and




the approximation method developed by  Tailing  (1971),  we estimate that,




in complete absence of all particulars, inclua L'u  chytopldn^ton, the




secchi depth would be at least 1C in (based  ^n  ',rf  +  "^  =




0.17 m~l; this is a minimal figure, since -.   is smaller for  Eore




transparent lakes).  This corresponds  co a  1%  light  level  of  27 m.  For




comparison, the average secchi depth in Lake  Tahoe,  a  lai-.e




exceptionally free of particulates and  soluble  organics,  ij  about 23 m




(Goldman 1974).




     Separation of the effect of  chlorophyll  from che  other  effects on




transparency involves some assumptions, but is  feasible.   First it is




useful to make a few qualitative  observations  concerning the  relative




importance of algae and nonliving suspended matter-   In i^'32  a major




and precipitous decline in transparency occurrea  just  at the  onse: of




stratification but before chlorophyll  levels  in the  upter  water oolunn




had begun to clitnb (Figure 16).   For example,  the secchi depch cf che




lake in 1982 changed from 4.1 m on May  24 zc  1,3  =  on  June 7.  Tha




effect was observable at the same ti-e  in 1931  but  was less  drv-stic.




This quick decrease in transparency coincided  wirr.  the serin;; runoff,




and, since the chlorophyll had not increased  much at  :nis  Tjint, the




sudden decrease in transparency must be attricutiola  tr no-living




particulate materials added to the lake by  spring runoff.   Tre




transparency of the lake prior to the  onset of  runoff  shews  trat tne

-------
                                                                       76





nixing of  rhe lake was not  responsible  for  the  decrease in




transparency.  All evidence  thus points  specifically  to the major role




of spring  runoff in decreasing  the  transparency of  the  lake in early




June.




     Transparency remained  low  in the lake  both years  into August,




despite the much earlier decline in amount  of surface  water entering




the lake.  Low transparency  after June must  therefore  have been in




large part attributable to  the  buildup of phytoplankton biomass.  This




is confirmed by data on chlorophyll, to  be  presented  in the next




section, and by microscopic  examination,  of  samples, which  showed a




major decline in non-living  particulates after  runoff  had  passed.




     In the last half of the stratification  season of both years,




beginning  in August, there was  a return  to  high transparency.   This was




accompanied by drastic declines in  the surface  chlorophyll which




continued  through the termination of stratification and into the fall




nixing season.  The reasons  for this return  to  higher  transparency are




related to the nutrition of  the algal populations, so  the  complete




explanation is deferred until the chlorophyll data are  discussed.




     Quantitative separation of the effect  of chlorophyll  from other




effects on transparency can be  achieved by  estimation  of the




chlorophyll-specific absorption coefficient, eg,  ^faich  is  related to




the extinction coefficient for  chlorophyll  as follows:   es = n /B




where nc is extinction due to chlorophyll (In units per m), es is




Che chlorophyll-specific coefficient (In units  •  m2/nig Chi a.) and B




is chlorophyll a_ concentration  (mgChla_/m3) .  ec is traditionally




treated as a constant for a given lake,  although  it is  known to vary

-------
some with depth  (due  to  changes  in  spectral  composition) and wita

seasonal changes in phytoplankton composition (Kirk 1975).  The

estimate of e_ for Lake  Dillon  is based  en linear regression of
             o
extinction coefficient (dependent variable)  on chlorophyll

concentration  (independent  variable).  The slope is then e_•  The
                                                -\
regression for Dillon yields  a  slope  of  0.016 2-/ing.,  wnich will be

taken as the value for es•  This  is within th.2 nidrange of literature
                                                       n
values, which  generally  fall  between  0.008 and 0.021  m^/mg and shew a

node around 0.015  (Tailing  1932).

     From the  value of es and the chlorophyll concentration on any

given day, the amount of light  extinction due to chlorophyll ("c) car.

be approximated.   If  the effect  of  pure  water and dissolved orgar.ics,

assumed constant at ~)v + r, ^ = 0.17 a   ,  is added to "c, the

remaining extinction must be  the non-chlorophyll particulate effect,

the last of the  four  components  of  extinction.  The separation nade in

this way is shown  in  Figure 17.

     Figure 17 confirms  the foregoing  interpretation  based on

qualitative considerations.   Low transparency cane suddenly in early

June and lasted  into August.  At first the effect was  attributable

almost solely  to particulate  materials brought in from the watershed,

but these materials rapidly settled out  and  were net  replaced as zhe

runoff passed  its peak.  Low  transparency after runoff was sustained by

buildup of algal biomass.   Low  transparency  persisted  until algal

biomass was reduced by nutritional deficiencies toward the end cf

August.  The increasing  importance of  phytoplankton in contributing tc

the low transparency of  the lake toward  the  -niddle of  the

-------
z:
o

H
O

2
H
X
LJ

H
X
soo


 80


 60


 40


 20


   0
                 !C8
                         mix
stratified
                                                             mix
               water,

               soluble materials
               particuiates

                   '	I	!	L
                J	L
           JFMAMJJAS

                                    1981
                                                  0    N    D
O
z

h-
X
LJ
 00


 80



 60


 40
                 ice
                         mix
              I     I
               water,
               soluble  materials
           particuiates
stratified
                                          J	L
           J    F   M   A
                               J    J

                               1982
mix
    A    S    0   N   D
  Figure 17.  Cumulative percent light extinction due to particuiates,

            chlorophyll, and soluble material fractions in Dillon for

            1981 (above) and 1982 (below).

-------
stratification season was accompanied by a  change  in  water  color  from




blue to green.  Whereas nonliving particulate materials  can  change  the




transparency of a lake considerably without changing  its color




significantly, phytoplankton biomass responsible for  an  equivalent




reduction in transparency has a marked effect on color  (Kucchinscn




1957).









Historical Changes in Transparency




     There is some historical information on  the transparency of  Lake




Dillon.  The secchi depth was measured by the National Eutrophicition




Survey in 1975.  In August 1975, the secchi depth, as determined  free




the mean of a number of measurements at different  sites  on  the  lake.




was 7.8 m.  In September of the same year the mean was 8.3 21.   5ir.cs




the secchi depths at the same times in 1981 and 1932 were only  about




half these values, it would appear that a major decrease in




transparency occurred between 1975 and 1981,  probably due to




eutrophication.  Secchi depth data are also available frcm v'esle-




Nelson, Colorado Division of Wildlife (personal communication).




Nelson's values average lower than readings obtained at  the  same  cima




by the Eutrophication Survey or the present study, possibly  because  of




differences in discs or in the location of data collection.  However,




his series spanning 1975-1982 for secchi depth en  or near August  25




shows some evidence of a trend toward lower transparency:  1975,  -.6  zi;




1978, 3.6 m; 1979, 2.0 m; 1980, 2.3 m; 1931,  3.1 m; 1982, 1.6 =.

-------
                                                                       80
Composition of Major Ior.3

     Table 13 shows the composition of major ions  in  water  from Lake

Dillon.  There are some seasonal changes in major  ions,  particularly

associated with the runoff in the month of June, but  these  are  not

great enough to have much biological significance,  so the amounts  of

major ions were not quantified on a routine basis.  Table 13  indicates

that the water of Laka Dillon is a bicarbonate  type exceptionally  rich

in sulfate.  Cations are dominated by calcium.  The total ionic

strength, as indicated by Table 13 and by the conductance measurements

that were made routinely throughout the course  of  the  study,  is

somewhat higher than might be expected for an undisturbed Rocky

Mountain watershed of similar geology.  The total  ionic  strength and

ion composition of the lakewater has undoubtedly been  augmented  to  some

extent by mining and dispersed earth disturbance,  as  will become

evident in the analysis of stream chemistry.  Expected conductance

computed from Table 13 and ionic mobilities is  179 umho/cm  at 25°C,

which checks well with the observed conductance (170  umho/ca).
concentration

Cations
Ca++
j*tg"H-
Na+
K+
Total cations
Anions
EC 03
SOT,
Cl"
NC3
Total anions
Total ions
Eg /I

21.3
3.58
3.45
0.54
23.89

47.1
39.1
3.90
0.93
91.03
119.92
meq/1

1.068
0.295
0.150
0.014
1.528

0.772
0.814
0.110
0.015
1.711
3.239
Table 13.   Major ion composition of Lake Dillon, based on analysis  of  a
           sample taken at the index station, 4 October 1982.

-------
Conductance

     Conductance of lakewater was slightly  higher  in 1S81 than 1932, as

shown in Table 14, which gives annual  averages  for the top, middle, ar.d

lower water column.  The difference between years  was  more pronounced

in the upper water column than in the  lover water  column.  Conductarxa

showed significant seasonal variation  and a degree of  vertical layering

under ice and during summer stracification.  These features are

illustrated by Figure 18, which shows  the change of  conductance through

time at the top of the water column  (0-5 m) and  in the bottom 5 m for

the two years of che study.  In both years  the  surface water --'as lover

in ionic solids than deep water under  ice.   This layering cust be set

up after ice cover, as fall mi:cing homogenizes  conductance.  Retention

of melt water near the surface is the  probable  explanation.  Surface

temperatures under ice were in the vicinity of  0°C at  the top and,

since water is less dense at 0°C than  at slightly  higher temperatures,

there would have been a slight resistance  to nixing  tc help retain the

layering as long as the ice was protecting  the  surface from wind.

Conductance, uinho/cm
1981
1982
1981
1982
Alkalinity, mg/1 CO^
1981
1982

0-5

168
152
7 .45
7.30
33.9
34.5
Decth, m
2C-25

167
153
1 . 36
7.59
34.4
TJi 1

35-40

1"S
172
7.^4
— :;/.
•\i_ -j
35.-
Table 14.  Means for conductance, ^H,  and  alVialinitv at three -ie
           the water column at  the  index station

-------
                                                                       82





     Vertical homogeneity of  conductance was  typical  of the spring




-ixing period, as expected.   With the onset of  thermal  stratification




in June, there once again was a divergence in conductance of surface




water and deep water-  The divergence appeared  rather  suddenly,




suggesting that spring snowmelt was at  first  held  in  greatest  amounts




in the upper water column after thermal stratification  began.   Not all




of the runoff entered the upper water column, however,  because the




lower water column also showed steady dilution.  Furthermore,  the




conductance of the upper water column was homogeneous  down to




approximately 30 m, a depth much greater than the  upper boundary  of the




thermocline.  Thus the dilution caused  by spring snowmelt affected the




entire water column but was more pronounced in  the upper 30 m  than the




lower 30 m.  This accounts for the higher average  conductance  of  the




lower water column as illustrated in Table 14.  With the fall  mixing,




vertical homogeneity of conductance was established as  shown in




Figure 13.




     It is possible to estimate the depth at  which entering streams




penetrated the lake, since a stream will tend to seek  the layer of




its own density.  This is not an important issue when  the lake is




mixing freely, but is potentially of interest during the stratification




season when the entering water may tend to flow selectively into  the




upper, middle, or lower water column.   In order to make such an




estimate,  we computed the density profile of  the water  column  on  each




sampling date during stratification.  Taking  into  account not  only




temperature,  but also dissolved and suspended solids, we also




calculated the density of the three rivers on each date.   By comparison

-------
<£>
Q.
LU
O
O
15
Q
2
O
O
      90  L
      50  -
  0
            JFMAMJJASOND
                  ce
                    mix
stratified      !  mix
90
      150  -
            JFMAMJJASOND
                               !982
           r.ductar.ce i- Laka Dillcr.,  _931-19?2

-------
                                                                       84





we could then show the depth at which  river  and  lake  density would be




matched, and this is the depth to which the  rivers  would  tend to flow.




The results are shown in Figure 19.  In both years  the  bulk of flow was




predominantly into the upper middle of the water column,  and in both




years the depth of penetration increased from June  to October-  In 1981




the penetration began at about 10 m, just below  the therraocline, and




increased to about 25 m by late stratification (due mainly to warming




of the lakewater).  In 1982, penetration was  not so deep;  it began at




about 5 m, and slowly passed down to about 15 m.  We  postulate that the




large flow at these depths, as well as heat  loss over the  spillway, was




responsible for holding the thermccline at a higher level  in 1982 than




1981.









pH and Alkalinity




     Table 14 shows that the pH at all depths was slightly higher in




19S2 than in 1981, and that in both years the mean  pH at  the bottom of




the water column was slightly lower than at  the  top, as expected




because of the influence of photosynthesis near  the top of the water




column.  Figure 20 shows the seasonal patterns near the surface (0-5 m)




and at the bottom of the water column.  The  pH levels at  top and bottom




diverged in a major way only after rapid photosynthesis began in late




June.  Photosynthesis elevated the surface pH values in proportion to




the rate of removal of free CC^, so the surface  pH more or less




paralleled chlorophyll.   The pH levels were  at no time extraordinary,




however,  and do not suggest extreme depletion of free C02,  which is




signified by pK levels above 9.0 (Tailing 1976).  In both  years there

-------
                                                T  5
  
-------
                                                          36
                   ice
mix
stratified
mix
    CL
         7  -
                        1	I	I	I	(
              J  F  M  A   M   J   JASOND
                                 198!
                    ce
mix
    CL
         8
         7  h
stratified        mix
                                             — 0-5 M
                                             - — 40-45 M
                                               J	L
              J   F
     J   J   A
     1982
         S   0   N  D
Figure 20.  pH in Lake Dillon, 1981-1982, (index station).

-------
was a return to lower pK levels as photosynthesis  dropped  off  after




July, and a resurgence after nutrient depletion was  relieved by




thickening of the mixed layer in fall.




     Alkalinity values fell within a relatively narrow  range,  as  shown




by Table 14.  Seasonal effects were slight  (Figure 21),  but  the  large




runoff of 1982 can be identified as the cause of some reduction  in




alkalinity.

-------
                                                     38
O
S
     38

     34

     30

     28
                ice
               mix
            stratified
                      mix
             !    J	  I	 !
J   F
A
                                  0-5 M
                                 J	!
                                              J	i
J   J   A    S   0   N   D

198!
                ice
               mix
            stratified
                      mix
     38

     34

     30

     26h
                              — 0-5  M
                              •--40-45 M
           J   F
            A   M   J    J    A

                     1982
                     S   0  N   D
Figure 21.  Alkalinity in Dillon, expressed as mg/1 C02  (index station).

-------
                                                                       39





          Phosphorus and Nitrogen  in Lake Water  and  Sediments









Phosphorus in Lake Water




     'The concentration of total phosphorus  and  the  contributions of




orchophosphate, dissolved organic  phosphorus, and  particulate




phosphorus to this total are  summarized  in  Table  15.   The  average




orthophosphate values were exceedingly low  in both  years  and  showed




virtually no trend with depth.  The averages  for  1932  were




unquestionably slightly higher, however.  This  is  explained mainly by




the greater flow of water into the lake  in  1982,  as  described core




fully below in connection with seasonal  trends.   Soluble  organic P was




similar to orthophosphate in  showing no  real  trend  with depth in the




averages and in the slightly  higher average for  1982.  Particulate P,




in contrast, was essentially  the same the two years  and showed a slight




tendency toward higher values near the surface  both  years,  probably due




to the ability of phytoplankton, which are  located  mostly  near the




surface, to sequester phosphorus.  As a  result  mainly  of  higher soluble




P levels, total P values were slightly higher in  1982  than in 1981.




Slightly less than half of the total P was  soluble  and slightly more




than half was particulate on  the average.




     Table 16 shows the results of a two-way  analysis  of  variance




(AXOVA) in which the total P and ? fractions  were  tested  for




statistically significant differences between years  and depths.  The




r-5 n layer was taken as representative  of  surface  water  and  the




40-^5 tr. layer was taken as representative of  deep water .   The years




differed significantly only for F04-p and total  ?,  and cepths

-------
                                                                      90
Concentration - ug/1
Soluble
PO^-P Organic P
Stratum
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
X
Mean %
1981
0.75
0.56
0.77
0.62
0.47
0.63
0.57
0.64
0.72
0.72
0.79
0.66
10%
1982
1.50
0.94
1.01
1.30
1.91
1.65
1.66
1.30
1.51
2.10
2.16
1.55
20%
1981
2.08
1.55
1.77
2.40
1.56
2.00
3.52
2.04
1.68
1.81
2. 60
2.09
31%
1982
2.36
2.49
4.96
2.57
1.84
1.53
3.21
2.11
1.31
1.17
1.48
2.28
29%
Particulate P
1981
4.80
4.98
4.46
4.31
4.15
3.76
3.68
3.51
3.70
3.47
4.00
4.07
60%
1982
5.54
4.96
4.87
4.37
3.99
3.54
3.58
3.38
3.83
3.27
2.73
4.00
51%
Total P
1981
7.63
6 .74
7.01
7.33
6.19
6.39
6.31
5.83
6.23
6.68
6.03
6.53

1982
9.15
8.20
10.64
8.01
7.45
6.16
8.19
6.56
6.40
6.80
6.33
7.67

Table 15.  Mean concentrations of total phosphorus and phosphorus
           fractions at various depths between the surface and bottom
           at the index station.  Fractions may not equal totals
           exactly due to occasional missing values for individual
           fractions.

-------
                    PO.-P
                      4
 Soluble    Particulate     Total
Organic P        P            ?
Difference bet.      Not            Not
   depths        significant    significant

Difference bet.   Highly           Not
   years         significant**  significant
              Highly        Highly
            significant** significant**
                          Significant*
significant
**p < 0.01       * p  <  0.05
Table 16.  Results of  a  two-way  ANOVA testing for differences in means
           between depths  (0-5 versus 40-45 m)  and between years (1981
           versus 1982).
differed  significantly  only  for  particulate  and  total P-   Mo

interactions of year with  depth  were  significant.

     Seasonal  changes in  the total  phosphorus  concentration and in the

contribution of the different  phosphorus  fractions  in the  upper water

column  (0 - 5  n) are summarized  in  Figure 22a.   Comparable information

is given  in Figure 22b  for the lower  water column  (40-45 m) .  There are

some notable trends both  in  total phosphorus  and in the  contributions

of the  fractions, especially in  the surface  layer.   These  trends can be

understood for the most part in  terms of  the  seasonal events occurring

in the  lake.

     Spring runoff coincided with some  of the  highest total phosphorus

concentrations in both  years.  Rising levels of  total phosphorus in

spring  were coincident  to  a  large extent  with  an increase  in the

contribution of particulate  phosphorus.   We  attribute this effect to

the transport  of large  amounts of particulate  material from the

watershed at the time of runoff.  Silt, which  is an effective

-------
                   ice
                                                                     mix
                      total  soluble  P
               total soluble  P
                                                                   N    D
Figure 22a.  Total  phosphorus and  phosphorus fractions  in the top 5 a
            of Lake Dillon.

-------
                ice
            mix
     8



     4
            stratified
                     mix
               total soluble P
          J    F   M    A    M
CO
z>
a:
o
i
0.
CO
O
X
0.
                   J    J    A    S    0    N    D

                    198!
ice
mix
stratified
mix
     2


     8
     0
          J    F   M    A    M    J     J

                                    1982
                             A    S    0    N    D
 •-ire 22b.   Total phosphorus  and phosphorus fractions  in ce

          (40-45 3) of Lake Dillon.

-------
                                                                      94






adsorption substrate  for phosphorus,  is  particularly likely to make a




major contribution  to the particulate fraction  of  phosphorus at this




time of the year.   The concentrations declined  rather quickly, however,




and in July reached levels that were more or  less  average  for the year.




The peak in total phosphorus associated  with  runoff  was  considerably




larger in 1982 than in 1981, as would be expected  from the relative




amounts of runoff,  and thus of particulate transport,  in the two years.




Inorganic particulates added with runoff can  be expected to sediment




rapidly after entering the lake.  This accounts for  the  decline in




total phosphorus of the surface waters after  runoff  has  peaked.




Particulate phosphorus was not brought to the surface  from the lake




bottom in significant amounts at the time of  fall  overturn, despite




complete mixing of  the water column.  From August  until  ice cover




returned at the end of December, the total phosphorus  hovered in the




vicinity of 6 to 8 ug/1, most of which was particulate and was




undoubtedly tied up in phytoplankton and bacteria.




     Orthophosphate was seldom a major contributor to  total phosphorus




in either 1981 or 1982.  For the most part, PO^-P  levels were in the




vicinity of 1 ug/1 at the surface and were somewhat  higher near the




bottom during stratification (but not as an annual average:  Table 15).




Higher PO^-P at depth during stratification probably reflects lower




biological demand there.  Because of uncertainties as  to the ability of




the standard molybdate test to distinguish between true  orthophosphate




and small organic phosphorus compounds, it is quite  possible that the




actual P04-P was even lower than the apparent PO^-P  as indicated by




the inolvbdate test.

-------
     Crthophosphate cone an.-rations cccasionally  rose  significantly




in the surface water-  In >'ay of 1931, orthcphosphate  rose  to about




4 ug/1, a concentration well above the annual  average.   It  seems almost




certain that this rise in PO/j-P was caused  by  spring  mixing,  when the




winter phytoplankton had bean dispersed from the  surface  and  the spring




phytoplankton had not been able to start  surface  accumulation because




of the movement of tie entire va;=r colisr.n.  Unfortunately  there was




only one May sample for 1982 because  of a ring of  ica  around  the lake




which prevented launching of the boat during most  of  May.




     There were also some minor but consistent increases  in the  amount




of surface PO^P between September and December  of  both years.   These




probably were caused by the incorporation of hypolimnetic water, with




higher PO^-P, into the mixed layer as the thermocline  descended.




     Soluble organic phosphorus made a significant  contribution  to




total P at all times of the year except at  the end  of  the




stratification season.  There was a definite and  consistent decline in




the total amount and percentage of soluble  organic  phosphorus at the




surface between the first of July and the end  of  August both  years.




This was almost certainly caused by strong  biological  demand  for




phosphorus, especially after the establishment of  large phytoplankton




populations by the end of June.  Bacteria may  have  been  involved as




well, since the downward trend was also notable  below  the thermocline.




Over a period of tine, the phytoplankton  and bacteria  populations




seemed to be able to sequester most of the  soluble  phosphorus,




including the organic fraction, in the form of bio-ass.   Alkaline

-------
phosphatase on cell surfaces nay account  for  this  ability to tap the

soluble organic pool (Fogg 1975, Nalewajko and  Lean  1930).

     Table 17 summarizes  the historical information  on surface total P

in the main body of the lake.  The late summer  values  show  no evidence

of change, but the early  summer values trend  upward.   The initial

spring phosphorus, which  is responsible for the growth burst of early

sutnmer, has increased, but the epilimnetic supply  is ultimately

depleted by sedimentation to much the same levels  now  as  in 1975,

according to the table.
Total P ug/1
Year
19751
19772
19782
19S1J
19823
June

5
7
11
12
July

6
6
7
10
August
6
60*
6
6
6
September
Q
—
_
6
/-
o
T
A National Eutrophication Survey
*- Summit County/EPA study
3 present study
^ Probably erroneous

Table 17.  Comparison of total P values from various  studies  of  Dillon
           (main body of the lake, surface samples).
Nitrogen in Lake Water

     Table 18 summarizes the information on total nitrogen  and  nitrogen

fractions at the index station.  As would be expected  in  visw of  the

presence of substantial "amounts of dissolved oxygen at all  depths

throughout the year, the amounts of nitrite were exceedingly  low,  and

nitrate was the dominant soluble inorganic form.  There was a

substantial depth gradient of average nitrate concentrations  in both

-------
Concentration - ug/1
>pth
:ratura (ai)
3-5
5-10
0-15
5-20
0-25
5-30
0-35
5-40
0-45
• 5-50
!ean
lean %
Nitrite N
1981
1.5
1.7
1.2
1.2
1.2
1.3
1.2
1.1
1.1
1.3
1.3
.4
1982
2.5
1.9
2.3
2.3
3.3
4.3
5.0
3.6
2.9
2.8
3.1
1
Nitrate N
1981
73
76
74
83
101
113
155
166
187
202
123
42
1982
67
66
67
79
106
139
174
188
224
259
137
31
Ammonia N
1931
19
14
13
13
13
18
15
13
13
13
15
5
1982
15
19
15
22
33
32
30
26
26
16
23
5
Soluble
Organic N
1981
127
104
104
135
109
1C7
108
126
113
222
126
43
1982
306
384
216
236
165
196
243
252
227
259
253
57
Total
Soluble N
1981
221
196
192
232
230
245
230
311
315
438
266
91
1982
391
472
300
388
308
372
452
470
430
537
417
93
Fartic .
N*
1981
23
"• Q
28
23
28
23
2?
23
28
28
28
10
1982
30
3C
30
30
20
20
20
30
20
3G
30
-
Total .S"
1 c ; 1 1 ° 3 2
0 , Q ^ ^ ^
22± 302
220 32?
2cC 41?
253 333
273 4C2
225 i>2
339 500
243 510
466 565
294 .47
100 100
'Values below 5 m set equal to annual mean  for  0-5 m

.'able 18.  Average total nitrogen and nitrogen  fractions  at  different depth strata in 1981
          and 1982 (index station) .  Slight  inconsistencies in the sums are due to
          occasional missing values for individual fractions.

-------
                                                                      98





years.  The seasonal analysis will show that  this was  caused  by the




removal of nitrate nitrogen by phytoplankton  in the surface waters




during the summer.  In 1982 the nitrate concentrations  were very




similar in the upper half of the water column to the 1981




concentrations, but were consistently somewhat greater  in  the lower




water column than in 1981.  Nitrate nitrogen  accounted  for most of  the




soluble inorganic nitrogen pool both years and made up  30-45% of the




total nitrogen.




     Ammonium nitrogen was present in measurable but very  low amounts,




accounting for only 5% of total nitrogen.  Preferential uptake of




ammonium by phytoplankton and continual oxidation of ammonium by




microbes undoubtedly maintained the equilibrium ammonium concentrations




at a low level.




     Soluble organic nitrogen was always an important  component of  the




total nitrogen pool.  In 1981 its percentage contribution  to  total




nitrogen was slightly lower than in 1982; the grand average for the two




years was about 50%.  Absolute levels of soluble organic nitrogen in




1982 were considerably higher at all depths than in 1981.  This is




probably explained by the larger amount of water entering  the  lake  in




1982.  It will be evident from the analysis of stream  chemistry in




later portions of this report that stream water entering Lake




Dillon contains large amounts of soluble organic nitrogen, much of




which is probably quite refractory, and that  the concentrations




increase with amount of runoff.




     Table 18 shows that soluble nitrogen, including both  organic! and




inorganic fractions, accounted for slightly more than  90%  of  the total

-------
nitrogen as an annual average.  Particulate  nitrogen was  only measured

in the surface water (0-5 m); the amounts  at  other  depths have been set

equal to the mean in the surface water.  This is  only an  approximation,

but errors associated with the approximation  are  not likaly to be

important in the computation of total  nitrogen  because of the small

contribution of particulate nitrogen to  the  total.   Total nitrogen,

which is shown in the last two columns of  Table 18,  was higher in 1962

than in 1981, as would be expected from  the  patterns in  the most

important nitrogen fractions.

     Table 19 summarizes the results of  a  two-way ANOVA  similar to the

one that was carried out on the phosphorus data.  The ANOVA confirms

statistically the suspected significance of  the difference between

years in concentration of soluble organic  nitrogen  and total nitrogen.

The difference between years in nitrate  was not significant because ic

appeared only in deep water rather than  over  the  entire water column.

The ANOVA confirms statistically the differences  in mean  nitrate

concentration with depth.
NO--N
Soluble
NH,-N Organic N
Total
N
Difference bet.    Highly          Not            Noc            Not
   depths        significant** significant    significant    significant

Difference bet.      Not           Not          Highly**       Highly**
   years         significant   significant    significant    significant

** p < 0.01                 ———
Table 19.  Results of a two-way ANOVA testing  for  differences in r^ear.s
           between depths (0-5 m versus 40-^5  m)  and  between  vears
           (1981 versus 1982).

-------
                                                                       100
     Figure 23 shows  the seasonal changes  in  concentration of all




nitrogen fractions at the surface of  the lake.   Some  missing values for




individual fractions hava been filled in by linear  extrapolation.




Figures 24 and 25 show the complete details over  time and  depth of the




distribution of nitrate, which is the most informative nitrogen




fraction from the viewpoint of phytoplankton  nutrition.   Figure 23




snows that ammonium was highest at the surface under  ice.   This




probably reflects the lower demand for inorganic  nitrogen  and lower




efficiency of nitrification at this time of year.   Soluble nitrogen




concentrations were highest in winter and  spring.   A  spring  increase




was especially pronounced in 1982, probably because of the larger




runoff that year.




     The inorganic nitrogen fractions began a precipitous  decline at




the lake surface as soon as stratification occurred.   There  is  little




doubt that this was caused by phytoplankton demand  for inorganic




nitrogen.  Complete depletion of inorganic nitrogen at the surface




occurred by the middle of July in 1981 and by the beginning  of  August




in 1982.  In both years, inorganic nitrogen began to  climb again  in




mid-September, in coincidence with major thickening of the mixed  layer




leading to complete breakdown of stratification.  These  trends  in




inorganic nitrogen are especially obvious  from the  time-depth diagrams




for nitrate (Figures 24, 25).  Both figures show very clearly that the




nitrate depletion was associated specifically with  the growth zone and




did not occur  below the thermocline where biological  demand  was much




lower.

-------
                ce
irnix i
t    i
                           L
                          f* T ^ ''
                          Sir
                            mix
    600-
    400
    200
       0
                 TOTAL  N
            PARTICULATE  N
         _ NH4-N
          JFMAMJ    JASON    D
                                !9Si
LJ
O
O
cr

S   800
    600
    400
    200-
     ice
mix
           stratified
             TOTAL

PARTICULATE  N
rr.ix
         NH4-N
         N02-N+N03-N
            I	!
          J   F   M
            A   M   J   J   A
                                982
                  S   0
    Figure 23.  Total nitrogen and nitrogen fractions  in the :cr ' rr. ^:
             Lake Dillon.

-------
H
M-
1J
rt-
fl)
rt
fB

H-
y
h-1
O
C
CO
 0)
 e
  *h

X
K
CL
Ld
Q
            0
                         ice
                                  mix
          10
         30
         40
         50
                               150
                 J    F
                                          J    J     A     3    0     N     D

                                          1981
                                                                                          o
                                                                                          I J

-------
TO
C
K>

Ol
a
n>

(X
n>
•d
rt
Q.
H-
P)
OP
i-<


I
p.
ft)


H-
O
H-
 
-------
                                                                       104
     Soluble organic nitrogen peaked very  sharply  in  1982 in




coincidence with runoff, suggesting that a major amount  of nitrogen




entering with the peak runoff was in soluble  organic  fora.  In 1981, a




year of low runoff, the organic fraction was  lost,  so we do not know




the natiire of the runoff spike.




     During summer stratification of both  years there was a reduction




in the organic nitrogen pool at the surface.  The  decline was  so




drastic in 1982 as to suggest the importance  of some  abiotic processes,




such as settling of flocculated organic matter-  There was no  evidence




of a steady drain on the soluble organic nitrogen  fraction as




phytoplankton completely depleted the inorganic nitrogen.  Thickening




of the epilimnicn in fall restored high levels of  soluble organic




nitrogen as soluble organic matter was brought to  the surface  from the




hypolitanion.




     No historical information is available on nitrogen  in the lake.




Total nitrogen was measured by the National Sutrophication Survey in




1975 but the nitrogen levels proved to be  below their detection limit.




     Table 20 summarizes the total nitrogen and total phosphorus data




as an average over the entire water column for both years and  shows




the molar ratios of nitrogen to phosphorus.   The ratio of aitrogen to




phosphorus is exceptionally high.  Municipal  sewage typically  has a




ratio of 6 to 14, and most inland waters fall between 15 and 40 (Stumm




and Baccini 1978).  The very high ratios of Dillon are partly  explained




by tertiary treatment for P, which greatly raises  the N  to ? ratio of

-------
effluents.  The ratio  of  N to  P for  phytoplankton bionass is typically

Total
Total
Total
Total

N ug/1
P ug/1
N, umole/1
P umole/1
N:P ratio (molar)
1981
294
6.6
21
0.21
100
1982
447
7.1
31
0.23
134
Composite
370
6
26
0
118

.3

1 n

Table 20. Summary of  average N  and  P  amounts  for  the  wnole water
          column at  the  index station.
between 10 and 17.  Thus  the  ratios  of  Dillon  are  suggestive  of

phosphorus control.  However,  the  differences  in  spatial  distribution

of nutrients and in the biological availability of  the  fractions  of  N

and P require other and more  direct  evidence of phosphorus  control.



Sediment Phosphorus

     Sediment samples from  eight different  sites  were  analyzed  for

particulate phosphorus and  interstitial  phosphorus.   The  collection

sites for sediment included the heads and mouths  of  the lake  arms as

well as the main body of  the  lake.   The  wet surface  sediments  from a

given site were mixed and a subsample equivalent  to  10  to 20 mg of dry

material was dried to constant weight in a  tared  container  and  then

analyzed for total phosphorus by the particulate  phosphorus procedure

described in the methods  section of  this report.  Anocher subsample of

wet sediment was centrifuged  for 30 minutes at 8,000 rpm, aftar which

5 ml of the supernatant was withdrawn and analvzed fcr  total  soluble

-------
                                                                       106




phosphorus according to  the procedures used  for  lake  water-   The




results of these analyses are reported in Table  21.




     There was surprisingly little variation in  the percentage




phosphorus per unit dry weight of sediment.  Such variation  as can be




discerned in the data cannot be related in any way  to  particular lake




arms or to particular depths.  Thus it appears that the  lake  bottom can




be treated as having 0.1% phosphorus on the average with  randon




variation spanning about -1-0.02%.  The percentage phosphorus  in the dry




sediments of the lake is not especially high.  For  example,  the summary




data of Brunskill et al. (1971) show an average  for 23 Wisconsin lakes




of 0.12% and for 30 eastern Ontario lakes of 0.11%.




     The total soluble phosphorus in interstitial water was high fay




comparison with the overlying lakewater, as expected  from the




decomposition of organic matter liberating phosphorus  in  the  sediments.




These values, although much higher than lakewater, are not




extraordinary, even for relatively unproductive  lakes.  For example,




the summary by Brunskill et al. (1971) on interstitial water  chemistry




for unproductive lakes of the Canadian Shield shows a  range of two




orders of magnitude in the interstitial total soluble  phosphorus and  a




mode of 220 ug/1 P.  In Lake Dillon the diffusion gradient between the




interstitial waters, at about 135 ug/1 of total  soluble P, and the




water column, at about 3 ug/1 soluble T?} is obviously  quite steep.  The




sediments must therefore play some role in supplying phosphorus to the




overlying water column,  but there is no evidence that  this internal




source is in any way unusual in Lake Dillon.

-------
Water
Depth (tn)
Collection Site
Lake
Mouth
Mouth
Mouth
Mouth
Head,
Head,
Head,
Mean
Center (Index)
, Dillon Bay (Main)
, Snake Arm (Main)
, Blue Arm (Main)
, Ter.mil e Arm (Main)
Snake Arm (Survey 2)
Blue Arm (Survey ")
Tenmile Artn (Survey 7)

60
32
33
30
33
23
13
15

Sediment ?
" ? in Dry
Sediment
0.
0.
0.
0.
0.
0.
0.
0.
0.
085
102
092
103
115
093
124
085
100
Interstitial P
Total Soluble,
us/1
203
113
103
1 :2
122
136
132
107
125
Table 21.   Particulate P and interstitial ? from surface sediments
           collected July 26, 1982 at eight sites on Lake Dillon.

-------
                                                                       108





                Particulates and Phytoplankton Biomass









Total Particulates in the Water Column




     The average values for total dry weight of particulates  retained




on Whatman GF/C filters are summarized in Table 22.   The  table  shows




that the upper water column had a considerably higher total particulate




concentration than the lower water column in both years.   The averages




for depths below 30 m were about half the averages for  the surface.   In




the upper 25 m, total particulatas were considerably higher in  1982




than in 1981 as an annual average.  In both years the upper 5 m was




slightly higher than any other depth, a layer from 5  to 15 ni was




distinctively high but not quita so high as the top 5 m,  and the




stratum from 15 to 25 m was a transition from the higher  concentrations




of the upper 15 m to the lower concentrations of the bottom 30  m of the




water column.




     Variations in total particulate concentrations with  depth




and between years are not really understandable without some detailed




consideration of seasonal variation.  Figures 26 and 27 show time-depth




diagrams of total particulate concentration, from which a number of




major seasonal variations are evident.  Total particulate




concentrations under ica were low at all depths.  In general, inorganic




particulates settle out very thoroughly under ice because  of the lack




of turbulence and very small amounts of incoming water-   Thus




particulates that are present under ice are likely to be  dominated by




phytoplankton.  Fine organic debris and bacteria may make  small




contributions, but microscopic examination of samples showed that

-------
                0
  OQ
  ri
  r1
  n>

  i j
VD >'•
r» U
M ro
   I
,-N Q.
U n>
00 T3
   0)
  C/Q
   t-l
   rT
   O
  X)
   tu
   r>
   c
   pi
   It
   n>
 a>
 e
  *b
X
L—
ou
UJ
o
               10k
              20 h
              30 h
      40  -
              50
                                                                   stratified
                      JFMAMJJ      ASON

-------
°9

ru

r-o
H
fD
TJ
rt
Zf
(W
H
U>
B

Ml
o
ft
H-
r>
e
(D
cn
o
rl
    e
3   3C
rt   L^x>
[ii   •
M   O.
•d
Oi
    Id
    o
          0
                                                                             mix
         10
        20
        30 -
        50
                     F
                                 A
J     J

1982
A
N      D
                                                                                         O

-------
Particulates - ng/1
Stratum (in)
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
45-50
50-55
1981
2.23
2.12
•> if
— • — -t
1.77
1.59
1.44
1.23
1.19
1.42
1.24
1.66
]_
3
3
3
0
2
-
1
1
1
1
1
982
.35
.;?
.64
.34
.21
.62
.57
.27
.98
.52
.40
Table 22.  Mean total particulates of Lake Dillon
           in different strata of the water  column
           (index station).

-------
                                                                       n:




these did not compare with  phytoplankton  in  accounting for particulates




under ice in Dillon.  As already noted, che  ice  was  not covered




completely with snow in the  first months  of  the  1931 ice cover.  The




conditions for development  of phytoplankton  in the upper water column




under ice were thus exceptionally good.   This is reflected in somewhat




higher total particulate concentrations in the upper water column under




ice in 1981.  Phytoplanktor,  also dominated particulates under the ice




in 1982 but were simply not  abundant enough  to create high particulate




levels.  As shown in Figures 26 and 27, total particulates in the upper




water column under ice reached concentrations two  to three times  as




high, in 1981 as in 1982, although the total  particulates in the lower




water column where phytoplankton were not present were comparable to




those in 1982.




     During spring mixing,  the winter phytoplankton  were dispersed




vertically and inorganic particulates from the watershed were added in




increasing amounts.  The stratification period,  beginning in early




June, was marked in both years by two peaks  of particulate




concentration occurring in  the epilimnion.   The  first peak coincided




with runoff, and was caused by the entry  of  large amounts of inorganic




particulates from the watershed.  As runoff  declined, the entry of




inorganic particulates declined and settling outstripped the addition




of particulates.  At the same time, as shown previously by Figure 19,




incoming water from the rivers passed to  greater depths as the summer




progressed.  Inorganic particulates entering deeper  water would settle




more quickly because of lower turbulence  in  deeper water-  The peak for




total particulates during runoff was considerably higher in 1982  than

-------
                                                                       11




in 1981, as would be expected.   la  both  years  there  was  a second peak




of surface participates in  July  after  the  peak  runoff  passed.   The




second peak of particulates was  of  essentially  the  same  magnitude as




the earlier one associated  with  runoff.   The second  peak was caused by




the buildup of phytoplankton biomass and  not by inorganic particulates.




This was confirmed by microscopic examination of  samples and by




chlorophyll analysis.




     The amounts of total particulates in  deep  water during




stratification were low by  comparison with the  surface values.   This is




principally explained by the smaller amount of  turbulence in water




balow the thermocline, which would  promote more rapid  settling  in deep




water.  In 1982 the total particulate  levels below  the thermccline were




consistently greater than in 1981,  as would be  expected  from the entry




and settling of greater amounts  of  inorganic particulate materials in




1982 with the greater runoff that year.




     In both 1981 and 1982  there was a pronounced decline of total




particulates in the surface water beginning in  the  last  half of July.




This was caused by a decline of  phytoplankton biomass  that  time.  This




phenomenon will be discussed later  at greater length.  Fall mixing




established relatively uniform distributions of particulates as




expected.









Chlorophyll a




     The distribution of chlorophyll a over time  and de~th  is  shown in




Figures 28 and 29.  There was a  significant amount of  chioroshvll under

-------
 09
 C
 m


 I 3
 o,
 H-
 w
 ftl
 a
 i-n
 o
 n
 rr
 M
 o
 t~t
 o
t3
 nr
'O
oo
M
i;
oo
0
                                                                                                mix
            50
                            F
                            A
J      J

1981
A
S     0     N      D

-------
 l_

 (B

 r j
 n.
 n>
-a
 n.
 H-
 (U
 0>
 g
 O
13
r'
cu
 e
   •%
X

a.
UJ
o
                    0
                  10   -
         20   -
                 30
                 40
                 50
                                                                                                                                     N
                                                                                                                                        D

-------
                                                                       116





ica in both years, but the amount was much higher  in  1981  than in 1982




because of the lack of snow cover and transparency of  the  ice  in




January and February of 1981.  In fact a pronounced chlorophyll peak




developed under ice in 1981, as shown in the  time-depth  diagram.  The




chlorophyll under ice in 1982 should probably be regarded  as more




typical, since the large amount of light reaching  the  water  in January




and February of 1981 was only possible because of  the  very unusual




weather conditions that year.




     The spring mixing in May of 1981 drastically  reduced  the  surface




chlorophyll to a concentration well below 1 ug/1.  Two factors




contributed to this decline.  First, there was a decline in  surface




chlorophyll that began well before the ice melted,  This may have been




caused either by light deprivation (due to light extinction  by spring




snow) or to nutrient depletion.  An attempt to separate  these  effects




will be made in a later chapter.  Secondly, the decline  was  accentuated




suddenly at the time of melt by the mixing of surface  waters,  already




low in chlorophyll, with deeper water having essentially no




chlorophyll.  This was taainly a dilution effect.   Dilution may have




been accentuated, however, by light deprivation due to deep mixing,




which moved the algal cells to such great depths that  they spent much




time in the dark.  This effect would have been short-lived, however,




since incipient layering quickly began to impede mixing.




     The light deprivation phenomenon can be shown more  clearly by use




of the concept of effective light climate (Raaberg 1976, 1979):
                        nt 2m

-------
where I is effective  light  climate  (mean  exposure  of  the average




cell in the mixed layer, expressed  as  langieys  per day PAR), z^ is




the thickness cf the  mixed  layer  (adjusted  for  morphometry), -v is




the extinction coefficient  (m~^), and  1^  is  penetrating PAS.




(langieys per day, PAR, with  surface correction).   In May,  during




mixing on Dillon, z^  was about  25 in and rlt.  was  about  0.3 m   .




On a cloudy day 1^ would have been  about  145 ly/day,  ana on a




bright day it would have been about twice this  amount .   Thus for a




cloudy day I would have been  11 ly/day and  for  a  bright day it




would have been 22 ly/day.  The presumed  threshold for  phytoplankron




growth is about 10-20 ly/day, so it is clear that  really rapid  growth




could not have occurred until there had been a  change in i^ cr  z^-




Spring changes in nt  were actually  upward,  reducing I.   Thus




growth could only have occurred through a major change  in z^-




Stratification accomplished the needed change in  zm,  bringing  I




into the range 50-120 ly/day  in early  June.




     The degree of decline  at the tine of spring  mixing in  1932 was not




evident because samples could not be taken  during  the month of  May, but




in all probability it was very  similar to the decline observed  in 1981.




In both 1981 and 1982, as expected  from effective  light climate, there




was a rapid resurgence of chlorophyll  concentration as  soon as  the




water column stabilized in  June.  As is typical in a  dimictic  lake, the




conditions for phytoplankton growth just after  establishment of




stratification in the spring were very good  because available  nutrient




concentrations were at a maximum as a  result of recent  complete mixing




of the water column,  while  daylength and water  temperature  were

-------
                                                                       118
increasing and the phytoplankton  cells were  held  within  the lighted




zone by the presence of a thermocline.




     In both years the chlorophyll concentrations reached their peak in




the first half of July.  In 1981  the increase  in  chlorophyll was more




or less uniform over the top 20 m until the  first of  July,  and thus




extended through the epilimion into the metalimnicn.   Subsequent to the




first of July, increases continued in the 5-10 m  stratum,  but not in




the upper 5 m.  Since the 5-10 m  stratum incorporated  the lower half of




the epilimnion and the upper meter cr two of the  metalimnion, the




annual maximum chlorophyll may have been partially metalimnetic.




Metalimnetic maxima are common in stratified lakes (e.g.,  Fee 1976,




Heaney and Tailing 1980, ?-k>ss 1972), since light  often penetrates




through the epilimnion and into the metalimnion in sufficient amounts




to support phytoplankton growth.  Often nutrients  are  depleted first




from the epilimnion, but growth continues in the  metalimnion for  some




time subsequent to the epilimnetic depletion,  thus producing a




metalimnetic maximum.  The sunlight reaching Dillon en a  clear day is




about 50% greater than at the same latitude at sea level  (Hutchinson




1957).  This and the minimal cloud cover over  the lake tend to promote




metalimnetic growth.




     In 1982 the summer chlorophyll peak coincided almost  exactly with




the timing of the similar peak in 1981, but the chlorophyll maximum was




more evenly distributed over the  top 15 m.  As in 1981,  phytoplankton




were present in quantity well below the epilimnion, to a  depth of about




20 m.   Both the transparency data (Figure 16)  and  the  primary




production data (see below) indicate that growth  was mainly

-------
confined to the upper  10 m, however,  so  the  cells  below this had




probably grown in the  upper 10 a  and  had  sunk  to greater depths.  The




more even vertical distribution over  the  top 10 m  in  1982 is probably




explained by the larger runoff in 1982  and  the tendency of runoff to




enter higher in the water  column  in  1982.   In  1982 significant amounts




of nutrients continued to  enter the mixed  layer well  after the first




tvo weeks of June (Figure  12, Figure  19),  and  this nay have retarded




the occurrence of nutrient depletion  in  the  uppermost  portion of the




water column.




     The size of the chlorophyll  maximum  was very  sinilar in the two




years.  The maximum chlorophyll per unit  volume was slightly higher  in




1981 than in 1982, but the maximum chlorophyll at  the  lake surface




(0-5 m) and the distribution of the maximum  chlorophyll levels over




depth were both greater in 1982 than  in  1981.  Thus ths surface




transparency of the lake was considerably  less during  July of 1982  than




in July of 1981, despite the similarity  in maximum chlorophyll




concentrations of the  two  years.   Because  1981 was a much more unusual




year hydrologically than 1932, we  would view the 1982  condition as  the




more typical.




     In both years a chlorophyll  decline  began in  the  middle of July.




Ve attribute this decline  to nutrient deficiency-   In  both years July




was the month when available phosphorus and  available  nitrogen were




first reduced to their suirmer minimum levels (Figures  22, 23).  The




disappearance cf critical macronutriants  is  only circumstantial




evidence of the nutrient status of phytcplankton,  however, because  of

-------
                                                                       120

Maximum chlorophyll concentration
on any day or depth (ug/1)
>?aximum summer chlorophyll concentration
on any day, surface (0-5 ra, ug/1)
Maximum summer chlorophyll concentration
nonthly average, surface (0-5 m, ug/1)
Average chlorophyll, sunmer, 0-5 m (post
runoff stratification)
1981 1932
Amount Time Amount Time
19.8 13 July 17.9 19 July
9.7 10 Aug 17. A l£. july
9.3 June 11.7 July
6.7 Jul-Oct 7.3 Jul-Oct
Table 23.  Information on chlorophyll in 1981 and 1982.

-------
the ability of phytoplankton  to  store  significant  amounts of




macronutrients (Lund 1965).   The hypothesis  that  the July decline in




phytoplankton biomass represents the effect  of  nutrient  stress will be




further supported by other types of evidence, especially enrichment




studies, to be presented  in later  parts  of  this report.




     The decline of chlorophyll  beginning in aid-July was different in




the two years.  In 1981 the decline of chlorophyll concentrations was




steady at all depzhs.  In 1982, however, the decline was steady in the




top 10 3i but there was a  very deep maximum of chlorophyll centered at




about 25 in in the first half  of  August.  The transparency data for the




sane time interval show that  significant photosynthesis  cannot have




occurred so deep in the water column (1% light  at  10-15  m:  Figure 16).




We therefore conclude that the 25-in maximum  represents sinking of




viable but nongrowing phytoplankton cells that  formerly  belonged to the




surface maximum.  In fact this sequence  of events  is suggested by the




shape of the curves on the time-depth diagram in  Figure  29.




     In the first half of September 1981 and 1982,  there was a




chlorophyll minimum.  As  in the  spring,  the  minimum  had  a dual cause.




There was a steady decline in chlorophyll until the  end  of  August prior




to mixing.  This was due  to severe nutrient  depletion, as will be snown




in the analysis of production and  enrichment studies. The  quick




thickening of the mixed layer in September accentuated this  ninimuin by




dilution.  The dilution renewed the nutrient supply,  and thus led to =




quick resurgence of chlorophyll from the minimum.   As the epilimnion




continued to thicken toward complete nixing  in  November, growth was

-------
                                                                       122
sustained by continual incorporation of available  nutrients  from deep




water, and this explains the steady increase in chlorophyll  after




mid-September-




     The increase in chlorophyll after mid-September  was  caused by




relief of nutrient deficiency, but this relief was  accompanied  by




declining light availability.  Available light (I,  effective light




climate) would have reached a minimum with complete mixing in November.




In fact there was a definite slowing of the chlorophyll increase after




October.  If the entire water column were mixing day  after day;




T could not have exceeded 30 ly/day, which is near  the  threshold of




growth.  Only strong winds will mix such a deep water column,




especially in the daytime.  Thus the mixing can be  sporadic  even when




water column stability is essentially zero.  This phenomenon probably




allowed growth in the upper 15 m.  Algae in deeper water  did not grow




but remained viable.  The combination of growth in  the upper water




column and viability below under sporadic mixing allowed  chlorophyll  to




increase even after the water column lost thermal structure,  especially




in 1932.









Phytoplankton Composition and Seasonality




     The phytoplankton composition of Lake Dillon has never  been




thoroughly studied in the past.  The present analysis of  phytoplankton




composition and seasonality in Lake Dillon is based on examination and




quantitative counts from sedimented whole samples,  in which




phytoplankton cells of all sizes were visible.

-------
     Table 24 summarizes  the composition  of  the  phytoplankton,  and




Figure 30 shows some of the more common taxa.  P.are  species  are omitted




from the table, and the remaining  species  are  divided  into  categories




according to their abundance.   The dominant  phytoplankton taxa  are from




three groups:  blue-green algae, diatoms,  and  microflagellates.




     By far the raost important  blue-green, and possibly  the  nost




important alga in the lake on an annual basis, is  Synecnococcus




lineare.  This species is often referred  to  as Rhabdcderma  liaeare in




the literature, but has recently been  renamed  by Komarek (1976).  The




cells of Synechococcus are very small; in  Lake Dillon, their  diameter




is very uniform at about  1 urn.  The cell  length  is more  variable,  but




averages about 15 um.  The cells are solitary  or joined  in  pairs or




units of 3 or 4, particularly during rapid growth, probably  as  a




result of recent cell division.  Because  of  its  small  size,  this alga




is frequently overlooked  completely in quantitative  counts,  even when




it is quite abundant.  As a result, very  little  is known of  its




biology.




     Lyngbya limnetica, the second blue-green  that ranks as  a dominant




for Lake Dillon, was considerably  less abundant  than Synechococcus on




an annual basis but did reach the  top  two  species  in abundance  at




certain times.  This species of Lyngbya is characterized by  extremely




tiny elongate cells having dimensions  and  shape  very similar  to those




of Synechococcus.  The cells are united in filaments,  which  typically




range from 5 to 15 cells.  The  filaments  are surrounded  by  a  sheath,




which is characteristic of the  genus.  This  species  is extremely

-------
                                     100
Figure 3C.
Scale drawings of Dillon phytoplankton:  Lyngbya  (A),
Oocystis (S, G), Synechococcus (C, L) , aicroflagellates
(D, H), Synedra (£), Rhizosolenia (F), Asterionella  (I),
Mor.oraphidiuin (J), Rhodomonas (K).

-------
           Taxon                                   Abundance*
Cyanophyceae
     Synechococcus lineare                             D
     Aphanothece nidulans                              C
     Lyngbya limnetica                                 D
     Lyngbya so.                                       C

Chlorophyceae
     Monoraphidium contortum                           C
     Monoraphidium setifonne                           5
     Scenedesiaus granulates                            C
     Scenedesmijs ecornis                               S
     Oocystis parva                                    C
     Klrchnerialla obesa                               C
     Cosmarium sp.                                     S
     Coccotnyxa                                         C

Chrysophyceae
     Dinobryon petiolatum                              S
     Dinobryon divergens                               S

Bacillariophyceae
     Asterionella for^iosa                              D
     Nitzschia sp.             ,                        C
     Synedra radians                                   D
     Rhizosolenia eriensis                             D
     Stephanodiscus sp.                                C

Cryptophyceae
     Rhodomonas ainuta                                 C
     Cryptomonas erosa                                 C

Unclassified
     Microflagellates                                  D
* D = Dominant, i.e., ranking the too one or  two species  in  abur.dar.ee
  at certain times of the year; C = Common, i.e., ranking  in  the  top  5
  species in abundance at some tine of the year; S = Secondary, i.e.,
  reaching significant numbers but never ranking high  in  abundance.

Table 2^.  Suanary of the phytoplankton composition of Lake  Dillon,
           1981-1982.

-------
                                                                       126





widespread, spanning the tropics to far north  temperate  latitudes,  and




is found in waters of widely divergent trophic status.




     Neither of the two dominant blue-greens nor any  of  the  other




blue-greens listed in Table 24 is characterized by gas vacuole




formation, nor is any of these species capable of nitrogen  fixation.




Large-celled vacuolate forms such as Microcystis, Anabaena,  and  large,




Oscillatoria were entirely absent in 1981 and  1982.




     Three diatom species are classified as dominant  in  Table 24.




Asterionella formosa is a large-celled colonial diatom common in the




spring growth of temperate lakes in all parts  of the  world.   Synedra




radians, a solitary species, is also a common  contributor to the diatom




growth of lakes, and is distributed from the tropics  throughout  the




temperate zone.  Neither Asterionella nor Synedra is  easily  identified




with any particular trophic state.  Rhizosolenia eriensis, a major




contributor to diatom biomass in  1981, is a such more unusual diatom




dominant, although it is distributed in a wide variety of lakes  at  all




latitudes.  It nay often be overlooked because of its extremely




delicate frustule, which is almost invisible even under  phase optics.




The frustule is subject to dissolution even by relatively mild




preservatives.  In the preserved Lake Dillon samples, the frustules




disappeared within a period of three weeks.  The presence of the diatom




was detected by the examination of fresh samples and, once its presence




was known, counts could be made at any later time from the protoplast,




which resists dissolution.




     The third group of species contributing to the dominants of Lake




Dillon phytoplankton consists of unclassified microflagellates.   These

-------
are extremely sniall Chrysophyceae  and  Chlorophyta that are generally




recognized as important under  certain  conditions  in temperate lakes but




are virtually impossible  to  deal with  taxononically.   The largest of




these taxa in Dillon was  about 8 urn  in diameter  and the smallest was




just 1 urn in diameter.




     The Chlorophyceae, Chrysophyceae, and  Cryptophyceae were




represented in the 1981-1982 Lake  Dillon  phytoplankton but did r.ot




contribute any dominants.  Among the chlorophytes,  Monoraphidi'jn




(formerly called Ankistrodesmus) was a steady  cor.cributor, as were :he




genera Scenedesmus, Oocystis,  and  Kirchneriella.   Besmids were in




general not well represented,  although one  unidentified Cosmariun:




species did make a significant appearance.   A  tiny  spherical




chlorophyte, Coccomyxa, was present  at times.  This is a




widely-distributed contributor to  the  so-called  u-algae (Lund 1961),




which in Dillon would also include the microflagellates.  Two Dinobryon




species, both common contributors  to coldwater phytoplankton, were




observed frequently but never  became very common.  The genera




Rhodomonas and Cryptomcnas, which  are  almost universal contributors




to the phytoplankton of lakes  and  are  extremely broad  in their




tolerance of trophic and  thermal conditions, made significant




contributions but never reached dominance.




     It is typical of dimictic  lakes to show major  seasonal changes in




composition of phytoplankton,  and  Lake Dillon  is  no exception in this




respect.  It is also common for dimictic  lakes to show considerable




variation in phytoplankton composition from  one year  to the next,




especially if basic environmental  conditions are  ^arkedlv different.

-------
At the same time, there is typically a degree of repeatability in  the




seasonal sequence of composition from one year to the next  (Lund 1971,




1979).  Lake Dillon shows some major differences between  1981  and  1982,




yet certain features were common to the two years.  1981  and 1982




probably approximate the maximum expected difference between two years




in sequence because of the very unusual hydrologic conditions  in 1981




and the development of large amounts of phytoplankton under the ice  in




that year.




     Figures 31 and 32 are a simplified representation of the  major




changes in composition that occurred in 1981 and 1982 in  Lake  Dillon.




The major chlorophyll peaks are shown from the chlorophyll  time-depth




diagrams, and the major contributors to these peaks are indicated.




Significant but subdominant contributors are also indicated in




parentheses.




     The phytoplankton under ice in 1981 included large contributions




by the delicate diatom Rhizosolenia, whose average abundance in the  top




25 m reached levels of about 16,000 cells/cc.  The tiny blue-green




Synechococcus was also a major contributor numerically, although it  is




small and thus contributed less per cell to biomass (or to  chlorophyll)




than a number of other species (Table 25).  Synechococcus reached




abundances of about 6,000 cells/cc under ice in 1981.  Smaller but




significant contributions were made by the chlorophyte Kirchneriella




(about 1200 cells/cc) and the diatom Asterionella, whose  peak  numerical




abundance was much lower (500-600 cells/cc), but which contributed




substantially to the biomass because of its larger cell size.   Although

-------
in
rr
o
3
    Tl
    h^-
    Og



    ID


    U
  O cn
  O (-•-
  :-> U
  rt TJ
  i-( i J
  H- H.
  U' i r,
  C I '•
  ft f\>

  O
  3 ^
  cn ft>
s.
       M-
       o
Ul O
•  c
  'i
  fD
a>
cn
(u
in

(i>

O.

rt
(X>
  (M O
  fl) t""tl
  (-•
    rt
  ("t D"
  Cii fD

  b O

  H- hj
  a o
  IX 1-1
  H- O

  cu rr

  01 M
  a. t '
       ro
       I
    •a
    ft
  rt>
  o
  CJ
    U  U
    o
 0)
-*—
 CO

 E
         CL
         LiJ
         Q
                 0
                10
               20
               40
             50
                             	1	7—n	1	i	r	1	1	IT	1       r

                             PEAK  UNDER  ICE                JULY PEAK
                                             Synechococcus,

                                              Asterionella
                                                                   O
                                                                     /
                                                                                Synedro


                                                                               FALL  INCREASE
                      Rhizosolenia,  Synechococcus

                      (Asterionella, Kirchneriella)
Syriecococcus,  Synedro

(Rhizosolenio,  Asterionello)
                                    MAM
                                                           J      J

                                                           1981
                                                                   A
                     L	I	

                        S      0
	   I	

N      D

-------
 00
 c
CM
ft)
H-1
 o'
 C
 a
 Cu
o
(D
H-
p
H-

03
CD


LJ
      a>
      OL

      UJ
              0
            10
           20
      x   30
40
           50  -
          Astefionella, Lyngbya


          Microflagellates


          (Monoraphidium,  Scenedesmus)
                                            TTTT"
                                            JULY  PEAK
                                                          Synedra
                                         (Lyngbya,  Microflagellates)
               Synechococcus


                  Synedra


               (Stephonodiscus,


                Cocconriyxa)





               FALL INCREASE
                         F     M     A     M
                                         J     J


                                         1982
A
0    N
D
                                                                                                      o

-------
Taxon
Synechococcus
Lyngbya
Asterionella
Synedra
Stephanodiscus
Rhizosolenia
Coccomyxa
Micro flagellates
Call Volume
um
12
7
^00
3 CO
150
25
0.2
3
Ratio to
Synechoccccus

1.0
0.6
33.
25.
13.
2.1
C.017
0.3
Table 25.   Sizes of important Lake Dillon phytoplankcon in absolute
           terms and relative to Svnechococcus.

-------
                                                                      132




Asterionella is principally a spring diatom,  significant  growth under




ice is well known (e.g., Moss 1972).




     The disappearance of ice cover virtually eliminated  the  large




Rhizosolenia population, but actually increased the  Synechococcus




population to about 12,000 cells/cc and brought up the Asterionella




population to about 1700 cells/cc.  The phytoplankton were  of  course




more evenly distributed vertically at this time than they were under




ice.




     The next major event in the phytoplankton dynamics of  1981 was the




July peak, at which time the maximum annual chlorophyll levels were




observed.  In terms of cell numbers, Synechococcus was the  dominant at




this time (175,000 cells/cc).  Asterionella had also reached higher




populations (about 3500 cells/cc).  Synedra had overtaken Asterionella,




however, with an abundance of about 9000 cells/cc.  Rhizosolenia (about




8500 cells/cc) had rebounded to about half its density under ice.




     The July peak was followed by a general decline and  by a  more  even




distribution of abundance over depth as nutrient depletion  became




pronounced.  As indicated in the analysis of chlorophyll  data,  there




was a resurgence of phytoplankton following the thickening  of  the  mixed




layer and renewal of nutrients.  This resurgence was accounted for




principally by increases in the abundance of Synedra.




     In 1982 the large phytoplankton populations under ice  did not




appear as they did in 1981.  The small populations that did develop




under ice were dominated numerically by microflagellates, which reached




abundances of about 2000 cells/cc, and by Lyngbya, which  reached




similar abundances.  Asterionella reached abundances cf about  1000

-------
cells/cc, which was quite  significant  in  terns  of  oercent bionass




because of the large cell  size  of  Asterionella.   The  chlorophytas




Monoraphidium and Scenedesinus also reached  abundances of about 1000




cells/cc.  The very large  abundances of Rhizosolenia  and Syr.echccoccus




that were observed under ice  in 1981 did  not  sake  their appearance in




1982.




     In 1982 the very  large  sunnier population of  Synecnococcus that




was so evident in 1981 was not  repeated,  but  the  large  sunmer growth of




Synedra that appeared  in 1981 was  repeated  in 1932  (18,000 cells/cc).




Other diatoms made only minor contributions.  Microflagellates and




Lyngbya made contributions of about 4000  cells/cc  each, and several




species of chlorophytes made  contributions  of 1000 or 2000 cells/cc




each.  The fall growth was dominated by Synedra  (6500 cell/cc) and




Synecnococcus (46,000  cells/cc), which appeared  in  large numbers  after




being rare throughout  the  earlier  part of the year.   Also appearing in




the 1982 fall growth were  large numbers of  Stephanodiscus (3500




cells/cc) and Coccony:
-------
                                                                      134
top 5 m of Lake Dillon over the two years of study.   Information  of




this type is sometimes useful in indicating the nutritional  status of




phytoplankton.  There are a number of complicating factors  that limit




the usefulness of the data for nutritional interpretations  in  any




particular situation, however-  First, the seston is  divided in an




indeterminate way between living and nonliving fractions.   The living




fraction in Dillon is principally composed of phytoplankton  (bacteria




and protozoa are minor contributors on a weight basis).  The nonliving




fraction is composed of detrital organic matter and inorganic




particles, principally silt or frustules from dead diatoms.  When the




phytoplankton contains a large contingent of diatoms, as is  the case




for Lake Dillon, the situation is further complicated by the presence




of heavy and largely inorganic frustules around each  living diatom




cell.  These frustules contribute 30 to 50% of the dry weight  of  diatom




cells, depending on season and species composition (e.g., Sailey-Watts




1973).




     The percentage of phosphorus and nitrogen in the ash-free dry




weight of phytoplankton organic matter or, equivalently, the ratios of




phosphorus and nitrogen to carbon in phytoplankton bioaass, are useful




indicators of the nutritional state of phytoplankton  (Nalewajko and




Lean 1980),  Under certain conditions when the ratio  of phytoplankton




to nonliving particles is quite high, field data on these ratios  are  a




good indication of the state of the phytoplankton in  a lake.  However,




the presence of significant amounts of debris, particularly organic




debris, confuses the interpretation because the ratio of elements in




phytoplankton and in organic debris is likely to be very different,

-------
thus distorting the apparent ratios for  phytoplankton.   Microscopic




examination of the sedimented phytoplankton  samples  showed  that Lake




Dillon contained organic debris  at virtually all  times  of  the year,




ranging from a minimum of 10 to  20% of total particulates  under ice or




at the height of stratification  to 30 or 90% during  spring  mixing,




spring runoff, and fall mixing.   Thus the  annual  averages  for element




ratios would be virtually meaningless.   It  is  worthwhile,  however,  to




examine the element ratios for portions  of  the year  when detritus




contribution to filterable organic matter  is known  to be least.  The




appropriate times include the period of  ice  cover and the  interval  frcm




the end of the first week in July, when  runoff has ceased  to  affect




surface water, to the middle of  September.




     The percentage of carbon in  particulate materials  filtered from




the top 5 m of Lake Dillon varied from a sharp minimum  of  5%  at the




peak of the 1982 runoff to .a maximum of  about  40% under ice during  1931




and 1982.  The ratio of carbon to dry weight for  phytcplanktcn, in  the




absence of diatom frustules, is  about 36%  (Stumm  and Morgan 1981)-




Carbon percentages between 35 and 40% on four  different occasions under




ice coincided with the presence  of large populations of nondiatomaccous




algae, as expected from the literature values  on  % C.   In  all other




seasons the percentage of carbon  fell well  below  this ,  and was




typically between 10 and 30%.  In 1981,  the  percentage  of  carbon was




typically higher than in 1982, probably  because of the  smaller  relative




contributions of diatoms in 1981  than in 1982.  The  sharp mini-urn of




percentage carbon at runoff in 1982 is attributable  to  the  inflow of




significant amounts of silt with  peak runoff.  The coincidence

-------
                                                                      136

low percentage carbon with low transparency confirms the interpretation

that has already been given concerning the control of transparency  by

inorganic participate material at this time of year.

     Table 26 shows phosphorus as a percentage of organic matter

(computed from carbon) for the two time spans during each year  when the

influence of nonliving material on the chemistry of seston was  likely

to have been lowest.  Percentage of phosphorus in phytoplankton

protoplasm as reported in the literature generally falls between  0.5

and 3.0% of dry weight (Parsons et al. 1977), although lower

concentrations are known and may sometimes be adequate to sustain high

growth rates (Lund 1970).  The observed ratios for Dillon were  highest

under ice, especially in 1982 when populations were small, and  lowest

during the post-runoff stratification.  This is to be expected  in view

of the greater likelihood that nutrient depletion during summer

stratification will be more extended and more severe than under ice
cover.
P as % N as %
organic organic
dry weight dry weight
1981
Ice Cover
Post-Runoff stratification
1982
Ice Cover
Post-Runoff stratification

1.0 13.8
0.8 8.1

1.4
0.6 4.9*
Chi. a as %
organic
dry weight

0.87
0.38

0.73
0.42
* with one unrealistically low value excluded
Table 26.  Composition of organic matter in the  top  5 m  of  Dillon at
           times when the contribution of non-phytoplankton sources  to
           organic matter was lowest.

-------
     Nitrogen percentages are also given  in  Table  26.   The  expected




percentages are between 4 and 9% for healthy phytoplankton  cells (Fogg




et al.  1973).  Three percent is considered  the  absolute zinimum for




phytoplankton growth (Fogg 1975).  Under  ice cover  during  1981 the




percentage was very high, perhaps unrealistically  so,  although




phytoplankton are known to have significant  storage  capabilities for




nitrogen (Fogg 1975) and are known to  exceed 10% N  or.  occasion (Lur.d




1965).   Too many data are missing for  the period of  ice cover  in 1982




to yield an estimate for that time.  During  post-runoff stratification




both years the percentages were lower  than under ice cover  in  1981,  and




fell within the expected range for healthy cells.   Significant changes




in nitrogen stores may have occurred during  the course  of  the




post-runoff stratification, but these  could  not be  deduced  from




individual determinations on field populations.  Aside  from the known




presence of some non-living material,  resolution of  changes over short




time intervals is not really possible  from individual  determinations of




element ratios because the degree of uncertainty attached  with any




given ratio determination.




     Chlorophyll a_ can be expected to  account  for 0.5  to 2% of




phytoplankton dry weight in most instances.   The median is  in  the




vicinity of I" (Wolk 1973, Taguchi 1976,  Gibson 197S,  Bailey-Watts




1978).   As shown in Table 26, the values  for Dillon  conform very well




to these expectations.  Under ice cover the  values  are  near the




midrange and in the post-runoff stratification  the  values  are  toward




the bottom of the range of expected values (O.i%).   low ratios cf




chlorophyll to organic matter are symptomatic of nitrogen  deficiency

-------
                                                                       138




(Fogg et al. 1973).  Nitrogen deficiency  is core  likely  in post-runoff




stratification than under ice, and  this may explain  the  difference




between the two seasons in  the ratio of chlorophyll  to carbon.  Other




factors, especially light adaptation, must also be considered,




however.




     The degree of certainty that can be  attached to  the ratio of




chlorophyll a to carbon is greater  than for any of the other  ratios,




since chlorophyll a_ is not likely to be affected  significantly by




nonphytoplankton organic matter.  For this reason, it is worthwhile to




examine trends in individual values over  the last half of the




stratification season.  Figure 33 shows the ratios of chlorophyll a to




carbon between the middle of July and the middle of  September for both




years.  In both years there was a -trend toward lower ratios as time




progressed.  This downward trend is suggestive of nutrient  stress,




specifically nitrogen deprivation, and is part of the circumstantial




evidence that concurs with the evidence of enrichment studies for




nitrogen deprivation provided by enrichment studies.

-------
    .80r
•?  .60
    .40
 o
i  .20
o
      0
             JULY
                                AUG
                                         S982
                                            198!
SEPT
Figure 33.  Chlorophyll _a as a percentage of total particr.
         weight.
                                        ulate dr'

-------
                                                                      140
       Photosynthesis and Oxygen Consumption in  the Water  Column









Photosynthesis




     The rate of photosynthesis, especially in relation  to light and




chlorophyll, is the most direct indicator of the physiological




condition of a phytopiankton community-  It is veil known  that  the




standing stock of phytopiankton, as measured by  chlorophyll




concentration for example, is not always a good  indicator  of  the vigor




of a phytopiankton population or phytopiankton community.   The  ratio of




photosynthesis to biomass tends to be highest when populations  are




first exposed to optimal growth conditions.  By  the time biomass has




reached its peak, the conditions for growth are  typically  less  optimal,




and the rate of photosynthesis may have actually declined.  Thus direct




measures of photosynthesis facilitate the interpretation of events  in




the phytopiankton community by giving information that cannot be




obtained from standing stock (chlorophyll) alone.




     As described in the Methods Section, primary production




measurements were made on Lake Dillon by the C-14 method.   These data




can be combined as needed with information on chlorophyll  and




phytopiankton composition on the same dates.  Four indexes  related  to




primary production will be of interest here.




     (1) Primary production per unit area.  This is a measure of the




total photosynthesis occurring under a unit area of the  lake  over an




entire day,  and is expressed as mgC/m'-./day.  Primary production per




unit area is affected mainly by the amount of light reaching  the




surtace on a given day, by the amount of nonphytoplankton  materials

-------
causing light extinction  in  the water  column,  by  the  size  of


phytoplankton populations, and by  the  physiological condition of


phytoplankton populations  (principally their  internal nutrient


inventories).


     (2) Column efficiency of photosynthesis.   This is  the ratio of


primary production per unit  area to  phocosynthetically  available


radiation (PAR) penetrating  the surface  (Lewis  1974,  Lewis and


Weibezahn 1976).  PAR is  the proportion  of  total  irradiance between


350-700 nin, estimated as  46% of total  irradiance.  Surface loss


(reflection) is estimated  as 10% (Tailing 1971).   Since  phytoplankton


biomass and sunlight can  boch be expressed  as  energy,  it would be


possible to express the column efficiency as a  true dinensionless


efficiency number, but, since there  is no particular  advantage in  this,

                                   f\
the units to be used here  are ngC/m~/langley.   This computation


removes most of the day-to-day variation caused by seasonal and


irregular differences in  daily insolation.  Main  causes  of variation  in


the column efficiency include extinction of light  in  the water column


by nonphytoplankton materials, size  of phytoplankton  populations,  and


physiological condition of the phytoplankton.   The estimate of coluisn


efficiency is more subject to error  under ice  than at  any  other  time  of


the year, since the effect of snow cover on light  penetration is


difficult to quantify.  For  present  purposes we have  used  an  absorption


coefficient for snow of 0.24 cm  .


     (3) Depth of the photosynthetic maximum.   Since  phytoplankton


populations ara typically not distributed evenly and  can show


inhibition of photosynthesis due to  high light  intensities near  the

-------
                                                                       142




surface In the middle of the day, the photosynthetic  maximum is usually




found somewhat below the surface.  The depth of  the maximum on a given




day can be approximated from the shape of the photosynthesis profile on




that day (Figure 34).  The depth of the photosynthetic maximum is




affected mainly by the amount of surface light,  its extinction with




depth, the vertical distribution of phytoplankton, and the  light




history of the phytoplankton (acclimation to high or  low intensities) .




     (A) Maximum photosynthesis per unit chlorophyll.  Photosynthesis




per unit time at the depth of maximum photosynthesis  is  divided by the




amount of chlorophyll at the same depth, yielding a measure  of the




photosynthesis per unit chlorophyll at that depth (mgC/mg chlorophyll




a_/hour), often represented as P^x1  Pmax ^s affected




principally by the physiological condition of the phytoplankton and by




temperature (Taguchi 1976, Harris 1978).  The Q-^Q (increase  caused




by 10°C temperature rise) of Pmax is about 2.0 (Harris 1978),  so




a 10°C rise in temperature will double Pmax if other  factors are




equal.  If temperature is taken into account, Pmax is the most




direct indicator of the nutrient status of phytoplankton (Ganf 1975,




Vollenweider 1965).  Pmax and the other three indexes for primary




production are presented in Figures 35 and 36.




     Production under ice during both winters was easily measurable, as




would be expected from the chlorophyll levels.   The absolute levels of




production per unit surface averaged only about  one-tenth the




production levels observed in the summer, however.  'This is  mainly




explained by low light availability under ice, since  the column




efficiency was not low.  Contrary to what would  be expected  from the

-------
    CL
    U
    Q
           0
                       PRODUCTION,   mg C / m:
Figure 34.  Production per  unit volume versus depth  for selected ti-es
          of 1982 as measured by  C-l-i uptake, shewing shapes cf
          ohotosvnthesis  curves.

-------
            J    F    M   A
                               J    J    A    S   0    N    D
                   ice
                        mix
stratified
mix
     2.0
        6 L


        2
CJ
 O
 en
 £
.8 -


A -


 0
                                    VI982
            J    F    M   A   M    J    J    ASOND

   Figure 35.  Primary production per unit area (above) and column
              efficiency of  photosynthesis (below).

-------
     0
LJ

Q
X
o

o
E
\
o
o>
E
     0
                                                            OND
                 ice
                                mix
stratified
mix
             i      i
               F    M    A    M    J     J    A
                                                              0
ri.rjre 36-  Cent- >?f raxi.^un photosy

           photcsyr.thesis per unit
                                   :tresis (a^ove

-------
much higher chlorophyll levels  of  1981,  the  production under ice in



1981 was higher than in 1982 only  on  the  first  date  of measure


(February 23).  This cannot be  explained  siniply by  the amounts of



incident sunlight on the dates  of  incubation, as  the column



efficiencies of photosynthesis  were similar  under ice in 1981 and 1932.



The higher chlorophyll levels of 1981 were thus not  paralleled by



higher production during the last  half  of the period of ice cover.  Cue


to thin ice, no data are available for  the first  two months of ice


cover, but it can be shown that high  levels  of  chlorophyll under ica in



1981 must have been caused primarily  by a burst of growth during


January and February.  The chlorophyll  present  at the time of first



sampling thus would represent the  accumulated bioinass from this earlier



growth which was declining by the  time  the first  sample was taken.



Consistent with this explanation,  the highest production figure under



ice was the first one in 1981 (23  February).  The total accumulation of



chlorophyll between January 1 and  the middle of March 1981 was at least

        **l
250 mg/m~.  From the ratio of chlorophyll to organic dry weight under



ice in 1981 (Table 26), we estimate that  this corresponded to 28 g of

                              O                          *)
phytoplankton dry weight per m~, or 11 g  of  carbon per m^.  Since



the period of synthesis extended over 85  days,  the average net



synthesis per day must have been about 120 mgC/ai-.   In view of the



observed decline by early March, synthesis during January and February


must have exceeded 120 mgC/in^/day  on  earlier dates.



     Tns pcax VaJ-ues were low under ice both years,  suggesting


nutrient limitation.  Nutrient limitation would almost certainly have



to be due to low amounts of available P,  since  M  was available in

-------
quantity under ice  (Figures  24,  25)•  When  light  is  available in such




low amounts, however, P;jjax must  be  interpreted  with  caution,




since light nay be  insufficient  to  saturate  photosynthesis.   The




consistency of apparent  surface  inhibition  (Figure  36)  indicates that




saturating intensities were  reached during  all  incubations,  however,




despite the low surface  intensities in winter-   Saturation at low




intensities is not  unusual for phytoplankton adapted  co  lew  light.




Saturation at intensities as  low as 30 uEinsteins/m  ,/sec are  known




(Harris 1979), and  the Dillon intensities even  under  snow exceeded




this.  We therefore conclude  that  the ?rita-A  values were  truly  low




under ice, even in  consideration of low  temperature,  and are  thus




indicative of P depletion.




     During the spring mixing, production per unit area  was




intermediate between production  of  the period of  ice  cover and




production of the period of  stratification.   Column  efficiency of




photosynthesis was  also  intermediate, but Pmax  was high.  Low




column efficiency is accounted for  by extreme vertical  dispersion of




the chlorophyll, which caused most  of the light  to be  taken  up by




nonchlorophyll absorbance (see Figure 17).   At  the sane  time,




Rr.ax was very high  because the phytoplankton, although  too




dispersed to harvest the light efficiently,  were  in  top  physiological




condition (no nutrient deprivation).  This  is expected  due to nutrient




redistribution with raixing.




     Production per unit area rose  shortly  after  the  onset of




stratification.  This is explained by increasing  amounts of  light due




to longer days, by  a larger  stock of chlorophyll, and  by reduced

-------
                                                                      148




vertical dispersion of chlorophyll after the cessation  of  mixing.




Column efficiency, however, decreased to a minimum in early  June of




both years.  This was a result of runoff, which temporarily  added  large




amounts of particulates capable of taking up most of the light  before




it reached phytoplankton chlorophyll.  The effect was much more




prolonged in 1982 because of higher runoff and shallower entry  of




runoff into the water column that year-  Pmax in 1981 declined




after the onset of stratification from its spring peak, suggesting  the




onset of some nutrient limitation very shortly after stratification




began.  In 1982 the June Pnax values were of similar levels,




although the decline was not documented because the lake could  not  be




sampled in May.




     In 1981, photosynthesis per unit area and column efficiency




climbed steadily until August, after which there was a  precipitous




decline.  Trie situation was very different in 1982, when the  rise of




both variables, which began much the sane way as in 1931,  leveled off




in June and began a slow decline in July.  The depth of maximum




fixation was much closer to the surface than in 1981 and ?„„,..
                                                           max



declined in parallel with total production.  Two factors explain the




difference between years:  nutrient depletion and interception  of light




by inorganic particulates.  Nutrient limitation bfcgan early  in




stratification both years, as shown by the decline in P-nax from




spring mixing.  In 1981, inorganic particulate loading  was snail after




June and passed into the water column below the producing  zone.  Low




chlorophyll at the surface, caused by nutrient depletion there, allowed




the chloropnyll peak to develop deeper in the water column where

-------
nutrients were more available.  This accounts  for  the  increasing depth


of maximum fixation.  In 1982, production  in deeper  water  was  net


possible because of lower transparency  caused  by much  higher


particulate input to the lake, and by the  proximity  of  this  input  to


the surface.  Surface nutrient depletion was thus  not  ccnpensated  by


downward shift in the depth of maximum  production  in 1932, and


production was consequently lower.


     In both years there was  a major thickening of the  epilimnicn  in


mid-September.  The immediate consequence  of this  was  to  raise


Pmax, because of the relief of nutrient depletion  by deepwatar


nutrients.  Production per unit area and column efficiency were  slower


to respond, since there was a dilution  effect  along  with  the


improvement of nutrient supply.  Sy the end of September,  however,


these variables also showed a peak reflecting  better nutrient  supply.


Biomass buildup led to a return cf lower Paax  and  a  downturn of


production in October.  Since nitrate was  abundant at  this time,


phosphorus depletion was the  probable cause of declining  ?-,ax-


     Among well-studied lakes exposed to similar climatic  conditions,


Lake Dillon has an annual production similar to that of Lake Erken,


Sweden (Nauwerck 1963).  Erken is considered mesotrophic.  In  general,


oligotrophic lakes are considered to be typified by  production  falling

               f\
below 300 mgC/m^/day during the growing season, and  mesotrophic  lakes

                                         r\
are typically between this and 1000 mgC/n~/day (Likens  1975).   Since

                                                               o
the average production of Lake Dillon was well above 300 o.gC/z-/cay


in 1981 and near 3CO ngC/n2/day in 1982, :he lake  -would be cost


reasonably classified as nesotrcphic on the basis  of production.

-------
                                                                      150





     The Pj^x values of Dillon are comparable  to  those  observed




in other lakes.  For example, Glooschenko  (1973)  found  that  Pmax




in Lake Huron varied between 0.5 and 3.5 in the ice-free  season; this




range is similar to Dillon's.




     The pattern of production in 1981 and 1982 was  essentially the




same:  spring maximum, early fall maximum, summer depression,  and  late




fall depression.  This is a common pattern for temperate  lakes




(diacmic:  Hutchinson 1967).  Despite the similarity  in pattern,




the two years differed considerably in total production.   The  lower




production in 1982 was caused by earlier termination  of the  spring rise




in production, which has already been explained in terms  of  nutrient




depletion and inorganic turbidity.  The difference between years can  be




traced back to hydrology.  Since 1981 was exceptional hydrologically,




1982 should be viewed as the more typical of the  two  years.




     Table 27 compares the primary production of  Lake Dillon with  other




lakes of the Colorado Rockies.  The comparisons cannot  be  considered




very definitive, since complete annual data are available  only for




Dillon.  The comparisons are based on single measurements  for  a given




day within two days of 31 August on the year indicated  in  the  table.




Since the comparative data were taken in the mid-1960's,  some  of the




production values may now be higher at the same time  of year as a




result of additional nutrient loading.  The table indicates  that the




production of Dillon, at least in the early fall, is  within  the




mid-range of values to be expected in lakes and reservoirs of




comparable size.

-------
   Lake                    Year        ngC/m-/day


Green Mountain             1963           1C8

Carter                     1964           153

Dillon                     1 = 32           279

Cranby                     1?62           351

Grand                      1963           337

Dillon                     1931           402

Shadow Mountain            1963           459

Estes                      1964           463

Horsetooth                 1964           774
Table 27.  Comparison of primary production, as
     measured by C-1&, between Dillon and other Colorado
     lakes.  All production figures are for single
     measurements made within two days of August 31
     in the indicated year.  All data except for
     Dillon are from Nelson (1971).

-------
                                                                      152





Oxygen Concentrations and Oxygen Consumption




     In an oiigotrophic lake of moderate to great depth,  the  oxygen




concentrations are expected to remain near saturation  throughout the




year over the entire water column.  Since temperature  affects




saturation, there will be some seasonal variation in oxygen




concentration at the surface, even in the most oiigotrophic lakes,  but




the concentration will always be near saturation at the  prevailing




temperature.  In lakes of high trophic status, oxygen  concentrations




well above saturation occur in the euphotic zone due to  the steady




supply of oxygen from photosynthesis, and significant  oxygen  depletion




develops in deep water due to the steady consumption of  oxygen  by




respiration.  Deviation of oxygen concentration between  surface and




deep water is greatest under ice or during summer stratification when




vertical exchange is minimal.  Mesotrophic lakes of comparable  size and




depth are intermediate in degree of surface supersaturation and




deepwater depletion.




     Oxygen depletion is of direct interest because of its  significance




to bottom fauna and deepwater organisms, including fishes,  and  because




of the major change in internal nutrient flux that occurs when  the




oxidized microzone at the mud surface becomes reduced  and thus  much




nore permeable to the chemical constituents of the interstitial waters




(Hutchinson 1957, Golterman 1976).  The deepwater oxygen concentrations




of a lake are also of indirect interest as an indicator  of  trophic




status,  provided that correct compensation is made for morphometric




variation.

-------
     The oxygen concentrations of Lake Dillon  in  19S1  and  1932 are




shown in the time-depth diagrams of Figure  37.  Table  28 shows the




oxygen concentrations corresponding to saturation at  the elevation of




Lake Dillon at various temperatures.   It  is  evident  from the  table and




from the time-depth diagrams that the  surface  waters  of  Lake  Dillon




were near saturation throughout 1981-1982.   Production was  never  high




enough to hold oxygen concentrations much above saturation.   The  most




extended surface deviations from saturation  occurred  under  ice,  whan




atmospheric exchange was  impeded.  Surface  oxygen concentration  was




slightly above saturation under ice in February of 1981.  This was the




result of the unusual winter burst of  production  in  January  and




February, when the ice was clear and free of snow that year.   Other




instances of supersaturation were very minor.  There were  also a  few




instances of subsaturation oxygen concentrations  near  the  surface.  A




significant depression of surface oxygen  concentration occurred  in




September 1981.  This was caused by thickening of the  mixed  layer and




incorporation of deeper,  less oxygenated  water with  surface  water-




Sept ember suppression of  surface oxygen due  to nixing  in 1982 was also




detectable, but was considerably smaller  because  the  speed  of




incorporation of deeper water was slower  in  September  of 1982 than in




September of 1981.  Complete mixing in the  first  week  of November




lowered the oxygen concentration at the surface also  (not  enough  to be




evident in Figure 37a, 37b).

-------
'I]
H-
fra
H
H-
S3
ro

Cb

-------
'] I
 c
 M
 IT)
M
h*.
11

 I
TJ
rl
o
t-t
BJ

as

u

I- '

o
      U>
      h-
      tL
      tiJ
             0
                           ICf;
           10  -
           20
           30
           40
           50
stratified
                  J     F    M     A    M     J     J
   A     S     0    N
                                                   1982

-------
                                                                      156
Oxvgen, i5T>m
Temperature, °C
0
5
10
15
5
50
57
64
72
6
59
68
77
86
7
69
79
90
101
8
79
91
103
115
Q
89
102
116
129
10
99
113
128
144
Table 28.  Percent saturation corresponding to various  temperatures
           and oxygen concentrations at the elevation of Lake  Dillon.
     Dissolved oxygen concentrations in deep water did not  remain  near

saturation throughout the year.  Saturation in deep water was  achieved

during spring and fall mixing, so that the water column entered  the

period of winter ice cover or the period of summer stratification  with

no oxygen deficit.  There was significant depletion of oxygen  in deep

water both under ice and during summer stratification.  The bottom

10 meters of the water column under ice declined to a winter minimum

between 6 and 7 ppm, as compared with a saturation concentration of

about 9 ppm at the prevailing temperatures.  Deepwater oxygen  depletion

during summer stratification was considerably more severe,  partly

because of the greater length of time over which depletion  could occur

and partly because of more rapid supply of organic materials subject  to

decay during the summer months.  Oxygen concentrations 5 m  above the

bottom declined to 4.4 ppm in 1981 and to 4.6 ppm in 1982.  Perhaps

even more significant than the minimum levels in the deepest water was

the great vertical extent over which notable oxygen depletion  occurred.

The bottom half of the water column was markedly affected in both

years.  Since deepwater oxygen concentrations at saturation would  have

-------
been about 9 ppm,  the  observed  oxygen levels were as low as 5C:« of




saturation.




     Salmonid  fishes and  many  species of invertebrates common to




coldwater oligotrophic lakes require  high oxygen concentrations.




Although the thresholds vary according to the circumstances and the




species, notable  stress begins  to  occur in some species in the vicinity




of  5 ppm (Alabaster and Lloyd  1980).   Thus Lake Dillon is experier_cir._




oxygen depletion  to a  degree sufficient to begin having biological




effects.  Major chemical  effects  are  typically not noticed until oxygen




is  completely  depleted at  the mud-water interface.




     Because oxygen depletion  is  so  important biologically, and because




it  is an indication of total biological activity in a lake, there have




been a number  of  efforts  to link  oxygen depletion quantitatively to the




productivity of lakes. Although  it  is generally accepted that the




linkage exists, the quantitative  complications are numerous.   First and




most obvious is the dependence  of  the degree of oxygen depletion on the




volume of oxygen-rich  water at  the beginning of the season during which




oxygen depletion  occurs.   This  so-called "norphometric effect"




(Hutchinson 1957)  is dependent  on  the ratio of the epilimnion volume to




hypolimnion volume in  a given lake.   To compensate for the different




ratios of layer volumes in various lakes, and thus achieve a  comparison




between lakes, the concept of the  areal hypolimnetic oxygen deficit




(AKOD) was introduced  (Strom 1951,  Hutchinson 1923).  The average




amount of oxygen  in the hypolimnion per unit surface of hypolinnicn is




computed by integration of oxygen  concentration with depth froz; the top




of the hypolimnion to  the lake  bottom.   The amount of oxygen  per unit

-------
area is plotted against  time and  the  slope  of  the  decline is determined




by linear regression  (Lasenby 1975).  The slope  is  the  AHOD,




expressed as mgO-j/m^/day-  The  terminology, which  is  dictated by




traditional usage, is somewhat  confusing in that AEOD is  not really a




deficit, but rather a depletion rate.




     Among different lakes, AHOD  is significantly  correlated with




trophic indicators such  as total  phosphorus, primary  production, and




transparency (Cornett and Rigler  1980).  These relationships are




potentially useful in predicting  the  degree of oxygen deficiency to be




expected in a given lake in response  to changes  in  trophic  status.




However, despite the corrections  for  major  differences  in layer volumes




and in the duration of stagnation that are inherent in  the  computation




of AHOD, a significant amount of  scatter does remain  in all of the




relationships between trophic indicators and AHOD.  Some  scatter is




removed by use of depth  in the  equations, but as yet  there  seesis to be




no simple method for reducing scatter beyond this  (Cornett  and Rigler




1980).




     The AHOD was calculated for Lake Dillon for the  19S1 and  1982




summer stagnations.  The decline of hypolimnetic 0? per unit area was




very linear both years,  thus the AHOD could be estimated  within narrow




limits.  In 1981 the AHOD was 710 mg02/in2/day (standard error,  40)




and in 1982 it was 630 mg02/m^/day (standard error, 37).   These are




considerably higher than the values for oligotrophic  lakes  given by




Lasenby (1975).

-------
     Three separate relationships were developed by  Ccrnect  and  Rigler




(1980) to predict AHCD from key trophic indicators.   Table  29  shows  the




information required to make the predictions and Table  30 shows  the




predicted and observed AHOD values for Dillon.  The  phosphorus equation




substantially underestimates AHCD for Dillon.  Dillon produces about




double the oxygen deficit that vould be expected from P concentrations.




This is due to the higher biological activity supported by  a given ?




concentration in Lake Dillon than in most lakes; this phenomenon will




be documented and explained in other sections of this report.  The




equation based on primary production is also a poor  predictor, at lease




for 1981 when hypolimnetic temperatures were exceptionally warm, but




this is not surprising since Cornett and Riglar found it to be the




least reliable of the relationships they tested.  Secchi depth produces




the best predictions.  1981 is underestimated by about  20% and 1982 is




estimated almost exactly-  Clearly the secchi depth  equation would be




the best choice for predicting AHOD in Dillon.

-------
                                                                               160
Year
1981
1982
Mean Total P
(0-5 m), ug/1
7.63
9,15
Mean Summer
Secchi depth (a) ,
25 June - 30 Set>t
3.14
2.78
Primary
Production
gC/m2/yr
110
57
Mean
Depth
m
24.1
24.1
Mean
Hypolianetic
Depth, m
23.7
23.7
Mean
Hypolimnetic
Temp., °C
8.9
6.9
Table 29.  Information required for prediction of AHOD by  the  equations  of Cornett and
           Rigler (1980).
                                                 Predicted AHOD, mgO /m"/day
          Year
Observed AHCD
       2
 mg02/m /day
                                                 Using Equation with  Indicated
                                                           Variables*
       Secchi, z
          Production,

            V TH
          1981

          1982
     710

     630
295

323
580

636
4841

 636
* P = mean total P concentration, z = mean depth, ZH = mean hypolimni-jn  depth,
T^ = volume weighted mean temperature of hypolimnion.

Table 30.  Observed areal hypolimnetic oxygen deficits in 1981 and  1982  compared  with
           those predicted from three separate equations developed  by  Cornett  and Rigler
           (1980).

-------
                      Nutrient Enrichment  Studies




     The purpose of a nutrient enrichment  study  is  to  determine  the




actual or potential limitation of phytoplanktcn  by  specific  nutrients.




The literature on nutrient enrichment contains a wide  variety  of




experimental designs and methods for such  studies.   This  variation in




methodology is in part explained by differences  in  the exact  purpose  of




the studies.  In general, nutrient enrichment studies  fail  into  one of




four categories:




     (1) Enrichment of laboratory algal cultures with  lakewater




filtrate.  For this type of enrichment, a  watsr  sample is removed  fros




the lake under study and is passed through a filter  sufficiently  fine




to remove the phytoplankton.  The water is then  added  to  standard  algal




cultures, along with a spike of phosphorus, nitrogen,  or  any other




individual nutrient or nutrient combination that might be of interest.




The response of the standard alga (often Salenastrum capricornutum in




the United States) is measured in terms of cell  number or chlorophyll.




The responses in comparison to the control (no nutrient additions) are




taken as an indication of the identity of  the potentially limiting




nutrient in the lake.  The special advantages of this  method are  that




it does not require extended field work and that it  provides a common




basis for the comparison of very different kinds of  lakes.  The




disadvantage is that standard laboratory cultures differ  considerably




from a natural algal assemblage.  For example, it well known that  algal




cultures, either through physiological acclimation or  genetic




adaptation, typically require much higher  amounts of nutrients to

-------
                                                                      162
produce significant growth than do field populations  (Nalewajko and




Lean 1980, Lund 1965).




     (2) Short-term in-situ enrichment.  Short-term enrichment  is




probably the most common type of enrichment study when  the  emphasis is




on the identity of actual (not potential) limiting nutrients  at a  given




time in a particular lake.  A sample is taken from the  growth zone and




placed in a number of beetles.  Nothing is added to certain bottles;




these serve as the control.  Other bottles are enriched with  individual




nutrients or nutrient mixtures.  All bottles are then inoculated with




C-14 and resuspended in the euphotic zone.  Production  is estimated




from C-14 uptake over a period of a faw hours.  In theory;  algae that




are limited by a given nutrient will show increased photosynthesis in




response to the addition of that nutrient.  Despite the extensive  use




of this method, a critical flaw in the underlying assumptions has




recently been demonstrated by Lean and Pick (1981).  For phosphorus,




and possibly for other nutrients, the response of nutrient-starved




cells over the first few hours after their exposure to  new  supplies of




the nutrient is the diversion of cellular resources to  uptake of the




nutrient rather than growth.  A few hours may therefore be  simply  too




short a tiae over which to measure the growth response  of phytoplankton




to the addition of a limiting nutrient.




     (3) In-situ enrichments lasting a few days.  This  type of




enrichment is identical to the short-term enrichment except that the




containers enclosing phytoplankton that have been subjected to  various




nutrient treatments are suspended at depth in the lake  for  one  to




several days instead of a few hours.  The growth response is  typically

-------
measured by change in chlorophyll with  respect  to  a control  rather than




by C-14, as the C-14 method is net valid  for  application  over  extended




time intervals.  The main disadvantage  to  this  method  is  the




artificiality of enclosure conditions as  compared  with open—water




conditions.  This disadvantage is conibacted  principally by the use of




large containers with soft sides rather than  small  bottles.




     (4) Long-term enclosure methods.   Very  large  corrals  or plastic




spheres are used in some instances for  datermination of nutrient




limitation.  Although these are exceedingly  useful  and informative




methods, they typically are far too expensive for  routine  application.




     Enrichment studies of Lake Dillon  in  1981  and  1982 were done by




the third method cited above, in-situ enrichment lasting a few days.




All enrichments were carried out at the index station. Cn the day of




each enrichment study, integrated samples  of  the top 5 m of  the water




column were removed from the lake and placed  in a  large chamber where




they could be homogenized.  These were  then  siphoned into  a  number of




soft plastic containers, each with a volume  of  10  1.   Certain  of  the




containers were left unaltered as a control.  Other containers were




enriched with KIUPO^ in amounts sufficient to increase the original




phosphorus by 100 ug/1 in the container.   Nitrate  enrichment was  by




means of KNOj to the 200 ug/1 level.  The  amoun". of enrichment was




intended to exceed any possible requirement of  either  N or ? over the




interval of incubation but to be below  the amount  that might have toxic




effects.  The number of replicate containers  for each  treatment and for




the control was typically three, and occasionally  four. All except one




enrichment study involved a nitrogen treatment, a  phcs-horus

-------
                                                                      164





treatment, a phosphorus plus nitrogen  treatment,  and  a control.  The




first enrichment study omitted the nitrogen  treatment.




     After the containers were filled  and  spiked  as needed,  they were




all resuspended at a depth of approximately  2 m in the  water column.




This depth corresponds very closely to the 'expected depth  of maximum




photosynthesis (Figure 36).  Because the containers were hooked to a




floating buoy, they moved up and down with the wave action at the lake




surface.  This and the incorporation of an air bubble  inside each of




the containers induced sufficient turbulence to keep  the algae inside




the containers in suspension.  The containers were left in place  for a




period of three to seven days, depending on  the time  of year-  When the




containers were removed from the lake, they were  taken  immediately to




the Snake River Wastewater Treatment Plant, where water from each




container was filtered for chlorophyll analysis as described in the




Methods section.  Some analyses were also done for other variables.




     Three enrichments were done in 1981.  In 1982 the  number of




enrichments was increased to seven because the 1981 data indicated a




shift in the limiting nutrient, and the timing of the shift  could only




be identified with certainty on the basis of more closely  spaced




enrichments.




     The chlorophyll concentrations for each enrichment were subjected




to analysis by one-way ANOVA in which the enrichment  treatment  served




as the variable of classification and the dependent variable was  the




final chlorophyll concentration.  The null hypothesis was  that




different enrichments were not significantly different  in  chlorophyll




from each other or from the control.

-------
     Table 31 summarizes the results  of  the  enrichment  studies.  The




one-way ANOVA tests showed that  the mean  response  to  the four




treatments was not homogeneous on  any one of the  ten  dates (P < 0.01).




In other words, the four treatments could be divided  into two or more




groups on any one of the ten dates.   In  order to  determine which cf the




treatments on any given date were  statistically distinct from each




other, the Student-Newman-Keuls  (SMK)  Multiple Range  Test vas applied




to the data-  Table 32 separates the  four treatments  into statistically




distinct groups according to the 3NK  Test.   The groups  are arranged in




the table in order of ascending  response  to  enrichment.




     Since there was always a statistically  significant  heterogeneity




among the four treatments, it is defensible  to identify  a limiting




nutrient for each of the ten dates.   This is accomplished by  the




combined use of Tables 31 and 32.  The limiting nutrient is here




operationally defined as the nutrient  producing greatest growth




response.  It is taken for granted that multiplicative  effects  nay




exist by which one nutrient interacts  with another  (Droop 1975).  The




degree of response to enrichment, which can  be expressed conveniently




as the ratio of the phosphorus plus nitrogen treatment  to the control,




varied considerably between dates, as  indicated in  the  last column of




Table 31.




     Certain features were common to  all  the  enrichment  studies.  The




? + N treatment always fell in the highest group of responses,  as  shown




in Table 31.  In six of the ten enrichments,  the response to  ?  T N was




higher than the response to either nutrient  alone.  Given the

-------
                                                                            166
Date
1981
20 July
24 August
23 September
1982
24 May
14 June
10 July
26 July
23 August
7 September
18 October

C

4.4
2.1
5,3

2.7
6.8
8.3
7.9
5.0
1.8
6.5
Treatment*
P N

5.7
2.2 5.7
5.8 5.0

3.9 3.2
11.7 6.8
17.1 8.9
7.1 11.9
4.9 12.0
2.0 2.3
13.4 6.8

P + N

20.5
20.6
11.1

3.8
11.1
20.5
39.0
25.4
2.6
14.5
Dominant
Limitation

N
N
P/N

P
P
P
N
N
P/N
P
Ratio
P+N/C

4.7
9.8
2.1

i /,
1.5
2.5
4.9
5.1
1.4
2.2
  C = Control, P = Phosphorus only, N = Nitrogen only,  P  -t- N =
  Phosphorus plus Nitrogen.

Table 31.  Chlorophyll concentrations (ug/1) for the  ten  enrichment  studies.

-------
Date
1981
20 July
24 August
28 September
1982
24 May
14 June
10 July
26 July
23 August
7 September
18 October
Group 1* Group 2 Group 2
(Lowest) (Highest)

c,
C, P,
C, P,

c,
c,
c,
c,
c,
c,
c,

P ? + N
N ? -r N
N ? + N

H P, N, r -r N
N P, P + N
N P P -r :;
P N ? -t- N
P N P ~ Y,
P P, N N, ? + N
N P, N + P
* C = Control,  N = Nitrogen only, ? = Phosphorus only, ? 1- M =
  Phosphorus plus Nitrogen

Table 32.    Enrichment treatments separated into statistically coherent
            groups based on the Student-Newman-Keuls Multiple Range
            Test (p < 0.05).  Groups are arranged frcn: left to right in
            order of increasing chlorophyll response.  A given
            treatment can belong to nore than one group if it is
            statistically inseparable from acre than one group.

-------
                                                                       163




relatively short duration of  the incubations,  such  a result suggests




that phosphorus and nitrogen  requirements are  closely balanced to the




supply of these nutrients, even though one or  the other  may be




identifiable at a particular  moment as the dominating control on




phytoplankton growth.  Another feature common  to the enrichment studies




is the appearance of the control (i.e., no enrichment) in  the lowest




statistical response group.   Since this is expected  given  the




assumptions of the studies, it suggests that the studies were




methodologically sound.




     The number of enrichments was much smaller in 1981  than  in 1982.




Where the enrichment studies  overlapped for the two  years,  however,  the




results are so similar as to  suggest a common  pattern.   In  July and




August of 1981 the response to phosphorus plus nitrogen was very strong




but the response to phosphorus only was very low or  nil.




Unfortunately, a nitrogen-only enrichment is not available  for 20 July




1981, but the low P response  and the high P -f  N response that year




suggest nitrogen limitation.  A much stronger  case can be made for




nitrogen limitation over the  comparable interval in  1982.   The 26 July




enrichment showed a definite nitrogen response, as did the  23 August




enrichment.  The 10 July enrichment, however,  showed  an unequivocal




response to phosphorus.  We therefore conclude that  there was a switch




fro:n phosphorus to nitrogen limitation in the  middle  of July  in 1982




and probably in 1981 also.




     Phosphorus limitation was unequivocal in  the June enrichment.  The




May enrichment, which coincided with spring mixing,  also produced a




definite phosphorus response, but the magnitude of the resoonse was

-------
much lower at this time, both  in  absolute  and  relative terns.   We




conclude that nutrient limitation was  net  pronounced  in May,  and was




possibly induced by confinement of  the samples in  the zone of  ideal




light for growth.  In both years, enrichment during  the fall  chickening




of the mixed layer produced  similar  results  to the  enrichments  at the




time of spring mixing.  The  absolute and relative magnitudes  of the




response were small, and the evidence  points toward  phosphorus




limitation more than nitrogen.  Under  field  conditions,  it is  likely




that neither element was immediately limiting  at  such times,  but




phosphorus was incipiently limiting and this was brought  out  by a few-




days of sample confinement.  The  October enrichment  in 1982 shows a




return to pronounced phosphorus control of phytoplankton  growth.




     The composite picture that emerges from Tables  31 and 32  is




summarized in Figure 38.  Although no  enrichments were done through the




ice, the presence of significant  inorganic nitrogen  under ice  and the




evidence of phosphorus limitation just before  ice formation and just




after disappearance of ice strongly  support  the  case  for  phosphorus




limitation under ice.  Thus  on a  calendar  year basis  the  period of




phosphorus limitation extended from  1  January  until  the  onset  of




nitrogen limitation, which appeared rather abruptly  in the middle of




July.  The coincidence of rhis switch  with the depletion  of nitrate is




shown in Figure 28.  Late summer  thickening of the mixed  layer  relieved




nitrogen limitation (mid September) .   September  can  be regarded as  a




month of transition, during which the  response co enrichment  was less




pronounced but indicative of phosphorus domination.   Unequivocal return

-------
                                                                     170
     240  -
o
 I
  X
O
                                                                       mix
                   F    M    A   M    J
A    S    0    N    D
  Figure 38.  Nutrient limitations as shewn by the  enrichment experiments
             superimposed  on the nitrate concentrations near the surface
             (0-5 m).

-------
to phosphorus limitation occurred  with  increased  thickening of the




mixed layer in October, leading  to  complete mixing  in November.




     The interpretation given  here  of  the  enrichment studies is




consistent with the inorganic  chemistry ar.d with  the primary production




data.  As shown in the previous  chapter, ?3ax>  an indicator cf




the nutritional status of phytoplankton, was highest during May  or




early June and in September.   These  ars periods when the  plankton were




abundantly supplied with nutrients,  and they coincide with  the minimal




responses to artificial enrichments.   ?max s'nows  evidence of




nutrient stress under ice cover  (?  depletion) and again  in  mid-summer




(N depletion).  In the fall  there  is again evidence  of phosphorus




deficiency in the ?max values.




     The enrichment studies  showed  that phytoplankton biomass  in Lake




Dillon was alternately controlled  by phosphorus and  nitrogen.




Switching of this type is known  from certain other  lakes  (Lake




Washington:  Edmondson 1972).  The  switching occurred only  during




summer stratification and appeared  to  be predictable in  its timing.   At




the beginning of stratification, the ratio of available  nitrogen to




available phosphorus was very  high.  This  was determined  principally by




the very high ratio of nitrogen  to  phosphorus in  the water  entering  the




lake.  After stratification  began,  however, the amour.t of incoming




water declined, and much of  it entered  the middle and deeper iayars  of




the lake, thus not greatly influencing  the surface  chemistry.




Biological processes then gradually changed the ratio of  available




phosphorus to available nitrogen in  the upper water  column.  From tre




organic nutrient chemistry,  ve believe  that the mechanism for  this

-------
                                                                      172
change has to do with the relative abilities of  phytoplankton to




use soluble organic nitrogen and soluble organic phosphorus.   As




stratification progressed, the relatively  large  amount  of  phosphorus




present in the soluble organic pool gradually disappeared,  presumably




because it vas taken up by the phytoplankton and microbes.   This was




not true of soluble organic nitrogen, which appeared  to  be  less




available to the plankton.  Because the phytoplankton and microbe cells




were able to divert more of the organic phosphorus to their metabolic




needs, the ratio of phosphorus to nitrogen inside the cells changed  to




such an extent that the phytoplankton become nitrogen limited.




Thickening of the mixed layer in September reintroduced  abundant




supplies of inorganic nitrogen, and thus returned the lake  to  the




phosphorus control that was typical of early stratification.




     The effectiveness of the phytoplankton in sequestering and  storing




phosphorus well beyond their needs is illustrated by  the particulate




phosphorus data that were taken as part of the enrichment studies.




These data are summarized in Table 33.  Particulate phosphorus differed




significantly among treatments on all dates (one-way  ANOVA, ?  <  0.01).




Phosphorus and phosphorus plus nitrogen treatments were  always lumped




together by the SNK criteria, and control and nitrogen treatments were




always lumped together by the same criteria.  This was true whether  or




not the dominant limitation was at a given time principally associated




with phosphorus or with nitrogen.  Thus the phytoplankton,  even  when




strongly nitrogen limited, took up substantial amounts of the  added




phosphorus.   This lack of linkage between the uptake  of  phosphorus and




the immediate nutritional requirement for phosphorus  illustrates the

-------
                       Particulate P - us/1

24
14
10
26
23
7
18
Date (1982)
May
June
July
July
August
September
October
C, X* ?, ? + X
5.0 17.1
11.9 33.2
6.2 30.2
3.9 2 '^ . ^
4.5 28.4
3.2 9.8
3.4 22.3
* C = Control, N = M added. ? = P added, P -f- X
  P + N added.

Table 33.  Particulate P concentrations in the
   enrichn:ent treatments.  C (control) and N
   (nitrogen treatment) are together because the
   test shows them to be consistently
   indistinguishable for particulate P.  The P
   (phosphorus) and P + N  (phosphorus plus nitrogen)
   treatments are together for the same reason.

-------
                                                                      174




significance of luxury consumption of phosphorus in the phytoplankton,




and the impossibility of evaluating phytoplankton nutrient status from




external nutrient conditions.

-------
               Horizontal Spatial  Variation  in  the  Lake




     \'~ to this T-oint most of  the  analysis has  dealt  with samples taken




from the index station in the  middle  of  the  lake.   The question arises




how typical the index station  is of the  lake as a whole.   This question




is answered here in two stages.  First,  an analysis is made of the




five-station heterogeneity series.  As described in the sections  en




scud-* design methods, this series  consisted  of  analyses cade or. a 3e:




of samples taken on 32 different dates at the four  main stations  and




the index station.  The top  (0-5 m) and  bottom  of the water column were




sampled at each station.  Since each  one of  the four  main stations was




located at the mouth of one  of the four  main arms of  the  lake, this




sample series gave information about  the degree of  variation to be




expected over the deep-water section  of  the  lake.




     The second stage of the analysis is based  on a treatment  of  the




14-station heterogeneity series.   This series included surface samples




(0-5 m) from 14 different sites taken on 10  different dates.  Unlike




the 5-station heterogeneity  series, these 1J stations extended over the




entire lake surface without  regard to depth,  and thus included




shallow-water as well as deep-water areas.









The 5-Station Heterogeneity  Study




     Samples taken at the same depth  at  multiple stations will always




show some degree of variation.  This variation  can  be divided  into




three components (Lewis 1980):  (1) error variance,  (2)  fixed




horizontal spatial variation,  and  (3) ephemeral  horizontal  soacial




variation.  Error variance is  that which would  be observed  in

-------
                                                                      176
replicates from the same site on a given date.   In effect  this  is




analytical error variance, although some handling variance  could be




included and, if the replicates were taken separately at the  same  site,




some true microscale patchiness might also be contributory  to  the  error




variance defined in this way.  In most cases, and certainly in  the case




of the variables of concern here, analytical error dominates  the




variance of replicates from a given site on a given date.




     Fixed spatial variation is associated with average differences




between stations.  For example, if a nearby shoreline or point  source




of nutrients continuously influences the chemistry of a particular




station, this would increase the variance of a set of samples including




that station.  In contrast, ephemeral spatial variation is  due  to




horizontal variation that is not stable over time.  There would be a




considerable difference among stations on any given date, but still no




difference in the average values for a given variable among the




different stations, if ephemeral variation were high and fixed




variation were nil.  Typically lakes can be expected to show  some




combination of fixed and ephemeral variation, but among these two




components ephemeral variation usually dominates except in  the  most




extreme cases of physical or chemical gradients.




     For the 5-station heterogeneity series, we first consider




fixed horizontal spatial variation, i.e., the possibility that  the five




stations were significantly different when averaged through time with




respect to their chemical or biological properties.  The analysis  is




framed as a statistical null hypothesis, which is that all  stations

-------
                                                                       177
have the same average  values.   The  hypothesis is tested by one-way




ANOVA.




     Table 34 shows  the mean  value  of 11 different variables for




samples taken in the 0-5 m  layer at each of the five stations.  The




table also shows the grand  mean for each variable over the 32 different




dates and five different stations and shows the outcome of the one-way




ANOVA testing for  significant  differences among stations.  A cursory




examination of the means for  the different stations shows that thev are
remarkably similar  to  each  other.   The  results of the one-way ANOVA




show that in all instances  the  small  amount  of variation that is




observed from one station  to  the  next is  statistically insignificant at




P a 0.05.  We therefore conclude  that,  from  the viewpoint of annual




averages, horizontal spatial  variation  at  the  surface over the deep




part of the lake, including the mouths  of  the  arms  of the lake, is




trivial.  This does not imply that  there  is  no variation between




stations on a given day (ephemeral  variation), but  it does indicate




that one of these stations  is as  good as  any other  in representing the




deep-water portion of  the lake  over the long run.




     The same analysis was  done for samples  taken at  the bottom of the




water column at each of the five  stations  and  the results are




summarized in Table 35.  The  list of  variables is slightly different




than for the surface samples  because  some  analyses  that  are meaningful




at the surface of the water column  are  not meaningful at the bottom




(e.g., secchi depth).  The  results  for  the bottom of  the water coluan




are tabulated separately because  they are  subject tc  a slightly




different kind of interpretation.   Because the bottom of the water

-------
                                                                               178
Station
Variable
Temperature, °C
Secchi , m
Conductance, umho/cm
?04-P, ug/.l
Total Soluble P, ug/1
Particulate P, ug/1
N03-N, ug/1
NH4-M, ug/1
Total soluble N, ug/1
Total particulates ,
mg/1
Chlorophyll a, ug/1
Index
(A)
9.5
3.0
161
1.2
3.4
5.1
80
16
287
2.9

8.3
Dillon
Bay (B)
10.2
2.6
162
1.0
3.2
5.1
71
12
292
3.1

8.0
Snake
Arm (C)
10.0
2.6
159
0.7
3.2
5.2
77
14
275
3.2

8.2
Blue Arm
(D)
10.2
2.7
162
0.9
2.9
4.9
69
13
265
3.0

8.6
Tennile
Arm (E)
10.3
2.7
162
1.0
4.0
5.0
69
13
258
2.8

8.0
Grand Significant
Mean Differences
10.0
2.7
161
1.0
3.3
5.1
73
14
275
3.0

8.2
none
none
none
none
none
none
none
none
none
none

none
Table 34.  Means for variables at the index station and 4 main stations on 32 sampling
           dates spread over the two-year study period (0-5 m layer).

-------

Variable
•emperature, °C
•onductance, umho/cm
•0,-P, ug/1
"4
otal Soluble P, ug/1
articulate P, ug/1
;o3-N, ug/1
•H^-N, ug/1
:'otal Soluble N, ug/1
'otal Particulates ,
!Dg/l

Index
(A)
4.3
179
1.0

3.0
3.1
201
18
287
1.8


Dillon
Bay (B)
5.1
167
1.3

3.7
7.8
132
25
327
4.9


Snake
Ana (C)
4.7
166
1.0

3.4
9.i
141
25
311
3.3

Statior
Bl ue Am
(D)
5.0
163
1.0

3.5
6.3
143
22
423
2.7

i
Tenmile Grand Signific^n;
Arc (E) Mean Dif^^rence^
5.0 i.3 none
171 17C none
1.2 1.10 r.cr.e

2.1 2.2- r.cr.e
8.6 7.0 yes
143 153 yes
25 22.0 none
351 2cO none
3.4 3.2 -'es

'able 35.   Means  for  variables at the index station and  4 main  stations  on 22 sampl:
          dates  spread  over the two-year study period  (bottom  of  water  colurn}.

-------
                                                                       180




column is found at different depths depending  on  the  station, a certain




amount of variation due to depth  is blended  into  the  true horizontal




variation between stations.  Thus while  the  bottom  samples at the index




station in the middle of summer were most likely  to have  been taken at




about 40 m, the bottom samples in Dillon Bay were more  likely to have




been taken at 20 n.  This potentially will cause  some variation due to




depth alone.  No attempt is made here to separate depth variation from




the variation due to horizontal patchiness,  since the focus  of




attention is on the nature of the environment  near  the  bottom of the




lake.




     Table 35 indicates that, despite the additional  possibility for




variation due to differences in depth between  stations, only three




variables show significant differences.  These  three  variables  are




nitrate nitrogen, total particulates, and particulate phosphorus.   In




the case of nitrate nitrogen, the variation  between stations is without




question caused by the much greater depth of the  water  column at the




index station.  Since nitrate nitrogen increased  with the depth during




stratification, as shown in the seasonal analysis of  data for the  index




station,  this is not an unexpected result.   The Student-Newman-Keuls




Multiple Range Test segregates the index station  from the other four




stations at the 5% probability level.  The results do indicate  the




significance of the rather large pool of nitrate  nitrogen very  deep in




the water column, as shown also by the higher  averages  for nitrate near




trie bottom at the index station.




     For total particulates and particulate  phosphorus, the  values for




the index station are very much lower than for other  stations.

-------
However, examination of  the  raw data  shews  "hat the higher averages at


the four main stations were  due er.tiraly  to  a handful of very high


values clustered in the  winter  of  1931.   Since there was verv little


water movement under ice in  the winter,  especially in the dry vear of


1981, these high values  were almost  certainly produced by overly close


approach of the sampler  to the  boctcm,  stirrir.g up sedimer.t and thus


inflating the total particulite -;r.d  ^?.:-tic;:late phosphorus value?.


This was not a problem in 1982.  Thus  ror tot"! particulates and for


particulate phosphorus,  we conclude  that  the average bottom values for


the four main stations in this  series  ar; spuriously high.


     Even though fixed horizontal  variation  is not significant either


at the surface or at the bottom of the  water column with the exception


of nitrate in deep water and the spurious cases of total particulate


and particulate phosphorus heterogeneity  in  deep water,  ephemeral


horizontal spatial heterogeneity may  still be significant.  A detailed


look at the magnitude of ephemeral spatial variance is possible by


separation of variance components.  The  separation will  only be done


for the surface sample series because  of  the much greater practical


significance of heterogeneity in the  growth  zone of the  lake.


     Total variance of surface  samples  at five stations  on a given date


(s*-^ can be broken down as  follows:


                       2     9   _i_   1
                      s"t =  s"e +  s~s

       1                              9
where s"^ is the error variance and  5% is the true variance
        c                              o


attributable to stations.  Since it has been shown that  the true


variance attributable to  stations  does not include any significant


contribution resulting from  fixed  (temporally stable)  differences

-------
                                                                       LS2
between stations, it follows  that s2s is entirely accounted  for by


ephemeral spatial variation.  Thus if separation of  the variance


components can be achieved, we will have a quantitative estimate of the


magnitude of the ephemeral spatial variation.


     The magnitude of error variance for the Dillon  samples  was


obtained as part of the quality assurance program by routine analysis


of replicates.  For each date and each variable, the total variance


among stations was obtained from the raw data and the replicate

                                                  i           ->
variance was subtracted from  this total, leaving s^-  From  s^s


we obtained the standard deviation attributable to ephemeral horizontal


spatial variation (sg).  We then calculated the coefficient  of


variation associated with this ephemeral spatial variation (ss/lz


' 100).  The mean values for  these statistics are reported in Table


36.


     The amount of ephemeral  spatial variation, both in absolute terras


(sg) and in relative terms (s/x" ' 100) is exceedingly small  for


all variables.  As might be expected, conductance, which is  a


biologically conservative property, shows the least  ephemeral spatial


variation, and some of the more biologically sensitive variables show


greater ephemeral spatial variation.  The overall conclusion is clear:


on the average, the degree of horizontal spatial variation over the


area covered by the main stations and the index station is very small


in relation to the absolute magnitude of the variable that is being


measured.   A measurement at any one of these stations on a given day


tails almost as much as there is to know about important chemical  and

-------
biological features cf the main  body  of  the  lake  and  the mouths of the




arms on that day-




     When the values of  ss or  the  coefficients  of variation




corresponding to sg are  plotted  against  time,  some trends are




evident.  For all of the variables, there  were  times  of  the year when




the ephemeral spatial variation  was essentially 0,  and  other tines of




the year when the ephemeral  spatial variation was very  definitely above




0, although still small  in absolute terms.  Table 36  incorporates a




summary for each variable of  the time spans  when  the  ephemeral  spatial




variation was maximal in relative  terms.   Certain patterns  are  evident




from the list of maxima  and  from the  raw data themselves.  First, all




variables with the exception  of  total soluble N showed  maxima in




relative variation under ice  or  over  a time  period that  overlaps with




ice cover and early spring mixing.  The nearly  universal tendency




toward maxima of relative heterogeneity under ice is  explained  by the




physical condition of the lake at  this time. Since the  lake surface is




not exposed to wind action,  any  heterogeneity that  is induced by




outside influences, or by biological  phenomena,  is  much  more likely to




persist than it would be in  the  ice-free condition  when  the surface




water can be homogenized by wind action.   Many  variables tended to show




a period of higher heterogeneity only toward the  er.d  of  ice cover, ar.d




this also makes sense in terms of  the physical  events that  were




occurring in the lake.  Whereas  watershed  influence on  the  lake was




•Vosolutely minimal from  the  formation of ice cover  in early January




into March, spring weather was accompanied by rapidly increasing

-------
Standard Deviation
Variable Among Stations*
Conductance, umho/cm
PO^-P, ug/1
Total Soluble P, ug/1
Particulate P, ug/1
N00-N, ug/1
NH4-N, ug/1
Total Soluble N, ug/1
Total Particulates, mg/1
Chlorophyll a, ug/1
1.9
0.20
0.43
0.26
4.4
1.5
33.
0.12
0.31
Coefficient of
Variation** among Tine of Peak
Stations (%)* Variation
1.2
8.3
11.6
4.5
6.4
5.9
8.7
3.4
3.9
April-May
Jan-April
Jan -July
Jan -April
August
March-April,
July
Jan -April
July
April-June
Jan-April,
July-August
* Error variance has been removed.
         ' 100.

Table 36.  Summary of variation among the 5 main
           of error (= analytical) variance, and
           for each variable.
stations (0-5 in), after removal
the months of peak variation

-------
external influence in  the  fora  of  snowmelt,  which at  first passed under




the ice cover, thus magnifying  the  sources of  heterogeneity.




     Another common tendency was for  relative  ephemeral spatial




variation to increase during late  summer.  This  was  true especially of




inorganic nitrogen and phosphorus  and  to  some  degree  also of




chlorophyll a_.  At this  tine of year  the  phytoplankton were under




extreme nutrient stress  so  that minor  physical phenomena could induce




measurable heterogeneity either in  the  amounts of soluble inorganic




nutrients or in the amount  of phytoplankton  chlorophyll.   It  is




important to note that, while these peaks of heterogeneity were




undoubtedly real, their  absolute magnitude was exceedingly small.  By




way of illustration, Figure 39  shows  the mean  chlorophyll values  for




the 5 stations over the  two-year period of study and  the accompanying




estimates of standard  deviation (ss).




     The general conclusion of  the  5—station heterogeneity study  is




that the main body of  the  lake, as  encompassed by the  index station and




the 5 main stations, can be treated as  a unit,  and  that any or.e station




en a given day is a good indicator  of  the condition of the entire




central region of the lake.









The 14-Station Heterogeneity Study




     The 14-station heterogeneity  study included not  only stations  ever




deep water but also stations near  the head of  each  one of the  bays , -2~;




a number of stations in the broad  Tertmile Am.   The locations  cf  these




stations have already been  shown (Figure 2).   The stations are




organized in Table 37 according to  location.   Table 37 also sumari;es

-------
                                                                     186
                                            stratified             mix
                                                          T	r
         J    F    M    A
                          J     J    A    S    0    N    D
rigure  39.
Means on various dates for chlorophyll for the 5-station
heterogeneity study.   For each mean,  the vertical  bar  shows
the standard deviation caused by true variation between
stations (error variance removed).

-------
l-.l-r  (Vnl.-r
     hi.ti-x  M,i( I MII  (A)
     2  Km  :,.  ol  In, lex  (I, S3)

III 1 Inn  H.iy
     Hi- iir  I 'I I Ion  (IS!)
     Mould  o(  B.iy  ( H)

Siuik c l< ( vt- r  Arm
     II.-U.I  of  nrw  (I.S2)
     ^kHllll  of  n nu  (C)

Hli if  Hiver  At iu
    II. -Ji.1  of  unu  (I.S4)
    Hoiilli  .. I  ,irm  (D)

Tc Mill f I f  A I 111
    H.',..!  ,<\   .11 in  i-.u.l  (
     MI mi D|   .Mm  KL-LI  i
    HI. hi )<• <>l   ji  .11  KOMI h  ( I.S7)
    Ml. I, I I.- 1,1   .1 ..... llll I h
    I.I l.i-l I '.on  11,1 y  t  LSI,)
    (loul h  "I  .urn  ( I  >
Secclil
(m)
2.
,S3) 2.
2.
2.
2.
2.
2.
2.
S'l) 2.
:;H) 2.
(I.S7) 2.
as1.) 2.
?.
2.
v .
.H Mini
8
5
J
4
2
4
8
6
0
1
4
i,
5
6
'•
-.-
Conductance
(umlio/cm)
164
158
163
167
163
158
164
163
177
20]
175
172
1 VI
] fi'i
1«,H

I'O. -P
4
(ug/1)
1.6
0.6
0.9
1 .2
0.8
0.6
0.8
0.8
1.7
0.7
0.8
0.8
0.9
0.7
0.9
NUM..
Total
Soluble
P (ug/1)
3.6
4.3
4.0
3.8
3.2
3.8
3.6
3.2
4.6
3.5
3.7
1-8
4.1
3.9
1.8
NiMH'
Particulate NO -N
5.2
5.4
6.1
5.2
5.7
5.3
6.8
5.4
7.5
7.5
6.0
6.2
6.4
5.1
6.0
llo ni!
105
93
76
75
86
83
78
89
53
132
83
83
78
82
86
None
NIIA-!J
12
11
15
15
14
12
34
13
82
30
18
12
15
11
20
Hone
Total
Soluble
N (ug/1)
227
225
348
251
235
252
256
231
434
361
215
251,
241
26U
267
ti in'
Part Iculatts
3.3
3.5
5.0
3.5
4.7
4.3
4.1
3.4
3.5
4.2
1. 1
3.4
3.1
2.9
3.7
None
Chlorophyll a
7
8
6
7
7
7
8
8
7
6
7
7
7
h
7
.3
.1
.7
.4
.3
.3
.6
.0
.1
2
.4
.2
.(>
.H
.3
Noiu
 I|(M ol   ] 4 -- si d I  I on  lit1 1 *.•
HI u.ly .
^i1 lit- 1 t y  tiltuly,  aliowl MJ-  I lie  nu-.mn  ovt-r  t lul  14  Hlallomi  (O-'i  ru)  lor  lite  10
                                                                                                                                                               ilates
                                                                                                                                                                                  t IILI  1 wo  yi_-.n Li

-------
                                                                      188





the mean values of each one of  the variables  for  each  of  the 14




stations over the 10 sampling dates covered by  the 14-station




heterogeneity study.  The data  for each variable  were  subjected to




analysis by one-way ANOVA according to a rationale identical to that




used in the 5-station heterogeneity study.  Once  again,  the  null




hypothesis was that the 14 stations are not significantly  different




from each other in the mean values for any given  variable.   In  no  case




could the null hypothesis be rejected at the  5% probability  level.  We




thus conclude that there are no significant differences in the  mean




values for any of these variables among the 14  stations, despite the




coverage of all the different arms and of both  shallow-water and deep-




water portions of the lake.




     Since the ability of a statistical test  to discriminate




differences is directly related to the amount of  data  available ,




extension of the 14-station heterogeneity series  to cover  30 or 40




different dates instead of 10 different dates might have shown  certain




significant differences between stations.  The  important point,




however, is that any such differences were so minor that they could not




be detected at all with a 10-date series, which would  have been




sensitive to consistent differences of any great magnitude between




stations.




     The raw data and the table of means are  suggestive of certain




pacterns that may be real, although they could  not be  verified




statistically.  For example,  it would appear  from Table 37 that the




Tenmile Arm averaged slightly higher in conductance than any of the




other arms of the lake or the main body of the  lake.   The higher

-------
conductance of incoming water  is  largely  offset  by efficient mixing,




however.  It is likely that a  more  extensive  data  set  would  be able to




show statistically that the Tenmile Arn has  a slightly higher




conductance on the average than  the rest  of  the  lake,  but the degree  of




difference is sufficiently small  that  biological consequences would re




minimal.




     Examination of  the raw data  also  indicates  the  tendency of




shallow-water stations to show higher  values  of  chlorophyll  and lower




transparencies during  late summer at the  time of most  aggravated




nutrient depletion (July-August).  However,  since  this pattern is




sustained for a very short time,  it is  difficult to  demonstrate




statistically and must be considered as suggestive rather than




definitive.  The pattern does  make  sense  biologically  insofar as there




will be slightly richer supplies  of inorganic nutrients  where there: is




more intimate mud-water contact,  as in  shallow-water areas,  and this




may induce a noticeable difference  at  times  when nutrient depletion is




extreme.




     Another pattern that is highly suggestive but cannot be proven




statistically is the tendency  of  the stations nearest  to wastewater




treatment plant outfalls to show  higher average  values for inorganic




nutrients, and especially for  ammonia.  In fact  it is  remarkable that




these stations, because of their  proximity to the  outfalls,  did not




show significant differences from the  other  stations cf  the  lake.




Examination of the raw data shows,  however,  chat the savage  effecz if,




although noticeable, relatively  small  in  magnitude and sporadic ir.




nature.  Vigorous vind-driven  circulation and variations ir.  the

-------
                                                                       190




effluents themselves caused the effect to be undetectable  more often




than not even at the stations closest to the discharges.




     The 14-station heterogeneity study argues very  strongly that the




lake, in terms of general chemistry, nutrient chemistry, and biological




variables, can be treated as a unit, and not as a  collection of




semi-isolated arms and bays that react at grossly  different  rates to




nutrient loading or seasonal events.  Although this  is  counterintuitive




to the apparent semi-isolated condition of the arras  and the  continuous




entry of water of different quality into each of the arms, it  is




obvious that mixing of the lake is sufficiently pronounced under  the




influence of the wind to homogenize the chemistry  and biological




variables very efficiently.  Perhaps this is not so  surprising, in view




of the relatively small size and high wind exposure  of  the lake.  Both




of these factors act against the development of pronounced or




persistent patchiness that would produce significant differences  in  the




arms or between stations.

-------
               Overview of Limnology  and  Trophic  Status









The Annual Cycle




     Figure 40 is a general  synopsis  of  the  annual  cycle based en the




foregoing chapters that have dealt  with  physical,  chemical,  and




biological variables in Lake Dillon.   The cycle  is  generalized for the




two years of record.  Layering  and  mixing patterns  are  the  key to




understanding chemical and biological  changes  through  the  year-  Since




layering and mixing are principally under the  control of seasonal




changes in air temperature and  solar  radiation,  the general  pattern




will be very much the same from one year  to  the next.  This  is well




illustrated by the similarity of layering and  mixing phenomena in 19?!




and 1982, which were meteorologically  quite  different in amount cf




runoff.




     The predictable duration and timing  of  ice  cover, spring  mixins,




fall mixing, and summer stratification enforce a  certain amount of




order in the other events in the lake.  We thus  expect  the  synopsis




shown in Figure 40 to have general  validity  even  if the  trophic scacus




of the lake should change in the future.   Chlorophyll maxima will




consistently occur in the month of  July,  and the  months  of  May and




September will show surface  chlorophyll minima.   The nutritional status




of phytoplankton will be most optimal  in  May and  September  and least




optimal in August.  Transparencies  will consistently be  low in June  cue




TO runoff and in July due to phytoplankton,  ar.d will be  highest ir.




  '~MSt because of low chlorophyll concentrations  resulting  frc-




r.utrient depletion at that time.

-------
                                                                      192
         M    A   M   J
                                                A   S   0    N   D
Layering }
Mixing
Temperature
Runoff
Transparency
Surface
Nitrate
Surface
Total P
Cniorophyli
Bottom 02
Primary
Production
P max
Nutrient
Limitation
	 1 	 ; 	 ; 	 1 	 1 	 i 	 1 	 1 ;
1= Increasing P= Extended Peak V= Variable
D= Decreasing M= Extended Minimum
L, 	 Strotificotion 	 *i

Ice Cover i Mixing! Stable Oescendinql Mixing
1 Thermociine iThermoclinel
M t t \ I a,P D !
1 !
M >t ! ^ 4P^ D r I M ,
1 1
i 1
V D M ! D 1 1
» i f > < > . » t j
1 1
P , , D > M , t ' L p ,
1
i
1
V 1 P D 1 1 V
-« «--< JXUX)
i
V r (Di ( 1 tP D I i 1 V
\ i
1
D P D i P I
1
I
M i DM ID
* * -1 * * l I
\ \
M ~i ! x D ! °
1
i '
Phosphorus Nitroqen Phosphorus
1 ^\.minimal ^^ '
1 limitation ,
	 : 	 1 	 1 	 1 	 1 i ! l ' ' :
J   F    M
                                  M
S   0   N   D
•igure  40.   Suraaary  of  seasonal  events,  generalized as a composite of
            1981  and  1982.

-------
     Despite the persistent annual  pattern,  certain features of the




annual cycle are subject to influence  by  irregular  variations,  and




others are subject to change  if  the nutrient supply to the lake should




change.  Runoff Is exceedingly variable from one  year  to the next and




it has certain direct influences  on the chemistry and  biologv cf the




lake.  It is runoff, not phytoplankton, that determines  the




transparency of the lake in June.   Higher runoff  will  thus neari a l<-wer




transparency in June.   In  fact,  under  present conditions,  the annual




minimum transparency in an average  year is  as likely to  occur ir. June




in connection with particulate loading as it is  to  occur in  July ir.




connection with the chlorophyll  maximum.




     Higher runoff will also  raise  the phosphorus peak that  occurs in




June and July, and this will  in  turn raise  the  peak and  average annual




total phosphorus, at least in the upper water column.    Counter to




intuition, the present  study  shows  that higher  runoff  than average is




likely to suppress primary production  even  though it raises  the pear.




total phosphorus concentration.   Inorganic  particulars  connected with




high amounts of runoff act against  higher production,  as they interfere




with use of light by the algae.   Higher runoff  also maintains a thir.rer




epilimnion and causes injection  of  the particulates closer to the lake




surface.  These factors enhance  the extinction  of light  by nonliving




particulates, thus discouraging  significant  production below about f c.




This accounts for the suppression of production  per unit area by




particulates .




     Biomass accumulation, as measured by chlorophyll  concentration; ,




is likely to be very sir.ilar  under  high and  low  runoff conciticr.s,

-------
                                                                      194
despite the difference in production.  The chlorophyll maximum in a




year of low runoff is likely to be situated well below the  surface,




while the maximum in a year of normal or above-normal runoff  will




probably include the upper 5m.  On the other hand, the total  amount  of




chlorophyll under a unit of lake surface will be pretty much  the  same




under the two conditions, if the two years of record are accurate




indicators.




     Under the present trophic condition, the phytoplankton of Lake




Dillon respond to enrichment with phosphorus plus nitrogen at  any time




of the year, although the degree of response is minimal in May and




September when the upper water column has been freshly enriched with




deep water containing supplies of available nutrients.  In 1981 and




1982, there was a critical switch in limiting nutrients from  phosphorus




to nitrogen in the middle of July, and this is probably a general




phenomenon for the lake in its present condition.  The phytoplankton




appear to use up the orthophosphate almost immediately after  the  onset




of stratification.  Growth continues nevertheless and is accompanied  by




steady decline in soluble organic phosphorus.  The algae at this  time




are phosphorus limited, and respond strongly to orthophosphate




addition.  However, they have not exhausted the phosphorus supply




because they are able to use soluble organic phosphorus, although they




obtain it more slowly and less efficiently than they would




orthophosphate and are thus somewhat suppressed in their growth rates




by the unavailability of orthophosphate.  In the middle of July,  growth




is essentially halted not by phosphorus but rather by inorganic




nitrogen, which is present in abundance at the beginning of

-------
stratification but  is  finally used  up by the middle of July.   The




effect of this nitrogen  exhaustion  is more  drastic  than the exhaustion




of crthophosphate,  implying  that  the  phytoplankton  are less able to use




the soluble organic nitrogen pool than to use the soluble organic




phosphorus pool.




     Although phosphorus  is  presently limiting over a much greater




segment of the year than  is  nitrogen, it is  inorganic nitrogen that




halts the progress  of  the summer  growth and  thus determines the maximum




amount of chlorophyll  in  the epilimnion.  Since the summer minimum of




transparency is a point  of major  practical  concern  resulting  from




eutrophication, it  is  fair to ask whether  the growth control  strategy




for algae should be based on nitrogen rather than phosphorus.   Aside




from economic considerations,  two biological factors should be weighed.




First, under the present  conditions,  the lake is very close to a




balance between phosphorus and  nitrogen limitation.  By the time




inorganic nitrogen  limitation occurs, phytoplankton have removed not




only all the orthophosphate  but virtually all of the soluble  organic




phosphorus as well.  Large responses  to enrichment  can be obtained only




with a combination of  phosphorus and  nitrogen.   Since the balance is so




close, the choice for  management  is essentially arbitrary and  car. be




based on economic considerations or convenience (cf-  OECD 1982).




     Another biological consideration is the possible encouragement of




nitrogen fixers if the present  nitrogen limitation  were exacerbated by




control of nitrogen input  with  the  same phosphorus  loading cr  gra-:er




phosphorus loading.  As the  balance is  shifted  more and -ore  strongly




toward nitrogen limitation,  there is  an increasing  ecologica_  ir.ce~.ci-.-5

-------
                                                                      196




for the development of algal taxa  that do not have high  nitrogen




requirements.  The algae that have lowest nitrogen requirements  are




those that can fix nitrogen, specifically the heterocystous




blue-greens such as Anabaena and Aphanizomenon.  Any  shift toward  the




dominance of such algae is usually considered undesirable, as  these




algae also have gas vacuoies which cause them to rise  to  the surface




and fora scums.  Thus while the July chlorophyll maximum  of  1982




gradually sank out of sight as shown in Figure 29, a  comparable  amount




of Anabaena biomass might well have come to the surface,  where it  would




have affected the appearance and surface transparency  of  the lake,




This and other considerations will be taken up again  later in




connection with predictions and modelling.









Trophic Status of the Lake




     The present trophic status of the lake is a matter of concern.




The trophic status of the lake can be evaluated on the basis of  any one




or a combination of several trophic indicators.  The most useful




indicators are those which can combine the practical  concerns of




eutrophication and biological fidelity.  Table 38 lists four such




indicators available for Lake Dillon.




     Lakes are traditionally classified as oligot::ophic, mesotrophic,




or eutrophic.  The assignment of a given lake to one of these




categories is to some degree arbitrary.  Using any single trophic




indicator, lakes could be arranged in a continuous spectrum  that lacks




sharp boundaries.  Furthermore, the use of several different trophic




indicators will typically provide slightly different perspectives  on

-------
                                                                      19;
     Indicator
Magnitude
Trophic Indication
Total Phosphorus (ug/1)
Spring surface maximum day
Spring surface maximum month
Annual average surface
July-October average 0-15 in
Transparency (m)
July secchi minimum day
July secchi average
July-October average
Chlorophyll a (ug/1)
Summer surface maximum day
Summer surface maximum month
July— October average
Primary Production
Annual (gC/mVyr)
Maximum daily (mgC/m^/day)
Summer oxygen
AHOD (mg02/m2/day)
5 m above bottom minimum (tng/1)
12-16 Lower mesotrophic or
11-13 upper oligotrophic
7-9
7-8
1.5-1.9 Mesotrophic
1.8-2.3
2.5-3.5
10-17 Mesotrophic
9-12
6-8
50-100 Mesotrophic
500-900
600-700 Lower eutrophic
4-5
Table 38.   Summary of trophic indicators for Dillon under conditions at
           the time of the study.  Range of magnitudes shews, in round
           figures, probable span with varying weather in different
           years .

-------
                                                                      198




the placement of a particular lake in a  trophic  sequence.   Thus  while




Table 38 attempts to assign a trophic status on  the basis  of  each of




the trophic indicators, principally on the basis of the  literature




survey by Welch (1980).  The categories  are of course  to some  degree




simply a reflection of tradition in classifying  lakes.




     Phosphorus is a widely used criterion for classifying  lakes




according to trophic status.  Although it has been recognized  for




decades that phosphorus is of great importance in determining  trophic




status, an attempt to make a quantitative linkage between  phosphorus




and trophic status dates back principally to Vollenweider  (1965).




Vollenweider, following Sawyer (1947), specified that  oligotrophic




lakes have total phosphorus concentrations below 10 ug/1,  that




mesotrophic lakes have total phosphorus  concentrations between 10 and




20 ug/1, and that eutrophic lakes have total phosphorus  concentrations




above 20 ug/1.  No simple rules are available according  to  which one




can select appropriate sampling times and the appropriate number of




samples for determination of total phosphorus, however,  Depending on




whether one uses annual or seasonal averages or seasonal peaks,  Dillon




might currently be classified as lower mesotrophic or  upper




oligotrophic according to Vollenweider*s criteria.  Furthermore,  it




should be remembered that Vollenweider's classification  has presently




been modified to take flushing rate into account, and  this  has not been




done for Table 38.  This matter will be  taken up more  fully along with




the model of Lake Dillon.




     Transparency, chlorophyll and primary production  are  also possible




indicators of trophic status.  If phosphorus is a major determinant of

-------
                                                                       ; QQ
trophic status, then transparency and  chlorophyll  are  sajor  synptons.


Secchi depths in Dillon have low values both  in  June and  July.   Table


38 gives the July secchi depth  as the  basis  for  judging  trophic  status


in this lake simply because the low  secchi depth values  in June  are due


to inorganic particulates rather than  chlorophyll.   Secchi values  in


July are low enough to bring the lake  into the mesotrophic category by


traditional criteria (Likens 1975, Welch  1980).  Similarly,  chlorophyll


£ maxima and growing season averages indicate mesotrophic  status.   In


fact the chlorophyll a_ values are higher  than in most  lakes  that have


comparable total phosphorus values (e.g., Kalff  and Knoechel  1975).


The reasons for this will be explored  in  connection with  the


development of the model.  Primary production, considering the short


growing season, is also in the  mesotrophic range.


     Areal hypolimnetic oxygen  deficit is also a potential indicator


but is less often used than the other  indicators of Table  23.  vcr:i-ei-

                                          ( ••)
(1941) set boundaries of 250 and 550 mgC^/iaVday dividing  the  three


trophic categories.  Thus according  to this  indicator, Lake  Dillon  is


just entering the eutrophic category.


     In summary, the best characterization of Lake Dillon  at  the


present time seems to be mesotrophic.  The lower mesotrophic  bcundarv


is an important one, as the visible changes  in the lake are most


notable to the casual observer  over  the mesotrophic range.  Lake Eillor.


has passed an important transition between trophic categories  and  car.


no Ic-ger be considered oligotrophic.

-------
                                                                      200





         Chemistry of Nutrient Sources as They Enter  the Lake









     Pathways followed by water and nutrients entering  Lake  Dillon can




be divided for present purposes into six categories:  (1) major  rivers,




(2) streams not joining a major river prior to reaching the  lake,  (3)




sewage effluents, (4) miscellaneous surface drainage  not accounted for




in other categories, (5) precipitation, and (6) groundwater.   All  cf




these categories except miscellaneous surface drainage  and groundwater,




which made minor contributions, were studied chemically at or  very near




their point of entry to the lake.  In the first category are the Snake




River, the Blue River, and Tenmile Creek.  These will take up  most of




the attention of the section, since they account for  the bulk  of




surface loading to the lake.  In the second category  are Soda  Creek and




Miner's Creek, both of which were sampled above the point sources  that




enter them near their mouths.  In the third category  are the effluents




of the Frisco WTP (which enters Miner's Creek near the lake)  and  Snake




River WWTP (which enters Soda Creek near the lake).   Both of these




effluents were sampled prior to entering streams.  Breckenridge




effluent was part of the Blue River sample, since  the effluent enters




above the river mouth, and Copper Mountain effluent was part of  the




Tenmile Creek sample for the same reason.  Discharge  data for  rivers




were obtained from U.S.G.S. data records.  For the effluents,  discharge




data were obtained from plant operators, and for Miner's Creek and Soda




"reek the data were based on our own measurements.  All chemistry  data




were from our own analyses.  The precipitation chemistry data  were




based on the collector near the Snake River WtfTP.

-------
     The chemistry of  the  nutrient  sources  that were sampled routir.elv




will be considered here.   Contributions  of  groundwater and




miscellaneous surface  drainage will be  approximated in the next




section, which gives total  transport, and  the  breakdown of surface




sources above their points  of entry to  the  lake will be made in a




subsequent section on  land  use in relation  to  nutrient yield.









24-hour Variation in River  Chemistry




     The mouths of the major rivers were sampled on 54 different dates




over the two-year period.   Since  the samples were taken mostly during




the daylight hours, a  question arises as to whether there  is any




regular periodicity in stream chemistry  that might  cause this  type of




sampling to introduce  bias.  Although no major periodicity was expected




a_ priori,  two 24-hour  studies were  undertaken  in order to  show whether




periodicities were of  any  concern.   A secondary purpose of these




studies was to show the degree of short-term irregular variation,  which




would contribute to random  noise  in the  data but would not bias the




results of a large data set.




     On 14 January 1981, samples were taken at each of the three river




mouths at  three-hour intervals beginning at 1600 and ending at the same




tine the next day.  These  samples were  analyzed for total  particulars,




ammonium,  nitrite plus nitrate, and orthophosphate.  The study was




repeated in Movember 1981 with a  slightly different sampling ir.iervs-




(four hours).  In che November sampling, analyses were mace for tots-




soluble phosphorus, total  soluble nitrogen, particulate phospr.crus , ir.i




total particulars.  In November 1981 the samples of the ~lue  ?_ver

-------
                                                                      202




were taken above die Breckenridge Wastewater Treatment  Plant,  and in




January 1981 they were taken below it.  In November 1981  the  Tenmile




Creek samples were taken above Frisco and the January  samples  were




taken below Frisco.




     The results of the two 24-hour studies are analyzed  here  in two




stages.  First, the data series for each variable at each river  mouth




on each date is tested for the existence of significant additional




variance above and beyond that which would be expected  from analytical




variance alone.  The presence of significant additional variance




indicates either irregular or patterned variation in the  stream,  or




some combination of these.  If additional variation is  present,  the




distinction between irregular and patterned variation will be  made on




the basis of subjective evaluation of pattern in the data, since the




time series is too short to allow statistical time series  analyses.




     Table 39 gives the mean for the variables over the 24-hour




sampling interval at each of the three sites on each of the two  dates.




In addition, the standard deviation of each series is  shown.   The ratio




of the square of the standard deviation to the error (= analytical)




variance is the F statistic, which is also given.  The  F  statistics  for




each set of analyses were compared with critical values for F  at  the




0.05 probability level (Rohlf and Sokal 1969).  On this basis, F values




indicating sample variance significantly in excess of  the  analytical




error variance were identified and have been marked with  asterisks in




the table.  In instances where the F statistic gives evidence  of




significant true variation in stream chemistry over the 24-hour  cycle,




the variation nay have been either patterned or irregular-

-------
                                 Snake  River
                             Mean
S.D.
                           Blue River
                    Mean
                      S.D.
                                                          Tenmlle  Creek
                                  Mean
                                   S.D.
January 1981

   I'ar t 1 on! a t es ,  wg/1
   Nil -N, ug/1

   NO,, I N03-tJ, ufi/l

   POA-P, ug/1
   Par Lieu]a re P,  ug/1

November I C)H1

   I'ar I I en 1 a I et;,  mp,/ 1

   Total  Soluble  N,  ug/1

   Tart leu I ate I',  uj',/1
                              1.46
                             10.3
                            324
                                .33
                              1 .71
1.40
6.4
8.0
 .61
1.13
6.32*
1.30
1.83
2.32
3.79*
                              1.62     .61    1.19
                            263      36.      0.21
                              1.1      .70    1.45
  0.98
434
577
  1.45
  3.6
                     1.84
                   118.
                     2.9
  0.44
223
417
  1.51
  0.53
                              0.53
                             39.
                                .2
   0.62
1570.*
4995.*
  14.*
   0.83
                                 0.90
                                 0.25
                                 0.11
                                                       1.11    1.84
                                                     495     81
                                                    1006
                                                        .77
                                                       2.63
85
  .68
 1.65
                            .79     -65
                         117.     20.
                           1.8      .6
 10.7*
209.*
207.*
  2.89*
  8.09*
                                                      1.34
                                                      0.06
                                                      1.07
* sample variance  fxct'eils error variance  significantly (p <  0.05)

T.-iMr 'I').   ^^1M^^;  ami  standard deviation  for  (he Lwn /••'i-liour  sampling  series,  and the F Hl.-iLisLtc  for  each
                                                                                                                  i )
                                                                                                                  o

-------
     Significant variation in stream chemistry above  the  error variance




did not occur for any of the analyses at any of the three rivers  in the




November sample series,"so we conclude without further  analysis that on




this date a sample taken at any time of day would have  represented




without significant bias the chemistry of the river from  which it was




taken.  In January, however, there was significant variation  in stream




chemistry over the 24-hour cycle for about two-thirds of  the




combinations of variables and collection sites.  Tenmile  Creek showed




significant variation for all variables.  The Blue River  showed




significant variation for three of the four variables and the  Snake




River showed significant variation for only one variable.




     Figure 41 depicts the stream chemistry values for  each combination




of sites and variables in the January series where there  was  evidence




of a significant variation in stream chemistry over the 24-hour




sampling period.  The data do not suggest any kind of repeated pattern.




For example, there was a definite minimum of ammonium on  the  Blue River




at 4 p.m. one day, but the next day the ammonium concentration was  high




at 4 p.m.  Although the data set is not large enough  to allow  rigorous




statistical analysis for pattern, we conclude with reasonable  certainty




that the added variance in each case is due to randomly timed




irregularities and not to predictable patterns in stream  chemistry.




This being the case, samples taken at any time of the day or night  will




represent the stream without bias, although a larger  sample series  is




required to yield a given confidence interval for transport when  added




variance is present.

-------
                                                                     205
               I  I  I   I  !  I  1   I  I  I   I  f  I  I
                                                                PM
Figure 41.   Data that showed statistical  evidence  of  significant diel
            variation during the study  of January  19S1  (SRI =  Snake River
            TMC1 = Tenmile Creek,  BR1 = Blue  River).

-------
                                                                      206

Chemistry of the Thrae Rivers as They Enter the Lake

     Table 40 summarizes the average concentrations of dissolved  and

suspended materials in the three rivers as they enter the  lake.

Averages can be obtained by unweighted averaging, time-weighted

averaging, or discharge-weighted averaging:

                                 _    n
          Unweighted averaging   C =  £ C./n
                                     1=1 "
                                    _    n       n
          Time weighted averaging   C =  £ C.t./ I t,
                                        1-1 X L i=l
                                         __    n       n
          Discharge weighted averaging   C =  I C.D./ I D
                                             1-1 X X 1-1
where C is mean concentration, C^ is the concentration at the

i*-^ tine interval, t^ is the length of the interval, and D^ is

the discharge over the interval.

     In the present case unweighted averaging is not very defensible

because the samples were taken more frequently at times of high

discharge when concentrations of dissolved and suspended constituents

were likely to be changing most rapidly.  A tine-weighted average would

give the best idea of the chemical composition of the rivers without

regard to seasonal changes in discharge, but the discharge-weighted

average is the best representation of the chemistry of the entire

annual flow accumulated in proportion to discharge, as it actually is

in the lake.  Thus Table 40 shows the discharge-weighted averages.

Tine-weighted averages might be useful for certain purposes, but are

not tabulated because they are in most instances very close to the

-------
1981                                                         1982

""., II, Ill'./l
MO -N, .if./ I
M *
Total SolnliU- N, iik'./l
Pun U nl jt c N. uK/l
I'O -1', uK /I
'lutji r.oiui.U' i\ UK /i
I'll I" t 1 . Ill Hi 0 I', llg/1
I'.M 11 1 I'.il 1 1 rill .11 I'll , lilf
H.irl IruUU' r, tiK/J
Alk;i 1 lull y , n.r / 1
(ill
( .Mnlui t ,nu .• , itnili.i/t in
III MI li.irr,'-, 1 /HIM:
I>1 HI li/irj-,'' , •• 1 u
* ( .Mil 1 I'll. Mil. 1 ITOI .1 I
Siiuke
Mean
0.9
95.
11.
183.
33.
1.2
2.0
4.2
;/! 't-0
252.
17.
7.0
101 .
1145.
44-
i urn ii. r;. <:.;;.
Kiver
S.E.
0.12
6.7
1.9
32.3
6.8
0.15
0.21
0. 53
0.70
38.7
1 .1
(1.05
2.55
*
*

Blue-
Mean
10.
227 .
187.
551.
28.
14.
14.
5.5
2.6
194.
44.
7.6
I6fl.
1 58 ri.
-,<,.

River
S.F..
1.92
43.7
36.0
104.
5.4
2.55
2.80
0.64
0.48
21 .4
0.04
O.O'I
3.14
A
A

Tenmlle
Mean
3.5
212.
26.
357.
66.
1.1
3.6
7.1
10.
691.
35.
7.3
100.
1839.
65.

Crock
S.E.
2.4
30.
13.
78.
16.
0.14
0.44
1.7
3.0
195.
2.18
0.07
25.3
*
A

Snake
Mean
1.4
97.
12.
234.
107.
2.7
6.3
6.9
6.1
451.
23.
7.4
84.
2334.
02.

Ulver
S.E.
0.58
10.6
1.6
30.6
34.9
0.50
1 .2
1.0
.81
91.0
1.9
.03
4.2
A
A

Blue
Mean
3.2
131.
107.
382.
73.
4.3
6.7
5.6
3.8
1 84 .
40.
7.8
134.
11 13.
110.

River
S.E.
0.79
17.0
32.5
74.7
17.9
0.68
0.69
0.54
0.37
24.3
1 .3
0.05
3.5
A
A

lunml
Mean
11 .
188.
94.
476.
89.
2.1
5.5
5.9
4.5
304.
39.
7.8
236.
329y.
117.

le Crc-uk
S.E.
1 -5
25.1
13.9
69.1
17.9
0.37
0.73
0.63
0.58
44.7
1 .9
0.05
18.0
A
*


-------
                                                                      208




discharge-weighted averages.  There are a few exceptions  to  this.   For




example, total soluble nitrogen can be as much as  25-50%  higher  when




expressed as a time-weighted average than it is when  expressed  as  a




discharge-weighted average.  Total participates and all particulate




constituents are consistently higher when expressed as




discharge-weighted averages than when expressed as time-weighted




averages.  This is explained by a general tendency of particulate  loads




of running waters to increase as a power function of discharge  (Leopold




et al. 1964, Bonnann et al. 1969).  In contrast, sone dissolved




constituents show a tendency toward higher means for  time-weighted




averages than for discharge-weighted averages because of  dilution




effects that operate especially on point sources, but also on non-point




sources and of certain constituents of natural soil-water systems




(Lewis and Grant 1979).




     A detailed consideration will be presented later of  the relative




contributions of point and non-point sources to the total transport  of




the three rivers.  For the time being, however, it suffices  to note




that the distribution of land uses and point sources differs among  the




three rivers.  The Snake River is not influenced by any raajor point




source, although it does reflect the influence of non-point  source




nutrient loading from ski  slopes and residential development.  The  Blue




River is influenced by a variety of non-point sources and also by a




major point source, the Breckenridge Wastewater Treatment Plant.




Tensile Creek is influenced by mining operations at Climax Molybdenum




Mine in the headwaters, plus a variety of non-point sources  between  the




headwaters and the mouth.  In addition, Tenmile Creek is  influenced  by

-------
                                                                      209




the Copper Mountain Wastewater  Treatment Plant,  which is situated about




10 km upstream from the  sampling  point  at the mouth (Figure 1).




     Table 40 shows, as  might be  expected from the foregoing synopsis




of the three watersheds,  that the soluble inorganic nitrogen and total




soluble nitrogen concentrations for  the Snake River in 1981 and  1982




were considerably lower  than for  the  other two rivers.  For all  three




rivers, half to three-quarters  of the soluble nitrogen was inorganic.




Particulate nitrogen was  generally present in concentrations much lower




than soluble nitrogen.   Only in the  Snake River  in 1982 did the




particulate nitrogen approach one-third of the total nitrogen.




     Contrasts among the  three  watersheds were not so extreme for




phosphorus compounds as  they were for nitrogen compounds.




Orthophosphate was considerably higher  in the Blue River than in either




of the other two rivers.   This  was largely due to  the proximity  cf  the




treatment plant to the collection point.  The 1981 concentration was




exceptionally high, and  requires  special explanation.  A scatterplot of




the concentration of orthophosphate  in  water  of  the Blue River  shows




that the very high average  for  1981 was caused by  exceptionally  high




values of orthophosphate  for the  interval September through November of




1981 (Figure 42).  These  high concentrations  were  not repeated  in the




following year, nor did  they appear at  the next  station upstream or. the




Blue River between September and  November of  1981.  It is  thus  clear




that some special phenomenon occurred at the  treatment plant between




September and November of 1981  that was responsible for the




exceptionally high average  of orthophosphate  in  1981.  Ve  postulated a




treatment plant malfunction, and  this was subsequently ccnfirred by

-------

0>
oT
1
o
CL

80
60
40

20
0
*
-
*
_
* **
* «
9
• «•••• J>+9 #«* •
: , : v,"*** ,,,,•,•, r ^r»- , ,•: ,v .*
JFMAMJJASONDJFMAMJJASOND
                        S98I                         1982
Figure  42.  Orthophosphate data  for the Blue River at  its  point of
           entry into Lake Dillon, showing the effect  of  malfunctions
           in tertiary treatment at the Breckenridge  WWT?;
           September-November 1981.

-------
                                                                      211
contact with the  plant  operators.   Although the plant had returned to




normal function by November  when values  were still high,  there was




evidently some residual effect.  Figure  42 is a dramatic  illustration




of the effectiveness of present  treatment  practices.




     The concentrations of  total soluble phosphorus and orthophosphate




both tended to be higher in  1982,  a wet  year, than in 1981,  a dry year




(discounting the high values for Blue River 1981).  A similar trend has




been shown for an undisturbed watershed  near the Continental Divide




(Lewis and Grant  1979).  Higher  streamwater concentrations in 1982 are




consistent with the higher  total soluble phosphorus concentrations ir.




the lake in 1982.  The  higher concentration of soluble phosphorus in




wet years is postulated by  Lewis and Grant (1979)  to  be caused by a




critical shift in the balance of biological phosphorus demand, which




sequesters P, and the influence of physical forces, which remove




soluble P from the zone of  demand.




     The particulate phosphorus and total  particulate concentrations of




Tenmile Creek were higher in 1981  than in  1982, contrary  to  what might




have been expected.  Examination of the  data shows that this was due




mostly to high particulate  concentrations  observed at peak runoff in




1981 but not in 1982.   This  phenomenon was observed not just at the




mouth of Tenmile Creek  but  at other points on the  Creek as well.  Its




explanation will be considered along with  the segment-by-segment




analysis of Tenmile Creek to be given in another section  of  this




report.




     Conductance, representing total soluble ionic raa.erials, shews the




expected dilution effect between years of  high discharge  and years c:

-------
                                                                      212





low discharge for the Snaka River and Blue River, but  not  for  Tenmile




Creek.  In addition, the data show clearly that the conductance  of




Tenmile Creek averaged considerably higher than that of  either of the




other two rivers.  Major sources of ionic dissolved solids  associated




with mining are located near the head of Tenmile.  These raise the




average conductance for Teninile and can produce irregularities that  are




not part of the natural seasonal cycle.




     Most of the dissolved and suspended materials showed  some kind  of




seasonal pattern in all three rivers.  The natural seasonal  pattern  is




more difficult to pick out for the Blue River and Tenmile  Creek  than it




is for the Snake River because of the influence of point sources.  We




therefore use the Snake River to illustrate the background  seasonal




patterns, and then describe how the other two rivers differ  frcoi the




Snake River.




     Figures 43-44 show the seasonal patterns of selected  variables




over the two years of study for the Snake River.  Some variables have




been omitted because their indications are redundant or  because  they




are not of special interest here.  Discharge is a key variable,  not




only because it embodies the seasonal changes that affect




concentration, but also because it is multiplied by concentrations in




the process of obtaining discharge-weighted concentration  averages.




The pattern shown by discharge in Figure 43 for the Snake  River  is




reflected very closely in the discharge patterns of the  other  two




rivers.  The discharge patterns were strongly dominated  by  seasonal




runoff associated with snowmelt.  In 1981, when total  annual discharge




was low, discharge peaked in the first week of June.   The  peak was

-------
                     1981                      1982
          JFMAMJJASONDlJFMAMJJASOND


o

0)
o»
a
0
t/>
5



o
0



-------
                                                            21
                       98!
           JFMAMJJASO
             1982
DlJFMAMJJASOND

0
-s_
.
— ,
1

SO
O


§L

z*
i
X
"Z.






— .
o»
^2_

2!
CO
H





200

150

100


50
0
40
32

24

IS

3


0

450

360

270

ISO

90

0
	 j 	 J 	 J 	 i 	 ! 	 ! 	 1 	 ! 	 1 1 1 | 1 1 1 1 ! ' ' ' ' ' '
-
•
-••• *.» . -
• •
O
* •• • •
u * - -
. • • • • .
* •» • • •
•
. * • • • •
*o . . • • •
~ t 1 1 1 1 1 1 1 1 ! 1 1 1 1 1 ! 1 1 1 . L 1 _. 1. ! ~

-
«
.
»* • • *

« * « * • •
— * * • • —
• * * • * • *
9 • * • • •
• • * • *
1 1 1 1 t J ! 1 1 1 1 I ! ! t 1 1 1 ! 1 1 ' ! !
* •
A A
855 - 825
• *
* •
— —4
* *
*
*
I
~l
•
9 •
* •
• * *
• « *
* •
• • • ** **
— *» » » * • • —
• * *
4
1 1 1 i ! t | | | | i ] i | ] | | , , , , i

JFMAMJJASONDJFMAMJJASOND
                      1981                       1982
Figure A4a.  Values of selected variables for  the Snake River as it
           enters Lake Dillon.

-------
              1981                     1982
   JFMAMJJASONDlJFMAMJJASOND
—
0
flfc
Q.

25
"o
CO
"o





o»
w
Q.
O
Q_



X.
o>
E
tf>
.£
"o
3
O
"•£
o
CL



10
8

6
4

2

0




10

5


0



8
6

4

2




— i — i — 1 — i — i — I — i — i — i — i 	 1 — 1 — i — i 	 ! — i — m — nr-T — ! — IT-: 	 1
A A A
13 25 18
-
•
• '. "
. " -
• * • * * *."*
- • •• • . -
• •• • • • «
i
1 1 1 1 1 ! 1 1 1 1 1 1 1 1 ! I 1 1 1 1 • i i
A
20'
'
•
— • -H
• • • • *
* •
• • • •
• • # • • • •
• ••**• •
i
1 1 1 1 1 1 1 1 1 1 I I ! 1 1 ! i 1 ! ! ' ! !
A A A
19 . ' 13 15
™ """
- • —


• • *
• * •• • •
* •
• • • • |
* 1 •
I 1 ! 1 I 1 1 I I I 1 I !' I '!'!!! 1 .
JFMAMJJASONDJFMAMJJASOND
             198!                     1982
b.  Values of  selected variables for rhe Snake River as it
   enters Lake Dillon.

-------
                                                                      216
broader in 1982 and the highest values did not occur until  the  end  of




June.  In both years there was a small early peak in May, probably  due




to early melt in the valley bottoms, and a shoulder or  low  peak in




August or September.  The amount of variation around the  seasonal




pattern was small, which provides an ideal situation for  the  estimation




of weighted mean concentrations and transport.




     Stream temperatures varied between 0 and 16°C.  The  highest




temperatures occurred during the last half of July.  Maximum




temperatures were lower in 1982 than in 1981 by 2-4°C.  The seasonal




change in temperatures was relatively smooth, although  some




irregularities were caused by variations in time of day when




temperature was taken.  Variation of temperature with season  and




between years affected the depth of penetration of water  entering the




lake, as already explained.




     Conductance showed the dilution effects that one would expect  to




be associated with changes in discharge for materials that  are  soluble




and little affected by biological activity.  For the Snake  River, the




conductance faithfully reflected the early small peak in  discharge




(May), the subsequent brief decline in discharge, and the major  peak of




discharge caused by spring runoff-  This sequence of events was  visible




both years for the Snake River and was also characteristic  of the other




two rivers.  However, the winter conductances of Tenmile  Creek  were




especially high because of the presence of especially potent  sources of




soluble materials in the upper part of this watershed,  at Climax




Molybdenum tailings ponds.

-------
     Nitrate concentrations  of the Snake River were high at the time of




runoff and then decreased  steadily until August, after which they




increased steadily until December or January.  Midwinter concentrations




were typically high.   The  low nitrate concentrations in the late sunnier




months were almost certainly due to high summer biological demand for




inorganic nitrogen, which  became all the more effective in reducing




nitrate levels as the  amount of water movement declined at the end of




summer.  With the onset of cool weather, and reduced biological demand,




nitrate concentrations again rose to higher winter levels.  The winter




nitrate levels of the  Snake  were somewhat higher than would be expected




for unaltered watersheds in  this region (cf. Lewis and Grant 1979,




1980b).  They probably reflect the influence of dispersed  nutrient




sources, especially from residential areas, which have an  especially




noticeable influence on winter concentrations because of the very lov




discharges in winter and the low biological demand for nitrogen at this




time.  The seasonal pattern  of nitrate concentrations observed in the




Snake River is, however, essentially what would be observed in an




unaltered watershed (Lewis and Grant 1980b).  The same pattern is




visible in the data for Blue River and Tenmile Creek, but  with




considerable scatter due to  the influence of point sources.  The late




summer minimum and winter  maximum are evident, however.




     Tenmile Creek was unusual in that nitrate concentrations were




higher at all times of the year than in other rivers.  In  1981 nitrate




nitrogen approached 1,000  ug/1 in winter (Figure 45).  The high nitrate




concentrations are explained by the large amounts of nitrate exported




from the site of Climax Molybdenum at the head of Tenaile  Creek.  This

-------
                                                             218
1000


 800


 600


 400


 200



    0
Z
 I
 fO

o
—9
             1  !  1   1  \  !   !  I  1   '  1  1   !  1  1   \  i  3   '-  1  1
            JFMAMJJASONDJFMAMJJASOND

                      1981                       1982
Figure 45.  Nitrate values for Tenmile Creek at its point of entry to

          Lake Dillon,  illustrating the very high nitrate levels at

          certain times of the  year.

-------
                                                                       O 1 Q





special nitrogen source will  be  considered further in the




segment-fay-segment analysis of Tenmile  Creek.




     For ammonium, there vas  also  a seasonal pattern of concentrations




in 1981 and 1982, but it was  not so clear as the pattern for nitrate.




In part this was due to the much lower  concentrations of ammonium than




of nitrate, especially in  the Snake River.  The Snake River data do




show indications of an early  spring maximum in concentration before




discharge had begun to increase  drastically, a late summer minimum




corresponding to the postulated  peak of biological demand for inorganic




nitrogen, and a fall increase after the stream water had begun to cool.




Two high values in early August  of 1981 for the Snake River departed




from the seasonal pattern  and were not  repeated in 1982.  These high




values were confirmed, however,  by other samples upstream and were




almost certainly caused by some  temporary or sporadic man-iaade source.




     Patterns in the ammonium data for  the Blue River and for Tenmile




Creek are scarcely discernible because  of the  large amount of




variability introduced by  point  sources.  Ammonium concentrations were




affected simultaneously not only by seasonal factors but by variations




in release from point sources and  variations in the conversion rate




from ammonium to nitrate.




     In all three rivers,  total  soluble nitrogan was accounted for




approximately half by organic nitrogen  and half by ammonium plus




nitrate.  Winter values of total soluble N tended to be higher than




summer values, in parallel with nitrate, a major contributor to to~a~




soluble X.  Seasonal pattern  was weak,  however, especially for the Blue




River and Tenmile Creek.

-------
                                                                      220
     Total soluble phosphorus concentrations showed  no  strong  seasonal




pattern in the Snake River, nor did orthophosphate concentrations.   The




average concentrations were often high at times of high discharge,




however.  The implication is that dilution effects associated  with




major seasonal increases in discharge were effectively  cancelled  by




increased transport at such times (cf. Lewis and Grant  1979).   There




were no clear seasonal patterns in concentration for  the  Blue  River or




Tenmile Creek, where the point sources introduced considerable




additional random variation.  For the Blue River, the increase  of




orthophosphate and total soluble phosphorus in September, October,  and




November of 1981 caused by treatment plant malfunctions has  already




been mentioned.  A similar but less exaggerated phenomenon was  observed




in Tenmile Creek in the first half of 1981, but was not repeated  in




1982 and was of obscure origin.




     Total particulates, particulate phosphorus, particulate nitrogen,




and particulate carbon are united by a tendency to show patterns




similar to each other.  In the Snake River there was a  trend toward




high concentrations around the time of maximum discharge.  Superimposed




on this was an occasional exceptionally high value associated with  some




kind of short-lived earth disturbance in the watershed, as shown  for




the Snake River in late August of 1981 and in the summer  of  1982.   Such




short-term spikes obviously can be extremely important  in total annual




transport of particulate materials.

-------
                                                                       79"
Chemistry of Miner's Creek and  Soda Creek




     Because of the relatively  small proportion of total land area




drained by Miner's Creek and  Soda  Creek,  the potential contribution of




these two watersheds to total loading of  the lake  with nutrients is




small.  Their chemistry is considered here mainly  for the sake of




completeness, since these two watersheds  are separate from the three




major river drainages.  Table 41 gives the discharge-weighted averages




for the chemistry of the two  streams and  the standard error for each




variable.  As in the case of  the three rivers,  the time-weighted means




were calculated but are not shown  in the  table. For  Miner's Creek and




Soda Creek, the time-weighted means are even closer to the




discharge-weighted means than for  the three major  rivers.




     The small size of contributions of these two  creeks to total




loading of the lake is evident  from the discharge  figures in Table 41.




The chemical data are unexceptional.  Soluble nitrogen concentrations




were in the range of those observed in the Snake River,  and




considerably below those of the  Blue River and  Tenmile Creek as




reported in Table 40.  This is  not  unexpected,  since  Miner's Creek and




Soda Creek at the sampling points  were not influenced by point sources




of nutrients, but were influenced  by some relatively  dispersed human




influences analogous to those characteristic of the Snake River




drainage.




     For Miner's Creek, the phosphorus concentrations, including both




soluble and particulate fractions,  were relatively low,  and were




comparable to the values observed  in the  Snake  River.  For Soda Creek,




however, both particulate and total soluble phosphorus concentrations

-------

N02-N, up,/!
fJO^-M. uc/1
v O > « '
Nil * — N uc/j-
Total Soluble N, ug/1
Participate N, ug/1
PO.-P. up/1
4
Total Soluble P, ug/1
Particulate P, ug/1
Total Pavticulotes , mg/1
Particulate Carbon, mg/1
Alkalinity, mg/1
pll
Conductance, uinho/cm
Discharge » I /sec
Discharge, cfs

Mi n e r ' s
Mean
0.8
32.2
14.9
141.
34.
1.9
3.2
6.4
6.9
320.
28.9
7.4
74.
88.
3.1
1981
Creek
S.E.
.11
4.3
3.0
34.
13.
0.23
0.34
1.2
1.9
70.0
1.7
0.08
3.4
10.
0.37


Soda Creek
Mean S.E.
1.2
12.7
12.2
90.0
56.
8.7
18.3
11.0
6.1
351.
59.0
8.0
168.
12.
0.42
.32
4.5
2.8
21.
17.
1.2
3.3
1.5
1.2
63.2
3.2
0.15
9.8
3.1
0.11

Miner's
Mean
1.4
20.2
10.0
102.
83.
2.9
5.1
5.8
7.7
531.
26.0
7.5
58.
246.
8.7
1982
Creek
S.E.
.20
4.1
1.9
26.
16.
0.47
0.60
0.76
1.7
116
1.2
0.05
2.6
47.
1.7


Soda Creek
Mean S.E.
2.3
63.
11.5
202.
100.
13.4
20.5
21.6
16.3
501.
62.3
8.0
142
54.
1.9
.23
14.
1.4
41.
29.
2.7
3.0
4.0
3.1
103.
2.4
0.08
8.5
15.
0.52
Table 41.  Discharge-weighted  means  for the two amall streams entering  the  ]ake  (sampling stations above WWTP
           cf II iicnts) .

-------
                                                                       123
were higher than  for any  of  the  three major rivers, and the discharge




was especially low in relation to  drainage area.   Conductance was also




notably higher for Soda Creek  than for Miner's Creek.   Water is




evidently used within the drainage,  and this results in reduced




discharge.  Water use and possibly spri-gs may account for high




conductance.









Frisco and Snake  River Treatment Plant Effluents




     Table 42 summarizes  the chemistry of the effluents from the Frisco




and Snake River wastewater  treatment plants.  Total soluble nitrogen




concentrations were 50 to 100  times  higher than the stream




concentrations and were relatively similar between plants  and between




years.  The phosphorus values  present quite a different picture.




First, the effluents of the  two  plants were rather different in the




partitioning of total phosphorus.   In the Snake River  Wastewater




Treatment Plant effluent, the  largest amount of phosphorus was




connected with the particulate fraction,  while in the  Frisco effluent




the largest fraction was  connected with the soluble components.  For




the Snake River Plant, the phosphorus concentrations were  about 10




times above those of the  receiving stream.  A few exceptionally high




concentrations were checked against  plant records.  High values




typically coincided with construction or  equipment problems.  The




average total P concentrations for the Frisco Plant were considerably




higher than for the Snake River  Plant, and there  was a rajor difference




between 1981 and  1982 for the Frisco Plant.   In 1982 the




discharge-weighted averages were exceptionally high at the Frisco




Plant.  Examination of the raw data  shows that the high averages were

-------
1981
Frisco WWTP
Effluent

NG^-N, ug/1
N03-N, ug/1
Nll^-M, ug/1
Total Soluble N, ug/1
Particuiate N, ug/1
P04-P, ug/1
Total Soluble P, ug/1
Particuiate P, ug/1
Total Particulates , mg/1
Particuiate Carbon, mg/1
Alkalinity, rag/1
pll
Conductance, uniho/cm
Discharge, 1/s
Discharge, cfs
Mean
135.
1579.
2787.
8911.
167.
53.6
76.7
17.6
1.3
741.
42.2
6.8
417
12.8
0.5
S.E.
15.8
407.
670.
1168.
23.
18.
19.
2.2
0.28
163.
2.9
0.08
29.
0.89
0.02
Snake WWTP
Effluent
Mean
116
3527.
1825.
6798.
496.
4.5
22.
45.9
6.0
4710.
50.5
6.5
450.
13.4
0.5
S.E.
17.
786.
620.
774.
160.
2.4
3.4
11.2
2.4
2384.
7.9
0.16
18.
0.92
0.04
]982
Frisco WWTP
Effluent
Mean
113.
1228.
2868.
7096.
-
496.
556.
57.0
5.1
-
65.8
7.2
449.
14.5
0.5
S.E.
25.
281.
543.
867.
-
169.
177.
16.
1.2
-
7.7
0.10
21.
0.50
0.02



Snake WWTP
Effluent
Mean
268.
2251.
1957.
8398.
-
24.4
38.6
120.
9.3
-
68.5
7.0
429.
14.6
0.5
S.E.
50.
910.
453.
1627.
-
12.
13.
32.
2.
-
6.
0.
20.
1.
0.









2

9
09

2
04
                                                                                                                            K>
Table 41'..   Dlscharge-wc I ghl ed inemir.  for  the Lwo effluents  discharging  near  the mouths  of  Miner's and
            r>od.i Crt-oks .

-------
                                                                      225
caused by discontinuous but repeated  occurrence  of  very high

concentrations (1000 to 3500 ugP/1).   Checks  of  the dates  of high

values with plant operators showed  that,  for  the most  part,  the  dates

of high P values coincided with known equipment  problems.


           *
Precipitation Chemistry at the Main Station

     The precipitation chemistry data consist  of the continuous  record

from the collector situated near the  Snake River Wastewater  Treatment

Plant and shorter records from two  other  locations.  We will consider

the more extensive record first and then  compare it  with the other  two

records.

     All of the precipitation chemistry data  to  be  dealt with here

refer to bulk precipitation, which  is  the total  of  all  wet and dry

deposition on a collecting surface.   A collecting surface  accumulates

dry deposition when there is no rain  or snow.  The  occurrence of  wet

precipitation augments the delivery of materials onto  the  collector

surface, especially for certain substances that  are  scrubbed out  of  the

atmosphere effectively by water droplets  (e.g.,  nitrate).  There  ara

typically changes in the concentration of dissolved materials in  wet

precipitation during the course of  a  rainstorm,  principally  as a  result

of the cleansing of the air by rain.   Because  of  the continuous

deposition of materials on a collector surface during  dry  weather and

the continuous change of concentrations during a  wet precipitation

event, an average concentration is  not very meaningful  for

precipitation chemistry.  For this  reason, we  express  all  the results

-------
for precipitation chemistry in terms of  loading  rates,  i.e.,  the amount




of material delivered to a unit surface  over a specified  time.




     Table 43 summarizes the means and standard  errors  for  the  loading




of chemical constituents analyzed in the  samples  from  the main




precipitation collector at the Snake River Wastewater  Treatment Plant.




The means are derived from 61 separate collections  spread over  the




two-year study period.  Since the collection intervals  were not all of




identical length, the means are time-weighted averages.




     The hydrogen-ion loading was lower  (4x) than loading reported for




the Como Creek watershed at similar elevation near  Ward, Colorado




(Grant and Lewis 1982), but precipitation was frequently below  pH  5.65,




indicating the presence of strong mineral acids.  Because of  the




dominant influence of stream chemistry on the chemistry of  a  lake  such




as Dillon, however, direct hydrogen ion  loading  from precipitation is




of minimal significance to the biology of the lake.  As expected,




nitrite was a minor contributor among the inorganic soluble nitrogen




species, but nitrate nitrogen was present in substantial amounts.   As




at Como Creek, which has been studied thoroughly  in this regard, the




precipitation at Dillon appears to be substantially augmented beyond




background in nitric acid as a result of  the presence of substantial




amounts of NGv in the atmosphere from fossil fuel combustion.   The




ammonium loading was relatively high but  secondary  to  the nitrate




loading as a nitrogen source.




     The loading rates for both PC^-p and total  soluble phosphorus




*ere considerably higher than at Como Creek (3-5x).  Sinca  the




mechanisms of phosphorus transport through the air  are  poorly known, no

-------
1981. - iag/m2/wk

N02-N
*
A
Total Soluble N
POA-P
Total Soluble P
Total Particulates
Particulate P
HCO~
H+
so4
Mean
0.079
2.616
3.479
18.380
0.732
1.012
187.
0.219
19.8
0.055
12.8
S.E.
0.008
0.284
0.353
1.776
0.087
0.143
8.01
0.015
8.60
0.006
0.86
1982 - ng/m2/wk
Mean
0.125
3.724
1.627
6.002
0.630
0.695
152.
0.178
23.4
0.040
46.7
S.E.
0.011
0.351
0.098
0.364
0.123
0.134
7.54
O.OC9
1.56
0.006
5. 7
Table 43.  Loading rates for bulk precipitation at the main collecting
           station, 1981 and 1982.

-------
                                                                      228
explanation can he offered.  However, the  observed  phosphorus transport




values are within the range of values reported  for  a  variety of sites




in the literature (Likens et al. 1977).  Particulate  phosphorus




transport was significantly lower than at  Como  Creek,  and  this




partially offset the higher loading for  soluble  fractions.   Sulfate




was a major ion, as would be expected by analogy with  the  Como Creek




studies.




     Figure 46 shows tine plots of loading rates for  some  of the




nutrient fractions that have a significant amount of  seasonal pattern.




The transport of both nitrate and ammonium is facilitated  by moisture,




as is evident from a comparison of the amount of rain  or snow and  the




loading rates for these ions.  Total soluble phosphorus  showed a




pronounced maximum in June of both years.  This early  summer maximum,




although quite high, appeared both years and at more  than  one station




and was thus most definitely not an error.









Precipitation Chemistry at Other Stations




     Phosphorus loading rates by bulk precipitation were determined at




two additional stations in 1982 as a comparison with  the main station




located near the Snake River Wastewater  Treatment Plant.   The first of




the comparison stations was located near Frisco.  A different type of




collector (tube collector, diameter 25 cm) was  used at this  station;




the dates of comparison were June 1 to November  17, 1982.   The means




r.r.d standard errors for the main station and the Frisco  station over




this interval are summarized in Table 44.  There was  no  significant




difference in the average loading rates  for the  two stations (p >

-------
Jtf
-v
E
E

E
„
z
'•0
O
Z

^
$
V.
w
E
o>
E
„
Z
1
X

JJ.
^
M
E
0
CL
^J
^^^
3
"o
"o
K
•t u
32

24
16

8

0
8

6

4


2

0



10
8

6
4

2
0


3

2
1




.
*

* •
4
— • __
*
• • 4
•*••*.. * * **
* • • • • •• ••• •
- -1 *i "i *i . ". i 1 . <• i P 1 •, , , • ,« , , , , . , -
A35 24ii.
30

*
— — '
• • 1
- * • u
* * • • • •
• *
* • • —
* • ••• ••• •
"" 1 *1 1 1 1 t 1 1 1 1* 1 1 1 1 1 1 1 1 l 1 • • ;
• • *
14AAA26
14

-

_ * —
_ • _
• •
* ... . .* * ~l
• ••» • •*•
" *| * 1 1 I 1 1 I 1*1 1 ! 1 1 !*! ! 1 !*! ' * ! ! ' ~*
1
i

•
-
• • •
I «| * ! *l ! ! ! » <••!•*! • ! • 1 ' j • 1 I ! ! «*• . :
JFMAMJJASONDJFMAMJJASONO
1981 1982

Figure -6.  Selected variables for bulk  precipitation  rear  the  Snake
            River WWTP.

-------
                                                                       130
0.05).  The period of collection  incorporated  the high phosphorus

loading pulse in June, which results  in  relatively high standard errors

for both stations.  However, the  pulse appeared  simultaneously at the

two stations, verifying our conclusion that  it was not an artifact.

1
1
Dates of Comparison
June - 17 November
July - 1 November
Total P
Main Station
mg/m /wk-
Mean S.E.
1.354 .969
.427 .065
Comparison
Station
Frisco
Raft
Total P
Comparison Station
9
mg/m*"/wk
Mean S.E.
1.561 .688
.533 0.099
Table 44.  Summary of comparison for phosphorus  loading  by bulk
           precipitation at the main station and  two  other stations.
     The second comparison station was situated on the  Denver  Water

Department's raft, which was anchored on  the water surface.  This

comparison station is of special interest because we wanted  to rule  out

the possibility that a significant amount of phosphorus  transport

observed at the Snake River station could be accounted  for by  localized

dust in that area.  The dates of comparison were July 1  through

November 1.  From July 1 through mid-September the raft  was  anchored  in

midlake, and from mid-September through October it was  anchored at  the

shoreline.  The means for the main station and the station on  the raft

ever these dates are summarized in Table 44 along with  their standard

errors.  The standard errors are much lower in this case because the

interval of sampling did not include the period of high  phosphorus

-------
                                                                       "1
loading in June.  There was no significant  difference  in  the  loading




rates between the two stations.




     From the data at the two comparison  stations,  we  have  every  reason




tc believe that the main precipitation  collection  station at  the  Snake




River Wastewater Treatment Plant  gave an  accurate  impression  of  the




bulk precipitation loading to the  surface of  Lake  Dillon.

-------
                                                                      232





                  Total Nutrient Loading of  the Lake









Water Budget




     Reliable estimations of total nutrient  loading of  the  lake over




the period of study presume the availability of sound information on




the amounts of water reaching the lake by various  pathways.   If the




flow measurements summarized in the previous chapter are  accurate,




their summation plus the unmeasured flows should be close to  the




calculated input determined by the Denver Water Department, which is




obtained by an indirect method based on measured outflow  and  change in




lake volume.  Table 45 summarizes the flows  from various  sources in




1981 and 1982.  The sources of the estimates for the rivers,  streams




and effluents are as given in the previous chapter.  Two  sources of




water were estimated without direct measurement:  miscellaneous  surface




runoff and groundwater.




     Miscellaneous surface runoff was the contribution  from a number of




small areas near the lake that did not drain through watersheds  where




there was gauging or discharge measurement.  These areas  total 4300 ha,




or 5.8% of the total land area of the Lake Dillon watershed.   Their




unmeasured contribution to runoff was estimated in two  parts.   The




first of these is the Meadow Creek drainage, which includes some




high-elevation areas.  This drainage was considered to  be




hydrologically similar to Miner's Creek.  The runoff was  thus  set at an




r.~cu"t .-r.qual to that of Miner's Creek, with  appropriate adjustment




tor area.  The second portion consists of smaller drainages that are




low-lying and without steep contours.  Because of low accumulation  of

-------
                                                                       233

Rivers
Snake
Blue
Tenmile
Subtotal
Strearas not on Rivers
Miner's
Soda
Subtotal
Effluents
Frisco
Snake
Subtotal
Diffuse Surface Drainage
Precipitation
Groundwater
Grand Total
1981 -
m /yr

40,100
49,900
58,000
148,000

2770
330
3150

403
423
826
3366
4336
8^41
168,119
Thousands
Acre-ft/yr

32.54
40.45
47.02
120.0

2.25
0.31
2.55

0.33
0.33
0.66
2.73
3.52
6.90
136.4
1982 - Thousands
m /yr Acre-ft/yr

81,800 66.35
98,200 79.61
104,000 84.31
284,000 230.3

7760 6.29
1700 1.38
9460 7.67

457 0.57
460 0.37
917 0.7i
10,385 8. -2
4057 3.29
84-1 6.90
317,260 257.-
Table 45.  Summary of water  flow  into  the lake in 1981 and 19?2

-------
                                                                       234




snowpack in such areas, the water yield was  expected  to  be small.  This




second portion was assumed to have  the water  yield  of Soda Creek




drainage, with appropriate adjustment for  area.   Any  errors in the




approximation procedure will obviously not have a very great effect on




the water budget of the lake because of the  relatively small proportion




of the watershed in the miscellaneous surface  drainage category.




     Groundwater also potentially supplies water, and thus nutrients,




to the lake.  A complete water budget was  constructed on  a monthly




basis for 1976-1979 from the outflow figures  of the Denver Water




Department, the inflow data of U.S.G.S gauges, measures of




precipitation, estimates of evaporation, and  an allowance for




miscellaneous surface drainage.  This water budget  should sum  to  zero




if all estimates were made with good accuracy  and if  there was no




groundwater entry into the lake.  Some variance around the zero  sum is




expected due to inaccuracies of the various estimates, but a trend




toward greater outflow than inflow would suggest  the  presence  of  an




unmeasured term, which would have to be groundwater.   The residuals




over the three years had an average of 575 acre-feet  per  month.   There




was no strong seasonal or annual pattern in the size  of the residual,




and the standard error was about 300 acre-feet per month.  This




residual may have been affected by the amount  of  drawdown as well as




other factors.  Since the irregularities in the residual  cannot be




pinned to a specific pattern, we treat the groundwater inflow  to  the




lake as constant at 575 acre-feet per month for both  years  of  the




s:ady.  Since this amount of flow is very  small by  comparison  with the




total flow to the lake,  any errors in the estimate within the  range of

-------
                                                                      235




one or two standard  errors  would  not  be significant to the overall




conclusions of the study.




     When all sources  of  water  for  the lake were sunned (Table 45), the




estimated inflow was 136,000 acre-feet for 1981 and 257,000 acre-feet




in 1982.  This figure  can be compared with the independently derived




estimate of the Denver Water Department for total input over the sane




interval.  First, however,  evaporation must be added to the Denver




Water Department estimate,  which  does not  include evaporation.  From




the averages of several years of  study by  the  Denver Water Department,




we estimate evaporation as  5900 acre-feet  per  year,  which  gives  a




corrected Denver Water Department input of 111,000 acre-feet per year




for 1981 and 223,000 acre-feet  per  year for 1982.  The deviation of




this from the totals based  on U.S.G.S.  data as described above is




10-20%, which is more  than  anticipated. Reexamination of  the  data did




not reveal any potential  errors of  this magnitude.  For nutrient




loading computations,  we  assume that  the U.S.G.S. data for input are




correct, although this assumption is  to some degree arbitrary.









Phosphorus Loading




     Table 46 summarizes  the phosphorus loading of Lake Dillon




according to the various  pathways of  nutrient  flow that were recognized




in the previous chapter and  in  the  section on  water balance.  For




rivers, streams, and effluents, the numbers in Table 46 vere octainec




by multiplving the discharge-weighted average  concentrations as




reported previously  (Tables  40, 41) tines  the  time-weighted average




discharge and dividing the  product  by lake  area.   The  estimate of

-------
Phosphorus -

Rivers
Snake
Blue
Term J 1 e
Subtotal
Streams not on Rivers
Miner's
,°3da
Subtotal
Effluents
Frisco
Snake
Subtotal
Miscell aneoua Surface
Drainage
Free ipi tat i on
Ground water
OR A Ml) TOTAL
PO -P
4
.036
.524
.048
.608
.004
.002
.006
.016
.001
.018
.011
.381
.006
1.030
Total
Soluble
P
.060
.524
.156
.740
.007
.005
.012
.023
.007
.030
.019
-520
.010
1.331
- kg/ha/yr
Part. P
.127
.206
.308
.641
.013
.003
.016
.005
.015
.020
.035
.114
.000
.826
- 1981
Total P
.187
.729
.464
1.380
.020
.008
.028
.029
.022
.051
.054
.634
.010
2.157
Phosphorus •
%
8.7
33.8
21.5
64.0
.9
.4
1.3
1.3
1.0
2.4
2.5
29.4
.5
100.1%
PO -P
4
.166
.316
.164
.646
.017
.017
.034
.170
.008
.178
.045
.328
.028
1.259
Total
Soluble
P
.386
.492
.428
1.306
.030
.026
.056
.190
.013
.204
.080
.361
.066
2.073
- kg/ha/yr
Part. P
.423
.411
.459
1.293
.034
.028
.061
.020
.041
.061
.091
.093
.000
1.599
- 1982
Total P
.808
.903
.887
2.598
.064
.054
.117
.210
.054
.265
.171
.454
.066
3.671

%
22.0
24.5
24.1
70.6
1.7
1.5
3.2
5.7
1.5
7.2
4.7
12.4
1.8
99.9%
Table 46.  Summary of  phosphorus  loading of the lake in 1981 and 1982.

-------
                                                                      237





loading for precipitation was  taken  directly from Table  43 of the




section on precipitation chemistry.   For  miscellaneous  surface




drainage, the water contribution has  already been obtained by




approximation as reported in Table 45.  The  concentration of phosphorus




fractions in this water must also be  approximated before an estimate of




loading can be made.  Phosphorus concentrations  in the miscellaneous




surface drainage category were  assumed  equal to  those observed in




Miner's Creek because of parallel land  usage,  and these  concentrations




were applied to the discharge  approximations made in connection with




Table 45 to produce the estimate of  loading  shown in Table 46.  A




similar approach was taken  for  groundwater.   The groundwater phosphorus




concentrations were assumed to  be equal to those of the  Snake River




drainage, a large watershed with mixed  land  use  but lacking point




sources.  It was also assumed,  however, that particulate phosphorus




would not be transported through groundwater and this fraction was




therefore set to zero.  The concentrations were  then multiplied by




the amount of groundwater as reported in  Table 45 to obtain the loading




of phosphorus via groundwater.  Both  the  miscellaneous  surface drainage




and groundwater categories were minor contributors, so  the details  of




these approximations are of no  great  concern.




     More than half of the  total phosphorus  from all sources entered




the lake in soluble form.  Table 46  shows that the bulk  of total




soluble phosphorus was in the  orthophosphate fraction rather than the




organic phosphorus fraction.   The particulate  fraction was significant




both years and was slightly higher proportionally in 1982 than in 1°?1,




probably because of higher discharge  in 1932.  The total phosphorus

-------
                                                                       233
transport in 1981 was about 60% of  the  total  phosphorus  transport in




1982.




     Trie rivers contributed about two-thirds  of  the  total phosphorus




reaching the lake.  The individual  rivers made very  different relative




contributions in the two years, however.  All were affected  by higher




discharge in 1932, but not identically.  Loading  from  the Snake River




was drastically affected by construction in the  river  bottom in the




second year.  Loading from Tenmile  Creek was  approximately proportional




to the difference in flows between  the  two years.  The Blue  River




loading was exaggerated in 1981 by  problems at the Breckenridge WWTP.




Other differences in sensitivity cf the three rivers to  changes in the




hydrologic conditions between years can be traced to their differing




land use patterns; these will be considered further  when land use is




analyzed.




     Precipitation accounted for a  large proportion  of the residual




loading beyond the contributions of the three major  rivers.   The




proportion of total phosphorus loading  attributable  to precipitation in




1981 was considerably larger than in 1982.  This  is  principally




explained by the much greater contributions from  surface runoff in




1932, which reduced the overall influence of  precipitation loading.




     The combined contribution of the Frisco  and  Snake River wastewater




treatment plants was only 2.4% in 1981 and 7.2%  in 1982.  The




contribution in 1982 should actually have been lower in  view of the




much higher phosphorus loading from non-point sources  that year, but




tr..'accent plant malfunctions produced an exceptionally high  effluent




contribution or, certain dates in 1982.  There are two other  major

-------
                                                                      239



wastewater treatment plants  in  the  watershed:   Breckenridge and Copper




Mountain.  The contributions of  all four  plants will be specified more




exactly when the nutrient  loadings  are  broken  down more completely




according to sources and land uses.  It is  obvious from Table 46 that




the bulk of phosphorus  entering  Lake Dillon does not originate from




point sources, however.









Nitrogen Loading




     Table 47 summarizes the Dillon nitrogen loading.   The  methods of




computation were similar to  those used  for  phosphorus.   Particulate




nitrogen was not measured  in the effluents  in  1982 because  it was such




a small contributor to  effluent  N.   The particulate N column in the




table was filled in for 1982 by  use of  the  1981 data.




     The inorganic fraction  of nitrogen was slightly larger than the




organic fraction for all sources combined.   Unlike phosphorus,  nitrogen




loading was mainly associated with  the  soluble fractions; particulate




nitrogen made a relatively small contribution.  The difference  between




years in total nitrogen loading  was very  similar on a proportional




basis to that observed  for phosphorus:  the loading for 1982 was al^osz




double that of 1981.  For  nitrogen, the dominating effect of the three




rivers on total loading was even more pronounced than it was for




phosphorus.  The influence of precipitation on nitrogen loading was




relatively small by comparison with phosphorus.   The contributions of




the two effluents on a  percentage basis were slightly larger than for




phosphorus, but still relatively minor  by comparison with total




loading (Figure 47).

-------


Rivers
Snake
Blue
Tenmlle
Subtotal
Streams
Miner's
Soda
Subtotal
Effluents
Fr laco
Snake
Subtotal
Miscellaneous Surface
Drainage*
Tec i pi tat Ion
Iround water*
CKAND TOTAL

NO -N

.02
.37
.15
.55

.002
.00
.002

.04
.04
.08
.005

.004
.002
.64

„,-.

2.86
8.49
9.20
20.55

.07
.004
.07

.48
1.12
1.59
.18

1.36
.23
23.98
Nl trogen
»v»

.33
6.99
1.1 3
8.45

.03
.003
.03

.84
.56
1.40
.08

1 .81
.03
11 .79
- kg/ha /y
Total
Soluble
N

5.49
20.61
15.49
41.59

.29
.03
.32

2.69
2.15
4.84
.79

9.56
.44
57 . 54
r - 1981
Part.
N

.99
1.05
2.86
4.90

.07
.02
.09

.05
.10
.21
.19

1.17*
.000
6.56

Totiil N

6.49
21.65
18.36
46.50

.36
.04
.41

2.74
2.31
5.05
.99

10.72
.44
64.11

'

10.1
33.8
28.6
72.5

0.6
0.1
0.7

4.3
3.6
7.9
1.5

16.7
.7
100.0%

"V

.09
.24
.86
1.17

.008
.003
.01

.04
.09
.13
.02

.07
.01
1.41
t>
MVN

7.16
7.98
13.81
28.95

.12
.08
.20

.42
.78
1.20
.32

1.140
.58
32.39
li trogen -
Nil, -N
4

.73
7.86
7.32
15.91

.06
.02
.07

.98
.67
1.66
.16

.85
.06
18.71
- kg/ha/yr
Total
Soluble
N

14.33
28.06
37.05
79.44

.59
.26
.85

2.43
2.89
5.32
1.59

3.12
1.16
91.48
- 1982
Part.
N

6.55
5.36
6.93
18.84

.48
.13
.61

.05*
.16*
.21*
1.30

.95*
.000
21.91

Total N

20.88
33.42
43.98
98.28

1.07
.38
1.46

2.48
3.05
5.53
2.&9

4.07
1.16
113.39

*

18.4
29.5
38.8
06.7

.9
0.3
1.2

2.2
2.7
4.9
2.5

3.6
1.0
99.9%
 estimated




}ble 47.  Summary of nitrogen lending of  the  lalce  In  1981  and  1982.

-------
      All other sources
 Frisco &  Snake
     effluent
Snake  R
  14%
                 Precipitation
                    21%
    Blue R.
      30%
                         Tennmle
                           23%
                 241
  P
loa
      All other sources
 Frisco S Snake
     effluent

  Precipitation
Snake  R
 13%
                                      Blue R.
                                       32%
                     Tenmile
                        34%
                   load
Figure £7.  Itenization of sources of ? and X loading for lake Zil-.cn
         at the points of entry to the lake (average 1931. 19821).

-------
Overview of Phosphorus and Nitrogen Loading

     Table 48 expresses the phosphorus and nitrogen  loading  in several

different ways.  Because of phosphorus removal at  the  point  sources and

lew background phosphorus in the watersheds, the ratio  of  nitrogen to

phosphorus in the loading was very high.  The tnolar  ratios of nitrogen

to phosphorus were 69 in 1981 and 66 in 1982.

1981
N
P
1982
N
P
Per unit
kg/ha/yr
85.3
2.9
134.0
4.2
lake area*
g/m /yr
8.5
0.29
13.4
0.42
Entire
kg/yr
85,600
2,900
151,400
4,800
lake
Ib/yr
189,000
6,300
333,800
10,700
*Based on mean lake areas of 1004 ha in 1981 and 1130 ha  in  1982.
Table 48.  Summary of N and P loading.
     Figure 48 puts the phosphorus loadings into perspective  by  use  of

the widely-used Vollenweider diagram (Vollenweider 1968), which

predicts the trophic status of a lake on the basis of its mean depth

and its phosphorus loading.  The position of Lake Dillon on the  diagras

is shown with relation to the position of a small selection of other

lakes.  The diagram is presently considered to provide  only a crude

approximation of trophic status because it does not allow the

phosphorus loading to be discounted for varying degrees of phosphorus

_-edir.entation and loss through outflow.  However, the diagram is  in

escieral agreement with the lake trophic indicators.  A more detailed

treatment of this matter will be given in connection with modelling.

-------
                                                   243
  1.0
I   0.5
O
_J
  O.I
                                 EUTROPHIC
                L.Erie, 1968.

              L.Washington,  I960
                   Dillon, 1982
                                MESOTROPHIC
                                , 198!
                                 OLIGOTROPHIC
                                                L. Tahoe
                                                 I
                   10                 100

                        Mean Depth,  m
                                                           1000
Figure 48.  Position of Lake Dillon on the original Vollenweider
         diagram (Vollenweider 1968).

-------
                                                                      244
                Nutrient Export in Relation  to Land  Use




     Nutrient export must be considered in relation  to land  use.   The




central information base for this purpose is  the  representative




watershed program, in which nutrient loss was measured for  individual




small watersheds dominated by particular land uses.   The  names  and




locations of the representative watersheds are shown in Figure  49.  The




watersheds represented eight different land uses:  undisturbed




(background), roads, interstate highways, residential development  on




sewer (nonpoint component), urban development on  sewer (nonpoint




component), residential development on septic, ski areas, and Climax




{•folybdenum operations.  The nutrient yields from  these representative




watersheds are considered in sequence below.  For each of the watershed




types except the undisturbed type, the background nutrient yield  is




subtracted from the total yield and the residual  is  related




quantitatively to the intensity of land use by some  index such  as  the




number of persons per unit area or the percentage of  affected area.




After the nonpoint sources have been treated  in this  way, consideration




is given to the contributions per capita from small  and large point




sources.




     Wacer yield enters into many of the nutrient yield relationships




developed below for the representative watersheds.   The water yields,




given as mm/year of runoff, are in all instances  based on discharge




measurements.  Water yields vary a great deal; the causes of this




variation are numerous.  There is a pronounced increase in water  yield




with elevation.  Long-term records of snow accumulation also show  that,




for a given elevation, the western portion of the catchment  tends  to

-------
•JO
c
50
(D
XI
i-l
ro
in
rt
H-
rt
n>
n
en
3"
n>
Q.
en
                      LASKEY GULCH

                           (W7B)



                   DILLON VALLEY
STRAIGHT CREEK


 PORCUPINE GULCH
              WILDERNEST
        	(W2A %8
       /       ~"—---_
WEST  TENMILE (W5B)
                                                        KEYSTONE GULCH

                                                            (W6B)
                             COPPER

                               MTN (W4B)
                                              GULCH  (W2A 81)
                                                                    milts
                                                                                     i j
                                                                                     4.~
                                                                                     Ui

-------
                                                                      246





receive more winter moisture  than  the  rest.   Slope,  vegetation, and




land use also have complex effects.  A certain  amount  of  water is also




pumped or diverted in various places (e.g.,  local  irrigation).  For




1981, the catchment-wide average runoff at lakeside  was 180 mm and in




1982 it was 340 mm.









Undisturbed Watersheds




     Porcupine Gulch and Laskey Gulch  provided  information  about




nutrient yield from undisturbed areas.  Table 49 summarizes  the amount




of runoff from these watersheds and the phosphorus and nitrogen yield




in each of the two years.  In these and a number of  other




representative watersheds the particulate nitrogen contribution was




computed from the percent carbon and total particulates of  the stream




by use of a common C:N ratio equal to  that of the  Snake River,  since




the amount of particulate N was often  below  detection.  The  standard




error for phosphorus is reported on the basis of weighted averaging.




For nitrogen, the estimation of particulate  nitrogen from carbon




precluded estimation of the standard error,  but the  ratio of  S.E.  to




mean would be roughly the same as for  phosphorus.




     The phosphorus yields of Porcupine Gulch and  Laskey  Gulch were




close to those reported in the literature for cold temperate  areas of




similar soil characteristics.  For example,  Wright (1974) reported 1.5




ng/m~/yr P export for the Boundary Waters Canoe area of Minnesota,




and ::'chindler et al. (1976) reported 5.0 mg/m2/yr  for  the




~'--:"&-.-imental Lakes Area, Ontario.  The data  also show great  similaritv

-------
                                                                           247
                    Water Yield
                      nan/yr
  P yield - mg/m /yr
Total   S.E.   % Part.
N yield - mg/m /yi
  Total   % Part.
Porcupine Gulch  (W7A)

   1981                  315

   1982                  541
 1.55   0.34     17

 3.52   0.43     32
    75

   126
 7

35
Laskey Gulch (W7B)
1981
1982
Keystone Gulch (W6B)
1981
1982
Upper Snake (SR3)
1981
1982

129
306

108
328

240
500

0.50
2.77

0.57
3.19

1.14
4.37

0.10
0.55

0.08
0.44

0.15
0.55

33
33

33
38

61
54

29
62

14
46

64
103

12
58

19
41

15
51
Table  49.  Yield of water,  total P,  and total N from two watersheds
          representing  background (undeveloped) conditions (W7A, W7B), and free.
          two watersheds with roads but otherwise essentially undeveloped (W6B,
          SR3).

-------
                                                                      248
to data for the Coino Creek watershed at  a  similar  elevation in Eoulder



County (Lewis and Grant 1979).


     Both the total phosphorus and nitrogen  export showed  considerable



variation between years and between watersheds.  Table  49  suggests,



however, that this variation was not random; it was  associated with  the



amount of water yield.  As shown by the  study of Lewis  and Grant



(1979) for the Como Creek watershed, the increase  in soluble  phosphorus



yield between a dry year and a wet one is  expected  to be slightly


higher than the increase in water yield.   A  quantitative relationship



was sought between the water yield and the nutrient  yields  reported  in



Table 49.  Since the Como Creek studies  suggest that  the relationships


are likely to be slightly curvilinear for  undisturbed watersheds,  the


following equation was used:


                            v  = -y b
                            "n   a *w


where YW equals the annual water yield, Yn equals  the annual



nutrient yield (P or N), and a and b are the critical parameters of  the



relationship between nutrient yield and water yield.  The  equation was


log transformed and then tested by linear  regression  for fit  to the



data.  The fit was excellent (P < 0.01).   A  similar  procedure  for


nitrogen also showed a good fit (P < 0.01).



     Casual examination of data from watersheds with  only  gravel roads



or dirt roads suggested that the effect of these roads might  be



sufficiently small that they would not raise the nutrient  export much


above background.  Since this would be an  advantage  in  that  it would



offer a larger number of watersheds from which the  terms relating  water



and nutrient yield could be determined for background conditions,  two

-------
                                                                      249
watersheds differing from background  only  by the presence of dirt




or gravel roads were examined.   These included Keystone Gulch,  which




was a representative watershed  (W6B), and  a segment of the upper Snake




(SR3).  The data for these watersheds are  shown in Table 49.  Keystone




Gulch actually has near the watershed crest a small septic field that




serves the upper lodge at Keystone  ski area, but it is of sufficiently




small size that it cannot account for more than 2% of N or P export




from Keystone Gulch, so the septic  effect  was ignored.




     The data analysis showed that  the two background watersheds can be




lumped with the two watersheds  containing  roads.  When the equations




for the two undisturbed watersheds  only were used to  project the yield




from the two watersheds with roads,  and the projections were subtracted




from the observed yields, two of the  nitrogen residuals and one of  the




phosphorus residuals were slightly  negative, suggesting that the yields




of watersheds with roads were very  near background.  The exponent (b)




derived for phosphorus yield on the  basis  of the two  undisturbed




watersheds was 1.39, and with all four watersheds it  was 1.37.   The




coefficient (a) was higher by 20% for all  four watersheds than  for  the




two undisturbed watersheds, but this  is a  small amount in view  of the




certainty of variation in background  caused by differences in




vegetation, exposure, elevation, and  slope.  Ve therefore conclude  that




all four watersheds should be treated together, and that the effect of




small roads in undeveloped areas is minor  enough to merge with




background for present purposes.  Thus eight watershed years are




available for determination of  background  yield equations for N and ?.

-------
                                                                      250



     Table 50 summarizes  the results  of  the  log-transform regressions



used to determine the coefficient and exponent  of  the  background yield



equations for the eight watershed years.   The goodness of fit is



excellent both for phosphorus and for nitrogen.  For phosphorus the



exponent is slightly greater than 1,  as  expected from  the Como Creek



work.  For nitrogen the exponent is lower  than  for phosphorus.  Once



again, this is expected for undisturbed  watersheds of  this  type from



the Como Creek studies (Lewis and Grant  1979).  The Corno  Creek work



shows that nitrate yield  typically increases more  slowly  than water



yield, but that particulate N yield increases faster than water yield.



The two effects partially cancel, producing  an  exponent in  this



instance only slightly greater  than 1.0.


                              __


                   Exponent  Error    Coefficient    Equation

	(b)      (M	(a)
Phosphorus*         1.372     0.20     0.000782    Y =  0.000782  Y1'3/2
                                                    n             w




Nitrogen*           1.154     0.17     0.0842      Y =  0.0842  Y1'154
                                                    n          w

      r\
* mg/m~/yr


Table 50.  Summary of statistical information  on  empirical


           establishment of the relation Y  = ai   for background  vield
                                          n   w          3

           (Yv = mm/yr water yield; Yn = mg/m2/yr nutrient


           yield).  Both equations are highly  significant (P<0.001).
Residential Area on Sewer



     A. number of developments and subdivisions  in  the Lake  Dillon



w--:nershed consist of clusters of homes or condominiums  served  by sewer.

-------
                                                                      25




These make a point-source contribution, which  will  be  considered  at  the




end of this chapter, and they also potentially make a  nonpoint




contribution to nutrient loading.  The nonpoint  contribution is




associated with the presence of pavement, with changes in  vegetation




and soil cover within the development, with  fertilizer application,




with more or less continuous minor earth disturbance,  and  with other




miscellaneous factors arising from human habitation.   We consider this




classification to apply to settlements or developments that  have  enough




vegetative cover to break up runoff and thus slow the  transport of




materials to streams.  This classification does  not apply  to  areas that




are very densely settled (more than 80% land coverage  by concrete or




dwellings).  Such densely settled areas are  treated below  as  urban area




on sewer.




     Several representative watersheds were  initially  chosen  for




quantification of the effect of residential  development served by




sewer.  All of these except one proved impossible to ase.   Some were




settled too sparsely and others lacked sufficiently coherent  drainage




to give good quantitative information.  The  information is  thus based




on a single watershed containing the developments of Wildernest and




Mesa Cortina, the composite of which we shall  call  Wildernest.  Data




are available only for 1982.  The data are summarized  in Table 51.




This watershed has an area of 246 ha, of which 126  ha  is taken up by




the residential development on sewer.  The watershed is occupied  by




an annual average of 662 persons per day.  It  is ideal for  present




purposes in that it is settled densely enough  to show  any  effects




rather clearly, but is not so densely settled  as to constitute an urban

-------

Water yield, mm/yr
Use Intensity, persons/ha
Phosphorus 0
Total yield mg/a~/yr
Std. Error
% Particulate»
2
Background, mg/ra /yr
2
Net yield, ng/m /yr
Net yield per use unit: g/person/yr
Nitrogen
Total yield mg/m /yr
% Particulate,,
Background, mg/m"/yr
2
Net yield, mg/a /yr
Net yield per use unit: g/persan/yr
Residential on Sewer
Wildernest (W2A)
246
2.69

3.29
0.40
29
1.49
1.30
5.63

131
39
43
83
308
Urban on Sewer
Dillon Valley
(W2B)
326
6.83

14.7
4.2
100
2.19
2.89*
4.22

456
55
67
218*
318
* Net yield after subtraction of interstate highway effect as well  as
  background.

Table 51.  Yield of water and nutrients from a watershed supporting residential
           area on sewer (1982 data, nonpoint contribution only), and a  watershed
           supporting urban developtaent on sewer (1982 data, nonpoint only).

-------
                                                                      253
area.  Most of  the dwellings  or  groups of dwellings are separated by
trees and most  areas are unpaved.
     Table 51 gives  the total water yield and total phosphorus and
nitrogen yields for  the watershed.   The background, as computed from
the background  equations,  has been  subtracted from these to give the
net yield attributable to  the land  use under consideration here.  Table
51 expresses this net yield in terms of the use intensity index, which
in this case is the  number of persons.  The net yields per use unit in
1982 were 5.63  g of  phosphorus per  person per year and 308 g of
nitrogen per person  per year.  These are relatively small in relation
to the sewage export per capita,  which would be in the vicinity of
400-600 g per year for phosphorus and about five times this amount for
nitrogen.
     The question arises whether  or not the yields shown in Table 51
are dependent on water yield.  It seems certain that there is a
positive relationship between nutrient yield and water yield, but this
relationship cannot  be quantified without more data.  It will be
assumed hereafter that total  phosphorus and total nitrogen yield have a
dependence on water  yield  of  the  same degree as documented for
background yields.  With the  exponent fixed by this means, the
coefficient can be determined algebraically from Table 51.  The
coefficient for P is 0.00295  and  for N is 0.536 to give g/person/yr.


Urban Area on Sewer
     Although we were not  able to sample in such a way as to segregate
cleanly any of  the cities  within  the watershed, we did find one snail

-------
                                                                       254
watershed segment almost exclusively on  sewer  that  could be treated as




an urban area.  This watershed  segment,  which  is  referred to here as




Dillon Valley, is shown in Figure 49.  Development  is  much more dense




in Dillon Valley than in Wildernest, which  accounts  for  the difference




in classification.  The watershed segment has  an  area  of 115 ha.   The




settled area is 65 ha, of which 62 ha is occupied by dwellings  or




roads.  The area is occupied by a time-weighted average  of 786  persons.




The segment is sufficiently small and complex  hydrologically that




discharge could not be estimated easily.  We therefore assumed  runoff




per unit area equal to Straight Creek.   Computation  of the urban  effect




also required subtraction of interstate  highway effects.   The basis of




this correction is given in the section  on  yield  from  interstate




highways.  The results are summarized in Table 51.




     The total yield of nitrogen and phosphorus for  Dillon Valley was




much higher per unit area than  for residential area  on sewer, but this




was expected because of the higher density  of  development.   On  a  per




capita basis, the yield above background was similar to  the  yield for




the residential area on sewer.  As with  the sewered  residential area,




we assume that the yield of both nitrogen and  phosphorus  is  related to




the water yield by an equation incorporating a common  exponent  (1.372




for P,  1.154 for N).  Once the exponents are thus set, the coefficients




of the equations can be determined algebraically  from  the  data  in Table




51, are 0.00150 for P and 0.410 for N to give  the yield  in units  of

-------
                                                                      255




Residential Areas on Septic  Systems




     Representative watersheds  for residential areas on septic systems




were South Barton Gulch  and  Illinois  Gulch.   Data are available both




years for South Barton Gulch and  in 1981  only for Illinois Gulch.  The




data are summarized in Table 52.




     South Barton Gulch  has  a moderate  density of development; 194




persons are distributed  over 817  ha of  watershed.  Illinois Gulch is




much more sparsely settled;  it  has a  time-weighted average of 25




persons distributed over a watershed  of 496  ha.  The three watershed




years provided a good range  of  water  yields.  Table 52 shows  the net




nutrient yield from each of  the watersheds after subtraction  of the




background.  The net yield is then expressed as per-capita phosphorus




and nitrogen contribution.   The data  for  nitrogen and for  phosphorus




suggest that water yield greatly  affected nutrient export  from these




watersheds.  This is to  be expected,  since nutrients stored by septic




systems in the soil are  carried out in  proportion to the  amount of




water percolating through the soil.




     The per capita phosphorus  yields were plotted against water yields




for each of the three years,  and  showed evidence of a strongly




curvilinear relationship.  By regression  following log transformation




it was established that  the  equation  of best fit is Yn =




3.44YW0'759 where Yn is  yield above background expressed  as




g/person/yr and Yw is water  yield (mra/yr).   The fit is good (P =




0.04, S.E. of exponent is 0.10).  The exponent is less than l.C,  unlike




the exponent for background  P yield.  The lower exponent  implies that




higher water yields were accompanied by P yields that were higher by a

-------
                                                                           256
South Barton Gulch Illinois Gulch
(W1A) (W2A/81)

Water yield, cm/yr
Use intensity, persons/ha
Phosphorus „
Total yield Eg/m /yr
Std. error
% Particulate,,
Background, mg/nVyr
2
Net yield, mg/m /yr
Net yield per use unit: g/person/yr
Nitrogen ^
Total yield ng/ra"/yr
% Particulate
Background, mg/ia /yr
2
Net yield, mg/m /yr
Net yield per use unit: g/person/yr
1981
190
0.147
4.14
1.61
74
1.04
3.06*
208

40
56
36
1.9*
129
1982
522
0.147
9.83
1.13
61
4.19
5.50*
374

397
72
115
275*
18722
1981
70
/ .1
0.05C
0.70
0.16
38
0.28
0.42
83

13
50
11.7
1.3
257
*  Contribution of 74 persons on residential sewer has been subtracted as well
   as background.

Table 52.  Yield of water, nitrogen, and phosphorus for watersheds containing
           residential septic systems.

-------
                                                                      257




less than equal proportion.  Individual  septic  fields  thus  appear  to




respond in a manner intermediate between that of  background  sources,




whose yield increases faster than water  flow, and a true  point  source,




whose yield would not be affected by water  flow.




     The regression procedure was repeated  for  nitrogen.  The  fit  with




a curvilinear model was poor; fit with a linear model  was much  better.




The resulting equation is Yn =  44.3YW-5205, where Yn is yield of




N as g/person/yr and Yw is mm/yr water yield.   The standard  error  for




the slope is 11.8 and the significance of the relation is P  = 0.08.




The cause of the negative intercept is clear from Table 52:  N yield




above background becomes vanishingly small  at water yields well above




0.  This is explained by biological uptake  and  retention  up  to  a




certain flushing threshold, by  denitrification, or by  a combination of




these.









Ski Slopes




     The water and nutrient yields from  Keystone  and Copper  Mountain




ski slopes (including lodges, trails, and lifts)  were  quantified in




1981 and 1982.  Contributions from the associated residential areas and




point sources are not considered here.   Table 53  summarizes  the data.




The yield of phosphorus was well above background.   The nitrogen yield




was also above background, but  less markedly so than phosphorus.   The




percentage of each watershed accounted for  by cleared  areas  was used as




a neasure of use intensity; for both resorts about 40% of the area is




cleared.  The number of skiers  would probably serve equally  well as ar.

-------
                                                                           258

Water yield, mm/yr
Use intensity, % open
Phosphorus „
Total yield, mg/m /yr
Std. Error
% Particulates
Background, mg/m /yr
2
Net yield, mg/m /yr
2
Net yield per use unit - mg/m used/yr
Nitrogen „
Total yield, mg/m /yr
% Particulate
2
Background, mg/ra /yr
Net yield, mg/m~/yr
2
Met yield per use unit, ag/n used/yr
Keys tone
(W4A)
1981
98
40

1.94
0.41
49
0.42
1.52
3.84

27
59
17
10
25
Copper Mountain
(W4B)
1982
195
40

5.51
1.30
67
1.08
4.43
11.20

75
53
37
33
96
1981
348
42

3.56
0.99
44
2.40
1.16
2.40

74
27
72
2
4
1982
329
42

6.29
1.49
47
2.2'
4,07
8.43

88
37
67
21
44
Table 53.  Yield of water and nutrients from watersheds supporting  ski  slopes
           (nonpoint contributions only).

-------
                                                                      259

intensity measure.  In 1981 and  1982  Copper  Mountain and Keystone had

between 500,000 and 700,000 skier  days  per year.

     The water yields from Copper  Mountain were higher than for

Keystone; this is explained partly by elevation and  partly by location

along the east-west moisture gradient.   There  is  no  indication in Table
            «
53 of the close relation between water  and nutrient  yield documented

for most other land uses.  The explanation is  not obvious, since the

mechanisms of nutrient release are complex.  Presence  of the  trails  and

slopes undoubtedly facilitates transport, even without human  activity,

especially since ski trails and  slopes  must  be cleared parallel to  the

gradient.  Possibly more important, however, is mechanical damage to

the ground surface when snowpack is minimal.   There  is also a certain

amount of spring fertilizer application on the slopes.  Yields at

either Keystone or Copper Mountain could have  been augmented  by

construction on or near the slopes.

     Since the yield from ski slopes  does not  show any indication of

sensitivity to water yield, we use a  mean.   The mean for the  four

watershed years in Table 53 is 6.5 mg/m^ of  cleared  area/year above

background for P and 42 for N.



Interstate Highways

     The effects of interstate highways were isolated  in the  upper

segments of Straight Creek and West Tenmile  Creek.  Table 54  su^narizes

the data and expresses yield above background  in  relation to  area of

roadway plus right of way.  For  phosphorus,  the yield  was unexpectedly

high in both watersheds.  The relation  of P yield to water yield

-------
                                                                            260
Straight Creek
(W2B2)

Water yield - mm/yr
Use intensity, % roads
Phosphorus 9
Total yield - mg/n*~/yr
Std. Error
% Particulate
2
Background - mg/m /yr
2
Net yield - mg/m /yr
2
Net yield per use unit - mg/m used/yr
Nitrogen „
Total yield - mg/m /yr
% Particulate
2
Background - mg/m /yr
2
Net yield - mg/m"/yr 9
Net yield per use, intensity - nig /m" used/yr
1982
326
2.97

4.09
0.84
60
2.19
1.90
64

102
61
68

34
1146
West Tenmile
(W5B)
1981
522
0.78

5.38
1.76
47
3.80
1.57*
200

181
18
115

66*
8426
1982
662
0.78

7.19
0.67
50
5.51
1.67*
213

123
36
151

-28*
-3574
* a small contribution from Vail pass public  toilet  septic  field has been subtracted.

Table 54.  Yield of water, N, and P from watersheds  with interstate highways but
           little else.

-------
                                                                       261




was tested  for  fit  to  a function of the form already used for other




nutrient sources.   The fit  was  reasonably good, and resulted in the




equation Yn = O.OOZOgy^'799 where Yn is nutrient yield per




m2 of highway plus  right  of way and YW is water yield (ran/yr).  The




slope has a standard error  of 0.53 and P = 0.09 (ideally P should be




below 0.05, but  the power of the test is low with N=3).  High yield is




probably due in  part to major disturbance of vegetative cover.  Also




the materials added to roads in the winter may contribute, as does the




fertilization associated  with revegetation programs now in progress.




     The nitrogen yield associated with roads was low by comparison




with other  sources  and less consistent than P,  as shown by Table 54.




The N yield above background for West Tenmile was negative one year.




This is quite possible due  to excess of uptake or denitrification over




yield, but  presents some  difficulties in prediction.  For purposes




of curve fitting, the  two yields from West Tenmile were averaged,




giving a low positive  yield.  The fit to an exponential equation was




then reasonably  good:   Yn = 1.71Y,,1-13* S'E' sl°Pe = °-40> p =




0.11, where the  units  are the same as for the phosphorus equation.









Mining




     The Climax  Molybdenum  mining operation is  a unique mixture of




nutrient sources and must therefore be treated  separately fror. other




land uses.  Water from Climax Molybdenum has a  high suspended and




dissolved solids content.   The  water is retained in tailings por.cs,




where sedimentation occurs,  and the ponds are purged during high water-




The ponds receive not  only  industrial wastewater but also domestic

-------
                                                                      262




effluent derived from the toilets and  kitchens  of  the  plant.   The




domestic effluent is not treated, except by  sedimentation in  the




pcnds.




     Table 55 summarizes the water and nutrient  yields from upper




Tenmile Creek where the active mines are located.   There  is little else




in this watershed except old mines and State Highway 91.   Roads,




tailings, and bare areas around the plant account  for  753 ha  of the




total 6447 ha.  Virtually all of the nutrient yield above background




can thus probably be attributed to Climax Molybdenum.




     The P yields were above background and  the  N  yields  were  even more




so in 1981 and 1982.  Neither seems related  to water yield.  Unlike




most watersheds, the total nutrient yields for upper Tenmile  were




higher in 1981 than 1982.  In 1981 the time-averaged work force at




Climax Molybdenum was 2650 persons and in 1982 it  was  1100 persons.




Tne yield of both N and P thus seems related to  size of work  force.




Table 55 expresses the yields both years on  a per  capita  basis; the




yield per capita was almost identical for the two  years.   We  will thus




treat N and P yield at Climax Molybdenum as  a function of work force




size.




     Tne ? yield per capita at Climax Molybdenum suggests that the ?




output above background is almost entirely a sewage contribution and




not an industrial effect, since the yield per capita is in the range of




what would be expecced from the domestic output  of  the workforce.  The




ponds appear to be effective in reducing particulate P losses  that




would result from earth disturbance.  N export is very high,  suggesting




major augmentation by industrial activity.  Ammonium nitrate  is used in

-------
                                                                       263
                                           Upper Tensile - W5A
                                         1981
                    1982
Water yield - nnn/yr

Use intensity - persons/ha

Phosphorus           „
435

  0.41
513

  0.17
Total yield - rag/m /yr
Std. Error
% Parciculate
2
Background - mg/m /yr
2
Net yield - mg/m /yr
Net yield per use unit -
g/person/yr
Nitrogen -
Total yield - mg/m /yr
% Particulate
2
Background - mg/m*"/yr
2
Net yield - rag/m /yr
Net yield per use unit -
g/person/yr
7.07
1.55
20
3.26
3.81
93


821
5

93
728
17710

5.84
0.64
58
4.09
1.75
103


491
15

113
378
22150

Table 55.  Yield of water,  N,  and P for the raining area on upper
           Tenmile Creek.   Use intensity is based on nunber of persons
           working at  Climax Molybdenum.

-------
blasting and may be the source of this augmentation.   Revegetation may

also contribute.



Point Sources

     There are two types of point sources in  the Lake  Dillon catchment:

large sources with tertiary treatment and small sources  with secondary

treatment.  We will establish here a contribution  per  capita for  each

category.  Large point sources with tertiary  treatment include  the

Snake River WWTP and Frisco WWTP, whose effluents  were sampled

directly.  The results for these two have already  been reported in the

chapter on total loading of trie lake, but have not yet been  converted

to a per capita basis.  In addition, data can be obtained  for the

Breckenridge WWTP as the difference between stations ER1 and BR2  on the

Blue River (just downstream and just upstream of the Breckenridge

WWTP outfall).  These data are summarized in Table 56.   The  Copper

Mountain WWT?; designed as a tertiary plant, was in a  transition  from

secondary to tertiary during the course of the study and will therefore

not be considered here as a representative of tertiary treatment.

                    PersonsP - g/person/yrN  -  g/person/yr
	Served*    1981	1982	1981  '     1982

Frisco WWT?          3300       11.7     84.8        1108           955
Snake River WWT?     3230        9.1     22.3        1003          1260
Breckenridge WWT?    6700      113.      44.0        1378          1479

* Annual average persons per day
Table 56.  Yield of P and N from three WWTP plants on  a  per  capita
           basis.  All three plants practice  tertiary  P  removal.



     Table 56 shows the effectiveness of the  tertiary  treatment,  since

all of the per capita ? yields are well below the  raw  per  capita

-------
contributions coming to point  sources  (400-600 g/person/yr).   At  the

same time, there was considerable variation  in P yield  per  capita

between years, even at the  same  plant.   This is explained  entirely by

breakdown or shutdown of tertiary treatment,  which  occurred  at all

plants on one or more occasions  over  the course of  the  study.   Although

it could be argued that tertiary sources should be  represented by their

yields when treatment facilities are  operating properly, which would

give yields about half as great  as  the  observed average (Table 57),  it

is probably realistic to assume  that  the plants will  be impaired  for  a

certain amount of time each year.

                                              g/person/yr

Phosphorus
   Secondary treatment                            270
   Tertiary treatment                              47

Nitrogen
   All treatment                                 1197

Table 57-Average point source  yields  on a  per capita  basis.



     Nitrogen yields from the  point sources  were much more uniform,

since they were not affected as  much  by variations  in plant  operation.

The Breckenridge plant practices ammonium removal,  but  this  does  not

affect total N yield per capita  (Tables 58,  57).

     The P contribution from small  point sources (package plants)

subject to secondary treatment was  estimated from data  taken  upstream

and downstream of such sources along  the Blue River above Goose Pasture

Tarn.  The results are reported  in  Table 57.

-------
                                                                      266




Overview of Nutrient Yields




     Export of N and P from various sources is  expressed  in  Figure 50




on a per capita basis wherever this is meaningful.   In  Figure  51 the




export rates are recast on an area basis, assuming runoff  is near  that




of an average year.  Among the point sources, the potency  of sources




without tertiary treatment is striking.  For nonpoint sources,  septic




systems stand out, as expected, as do interstate highways, which might




not have been expected.

-------
                       P,  g/person/yr                   25:

    0       40      80      120     160    200     240     280
     _ Plant with Secondary P Treatment
   P
   N
    Plant with Tertiary  P Treatment
P
N
       Climax  molybdenum,  per worker year
   P
   N
       \ \ V\]
                      1
       Septic

   N
       Urban on sewer

   N


       Residential on sewer
   pg
   Nf
       Ski  Slopes
                                       \ \ I Point source
                                          I   I Non-point source !
                     6       9       12       15       18      2!
                        N,  kg/person/yr

Figure 50.  Nutrient yield from various sources expressed  on a pe.
          carita basis,  assuming 300 :m runoff.   Sources that car.r.cr
          be ixoressed on a per capita basis are  r.et shown.

-------
                   P,    mg/mVyr

0      20     40    60     80    100    120
                                                        140
      Urban with Tertiary P Treatment,  12 persons/ha
   p|\\ \\\\\\\\1i
      Residential with Tertiary P Treatment,  2.7 persons/ha
      Residential with Secondary P Treatment,  2.7 persons/ha
   P i\ \ \V\ \ \\X\\\\\ \T1
   ?
   N
   P
   P
   PQ
   P
   N
      Residential on Septic, 0.5 persons/ha
      CMmax , 753 ha in use
      interstate, pavement plus right cf way
      Ski areas,  40 % cleared
      Undisturbed
                                       Point source

                                       Non-point source
   0        2000      4000      6000      3000

                      N,     mg/m2/yr

Figure 51.  Nutrient yield from various sources expressed on an areal
          basis, assuming runoff  of 300 mm.

-------
                                                                       269




          Separation of Nutrient  Sources  within the Watershed




     Chemistry and discharge measurements were taken at a number of




points along each one of  the three  major  rivers (Figure 3).  Each of




the river drainages was thus divided  into a  number of watershed




segments.  If the land use within any given  segment is known




quantitatively, the equations  that  have been developed in the previous




chapter for nutrient yield associated with various land uses should




provide a prediction of the nutrient  yield from that segment.  Three




important factors may cause deviations between the observed and




predicted nutrient yields from any  given  watershed segment.




     (1) Estimation errors.  The  observed yields for any given segment




are based on chemistry and discharge  measurements, both of  which are




subject to analytical error.   We  may  also include under the heading of




error variance the confidence  limits  around  any one of the  curves used




to predict nutrient yield from land use.   These confidence  limits




reflect not only error variance in  the original data from which the




curves were developed but also a  certain  amount of individuality ir.




watersheds with respect to soils, vegetation,  exposure,  slope,  and




elevation.  If these sources of error were the only cause of deviations




between predicted and observed, we  would  expect to see a random




assortment of positive and negative deviations, between the  observed and




predicted values.




     (2) Unexpected sources.   The second  cause of deviation between




•'bserved and expected yields for  individual  segments is  the preser.c.  of




nutrient sources that are either  unquantifiable or unknown.  In the cr.e

-------
                                                                       170
instance where such a source proved to be  important,  it  was  possible to




determine the source of unusual and unexpected nutrient  yield.




     (3) Storage.  The third factor affecting  the  agreement  between




observed and predicted yields is a storage effect  associated with areas




of low relief in the river bottoms.  This  has  not  been incorporated




into the equations used for prediction.  The equations were  developed




on the basis of small watersheds off the main  channel where  individual




land uses could be isolated.  When nutrients flow  to  lower elevation,




and particularly into a much broader stream channel incorporating




low-lying areas and wetlands, significant  amounts  of  nutrient can be




stored in years of low flow and purged in  years of high  flow.




Mechanisms of phosphorus storage include sedimentation of particulate




phosphorus, biological uptake of soluble phosphorus,  and ligation or




inorganic precipitation of soluble phosphorus.  For nitrogen, the




possibilities are more complex, including  all  of the  possibilities  for




phosphorus plus cenitrification, by which  oxidized nitrogen  species  are




converted to nitrogen gas, which is subsequently released into  the




atmosphere.  All of these phenomena are most likely in the river




bottoms and associated flood plains and wetlands.




     Nutrients lost by sedimentation and ligation  or uptake  in  years of




low water may be returned to the open channel  and  thus enter the  lake




during a year of high water when the physical  forces  at work in the




river bottoms are sufficient to move large amounts of accumulated




materials.  Thus the river bottoms can act as  a storage  site that




retains a portion of the nutrient yield in dry years  and gives  back




much or all of this retained portion in wet years.  These storage and

-------
release phenomena are most  correctly regarded as temporary changes in




the watershed nutrient  inventory from one year to the next, and will be




referred to below as the  "inventory change."  If inventory change is




important in determining  the  yield of nutrients to Lake Dillon, we




should see in dry years a consistently lower ratio of observed to




predicted yield in river  segments incorporating areas where the river




water moves through wetlands, gravel beds, or ponds.  If much or all of




the storage (increase in  inventory) is given up in a wet year (decrease




in inventory), we would expect consistent increase in the ratio of




observed to predicted yields  from these segments in a wet year.  If




nutrient inventory change is  shown by the data from individual streaa




segments, then additional equations must be developed to account for




year-to-year inventory  change of nutrients due to their accumulation in




low-lying areas in dry  years  and their release from these areas in wet




years.









Snake River Drainage




     The Snake River drainage was divided into four segments as shown




in Figure 52.  Histograms of  the observed and predicted yields are




superimposed on each one  of the stream segments.  Phosphorus yields are




presented on the left side  of the river channels and nitrogen yields or.




the right side of the river channels.  For each year and for each




segment, one histogram  bar  shews the observed yield fron the segner.t,




ard the adjacent bar shews  the predicted yield.  The scales for t':.e




yields are all the sanie for a given nutrient, but differ between the




two nutrients because of  the  larger yields of nitrogen.

-------
        OQ
        C
        i-f
        fD
fu  O pd O
3  3" H- cr"
'<:  fu <  C/i
   3 it)
   3
OQ
H-
<
(D
        fD
   ft)
   M s;
  -  |D
     rr
   fu (D
en  3 n
fD

«8
fD
     0)  O.
   3 fD -a
   H- a. t-t
   o    H-
  OQ 'rJ n
   fD ^ rt
   3 O  ft)
     cn  fi,
   H- 'a
   cn 3^ '•^
     o  H-
   o ^i  m
   3 £  M
     cn  fx
   rr    w
   D" H-
   (D W  Hi

   i-i O  O

  OQ
   3* rt  CO
   rt 3" ft)
     (t> 00
   O    S
   HI h-- a,
     ft  3
   rt t-ti rt
   3J rt  Ln
   fl)
     O  O
   D Hi Hi
   3'
   fu (t  rt

   3 (6  fD
   (D
   I-1 cn  cn
     rt  3
   H* M  [11
  3 fD  JT"
     P  ft)
                                      PHOSPHORUS
                                          3
                                          \
                                       E
                                          0
                                                                                      NITROGEN
                                                                                           - 150
                                     SNAKE

                                     RIVER
                                                         OBSERVED


                                                         PREDICTED
                                                                                                                                      ro
                                                                                                                                      ~-j
                                                                                                                                      10

-------
     For phosphorus  the agreement  between predicted and observed yields




from individual stream segments  is  generally very good.  One notable




exception is  the phosphorus  yield  in 1982 from segment SR2,  which




incorporates  Keystone Gulch,  Keystone Ski Area,  and Keystone




Development.  The observed phosphorus yield from this segment was




extraordinarily high in 1982;  it was far  above any expected  random




deviations from the  predicted.   This exceptionally high observed




phosphorus yield is explained  by construction activities.  Keystone Ski




Area installed six 300-hp pumps  in  the river bed during 1982 and




constructed an associated pipeline  to carry the  water from thesa pucps




to ski slopes, where it was  to be  used for snowmaking.   Construction




was carried out at several other locations in various watersheds during




the course of the study but  did  not produce such extraordinary yields




of phosphorus.  The especially high yield in this instance is




undoubtedly accounted for by work  in or near the river  bed,  where water




movement greatly facilitated  the transport of phosphorus-bearing




materials loosened by the construction activities.  Since  the




prediction equations do not  take into account any such  special




activities, the gap between  observed and  predicted P of 11 mg/m'Vyr




can be accounted for specifically by the  construction.   Total 1982




transport due to the construction  in this segment was 295 kg of




phosphorus.   This was approximately 7% of the runoff yield of




phosphorus for the entire Lake Dillon watershed  during  1982.  Expressed




in another way-, it was equivalent to the  annual  yield of phosphorus




from a tertiary point source serving 6300 persons, assuming  efficiency




of point source operation as shown  in Table  57.

-------
                                                                      274
     There is little evidence for significant valley  bottom  inventory




changes affecting phosphorus yield in the four Snake  River segments.




The inventory change effect is unlikely in SR3 and SR4 because  of  the




steep relief in these segments.  The effect is most likely in segment




SR2, where it would be obscured by the large construction effect in




1982.




     Predictions for nitrogen are also in good agreement with the




observed values with the exception of the 1982 value  for segment SR2.




Once again, we associate the extraordinarily high yield in 1982 for SR2




with the Keystone construction activities.









Blue River Drainage




     The Blue River drainage was originally divided into seven




segments (BR1-BR7).  Figure 53 shows the segment yields based on a




total of four segments.  Segment BR1 has been omitted because this




small segment isolates only the Ereckenridge Wastewater Treatment




Plant, whose yield has already been discussed in the  previous chapter.




Some of the smaller segments between BR2 and BR7 have been combined for




present purposes because a considerable land area is  required to show




a quantifiable effect on the chemistry and discharge  of a river as




large as the Blue River (BR2 in Figure 53 includes the original 3R2 and




ER3; BR4 in Figure 53 includes original BR4 and BR5).  In addition,




diffuse drainage in the vicinity of Breckenridge made impossible any




reliable determinations of discharge there.




     There were no extraordinary departures of observed from predicted




phosphorus yield in the Blue River drainage.  The Blue River drainage

-------
                                                                    275
PHOSPHORUS
   3r
 0>
 E
   0
NITROGEN
       150
                                             BLUE  RIVER


                                                   Observed

                                                   Predicted
          53.  Observed and  predicted yields  from segments  of  the 31ue
              River watershed.  Phosphorus is on the left  of  the stra?.:
              channels, nitrogen on the right.

-------
                                                                       276
is a good place to look for valley bottom inventory  change because the




main channel, even as high as BR7, incorporates  extensive  wetlands,




small ponds, and accumulated old tailings in  the river  bed.   In every




segment of the Blue River drainage, the ratio of  observed  to  predicted




phosphorus was lower in 1981 than in 1982.  This  is  explained by tha




inventory change effect, which leads to the entrapment  of  a  certain




amount of the yield in dry years, and a release  of a portion  of this




material during wet years.




     In BR7, the nutrient sources drain through  extensive  areas of low




relief.  The predictions were consistently above  the  observed yields in




this portion of the watershed for phosphorus, and the ratio  of observed




to predicted was higher in 1982 than in 1981.  The data  thus  suggest




that significant retention occurred even in the  wetter  year of 1932,




although it was especially pronounced in 1981.   The  same was  true  of




BR6, which incorporates Goose Pasture Tarn as well as an extensive




diffusely drained area above the Tarn.




     Segment BR4 has a much different character  from  that  of  BR7 and




BR6.  The river bottom has been extensively modified, both by mining




and by settlement around the town of Breckenridge.   The  phosphorus data




show phosphorus accumulation in low-lying areas  in 1981 and a strong




purging of the stored material in 1982.  The behavior of Segment BR2




was more similar to that of BR6 than BRA.




     There was considerable difference between the yield patterns  of




nitrogen and phosphorus in the Blue River drainage.   For several




reasons, phosphorus and nitrogen cannot be assumed to behave  similarly




'-" any given watershed segment.  First, a great  deal  of phosphorus is

-------
                                                                      277




transported in particulate  form,  whereas  the portion of nitrogen




transported in particulate  form is  generally much smaller.   Thus the




factors that lead to entrapment of  significant amounts of particulate




phosphorus in a given watershed segment will not  necessarily have a




similarly strong effect on  total  nitrogen yield from that segment.




Furthermore, nitrogen yields can  be affected by denitrification losses.




Denitrification is especially  likely in low-lying areas with diffuse




drainage where anoxia can occur.  Finally,  the biological uptake of




phosphorus and nitrogen is  likely to differ-




     The ratio of observed  to  predicted nitrogen  yield for  individual




segments was higher in 1982 than  in 1981  for all  segments except BR7.




The data thus indicate operation  of the inventory change effect that




was observed for phosphorus.   However, it is clear that other factors




complicated the yield of nitrogen from most of the segments.  Negative




yields (i.e., N sources plus incoming N greater than outgoing I,") were




observed for segments BR2 and  BR6 in 1981.   Yields in segment BR6 were




exceptionally low even in 1982 by comparison with the prediction.  The




major deviations of segment BR6 from the  predictions are almost




certainly connected with losses of  nitrogen associated with Goose




Pasture Tarn.  Concentrations  of  nitrogen above and  below the Tarn




indicate that it is a major nitrogen sink.   Nitrogen loss in the larr.




is probably due to a combination  of uptake  by weedbeds,  sedimer.taticr.,




and denitrification in the  mud.

-------
                                                                       273





Tenmile Creek Drainage




     The data for Tenmile Creek are  shown  in  Figure  54.   Three segments




are shown.  Smaller segments have been  combined  for  reasons similar to




those mentioned for the Slue River.  The upper segment  is the site of




Climax Molybdenum, for which predictions are  good.   The  middle segment




(TM4) has numerous dispersed sources of nutrients, and  is also the




discharge point for the Copper Mountain Wastewater Treatment Plant.




The presence of the wastewater treatment plant complicates  the




interpretation of the data.  Tne Copper Mountain plant was  doing




secondary treatment during almost all of 1981, and the effluent was for




the most part pumped into a pond instead of into the  river  directly.




Tertiary treatment was in force, but was considered not  fully effective




according to plant operators, during the first ten months of 1982,




after which secondary treatment was resumed for  the months  of November




and December.  The presence of the wastewater treatment  plant, and  the




problems associated with its new treatment practices, make  the




estimates of other factors difficult because  the plant introduced  nuch




more irregular variation than would be expected  in other  watershed




segments.  There was major overestimation  of  phosphorus  for  TM4 in  1981




but not in 1982.  The 1981 overestimate is probably due  to  storage  in




the pond and in the river bottom in  the lower half of the segment.




Storage would be very likely in the poorly drained areas  known as




Curtain Ponds and Wheeler Flats.  However, segregation of this storage




and purging effect is more difficult in T>!4 than in any  other location




because of the changing operations of the  point  source.

-------
                                                           279
  PHOSPHORUS
    3r
NITROGEN


-

-

-
150

100

50

0

w
>•%
CVJ^
£
\
E

          81  p  82
                                            ENMILE  CREEK
 Observed

i Predicted
54.   Observed and predicted yields froci  segments of the  Ter.r.ile
     Creek watershed.   Phosphorus is  on  the left of the  strea~
     channels, nitrogen on the right.

-------
                                                                       ISO
     Segment TM1 incorporates a large  portion  of  the  town of Frisco.




Since the town is located just adjacent  to  the  lake,  the  drainage of




Tenmile Creek near its mouth is modified  in a  major way by the presence




of the city.  There is a large area of flat relief in the vicinity of




Frisco.  Probably because of this flat area, the  phosphorus and




nitrogen yield data for TML show evidence of inventory change between




the dry and wet years.  Both phosphorus  and nitrogen  yields for 1981




were negative, indicating that the areas  of low relief within this




drainage stored not only some of the yield  from segment IM1 but also




some of the yield of upstream segments.   However, the observed storage




in 1981 was almost exactly compensated by excess  of observed over




predicted yield in 1982, indicating a  purging  effect.   The purging




effect was less pronounced for nitrogen,  possibly because of permanent




nitrogen losses to denitrificaticn or biological  uptake.









Relationship between Runoff and Nutrient  Storage




     Since the data show that nutrients originating at various points




in the watershed are stored in the river  valleys  during years of  low




runoff and purged from these areas during years of high runoff, the




yield equations cannot predict total output from  the  three rivers until




they have been coupled with an inventory  change function.  An estimate




can be made of the function from the data at hand.  Inventory change




will be positive in those years when part of the  yield from areas of




steeper relief accumulates in areas of low  relief and will be negative




wner. part of this stored inventory is released  during  wetter years.

-------
                                                                       281





For nitrogen, the  function  will  be  considered to include any effects of




denitrification.




     The first step  in estimating  the inventory change function is to




sum up the observed  and predicted  total  yields from runoff for each of




the two years of the  study.   The lakeside point sources (Frisco, Snake




River, and Breckenridge WWTP)  are  excluded from the summation, since




they are so close  to  the  lake  that  the nutrients they give off are not




subject to storage.   The  summations  are  corrected for major




discrepancies from the predicted values  that are known not to be




connected with storage.   In the  two  years of record,  the only example




of this is the boost  of yield  in the Snake River watershed during 1982




caused by construction in the  vicinity of Keystone.  After this




correction, the remaining discrepancy between observed and predicted




yields from all sources combined is  assumed to be due to inventory




change.  Since inventory  change  is  dependent on amount of runoff, we




postulate a decline  from  positive  to negative inventory change as




runoff goes from the  lowest  to the  highest annual values.  A coroll.-sr/




is that the average  change  in  inventory  over a number of years is not




very great.  Even  if  the  watershed  is aggrading, as it may be with net




accumulation of vegetation  and filling in of excavated areas,  the




resulting departure  from  equilibrium over a number of years would




amount to a relatively small percentage  of nitrogen and phosphorus flux




(cf.  Bormann and Likens 1979).




     Figure 55 shows  the  apparent change in nutrient  inventory for tn=




two years of record for both nitrogen and phosphorus.  The nutrient




inventory change is  expressed  as a percent of watershed overland

-------
                                                               282
 o>
 c
 o
JZ
o
 >»
 V.
 o
•«••



 >
 o
 13
 C
                          Average  Runoff

                                 t
P: y»- 0.40x  + 87.6



  : y*-0.57 x  + 112.4
         100                       200                       300


           GAUGED  RUNOFF,   thousands  of  acre  feet/yr
   Figure 55.  Linear plot  of the  inventory change functions  for nitrogen

             and phosphorus.  Positive inventory change indicates

             storage.  Inventory change is expressed  as a percentage of

             the summed runoff nutrient sources.

-------
                                                                       283

nutrient yield, excluding  lakeside WWTP sources.  The percentages were

calculated as follows:

                              (Y    - Y   )
                       wn  =  _J^	£•£.. x 100
                                   n,o



where Wn is nutrient inventory  change (%), Yn o is observed

overland yield, and Yn e is  the expected overland yield from

summation of all sources with the  assumption that inventory change is

zero.  Figure 55 shows the relationship of Wn to discharge.  The two

points (1981, 1982) for each of the two nutrients are joined by

straight lines.  Although  the form of the relationship is  not

necessarily linear, numerous additional points would  be required to

show curvature.  Unless there is very marked curvature, the line will

provide reasonable approximations.

     The watershed appears to be aggrading with respect to phosphorus,

since an average water year  is  accompanied by a 10-15% retention of

yield in valley bottoms.   Nitrogen also shows some net retention in an

average year, but less than  phosphorus.  The equations for the  lines,

as indicated in Figure 55, will be used to represent  the expected

change in inventory in any particular year.



Comparison of Observed and Predicted Total P and N Loading

     The land-use equations, point  source equations,  and

inventory-change function  lead  to  predictions of total runoff yield.

All identifiable sources are dealt  with by these equations except rnajor

construction in or near stream  beds,  such as that observed near

-------
                                                                       284
Keystone in 1982.  Although the frequency of  this  activity is  largely

unpredictable, it will surely be above zero.  As an  approximation of

the effect, we use the average increase in yield over  the  two  years of

record expressed as percent of other runoff.  This boosts  the  P

prediction by 3.5% and the N prediction by 2.2%.

     When the equations for yield from all watershed sources are  used,

and a mean figure is used for contributions from precipitation and

groundwater, the result is an overall predicted loading of  N and  P for

the lake in 1981 and 1982.  These predicted loadings are shown in Table

58, which also shows the observed loading for comparison.   The observed

and predicted values agree very well (<10% difference).  Deviations of

predicted from observed are caused mostly by  the irregular  nature of

failures in tertiary treatment plants, and to some extent  by irregular

variation in other sources.

                    P load, kg/yr	          N load, kg/yr
  Year	Observed    Predicted	Observed    Predicted	

  1981            2900      -  2700          85,600       83,500
  1982            4800        4600         151,400      154,300

Table 58.  Comparison of predicted and observed N and  P loads  for the
           two years of study.



Itamization of Lake Nutrient Sources for 1981 and 1932

     In the chapter dealing with total nutrient loading of  the

lake, nutrient sources were itemized according to  the  major pathways  by

which they entered the lake (Tables 46, 47).  On the basis  of  the

r-;trient yield equations, nutrient sources can be further  broken  down

according to land use.  In making this breakdown, we use the values

-------
                                                                      285




reported in Tables 46 and  47  for  precipitation,  groundwater,  and




effluent contributions from the Frisco  and Snake River  plants.   The




contributions of the Breckenridge Wastewater Treatment  Plant  are taken




from the analysis of the watershed segment BR1  as reported in Table 56.




The contributions of all other  sources  are obtained by  application of




the yield equations and the inventory change equations  to  the land use




data and water yield data  for 1981 and  1982.  Since the ability of




these equations to predict observed yield  accurately for 1981 and  1982




has already been demonstrated,  it is expected that the  sum of




individual sources obtained by  this means  will  be very  close  to, but




not exactly equal to, the  observed loading.




     Table 59 shows the complete  breakdown of sources.   The table  gives




the kilograms of nitrogen  and phosphorus  from each source  reaching the




lake in 1981 and 1982, and also expresses  these  as a fraction of total




annual loading.  Although  data  for both years are given, it should be




noted that the distribution of  contributions for 1982 is much more




typical of the median or average  situation under present land use




conditions than is the distribution for 1981.  There are several




reasons for this.  First,  the runoff for  1981 was exceptionally low.




The sum of flows for the 22 years of record  at  the four U.S.G.S. gauges




averages 185,600 acre feet per  year. The  gauged flow for  1981  was




116,500, or 63% of the average.   Although  the runoff for 1982 was  above




the average (210,900), it  was only slightly  so  (14%) and thus




represents more accurately the median distribution of various




contributions to the total nutrient loading  of  the lake.  In  addition,




the wastewater treatment plant  at  Copper Mountain was operating under

-------
                                                Phosphorus
                                                      Nitrogen
                                       1981
                1982
                                    1981
                   1982
                           Mean %
Tertiary Outfall.-,
  Frisco
  Snake River
  Breckenridfie
  Copper Mountain
     Subtotal

Secondary Outfalls
  Package plants
  Copper Mountain
     Subtotal

Climax Molybdenum

Keystone Construction
 49   1.6
322  10.8
371  12.4
152
5.1
49
47
96
104
295
1.1
1.0
2.1
2.2
6.4
1.3
5.9
7.2
3.4
3.2
  198
 1305
 1503
                                         0.2
                                         1.5
                                         1.7
28583  33.0
                                                                                                          Mean
39
29
757
-
825
1.3
1.0
25.3
-
27.6
280
72
295
41
688
6.1
1.6
6.4
0.9
14.9
3.7
1.3
15.9
0.5
21.3
3656
3240
9233
-
16129
4.2
3.8
10.7
-
18.7
3152
4070
9909
1893
19024
2.1
2.7
6.5
1.2
12.5
3.2
3.2
8.6
0.6
15.6
198
378
576
23736
3883
0.1
0.2
0.4
15.6
2.6
0.2
0.8
1.0
24.3
1.3
Dispersed Nonpoint Sources
Residential
Urban areas
Septic
Interstate
Ski slopes
Background
Subtotal
Preci pi tation
Croundwater
areas (sewered)
(sewered)

highway

runoff



GRAND TOTAL
9
9
161
21
36
545
781
846
13
2988
0.3
0.3
5.4
0.7
1.2
18.2
26.1
28.2
0.4
99.8
32
31
383
132
59
2105
2742
606
88
4619
.7
.7
8.3
2.9
1.3
45.6
59.4
13.1
1.9
100.0
.5
.5
6.9
1.8
1.3
31.8
42.8
20.7
1.1
99.7
487
761
6713
461
214
16577
25213
14311
587
86326
.6
.9
7.8
0.5
0.2
19.2
29.2
16.6
0.7
99.9
1901
2702
21750
2198
428
69015
97994
5433
1549
152196
1.3
1.8
14.3
1.4
0.3
45.4
64.4
3.6
1.0
100.1
1.0
1.4
11.1
1.0
0.2
32.3
46.8
10.0
0.8
100.0
                                                                                                                     10
                                                                                                                     CO
Tali! c S9.  Hreakdown of nutrient  contributions  to Dillon in 1981 and 19S2.

-------
                                                                      287
secondary treatment during  1981,  whereas  the standard of operation in




the future, and for most of  1982,  is  tertiary treatment.  Finally, the




performance of the largest wastewater plant  (Breckenridge) in 1981 was




unusually poor due to equipment shutdown  and malfunction.




     The various nutrient sources  have been  grouped in a number of ways




in Table 60to emphasize different aspects of the  distribution of  the




load among sources.   In 1981 the  contribution of wastewater treatment




plants to the total loading  was exaggerated  by low runoff, which




increased the relative role  of  these  sources, and  by the difficulties




at the Breckenridge Wastewater  Treatment  Plant.  The 1982 situation,




which was more characteristic,  showed about  16%  of the phosphorus  and




about 13% of the nitrogen coming  from the four major wastewater




treatment plants combined.   Package plants made  only a minor  addition




to this total.  Thus  it is clear  that well over  three-quarters of




phosphorus and nitrogen loading under the present  land use conditions




comes from nonpoint sources  of  natural and man-made origin.




     Climax Molybdenum contributed about  five percent of the  total




phosphorus loading in 1981 and  about  half as much  in 1982.  The decline




is explained by the reduced  pace  of operations,  and finally the




shutdown in 1982, which drastically reduced  the  nutrient output froc




that source.  Thus for Climax the  1982 data  probably underestimate the




equilibrium contribution, assuming that the  mines  ultimately  resume




operation.  An equilibrium number  would thus be  more in the  vicinity  of




five percent than two percent.  Although  Climax  has not traditicnall..




been thought of as a  point source  for nutrients, it actually  behaves




much the way a point  source  would,  especially in the apparent  direct

-------
                                                                       288
P load, %
Categories
Natural vs. Man-made
Natural
Man-niade
Point vs . Non-point
Wastewater plants
Package Plants and
All other sources
Hunan Was te ( WWT? ' s ,
vs. all others
Human waste
Other sources
1981 1982 Mean

47
53

38
Climax 7
55
Package Plants and

45
55

54
46

16
3
81
Septic)

26
74

50
50

27
5
68


36
64
N
1981

36
64

20
33
47


28
72
load, %
1982 Mean

50 43
50 57

13 16
16 24
71 60


28 23
72 72
Table 60.  Percentage contributions to loading of the lake aggregated  in
           three different ways.  The 1982 data are most  representative of
           a median year.

-------
                                                                      289




dependency of the nutrient  yield  from the  source  on the size of the




workforce.  It has already  been noted that the  nitrogen yield from




Climax is exceptional, and  this is  shown very clearly in Tables 59 and




60.  The nitrogen yield of  Climax Molybdenum is greater than all four




wastewater treatment plants combined.




     Among the dispersed nonpoint sources, the  background contribution,




including everything that would be  expected in  the  absence of human




habitation, is considerably greater than all others combined.  Under




present conditions, it accounts for about  half  the  total loading of




nitrogen and phosphorus.  Among the dispersed nonpoint  sources  related




to human activity, septic systems stand  out clearly as  the highest




contributor.  All other sources account  for contributions of a  fraction




of a percent to a few percent each.  Although no  single dispersed




source related to human activity  is of outstanding  importance except




for septic systems, the aggregate effect of the dispersed sources is




significant.




     Table 60 shows that the natural background and the sources




associated with human presence share the phosphorus loading  almost




equally for phosphorus, and for nitrogen are only slightly skewed




toward the man-made sources.  In  general Table  60 shows that, under the




present land-use conditions, in a median year hydrologically, both  the




phosphorus and nitrogen loading of  the lake can be  considered to be




divided approximately into  four quarters.   The  first  two of  these




quarters are taken up by the natural sources.   The  third quarter is




taken up by sources associated directly  with human  waste (wastawater




treatment plants, package plants, septic systems),  and  the fourth

-------
                                                                       290




quarter is accounted for by human activity  not  related to the




processing of human waste.  In this last category  we  have a wide




assortment of mechanisms including earth disturbance,  change in




vegetative cover, fertilization, and others.




     Figure 56 gives a visual impression of  the  nutrient  contributions




from various sources.  The 1982 data are shown  in  the  figure because




these are most representative of the median  condition  at  the present




time.









Possibilities for the Reduction of Nutrient  Loading




     The foregoing analysis allows us to suggest a number of ways  the




nutrient load reaching the lake might be reduced.  The maximum




imaginable reduction under the current land  use  conditions  would be




about 25%, assuming virtually no limit on financial investment and




that no water is diverted away from the lake.  Approximately 50% of the




present loading (i.e.,  the natural background) is  completely




uncontrollable and roughly half of the remaining portion  would probably




be uncontrollable even with maximum investment.  Thus  the following




alternatives in aggregate probably represent no more than 25% of the




total present loading.




     A.  Major point sources.  There are four major wastewater




treatment plants in the drainage.  We have assumed up  to  this point




that all four plants will operate much as did the  Breckenridge,  Frisco,




and Snake River plants  in 1981 and 1982.  Since  the Copper  Mountain




P^ant -.ras changing ever to tertiary treatment during this interval, we




taKe t'r.e ctrier three T?lants as a better indication of  the immediate

-------
                                                    29:
All other sources
           Precipitation
            13%
                   Background  Runoff
                             Major treatment plants
                                               Septic
                                         PHOSPHORUS
All other sources
                   Background  Runoff
                            Major treatment plants
                                         NITROGEN
ngure
Percentage of loading due to various sources in 1982.
sources over 5% are shown separately.

-------
                                                                       192
future situation for all four wastewater  treatment  plants.   The three




operating tertiary plants added at least  twice as much  phosphorus to




the lake in 1981 and 1982 as they might have  if  they  had  been able to




operate continuously without equipment failures.  Thus  a  reduction of




something of the order of 8% in total phosphorus loading  to  the lake




might be achieved if tertiary plant failures  could  be held at a level




well below the level observed in 1931 and 1982,  or  if there  were some




kind of fail-safe system for each plant.  Perhaps the disruptions of




operation will be far less numerous when  all  of  the plants have reached




their capacity and thus are no longer under construction  or  expansion.




If so, then a significant reduction of nutrient  loading from this




source may occur-  At any rate, it may be important to  pay closer




attention to the occurrence and duration  of plant failures and to




embark upon a more vigorous program by which  these  can  be minimized  or




avoided.




     B.  Climax Molybdenum.  Climax Molybdenum is a significant source




of phosphorus and behaves essentially like a  point  source.   With the




workforce at 1981 levels, the phosphorus  contribution from this source




will be roughly five percent.  If tertiary treatment  standards  were




applied to this source for phosphorus, the contribution could be




reduced considerably, perhaps to as little as one percent of  the total




phosphorus loading of the lake.




     C.  Construction.  The 1982 data on  the  Keystone construction in




the -naVe River bottom illustrates the potential exacerbation of normal




phosphorus loading by heavy construction  in or very near  the  river




beds.   Although such construction is not  routine, it will probably

-------
                                                                      293




continue to occur at irregular  intervals.   Precautions to prevent the




dispersion of disturbed earth and  sediments from entering the stream




channel might easily reduce  this source  by half.




     D.  Sewered residential areas.   Runoff control in sewered




residential areas, including especially  the densely settled urbanized




zones, could potentially reduce nutrient input  to the  lake by one




percent if applied to all sewered  areas.  Although reductions of this




magnitude may seem scarcely  worth  bothering with, it is obvious  that




any and all sources of nutrient reduction should be considered if




additional development of the watershed  is to  be expected.  The  data




show that nutrients associated  with  sewered residential areas, and




particularly with urban areas,  are transported  specifically in the




wettest years, and presumably at the time of greatest  runoff.




Technology for the ponding of runoff of  this type is available and




would be of interest in reduction  of the nutrients from this source.




However, removal of the trapped nutrients  to prevent later transport to




the lake would be essential  if  the control program were to be really




successful.




     E.  Septic Systems.  Septic systems are a  major contributor of




phosphorus.  If septic systems  could be  replaced by sewer, the yield




from these areas could probably be reduced by half (from about eight




percent to about four percent of the total). Greater  reduction  than




this is probably unrealistic for several reasons.  First, a certain




fraction of the contribution is associated with the presence of  a




residences and not with the  presence of  septic  fields.  Secondly,




abandoned septic fields will presumably  continue to yield phosphorus

-------
                                                                      294





and nitrogen for some time to come, and  the  decline  in  yield  may not be




immediate.  Finally, tartiary treatment  following  the addition of sewer




cannot be expected to remove all of the  phosphorus.




     Present and future yield of septic  systems  could probably be




reduced significantly by proper maintenance  of  septic fields.  This  has




not been a local government priority in  the  past but perhaps  could  be




in the future as part of a nonpoint control  program.




     F.  Interstate highway.  Interstate highway contributed  2.8




percent of the total phosphorus yield in 1982.   This is  relatively  high




in view of the fact that interstate highway  does not cover  a  very large




percentage of the watershed and does not require the disposal of any




human waste.  In other words, interstate highway is a surprisingly  rich




source of phosphorus on a unit area basis.   It  is  difficult to




prescribe control measures for this source,  since  the mechanism by




which phosphorus yield comes from interstate highways is  not  clear.




The application of phosphorus for revegetation  should be  computed to




see if it is potentially significant.  Possibly  some very localized




sampling would be of use in showing more precisely the nature of the




source connected with interstate highway.  Containment systems may  be




feasible, particularly if the source is related  to the transport of




particulates from eroding areas or with heavy use  of




phosphorus-containing fertilizers (this may  decline with  time).   Quite




possibly this source could be cut in half, although  the  expense may  be




considerable if erosion control is involved.




     G.  Ski slopes.  The yields from ski slopes accounted  for slightly




more than one percent of the phosphorus load in 1982.  This is

-------
                                                                      295
obviously not a major increment, but  It  Is  still  vulnerable  to  some




control because of its localized nature.  As  with interstate  highway,




the prescription is difficult because exact mechanisms  of  increased




nutrient yield are difficult to pinpoint  and  in an all  probability are




diverse.  Fertilizer use, mechanical  action of  skiers on  the  slopes,




and erosion aggravated by construction may  all  play a part.   Ponding  of




runoff by a rationale similar to that which is  used in  urban  areas is




one possibility and might conceivably reduce  the  phosphorus yield  to




half its present amount.




     One final possibility is the addition  to Dillon of significant




amounts of water lower in phosphorus  than the average runoff  currently




entering the lake.  This would in effect  lower  the mean phosphorus




concentration of entering water and thus  lower  the equilibrium




phosphorus concentrations in the lake.   Although  this alternative  is




almost certainly not feasible solely  as  a mechanism for the control of




trophic status of Lake Dillon, it may be  considered as  a potential side




benefit to various water diversion projects,  but  only if those  projects




involve the use of water having very  low phosphorus concentrations.

-------
                                                                      296





                     The Dillon Clean Lakes Model




     A major purpose of the present study was  to develop  a  model that




could assimilate information on changes in land use  and  from this




predict changes in the lake, and especially those  changes of aesthetic




or economic importance.  We have developed such a  model  from the




information on the lake and watershed nutrient yields.  We  will  refer




to this model as the Dillon Clean Lakes Model.




     The Dillon Clean Lakes Model consists of  three  parts,  which are




shown in diagrammatic form in Figure 57-  These are  the land use




component, the trophic status component, and the effects  component.









The Land-Use Component of the Model




     The land-use component requires as input  a matrix of land use




information.  The land-use matrix specifies the intensity of different




types of use (on an area or population basis)  for  each of 19 watershed




segments (Figure 58).  The land-use component  also requires  the  amount




of gauged runoff (sun of the four U.S.G.S. gauges) for the  year  that  is




being modelled, the amount of water to be pumped cr  diverted from other




watersheds, and the mean concentration of this pumped or  diverted water




on a nonth-by-month basis.  The land-use component accepts  as input the




water yield from ail watershed segments, but will  compute these  yields




from the total gauged runoff if the yields are not supplied.




     The land use computations are organized on a  segment-by-segment




basis.  The nutrient export from each segment  is computed from the




water yield and the land-use matrix by use of  the  equations  that were




developed in the analysis of nutrient yields.  The yields are sunmed

-------
 CW
 c
O rt
M M
TO C
(U O
'J '<
  C
f '-I
fa ft,
f<"
n> o
en 11,
                 INPUT


                    Gauged runoff —

            Water yield by segment^-

      Amount  of diversion by month—.


             Mean concentration of

                diversion by month


     Land use  matrix (by segments) -—"
                                                MODEL
                Lake level


Total  runoff Input by month


    Total outflow by month
                                         Land use component
                                        Amount of diversion
                                                                   Gauged runoff
                                             by month     Annual P load
                                              Trophic status component
                                         Mean  P cone July-Oct
                                                          Gauged runoff
                                                     CHLa
                                                   Effects component
                                                      Plot  f)«i c k a go
          OUTPUT


 e


-*- Kg  yield by segments, P & N


-*- Kg  yield by sources, PAN


-»•- Loading by sources,  PAN


-»- Total loading, PAN
   Mean total  P  cone

   Mean chlorophyll a
                                                                          ^ Secchi (mean, min)


                                                                          *- AHOD

                                                                          »- Minimum oxygen


                                                                          *•• Carlson  index



                                                                          »- Plot % chango

-------
   H-
  OQ
  Ln
  CO
 tn
 (T)
 OQ
 0
 c
 en
 (T>
 O.

 H-
 a


 b
 Cb
 fD
 M
 M
 H-
 3
OQ
                                                                                                                                     miles
                                                                                                                                                                                   to
                                                                                                                                                                                   VO
                                                                                                                                                                                   00

-------
                                                                      299




for all segments and  the  inventory  change function is applied in order




to correct for changes  in the  phosphorus  or nitrogen inventories in the




valley bottoms according  to  the  amount  of runoff.   Other sources are




then added, including the three  lakeside  WWTP's,  groundwater,




precipitation, diversions, and irregular  construction activities.




     Output from the  land use  component of the Dillon Clean Lakes Model




includes the nutrient yield  by watershed  segments  and by source




categories for both P and N.   The total annual P  and N reaching the




lake is also part of  the  output.









The Trophic-Status Component of  the Model




     The land-use component  of the  Dillon Clean Lakes Model passes




certain critical information to  the trophic status component of the




model.  This information  includes the monthly  schedule of  diversions or




pumping, the concentrations  of phosphorus and  nitrogen en  a monthly




basis for pumped or diverted water, the annual phosphorus  load  by




runoff and precipitation,  and  the amount  of gauged runoff  (i.e., su~ of




the four U.S.G.S. gauges).   Additional  input  required at this point but




not supplied by the land  use component  of the  model includes the total




surface water and ground  water entering the lake  on a monthly basis




(this will be slightly  larger  than  the  gauged  runoff), the monthly




average lake level, and the amount  of water leaving the lake on a




monthly basis.  The output includes the mean  total phosphorus




concentration of the  top  15 in  during the  period July through October




and the mean chlorophyll  a_ in  the top 5 m over the sar.e interval.




     Because the trophic  status of  lakes  has been  the subject of nuch

-------
                                                                       300





modelling effort, there is considerably more  latitude  in  the  choice of




approaches for the trophic-status component of  the  Dillon Clean Lakes




Model than for the land-use component.  Whereas with the  land-use




component there is little basis for anything  other  than a




simple empirical approach to modelling, the trophic-status  component




could be based in a more complex and fundamental way on current




knowledge of biological, physical, and chemical processes in  lakes.   In




discussing the alternative approaches, we can recognize two different




classes of models:  process models and empirical models.




     A process model assumes that the critical biological,  chemical,




and physical processes that have an important bearing  on  the  variable




of interest (in this case, trophic status and its derivatives)  are




sufficiently well known to be represented by  equations.   For  example,




in a process model dealing with algal growth, logical  functions  to  be




incorporated in the model would include nutrient uptake,  sedimentation




cf algae through the metalimnion, and decomposition of algae  to  release




nutrients to the water.  Since even the simplest natural  planktonic




systems are exceedingly complex at the elementary functional  level, any




such approach inevitably incorporates very large systems  of coupled




equations.  An example is the Lake George Model, CLEANER  (Park  et al.




1979).




     Process models have value as exploratory tools for basic ecosystem




research, but their use for prediction, and thus for problem  solving,




is probably unwarranted and ill-advised at present.  Opinions are by  no




means undivided en this subject, but a consensus seems to be  developing




about the practical restrictions on process models  (e.g., Tailing 1979,

-------
                                                                      301




Harris 1980, Hobble and Tiwari  1977).   The  main reasons for the




limitations on usefulness of process models is  that  the critical




phenomena controlling biological  variables  in ecosystems are too poorly




understood, too labile, and too complex to  be treated  in this  manner.




It is sometimes stated that the use of  process  models  on biological




systems only awaits the accumulation of further basic  information.   The




implication is that the ecology of biological systems  is in a




relatively primitive state and, as it matures,  will  logically  support




complex models that produce satisfyingly exact  answers  to  practical




problems.  This line of reasoning is probably misleading in several




ways.  First, the fund of information presently available  is quite




vast, especially for planktonic systems. It  seems unlikely that any




short-term improvement in this  fund of  information will drastically




alter the feasibility of process models. Secondly,  other  disciplines




in which the use of models to deal with large and complex  systems has a




long history have encountered much the  same problems in the use of




process models as ecosystem biology.  Economics and meteorology are two




such disciplines.  The motivations for  predicting either the economy or




the weather well in advance are obviously enormous,  and yet the success




of models to accomplish this is very modest;  available  models  deal  best




with short-range predictions that assume nothing unusual happens.




     There exists in the study of ecosystems, and particularly of




lakes, an alternative to the process model.   Lakes are  so  numerous  that




empirical experience can be accumulated on  their responses  to  widely




varying conditions.  This experience can serve  as the  basis for models.




The philosophy of such models is entirely different from that  of

-------
                                                                      302
process models.  The dependent variable of  interest  is  examined  from




the viewpoint of a very small number of master variables.   While these




master variables are chosen for their known direct and  indirect




relationships to the dependent variable, no attempt  is  made to dissect




the numerous separate functions that contribute  to the  overall




significance of such a master variable.  This approach  was  first




popularized for lakes by Vollenweider (1968), who related trophic




status to phosphorus concentration and phosphorus concentration  to




phosphorus loading and mean depth.




     Empirical models inevitably evolve, taking  on more complexity -




While the general validity of such a model  can be established by




sampling large numbers of systems and comparing  the  observed  to  the




predicted values for the variables of interest,  there will  always be  a




certain amount of scatter in the agreement  of observations  and




predictions.  An obvious challenge then is  to reduce the scatter by




introduction of new concepts involving the  same  master  variables or




addition of other master variables.  The model thus  tends to  become




more complex, and in principle converges with the process model  by




building backward from the level of master  variables to  the level of




processes.  It is doubtful that this convergence will ever  be




completed, however, as there are severe practical limits to the




reduction of scatter in the relationship between observed and predicted




values.  Even in their most complex forms,  present modifications of  the




Vcllenweider model are very distant in rationale and complexity  from




a process model such as that developed for  Lake  George.

-------
                                                                      303




     For reasons outlined above, we  have  accepted  as a philosophical




basis for the modelling of  the  trophic  status  of Lake Dillon an




empirical modelling approach.   We  have  resolved  to use a Vollenweider




type of model, but to consider  modifications as appropriate  in the




particular circumstances of Lake Dillon.




     The first and most obvious improvement  in the original




Vollenweider model, which was based  solely on  phosphorus loading  and




mean depth, involved corrections for the  flow  of water through lakes.




Obviously two lakes with equal  phosphorus loadings and mean  depths  but




very different flushing rates might  easily differ  in trophic status.




It seems intuitively obvious that  a  certain  amount of the phosphorus




income should be discounted for losses  through the outlet.   This




discounting concept has been considered by a large number of students




of eutrophication (Welch 1980).




     It is a short step from corrections  based on  flushing rate to




corrections based on measured phosphorus  retention coefficients.  This




approach has been taken by  Larsen  and Mercier  (1976) and by  Dillon  and




Rigler (1974).  Since phosphorus retention and flushing rate interact,




a number of formulations are possible,  as described by Vollenweider et




al.  (1980).  Underlying all of these approaches is a general  mass




balance equation by which a steady-state  phosphorus concentration is




defined in terms of phosphorus  gains and  phosphorus losses.   In a lake




such as Dillon, where flowthrough  is substantial,  there is little doubt




of the need to incorporate  some kind of correction for the hydraulic




residence time, or the sedimentation coefficient for phosphorus,  or




both of these.

-------
                                                                       304





     The approach  to be  taken  here  can  be  credited to Vollenweider




 (1969), although it is a  specific derivative  of general mass balance




 equations.   Since  this formulation  was  proposed by Vollenweider




 subsequent to his  well-known original formulation, which did not take




 into account hydraulic residence time,  we  refer to it as the "modified"




 Vollenweider model.  The  critical equation is  as follows:




                      Cp  = Lp/1T(l/tw +  s)                (16-1)




 where C_ equals the phosphorus  concentration  of the lake,  L_ is




 phosphorus loading of the lake, ~z is the mean  lake depth,  s is the




 phosphorus sedimentation  coefficient and t^ is  the residence time for




 water (Vollenweider et al. 1980).   Vollenweider was able to reduce the




 number of parameters in the equation by assuming a sedimentation rate




 of 10-20 m per year.  The simplified equation  thug became




                Cp = Lptw/z-(l + tw°'5).          (16-2)




     Exact meaning of the model parameters is  to some degree open to




 definition by the  user and is not fully standardized.  For Lake Dillon,




 we defined C  as the mean phosphorus concentration of the  upper 15 m




 of the water column from  the interval between  July 1  and October 30.




 The depth zone 0-15 m corresponds to the maximum extent of the growth




 zone.  Since changes in transparency and chlorophyll  a  in  the upper




 water column are ultimately of interest, this was  considered to be the




most advisable choice for the definition of C_•  However,  had C




 been defined in any other reasonable way,  major differences would r.ot




 have rosuited.  It was considered advisable to  define the  time interval




 for -determination  of Cp in such a way that the  spring runoff, which




 incorporates an exceptionally large fraction of heavy particulate

-------
                                                                       305
material  that  settles  out  rapidly,  would not unduly influence C -
This prevents  the model  from being unacceptably sensitive to variations


between years  in the  amount  of  runoff.


     The hydraulic  residence time, t^, can also be defined in a


number of ways.  In a lake  that remains at steady state, all
           *

definitions are essentially  equivalent.  In the case of a lake such as


Dillon, outflow and inflow are  not always equal for a given year.  For


present purposes we use  as  our  definition of t^, the ratio of lake


volume to  total inflow.   This is partly necessitated by discrepancies


between the U.S.G.S.  and Denver Water Department figures on inflow and


outflow, which have forced us to choose one or the other of these data


sources.   For  consistency, we use the U.S.G.S. inflow figures.


     Use of the simplified  formula (eq. 2) for the modified


Vollenweider model  predicts  P values that are reasonable but are


consistently too high for the two years of record.  Since overestimates


occurred both  in a  low-water and a high-water year, the implication is


that the simplified formula  underestimates the sedimentation


coefficients for Dillon,  and that the performance of the model could be


improved by use of  a  more correct sedimentation coefficient, which in


turn necessitates use of the more complex version of the relationship


(equation 16-1).  For this reason, sedimentation velocities (v.  where v


= z"'s) were approximated  according to the procedure outlined by


Higgins and Kim (1981).   This procedure relies on knowledge of the


inflow and outflow  concentrations.  The resulting estimates of


sedimentation  velocities  for the two years were between 13 and 15 = per


year.  These sedimentation velocities were used to compute values of s

-------
                                                                      306
(T was 23 m both years), which were in turn used  in  the  Vollenweider




equation (16-1).  The results were better but still  consistently




somewhat high for both years, implying that the actual sedimentation




coefficient was still being underestimated.  Since the estimation




procedure from Higgins and Kim assumes a steady-state condition, which




is not ever literally applicable, the estimate is subject  to  unknown




error.  Furthermore, the outflow concentrations could not  be  determined




directly for the water going through the Roberts  Tunnel, which was  not




available for sampling, and this may have resulted in some estimation




error.




     Since the sedimentation coefficients were underestimated even




after correction by the procedure outlined by Higgins and  Kim, another




approach was taken.  Each year was treated independently and  the




equation was solved for s using the observed values  of the  other




parameters for the year in question.  The resulting  estimates of




sedimentation velocity were 24 m/yr for 1981 and  32  m/yr for  1982.   The




larger coefficient for 1982 is quite reasonable in view  of  the fact




that heavy particulates are moved in nuch larger  quantity  in  years  of




high runoff than in years of low runoff (see Figure  22a  in the section




on lake phosphorus).  The question then arises how the sedimentation




coefficients should be set for purposes of prediction in years of




different runoff.  The positive relationship between sedimentation




coefficient and amount of runoff seems justified  from the  information




at hand but the exact form of the relationship cannot be determined




from two points.  It is therefore assumed that the relationship  is




linear.  Nonlinearity will result in a certain amount of error,  but  the

-------
                                                                       307
assumption of  linearity is  much better than assuming constancy of




sedimentation, which  is clearly incorrect.




     A second  element of the  trophic status component of the model is




the relationship between phosphorus  and chlorophyll.  A number of




statistically  derived relationships  are presented in the literature.




These show consistently a statistically significant trend toward higher




summer chlorophyll  a_ with higher total phosphorus in the mixed layer.




There is, however,  quite a  bit  of scatter and this has been the subject




of considerable discussion  and  analysis.




     The general equations  that are  available in the literature (e.g.,




Dillon and Rigler 1974, Carlson 1977,  Jones and Bachmann 1976, Jones




and Lee 1982,  Lambou  et al. 1982) do a uniformly poor job of predicting




Lake Dillon  chlorophyll a.  Observations are 3-6 times higher than




predicted values; Dillon produces considerably more chlorophyll per




unit of P than most  lakes.  This is  unusual but not unique,  as shown by




the scatter  of points in the  above-cited publications, but the




explanation  is not  immediately  obvious.




     A recent  study by Smith  (1982)  appears to explain the unusual




phosphorus-chlorophyll relationship  of Lake Dillon.  Smith assembled




data from several surveys in  which both the nitrogen and phosphorus




concentrations were known.  He  treated nitrogen as an added  variable in




studying the relationship between phosphorus and chlorophyll.  He




showed that  nitrogen  is  a very  important covariate, greatly  increasing




the amount of variance that can be explained even when phosphorus




clearly limits growth rates.  His conclusion is that increases in the




nitrogen to  phosphorus ratio  boost the amount  of chlorophyll that will

-------
                                                                      308
be produced for a given amount of phosphorus,  even  when phosphorus is




limiting.  The mechanism by which this occurs  is unknown,  but  the




reduction of variance by incorporation of nitrogen  into the  equations




is impressive.  It is worth noting that OECD (1982) found  no added




variance component explained by nitrogen.  Smith and OECD  did  use




different data bases; possibly Smith's was better suited for




statistical illustration of the role of nitrogen.   For  present




purposes, we accept Smith's conclusions as valid and attempt to  apply




then to Lake Dillon.




     The lakes studied by Smith varied widely  in the ratio of  nitrogen




to phosphorus.  Smith developed equations for  predicting chlorophyll  in




lakes of differing nitrogen to phosphorus ratios.   His  equations show




that the effect of increasing the nitrogen to  phosphorus ratio is to




shift the line relating phosphorus to chlorophyll upward (Figure 59).




At low to moderate TN:TP ratios, the upward shift occurs without




changes in slope; the relationship thus takes  the form  of  a  family of




parallel lines.  In the highest category used  by Smith, however




(TN/TP > 35 by weight), the slope increases, and simultaneously  becomes




insensitive to further change in the TN/TP ratio.   For  lakes with low




TN:TP ratios, it is essential to use N as a dependent variable as well




as p.  For lakes with a high TNtTP ratio, N can be  dropped from  the




equation, but the coefficient on P will differ from that of  lakes




having low TN:TP ratios.




     LaV.3 Dillon, with TN/T? = 54 (weight basis), falls in the very




highest category of nitrogen to phosphorus ratios among the  lakes




studied by Smith.  In this highest category, as the influence  of

-------
                                                        3C9
        00
to
 E
 \
 e

 01
 -I
 o
         0
        O.I
                DILLON  RIGLER
              LAKE  DILLON-V
                                    SMITH,
                                     TN:TP=IO
                               SMITH,
                                TN:TP=25
                          10            100
                          TOTAL  P,  mg/m3
                                                    000
Figure 59.  Log-log plot of total P and chlorophyll a_, showing lines
          for two of  Smith's (1982)  lower TN:TP categories, the
          Dillon-Rigler line, and the Lake Dillon line.

-------
                                                                      310
further change in TN:TP becomes negligible,  the  coefficient  on ? is




high and convergent with that of the Dillon-Rigler  equation  (1974).




This convergence is expected, since the Dillon-Rigler  equation was




developed without inclusion of N.  However,  Smith's  study  shows that




the intercept may continue to be affected as TN:TP  is  raised to high




levels, even though the slope stabilizes.  We thus  approach  the Dillon




data with the following points in mind:  (1) the TN:TP ratio is so high




in Dillon that N need not be used in the equation,  (2) the appropriate




slope can be taken from the Dillon-Rigler equation,  (3)  the  intercept




should be determined uniquely for Dillon.  We therefore  set  the slope




of the phosphorus-chlorophyll equation equal to  that of  the




Dillon-Rigler equation, but assume that the  intercept  will be




determined by the nitrogen to phosphorus ratio.  We  determine the




intercept empirically by solving for the intercept using the slope from




the Dillon-Rigler equation and the observed  chlorophyll  a values.  The




resulting equation is as follows:




        logB = 1.449 log(Cp) - 0.398                 (3)




where B is chlorophyll a_ (July-October, ug/1, 0-5 in) and Cp  is  total P




(0-15 m, July-October, ug/1).  This equation performs  well for  both




years (predictions will be given below).









Effects Component of the Model




     Total phosphorus and chlorophyll a_, as  predicted  by the trophic




status component of the Lake Dillon Clean Lakes Model, have  a number of




correlates of economic and aesthetic importance.  Since  public




attention and the attention of lake managers and users is likely  to  be

-------
                                                                       311




focussed on some of  these  correlates,  it is  useful to be able to make




predictions for them.




     Transparency is perhaps  of  greatest concern as a correlate of




eutrophication.  This is the  aspect  of eutrophication most obvious to




the casual observer  and most  intimately connected with the aesthetic




appeal of a "lake.  The most common measure of transparency is the




secchi depth.  Since a great  deal of information is available from the




literature on the relationship between secchi depth and chlorophyll a,




and since there is a complete set of secchi  depth measurements for Lake




Dillon in 1981-1982, we have  designed  the effects component of the




model to predict secchi depth.   The  prediction is based on chlorophyll




a, which is the variable most directly responsible for changes in




transparency as eutrophication occurs.  The  interval of interest is




July 1 through October, corresponding  with the post-runoff




stratification season.  As in other  instances, we have excluded the




month of June from the predictions because all predictions are




complicated in this month  by  the dominance of inorganic particulate




matter in the water  column as a  result of runoff.  The observed mean




secchi depth over the July-October interval  in 1981 was 3.15 m and in




1982 it was 2.76 m.




     The literature offers a  number  of equations relating secchi depth




to chlorophyll a.  There are  two drawbacks to the use of these very




general equations on Lake  Dillon. First, there is a great deal of




scatter around any of the  lines  taken  from the literature, as would be




expected for the comparison of very  different kinds of lakes.




Secondly, Lake Dillon is not  located in the  densest cluster of points

-------
                                                                      si;


used for any of these equations, since its  chlorophyll  a_ values are


still relatively low.  Thus the general equations,  which for


transparency are particularly likely to go  astray when  extrapolated


from lakes of low transparency to those of  high  transparency,-  may not


be especially appropriate for Lake Dillon.


     The relationship between secchi depth  and chlorophyll £ is always


nonlinear.  Transparency as measured by secchi depth  decreases at a


slower rate than chlorophyll increases.  The  fit of points  to  an


empirical transparency/chlorophyll a_ relationship is  typically rendered


linear by log transformation.  The log-transformed  relationship thus


has a slope and an intercept.  Problems related  to  the  determination of


the slope and the intercept are different.  The  slope of the


relationship should be much more universal  than  the intercept.  The


slope shows the increment of change in secchi depth for an increment of


change in chlorophyll a_, and lakes cannot be  expected to differ widely


with respect to this relationship.  It is thought that  the package  size


of chlorophyll a_ (i.e., the average cell size for phytoplankton)


results in different values of chlorophyll-specific extinction (Harris


1978), and this may be responsible for a certain amount of variation in


the slope.  However, we have already shown  that  Lake  Dillon  is very


average with respect to its chlorophyll-specific extinction  (es =

       f)
0.015 m^/mg chlorophyll a_) .  We therefore believe it  is justified to


use a slope from one of the equations in the  literature. We choose for


this purpose the slope of the equation of Carlson (0.68), which is


b.a^d on 147 lakes  and has an r value of 0.93 (Carlson  1977).   The


Mational Eutrophication Survey data have a  somewhat higher slope  (0.86:

-------
                                                                      313

Lambou et al. 1982), but the observed  values  for  Lake  Dillon for the

two years of record suggested much closer  agreement  with  the Carlson

coefficient than with the National Eutrophication Survey  coefficient.

We therefore adopt the Carlson coefficient.

     The intercept of the log-transformed  relationship between secchi
           «
depth and chlorophyll a_ is likely to vary  much more  extensively between

lakes than the slope, particularly if  one  considers  reservoirs and

natural lakes together.  The intercept  for a  given lake will be

determined largely by a combination of  nonchlorophyll  factors

contributing to the extinction of light.   In  very transparent lakes,

this includes principally the extinction effect of pure water and of

dissolved substances.  In lakes  that receive  substantial  amounts of

inorganic particulates, extinction by  particulates becomes  a major

determinant of the intercept.

     It is feasible to customize the intercept for a prediction

equation to a given lake by using a slope  from a  series of  lakes and

the observed values for the lake in question  to solve  for the

intercept.  We did this separately for  1981 and 1982 on the Lake Dillon

data using the Carlson slope.  The intercept  for  1981  was 2.44 and  Che

intercept for 1982 was 2.37.  These intercepts were  so  similar chat  a

common intercept of 2.4 was adopted for use with  the Dillon data.   This

intercept is higher than the intercept  of  the Carlson  equation (2.04),

probably because the Carlson equation  is based on natural lakes lacking

the inorganic particulate input  that one sees in  Lake  Dillon.   The

intercept is much closer to but  slightly lower than  the intercept for

-------
                                                                       314



the National Eutrophicatlon Survey (2.56),  which  includes a large




number of reservoirs.




     The equation resulting from the above  analyses  is  as follows:




                         InSD = 2.4 - 0.681nB




where SD is secchi depth in meters and B  is  chlorophyll a_ in ug/1.  The




prediction of 1981 and 1982 chlorophyll from this  equation is excellent




(see below), suggesting that use of the Carlson slope is realistic.




     A rough approximation can also be made  of the minimum secchi depth




from the average secchi depth.  The minimum  secchi depth in this  case




applies to the time of maximum phytoplankton standing stock.   There  is




another seasonal minimum associated with  runoff, which  is not of




concern here.  In 1981 the ratio of minimum  secchi depth to average




secchi depth was 0.76 and in 1982 it was  0.65.  As an approximation  we




take the ratio to be 0.70.  There is no extensive  literature  on this




subject and the prediction is therefore less secure  than the  other




predictions mentioned up to this point.   However,  the use of  this ratio




will give some indication of the transparency of the lake at  that




particular time of the year when transparency problems  may be most




aggravated (July).




     Oxygen depletion in deep water during  the summer is also of




concern in connection with eutrophication.   In the section on oxygen,




we presented an analysis of the areal hypolimnetic oxygen deficit and




showed good agreement between the predicted  AHOD and an equation




.eve loped by Cornett and Rigler (1980) based on secchi  depth  and  mean




     ..epch.  We use this equation in the  effects component of the




Dillon Clean Lakes Model, but we believe  it  is feasible to improve the

-------
                                                                       315




prediction by making a. correction  for  the amount of water entering the




lake in a given year.  In  1981  the equation underpredicted the AHOD by




130 mg/m2/day (20%) and in 1982 it overpredicted by 6 mg/m2/day




(1%).  The difference in the  tvx> years  is almost certainly due to the




different amounts of oxygen transported by advection to deep water in




the years of high runoff and  low runoff.   In a  wet  year,  advective




transport is greater and the  areal hypolimnetic oxygen deficit is thus




lower.  We correct by making  the assumption that runoff is proportional




to the differential between the observed  and actual hypolimnetic oxygen




deficits.  This results in a  small correction that  is added to the




model.




     Although the AHOD is  of  some  direct  interest,  the minimum oxygen




level in deep water is of  more  specific concern.  There are no general




equations for prediction of minimum oxygen content, since this




prediction would involve the  shape of the lake  and  the duration of the




stratification period.  However, for a  given lake,  there  is obviously a




very close relationship between the AHOD and the amount of oxygen




depletion.  The relationship  should be  very close to linear if the




duration of the stratification  period is  constant.   For Lake Dillon  the




initial oxygen can safely  be  assumed equal to 9.0 mg/1, the saturation




concentration at spring overturn.   The  oxygen depletion is equal to  the




difference between 9.0 mg/1 and the minimum oxygen  observed toward the




end of stratification.  We deal specifically with a point 5 m over the




bottom of the index station,  and use the  degree of  depletion (mg/1)  in




each of the two years as the  dependent  variable and the AHCD




(mg/m^/day) as the independent  variable.   This  yields a slope of

-------
                                                                       316
0.00698.  This slope, when multiplied  times  the  AHOD,  gives the




predicted amount of depletion.  The amount of  oxygen  remaining is then




the difference between the depletion and  9.0.  The  lower bound of




remaining oxygen is set to zero, even  for depletion more than severe




enough to remove all oxygen.









Comparing Predictions and Observations for 1981  and 1982




     The complete Dillon Clean Lakes Model was run  on  land  use and




runoff data for 1981 and for 1982.  In these runs,  the unusual features




of 1981 or 1982 that would not be obvious from runoff  or from land use




data of the type accepted by the model for prediction  purposes were




ignored.  For example, the output of wastewater  treatment plants  was




computed on the basis of the number of persons served  and the




generalized equation described in the  section  on land  use rather  than




the known yields from these plants.  Similarly,  the construction  in the




Snake River bottom near Keystone in 1982 was not treated in any special




way in the model for 1982.  Thus the agreement between predictions and




observations for the two model runs gives an idea of the deviation




between observed and expected that could be produced by randomly




occurring events whose timing cannot be anticipated.




     The predictions and observations  are summarized in Table 61.   For




all variables the agreement is within  5%, except for the minimum




deep-water oxygen,  which is within 10%,  The performance of the model




on theso years of known characteristics is thus  excellent,  despite the




occurrence of a number of events whose timing cannot be  anticipated by




such a model.

-------
                                                                      317
1981

Total P, ug/1*
Chlorophyll a, ug/1*
Mean secchi , m*
Minimum secchi , m
2
AHOD, mg/m /day
Minimum deepwater 0«
Predicted
6.7
6.3
3.2
2.2
706
4.1
Observed
7.0
6.7
3.2
2.4
710
4.4
1982
Predicted
7.1
6.9
3.0
2.1
590
4.9
Observed
7.4
7.3
2.8
1.8
63C
4.6
*  Mean, July-October
Table 61.  Comparison of predictions from  the Dillon  Clean Lakes Model
           with observations for 1981 and  1982.

-------
                                                                       313
                 Predictions for Development  Scenarios




     The Dillon Clean Lakes Model was applied  to  10 different




scenarios provided by the Dillon Clean Lakes  Steering Committee.   We




report below the results of modelling for each scenario.   The 10




scenarios comprise five sets of assumptions for development  and  land




use, each of which was applied to hydrologic assumptions  for a dry year




and for a wet year.  Predictions in each case  include the total  loading




of the lake, the source distribution of the loading, and  the response




of the lake to the predicted amount of loading.




     The dry-year hydrologic conditions were always  set identical  to




1981.  Since the runoff for 1981 was among the  lowest for the period of




record, 1981 provides a good estimate of the limiting hydrologic




conditions for a dry year.  For the wet year,  we  used the hydrologic




conditions of 1982.  1982, although definitely  above average,  was  not




so extreme hydrologically as was 1981.  Use of  the  hydrologic




conditions of 1981 and 1982 is advantageous since the behavior of  the




lake in these two years under present land use  is very well  documented




and can thus be compared easily to future years of  similar hydrologic




conditions but very different land uses.  We have also used  the




observed lake level changes for 1981 and 1982  along with  the hydrologic




data for these two years.  Thus we have assumed that the  schedule  of




water release from the lake remains more or less  as  it is presently.




Predictions might be different if the Denver Water  Department began to




release water in a very different way, e.g., by more excessive summer




drawdown or more extended periods of drawdown.

-------
                                                                       319




     The predictions  of  trophic  indicators are translated into a




trophic status category.   For  this,  certain conventions are necessary.




We use chlorophyll as  the  key  trophic indicator,  since algal biomass is




the central concern of changing  trophic status.   We set limits on the




trophic categories from  a  midrange of boundaries  accepted by various




sources.  According to the review by Welch (1980),  the oligotrophic




boundary will be somewhere between 2 and 4 ug/1  and the eutrophic




boundary will be between 6-10  ug/1-   We therefore set the mesotrophic




range for present purposes at  3-8 ug/1 chlorophyll  a_ (this is also the




range used by OECD 1982).   One other convention  is  the adoption of an




annual time frame for  trophic  classification.  Trophic classes make no




sense at intervals less  than one year,  so this is the maximum




resolution possible.   There is some  merit to the  argument that status




should be based on a  run of years, or on a year  of  median conditions.




However, since we are  interested in  variation  between years, we adopt




for modelling purposes an  annual interval.




     Subsequent to the presentation  of model runs is  a more generalized




graphical presentation based on  the  logic of the  model.









Model Runs for Present Land Use  Patterns




     Before any of the scenario  data were used,  the model was run




twice (wet year/dry year)  with a land use input matrix representing




present conditions.  The purpose of  these model  runs  was  not to check




the behavior of the model,  which has already been discussed, but rather




to define the loading and  trophic  condition of the  lake under




standardized conditions  subsequent to 1982.  These  nodel  runs differ

-------
                                                                      320




from the predictions discussed in previous  sections  in  that  Copper




Mountain is assumed to have transferred completely to tertiary




treatment, and to serve 2500 persons, slightly more  than  1981-1982.




Climax Molybdenum is set at the 1981 employment  level (2650  persons).




Also, no allowances are made for any of the  special  events that




occurred in 1981 and 1982.  Thus these two model runs for present




conditions give for comparative purposes the output  of  the model under




the assumption that present land use stabilizes  (by  resumption of




activities at Climax and completion of the  transfer  of  Copper  Mountain




to tertiary) but that no growth occurs.  The results of the  two model




runs on stabilized present conditions are shown  in Tables 62 and 63.




As expected, the predicted loading and response  of the  lake  are very




similar to those actually observed in 1981  and 1982.









Scenarios 1A, IB;  Low Growth




     These scenarios specify that wasteload allocations for  point




sources will be met, and that septic areas will  be built out to 70%.




This means that 70% of the lots will be occupied by  dwellings  used in




the same way as present dwellings.




     The point sources are at present meeting their  wasteload




allocations, although there is some random  variation from one  year to




the next.  We therefore leave constant at the 1981-1982 levels the




number of persons served by the Breckenridge, Snake  River, and Frisco




wastewa:er treatment plants.  We assume that Copper  Mountain serves




2:iV persons, which is slightly above the 1981-1982  level, and that




complete tertiary treatment for Copper Mountain  is put  into  effect.  We

-------
*
Scenario Conditions
Wet Year (similar to 1982)
Current conditions***
Low Growth (1A)
I/)W Orowth with Diversions (2A)
High Growth (3A)
Hl};h Growth, Best Controls (4A)
High Growth, Cost-effective
Controls (5A)
Dry Year (similar to 1981)
Current conditions***
Low Growth (IB)
Low Growth with Diversions (2B)
High Growth (315)
Hij-,11 Growth, Best Controls (4P>)
High Growth, Cost-effective
Controls (5B)
Total P
Load , kg

4678
5049
50048
6389
3657

5001

2459
2618
47618
3190
2021

2596
Percent
WWTP*

17.4
16.2
1.7
12.8
22.3

16.3

30.0
28.3
1.6
23.2
36.6

28.5
Percent
Septic

8.2
14.7
1.5
31.9
0

20.4

6.6
12.0
0.7
27.2
0

16.7
Percent
Diversion

0
0
89.9
0
0

0

0
0
94.5
0
0

0
Percent
Background**
«
60.5
56.1
6.7
44.4
77.5

56.6

51.7
48.5
2.6
39.9
62.9

49.0
Percent
Other

13.9
13.0
0.2
10.9
0.2

6.7

11.7
11.2
0.6
9.7
0.5

5.8
*   Includes pnrkapc-  plants
A*  Includes hack}*,round runoff  and  precipitation
*** Assumes Copper Mountain  on  tertiary,  Climax at  full  employment 1981 levels.
Tabl
e 62.  Results of  the application of  the  Dillon  Clean Ibices  Model  to scenario data

-------

Scenario conditions

Wet Year (similar to 1982)
Current conditions***
Low Growth (1A)
Low Crowth with Diversions (2A)
High Growth (3A)
High Crowth, Best Controls (4A)
High Growth, Cost-effective
Controls (5A)
Dry Year (similar to 1981)
Current conditions***
Low Crowth (IB)
Low Growth with Diversions (2B)
High Growth (3B)
High Growth, Best Controls (4B)
High Growth, Cost-effective
Controls (5B)
Lake Total
P, ug/1


7.3
7.9
57.
10.0
5.7

7.8

6.1
6.4
74.
7.9
5.0

6.4
Chlorophyll
ug/1


7.1
8.0
141.
11.2
5.0

7.9

5.4
6.0
206.
7.9
4.1

5.9
Mean
Secchi , m


2.9
2.7
0.38
2.1
3.7

2.7

3.5
3.3
0.29
2.7
4.2

3.3
Minimum
Secchi , m


2.0
1.9
0.27
1.5
2.6

1.9

2.4
2.3
0.21
1.9
4.8

3.0
Minimum
0 , mg/1
2

4.8
4.6
0.0
3.8
5.4

4.6

4.3
4.2
0.0
3.6
4.8

4.2

Trophic Status


Mesotrophic
Mesotrophic
Eutrophic
Eutrophic
Mesotrophic

Mesotrophic

Mesotrophic
Mesotrophic
Eutrophic
Mesotrophic
Mesotrophic

Mesotrophic
* Includes package plants
** Tnr.ludftS harkfrnimrl runoff and nrec \ nitat I o r\
*** Assumes Copper Mountain on tertiary, Climax  at  full  employment  1981  levels.




Table 63.  Results of the application of the  Dillon  Clean  I-akes  Model  to scenario  data.
                                                                                                                CO
                                                                                                                ho
                                                                                                                to

-------
                                                                       323
assume  that  Climax Molybdenum is  operational with 2650 persons in the




work force.   The number  of  persons  on septic systems represented by 70%




buildout would be 3052  (full-time equivalent residents).  These would




be distributed geographically by  proportions identical to septic




distributions for 1981-1982.   All other land uses remain unchanged.




     The model was run  first  for  a  wet year (1A) and then for a dry




year (IB).   The results  are reported in Tables 62 and 63.  Figures 60




and 61  give  a graphical  representation of the degree of change as




compared with present conditions.  In the wet year,  the predicted




loading is higher by slightly less  than 10% as a result of low growth.




This causes  an increase  of  chlorophyll to 8 ug/1.  As explained




previously,  we have taken the range of chlorophyll values between




3 ug/1  and 8 ug/1 as indicating a mesotrophic condition.  Thus in a wet




year the low growth scenario  pushes the chlorophyll  a_ exactly to the




upper limit  of the mesotrophic span in a wet year.  There is  an




accompanying  decrease in transparency and a decrease in the minimum




oxygen  at the bottom of  the lake.




     For the  dry year,  the  degree of change in total loading  is




smaller.  Furthermore,  the  lake is  much more toward  the middle of the




mesotrophic  range for chlorophyll and other trophic  indicators during a




dry year, except for deepwater oxygen.  Thus increasing the loading




slightly to  the degree specified  by the low-growth scenarios  does not




bring the chlorophyll level so close to the upper boundary of the




mesotrophic  range.  For  a dry year,  the lake remains solidly




raesotrophic.

-------
                                                                  324
                      WET  YR  LO  GRQWTH(lfi)  VS PRESENT
               100
         CJ
         Q£
         UJ
         0_

         UJ
         to
60-
                2C+
               -20--
         UJ
         UJ
         Q_
         UJ
         UJ
         cn
               -so-
              -JOO
                 0       l.OCO
                   TOTflL P
                                -f-
                    2.000
                  CHLfi
                                               -t-
 3.000
SECCHI
4.000      5.000
BOTTOM  02
Figure  60.  Graphical summary  of Dillon Clean Lakes  Model  outputs for
           four key lake  characteristics.  Vertical lines  show
           deviation from present conditions; horizontal  lines with
           "+" mark the oligotrophic and eutrophic  boundaries for
           chlorophyll a.

-------
                      DRY  YR  LO  GROWTH!IB) VS PRESENT
               100
         UJ
         (_)
         a:
         UJ
         o_
          *
         LJ
         V3
         ce
         o
         3
       SO-
                20-
               -20-
UJ

tt:
LU
0-
 »
ce
UJ
         UJ
         CD
               -so-
              -100
                            •4-
                 0       1.000

                   TOTflL P
  2.000

CHLfl
                                     3.000

                                    SECCHI
4.000      5.CCO

BOTTOM  02
Figure  61.
   Graphical  summary of Dillon Clean Lakes Model outputs for
   four  key lake characteristics.   Vertical lines show
   deviation  from present conditions; horizontal lines with
   "+" mark the oligotrophic and eutrophic boundaries for
   chlorophyll a.

-------
                                                                      326
Scenarios 2A, 23;  Low Growth with Diversions




     In these scenarios it is assumed that 10,000 acre-feet/year  is




diverted into Lake Dillon from Straight Creek and 183,000




acre-feet/year is diverted into Dillon by the Eagle-Colorado  Project.




The total phosphorus concentrations are set at 30 ug/1  for  Straight




Creek and to 200 ug/1 for Eagle-Colorado.  The amounts  of water




diverted are assumed to apply to years of average or above-average




moisture.  For the dry-year run of the model (2B), the  amounts of




diversion are scaled down in proportion to the gauged runoff  (sum of




the four U.S.G.S. gauges in the watershed).  The diverted water is




assumed to enter the lake on a schedule identical to the Tenmile  Creek




hydrograph.  All other conditions are set identical to  scenarios  1A and




IB.




     The results of the two model runs (2A, 2B) are summarized in




Tables 62 and 63.  Figures 62 and 63 give a graphical representation.




The phosphorus loading coming in by way of diversion is roughly ten




times the loading under present conditions without diversions.  The




lake shows drastic increases in total phosphorus concentrations and in




chlorophyll levels.  Complete oxygen depletion is predicted for the




hypolimnion, and predicted secchi depth values are extremely  low.  The




lake under these circumstances would be unequivocally upper eutrophic




(sometimes called hypereutrophic).  Nuisance blooms would be  almost a




certainty under these conditions.




     The chlorophyll levels predicted for the diversion scenarios are




so high that they probably exceed the theoretical maximum chlorophyll




possible in freshwaters.  About 300 mg/ni2 of chlorophyll a  absorbs

-------
                                                                  327
               WET  YR  10  GROWTH  W DIVERSION!2fl)  VS  PRESEN
                too
          UJ
          O
          to
          01
          O
                 60-
                 20-
                -20-
          UJ
          LU
          Q_
          LU
          CD
               -60-
               -100
                  0       1.000

                    TOTflL  P
 2.000
CHLfl
 3.000

SECCHI
4.000      5.000

BOTTOM  02
Figure  62.  Graphical summary of Pillon Clean Lakes Model outputs  for
           four key lake  characteristics.  Vertical lines show
           deviation from present conditions; horizontal lines vit'-.
           "+" nark the oligotrophic and eutrophic boundaries for
           chlorophyll a_.

-------
                                                                  328
              DRY  YR LO GROWTH  W 01 VERSIONS^23)  VS  PRESENT
               10£h	
         LJ
         CJ
         o:
         UJ
         Q_

         UJ
         CO
         cz
         o
SO-
               -20-
         LU
         Qi
         UJ
         UJ
         CD
               -SO-
              -100
                     -4-
                  0      1.000
                   TOTflL  P
 2.000
CHLR
                              3.000
                             SECCHI
                                                               •4-
4.000      5.000
BOTTOM  02
Figure  63.  Graphical summary of Dillon Clean  Lakes Model outputs for
           four key lake  characteristics.   Vertical lines show
           deviation from present conditions;  horizontal lines with
           "+" mark the oligotrophic and eutrophic boundaries for
           chlorophyll a.

-------
                                                                       329
99% of the light  (Margalef  1978,  Tailing 1982).   At this point light,


rather than nutrients, becomes  limiting  and further addition of


nutrients does not  boost  algal  biomass.   In the  upper 5 m of Lake


Dillon, to which  algal growth would  be essentially confined under the


low transparency  conditions of  these scenarios,  we could expect to find


at a maximum about  60 ug/1  of chlorophyll a_,  which would account for

        n
300 mg/m* summed  up over  the 5-m  mixed layer.  Thus the chlorophyll


predictions, which  are in excess  of  100  ug/1,  should not be taken too


literally.  They  merely indicate  that  the lake reaches  the  theoretical


maximum biomass possible  under  the light limitation conditions

prevailing in Lake  Dillon.   The exact  concentrations are of course of


little interest once they reach such high levels.


     The lake under the conditions of  scenarios  2A and  2B would have


the qualities of  a  sewage lagoon.  The lake would  thus  differ


drastically from  its present appearance  and would  be very different


biologically because of the extreme  depletion  of oxygen in  deep water


and very large crops of algae.

     The phosphorus concentrations in  the diversion water are  the cause


of the extreme trophic response of the lake to diversion.  Accurate


knowledge of the  total phosphorus concentrations in the diversion water


is absolutely essential if  accurate  predictions  are to  be made.  For


scenarios 2A and  2B, by far the greatest uncertainty of the prediction


has to do with the  phosphorus values that are  fed  into  the  model for


the diversion water.  The concentrations were  supplied  by Ton  Elncre


and were derived  by him from existing measurements not  taken in

connection with the Dillon  Clean Lakes Study-  Although the present

-------
                                                                      330
modelling exercise certainly shows that the worst-case  situation  for




phosphorus in diversion water would result in a very undesirable




degradation of the lake, sound predictions beyond  this  will  require




extensive and careful study of total phosphorus in diversion water.









Scenarios 3A, 3B:  High Growth




     These scenarios assume that current wasteload allocations  are met




for the wastewater treatment plants.  This implies essentially  no




expansion beyond scenarios 1A, IB of the number of persons served by




any one of the four wastewater treatment plants.  The number of persons




served by these plants is therefore allowed to remain as for scenarios




1A, IB and Climax Molybdenum is assumed to operate with a workforce of




2650 persons.  The high-growth scenarios assume that septic  systems




will serve a peak population of 37,000 persons.  This corresponds to  a




time-weighted average of 8409 full-time equivalent residents.   It is




assumed that this number of persons is distributed within the watershed




according to the same pattern of present septic system  users.   However,




it should be noted that this number of persons cannot be housed on the




existing developments served by septic, even if all of  these




developments are built out to 100%.  Thus these scenarios assume  that




additional developments will be built and served by septic systems, or




that the present developments will have a higher density of  housing,  or




some combination of these possibilities




     Tables 62 and 63 and Figures 64 and 65 summarize the results of




tha -cdel runs for high growth in a wet year (3A) and in a dry  year




(3B).  For the wet year, the scenarios indicate an increase  in  total

-------
                 100
                        WET  YR  HI  GROWTH  (3fl)  VS  PRESENT
           CJ
           ce
           UJ
           o_
          • «
           UJ
           VJ
 SO-
                 20-
                 -20-
           LU
           o
           0£
           UJ
           Q_

           os:
           UJ
           CD
-60-
                        •4-
                   0       1.000

                     TOTflL P
                     2.000

                   CHLfi
 3.000

SECCHI
4.000

BOTTOM
 5.000

02
Figure  64.  Graphical summary of Dillon  Clean Lakes Model outputs  for
           four key lake characteristics.  Vertical lines show
           deviation from present conditions; horizontal lines with
           "+" mark the oligotrophic  and eutrophic boundaries  for

           chlorophyll a.

-------
                                                                   322
                        DRY  YR  HI  GROHTH(3B)  VS  PRESEN'
                100
          UJ
          CJ
          as
          UJ
          Q_
           *
          UJ
          to
 so-
                 20+
                -2C--
          UJ
          O
          Oi
          UJ
          0-
           *k
          c:
          UJ
          UJ
          03
-SO--
               -100
                   0       1.000
                    TOTflL P
                                       •4-
                     2.000
                   CHLfl
 3.000
SECCHI
4.000      S.OCO
BOTTOM  02
Figure  65.   Graphical summary of Dillon  Clean Lakes Model outputs for
            four key lake characteristics.   Vertical lines show
            deviation from present conditions; horizontal lines with
            "+" mark the oligotrophic  and eutrophic boundaries for
            chlorophyll a.

-------
                                                                       333
wet-year loading  of  approximately 50% over present conditions.  There




is a drastic increase  in  the  percentage  of loading caused by septic




systems, as would be expected.   Chlorophyll is predicted at 11.2 ug/1,




which is well over the boundary  from mesotrophic to eutrophic.




Transparencies and minimum  oxygen in deep water similarly reflect a




higher trophic status.  Although not nearly so drastic as the changes




of scenarios 2A and  2B, the changes  of scenario 3A with respect to the




present condition would almost certainly be significant enough for a




casual observer to notice.




     For the dry  year, the  effects of high growth are  of smaller




magnitude but still  significant.  Chlorophyll  is predicted  to fall at




7.9 ug/1, just below the  upper mesotrophic boundary at 8 ug/1.  Thus




the lake could probably be  classified in the very dryest of years  as




mesotrophic, but at  all other times  would be eutrophic,  and in wet




years would be especially so.









Scenarios 4A, 4B;  High Growth with  State of the Art Controls




     This scenario specifies  that  wasteload allocations  are met  by




point sources, and that no  diversions enter the lake.   All  sources




except the wastewater treatment  plants,  package plants,  groundwater,




precipitation, and background are  set to 0.  Thus it is  assuned  that




all nonpoint sources are  completely  eliminated.




     The results of  the model runs are shown in Tables 62 and 63 for  a




wet year (^A) and for a dry year  (4B).   Figures 66 and 67 show the




corresponding graphs.  In either  a wet or a dry year,  the model




predicts significant improvement of  the  trophic status of the lake,

-------
                                                                  334
           WET  YR HI GRO  STfiTE  OF  THE  flRT(4fl)  VS  PRESEN'
             lOOi	——	
       UJ
       o
       ce
       LU
       Q_
        »
       UJ
       CO
60--
              20--
             -20-
       UJ
       C_)
       ct:
       UJ
       o:
       UJ
       UJ
       CD
             -sof
            -100
           •4-
                                   •4-
               0       1.000
                 TOTflL P
                    2.000
                  CHLfi
3.000
SECCHI
4.000      5.000
BOTTOM 02
7igure 66.  Graphical  summary of Dillon Clean  Lakes  Model outpucs for
           four key lake characteristics.   Vertical lines show
           deviation  from  present conditions; horizontal lines with
           "+" mark the oligotrophic and eutrophic  boundaries for
           chlorophyll a.

-------
                                                                   335

               YR  HI  GROWTH STflTE OF  THE flRTUB)  VS  PRESENT
       UJ
       u
       or
       UJ
       Q_

       UJ *
       V3
       OC
       O
 sc
             -20
       ce
       UJ
       Q_

       o:
       UJ
       UJ
       CD
-60
           -100
            •4-
               0       1.000

                TOTflL  P
                    2.000

                   CHLR
 3.000

SECCHI
4.000

BOTTOM
 s.coo
02
Figure  67-  Graphical summary of  Dillon Clean Lakes Model outputs for
           four key lake characteristics.  Vertical lines show
           deviation from present conditions; horizontal lines with
           "T" mark the oligotrophic and eutrophic boundaries for
           chlorophyll a.

-------
                                                                      336




despite high growth.  Chlorophyll values show lower mesotrophic  status




for the lake under this scenario.  Transparency, bottom oxygen,  and




other indicators are correspondingly improved.




     As attractive as this scenario seems, it is almost certainly




unrealistic.  It implies that there would be no yield whatever above




background for septic systems, for Climax Molybdenum, and  for the many




distributed sources of nutrients associated with human activity.  The




high-growth conditions imply that about 8000 persons (annual average)




will be served by septic systems.  This is essential because the




wasteload allocations cannot be held constant if additional population




is to be served on sewer.  Thus a substantial septic contribution or a




change in the wasteload allocation seems inevitable as an  accompaniment




to high growth, but scenarios 4A and 43 allow neither-  The results




should therefore be viewed as representing an unrealistically high




degree of control over phosphorus loading, regardless of financial




considerations.









Scenarios 5A, 53:  High Growth with Cost-Effective Controls




     Scenarios 5A and 53 are intended as counterparts to scenarios 4A




and 4B, but with controls set at levels that are more likely to  be




realized.  The assumptions are identical to those of scenario AA and




4B, except that all nonpoint sources, which were set to 0  in 4A  and 4B,




are allowed to assume half their uncontrolled value.  The  results




appear :'n Table 62 and 63 and Figures 68 and 69.




     ioc a wet year, the model predicts loading that is slightly more




than 5% above present conditions.  This causes a slight increase in the

-------
                                                                  337
       WET YR HI  GRO  COST  EFFECTIVE CONTR(Sfl)  VS PRESEN
           lOCh
     LU
     CJ
     ct:
     UJ
     D_

     UJ
     CO
     CC
     o
 60--
            20-
           -20-
     UJ
     o
     a:
     UJ
     a_
      *
     ce
     UJ
     UJ
     in
-60--
          -100
             0       1.000

               TOTflL  P
                    2.000
                   CHLfl
 3.000
SECCHI
4.000      5.000

BOTTOM  02
Figure  68.  Graphical summary of Dillon Clean Lakes  Model outputs  fr.
           four key lake characteristics.  Vertical lines show
           deviation from present conditions; horizontal lines wit:
           "+" mark the oligotrophic and eutrophic  boundaries for
           chlorophyll a.

-------
                                                                  338
        DRY  YR  HI  GRG  COST  EFFECTIVE CONT(53)  VS  PRESENT
      UJ
      O
      a:
      LU
      a.
      UJ
      03
      ce:
      a
 SO--
             20f
            -20f
      2:
      UJ
      CJ
      a:
      UJ
      a.
       &
      cc.
      UJ
      UJ
      C3
-50-
           -ICO
               0       1.000
                TOTflL P
                    2.000
                   CHIP
 3.000
SECCHI
4.000      5.000
BOTTOM  02
Figure  69.  Graphical sunmary of Dillon Clean Lakes Model outputs  for
           four key lake characteristics.   Vertical lines show
           deviation from present conditions; horizontal lines with
           "+" mark the oligotrophic and  eutrophic boundaries  for
           chlorophyll a.

-------
                                                                       339
expected chlorophyll  level,  which  reaches 7.9,  very close to the upper


mesotrophic boundary  at  8.0.  For  a  dry year,  there would also be a


slight increase in chlorophyll above the present conditions, but the


dry-year increase would  not  bring  the chlorophyll so close to the upper


mesotrophic boundary.  Thus  the model shows  that, if the controls could
          *

be accomplished, the  lake  would in most years  fall in the mesotrophic


range.


     Although  scenarios  5A and 5B  are labelled  as cost-effective, they


may actually imply intolerable expenditures.   For example, the


scenarios assume that  phosphorus yield from  the expected 8000 persons


on septic systems is  somehow reduced by 50%, and they assume the sane


for Climax Molybdenum.   They also  assume that  such diffuse sources  as


the interstate highway and ski slopes can be treated in such a manner


as to reduce yields by one-half.  Thus the realism of this scenario is


open to some skepticism, but it certainly approaches the reality of


maximum-investment controls  much more closely  than scenarios 4A and


4B.





Generalized Graphical  Predictions


     A large number of relationships can be  developed in graphical  forn


using the logic that  is  inherent in  the model.   These graphical


relationships are necessarily less specific  in  their predictive


capabilities than actual model runs, since simplifying assumptions  are


required for a generalized graph that would  not be required for a nodel


run.   Nevertheless, generalized graphs give  an  overall impression of


the sensitivity of the lake  and its  watershed  to development of

-------
                                                                      340





different kinds.  We have developed for a graphical  presentation three




kinds of relationships:  1) the relationship between  total  phosphorus




loading and chlorophyll, 2) the relationship between  number of  persons




on sewer and total phosphorus loading, and 3) the relationship  between




the number of persons on septic system and total phosphorus loading.




     Figure 70 shows the general relationship between total phosphorus




loading and chlorophyll concentration.  As in the modelling exercises,




the chlorophyll concentration referred to in the graph is the average




for the 0-5 meter layer over the post-runoff stratification interval.




The relationship of chlorophyll to phosphorus loading is depicted  for




two different sets of conditions.  The first set of conditions  is  based




on the 1981 study year.  In the preparation of this curve it was




assumed that the runoff, water level, and lake volume changes were




identical to those of 1981.  The second curve is based on 1982  and




involves similar assumptions for that year.  Since 1981 was  a very  dry




year and 1982 was a year of above-average but not extreme wetness,  a




median year would fall between the two curves but nearer to  the  1982




curve than the 1981 curve.  Changes in lake operation (water levels)




for either of these conditions would change the positions of the




curves, but the curves would not move dramatically unless there  were




very significant departures from the present mode of  lake operation.




     Figure 70 indicates along with the two curves the lower limit  of




loading for the two hydrologic conditions.  These represent  the  natural




background for the two years depicted in the graph.   Also indicated in




Fi:r:rs 70 are the observed loadings in 1981 and 1982  and the trophic

-------
   co
   c
   1-1
   O)
I—• T3 cn >•
vo £J" fl> 
M cn tn r«
 T3 O &1
(u :r 3 TO
0O   ID

 COO
H en | D*
vo  o> M
oo |-™*   o
N> O 0 fl
• (11 x—' O
 O-   T)
 H- H- y
OQ


g
Cu
n>
   t-h
   o
  O
O. O

O

O W
W
H- CO
  HI 3
n e: o
o 3 HI
3 o HI
O. rt
H- H- in
rt O ft
H- 3 i-l
O  (a
     Hi
 tn rt h1-
 H- o O
 E;l rt 0)
 H- (1) (t
 (_• I--1 H-
 Co   O
 M   TJ

 rt
 O
          39

          36


          33


          30


          37
             0
          :24
         o»
          '21
         O

         o 15
            9


            6


            3

            0
                     4000
                                  Total  P load - Ib/yr

                                8000     12000    16000
                                             18000
                                                            Wet year
                                                             (1982)
                         Bsr.^   _  /^             Eutr°phlc
                        ^-Observed load      f
Dry year    /     .^        '^82       Mesotrophic
bac^k^roupd	^^_	I	

              -Wet year background        Oligotrophic
             0   1000     3000      5000      7000

                                 Total P loading - kg/yr
                                                            9000

-------
classifications that correspond to various chlorophyll  levels.   It is




clear from the graph, as it was from the modelling  exercises,  that




minor increases in loading either in a dry or wet year  will  push the




lake from the upper mesotrophic category where  it presently  lies into




the eutrophic category.




     Figure 71 shows the relationship of number  of  persons on  sewer to




the phosphorus loading in kilograms per year.   Number of  persons is




annual full-time equivalent (1 full time equivalent  equals ca.  100




gallons per day effluent).  As with Figure 70,  the  graph  includes  two




curves, one of which is based on conditions identical to  1981  and  the




second on conditions identical to 1982.  The intersection of each  of




these lines with the y-axis represents the phosphorus loading  to the




lake if there were no contributions whatever from sewered areas.  This




can be considered a sort of non-sewer background for the watershed.




Contributions of sewered areas above this background are  then  added  on




the basis of contributions of sewered areas through  point and  nonpoint




sources.  Point sources are assumed to consist of waste water  treatment




plants utilizing tertiary treatment.  Tertiary  treatment  is  assumed  to




function at the same overall efficiency as observed  in  1981-1982.   In




addition to this point-source contribution, the  non-point source




contribution corresponding to urban area on sewer is also added  to get




the total per capita contribution from sewered  areas.   Since the




nonpoint contribution is dependent on amount of  runoff, the  nonpoint




contributions for years of different wetness are different,  but  in both




cases nonpoint contributions were below 10% of  the  point  source




contributions.

-------
t—i rt to
vo H a
GO fD a.
H' to
I  rt to
h~* 0 0)
vo n> en
oo a c
tj rt B
  ft) to
  HI rt
    to to

    ex, ru
    o (D
    a M
     I o.
    •o
    o n
    H- o
    a a
    rt Co
      rt
    n to
    §s
n
a "0
o o
V<* )-j.
  3
C en
to O
M C
rt O
O ft)
  (O
  rt o

  O H-

  10 0>
  n>
  H rt
  < (D
  ft) >-t
  0- rt
   H-
  n.
  c
  M
  H-

  00
cr H
c a1
rt fD
p-

a ^
W O

to fo'
CO f)
CO rt
O H-
n o
H- 3
to co
rt
m H-
o. a
  o
C 1 '
H- C
rt O.
rr m

in cr
(D O
*I rt
fo :r
H
ro rt
o- rr
  ID
     00
     C
     t-t
     (D
   "O  ?3
    a* (o
    o  H'
    en  to
    _  rt
    a* H-
    o  o

    ss
    o)  ^r

    MT)
    o
    to  o
    o. H,
    H-
    a  a
   OP  c
      a
    O  D'
    MI ro
to o
?i* Hi
m

O ro

M Co
M O
O 3
3 en

P- O
Ml 3

O 01
rt fD
:r <;
ID n
ft)
cu
in
        O O
        to
        Cu rt
        P- 3*
        a fD
        oo
         rt
        rn o
        O rt
        f. ft)
        f1 t—'
        O

        tfi
      o  5000
      0)
         S  4000
         ex

         en
         _v:
          i
         T3
         u
         o
         o_  2000-
             1000-
                         20    30   40   50   60   70   80   90

                           THOUSANDS  OF PERSONS  ON  SEWER
                                                                                    200OO


                                                                                    18000
                                                          Wet  year
                                                          (1982)
                     Observed load
                         1982
                                    Observed load
                                         1981
                                                                              0
                                                                            100
                                                                                       a
                                                                                            a.
                                                                                            (/>
                                                                                            .a

                                                                                            i
                                                                                             CO
                                                                                             -F-
                                                                                             UJ

-------
                                                                      344




     The observed loadings for 1981 and 1982 are  indicated  on  Figure 71




for reference purposes.  The observed loadings correspond to




populations on sewer of about 20,000.  The actual  number of persons on




sewer in 1981 and 1982 was actually closer to 15,000.   The  observed




values correspond to a slightly higher population  because part of  the




population served by sewer (Copper Mountain) was not receiving tertiary




treatment in 1981-1982.  In the projections, it is  assumed  that all




sewered areas are receiving tertiary treatment, which  should be true




for future years.




     Figure 72 has a similar rationale to that of  Figure 71 except  that




phosphorus loading is related to contributions from septic  systens




rather than sewers.  Lines representing dry year and wet year




conditions differ in slope because of the higher yield  of phosphorus




from septic sources when runoff is high.  In addition,  the  inventory




change function dictates that a larger percentage  of the phosphorus




originating in septic areas will reach the lake in  a wet year  than  in a




dry year.  It should be noted that the loading scale (y-axis)  covers




three times the range of the scale in Figure 71 for sewer.   Observed




loadings in 1981 and 1982 are only a small fraction of  the  total




potential loading if the population on septic were  drastically




increased.









Overview of Scenario Studies




     The modelling and graphic studies indicate that Lake Dillon will




r.ove into the eutrophic category if diversion water rich in phosphorus




is added in quantity to the lake or if high growth occurs without  the




adoption of nonpoint source controls cr other measures  not  now in

-------
     oq
     c
a* n rt pd
to o o m
cn 3 rt M
n cn to to
CX rt M rt
  to  H-
O 3 T) O
3 rt 3' 3
    O cn
to to cn 3*
< rt -a H-
fl>   3" X)
1 it O
to 3- M O
OQ fD C if>
<0 H- tn

t!   (-• O
to (-< O T>
It vo CO C
fD 00 C^ H4
1 !-• H- to

3-0 tS H.
•-* IJ uw r**
(D r<  O
&•   to 3
  M cn cn
1 vo cn
g oo c cn
P to Q ft>
O   p- i-i
i-h M 3 <
ni n> CTO n>
  <  ex
M> (D O
O M rt cr1
M cn 3- •<
  •  ID
OO H h-"U
I-1 3" O rt
  fD CU M*
cu  a. n
3 -a H-
CX i-1 3 in
  o cro "^
M LJ. tn cn
oo o cu
  o
  3
     fD
     B
   fD tn
          2700O
         24000 -
          21000 -
          18000 -
         >> 15000-

         O)
         Q.

         cn|2000

           9000-
           6000-
           2000-
                0
                                                       Wet year
                                                       (1982)
   Observed
 t load
1982
  tObserved load
   1981
    10   20
   rt> o

   p-
                                30   40    50   60   70   80
                              Thousands of persons on septic
  60000


  54000


'-J48000


- 42000


- 36000


- 30000
*•

- 24000


- 18000


- 12000


-6000
                                                                90  100
                                                                               2
                                                                                '
                                                                               •a
                                                                               o
                                                                                       U)
                                                                                       4>
                                                                                       Ui

-------
                                                                      246




practice to reduce the phosphorus loading of the lake.  Under  low




growth or high growth with additional controls, the condition  of the




lake could be held within the mesotrophic range, but might  suffer  some




trophic degradation or remain more or less the same, depending on  the




exact conditions.

-------
                              Literature  Cited









Alabaster, J.S., and R. Lloyd. 1980. Water  Quality  Criteria  for  Freshwater




     Fish. Butterworth, Boston. 297 pp.




American Public Health Assn.  1976. Standard Methods  for  the  Examination of




     Water and Wastewater. American Public  Health Assn.  Washington,  D.C.




     1193 pp.




Bailey-Watts, A.E. 1978. A nine year study  of the phytoplankton  of the




     eutrophic and non-stratifying Loch Leven (Kinross,  Scotland). J. Ecol.




     66: 741-771.




Bendschneider, K., and R.J. Robinson.  1952. A new spectrophotometric method




     for the determination of nitrate  in  sea water-  J. Mar.  Res.  11: 87-96.




Bormann, F.H., G.E. Likens and J.S. Eaton.  1969. Biotic  regulation of




     particulate and solution losses from a forest  ecosystem.  BioScience




     19: 600-610.




Bormann, F.H. and G.E. Likens. 1979. Pattern and Process in  a Forested




     Ecosystem. Springer-Verlag, New York.  253 pp.




Carlson, R.E. 1977. A trophic state index for lakes. LLmnol. Oceanogr. 22:




     361-368.




Cornett, J.R. and F.H. Rigler. 1980. Areal  hypolimnetic  oxygen deficit:  Ar.




     empirical test of the model. Limnol. Oceanogr-  25:  672-679.




Dillon, P.J. and F.H. Rigler. 1974. The phosphorus-chlorophyll relationship




     in lakes. Limnol. Oceanogr. 19: 767-73.




Ednondson, W.T. 1972. Nutrients and phytoplankton in Lake Washington, pp.




     172-188 _In_ G.E. Likens (ed.) Nutrients and Eutrophication.  American




     Society of Limnology and Oceanography  Special  Synposia, Vol.  1.

-------
                                                                       348




Fee, E.J. 1976. The vertical and seasonal distribution of chlorophyll  in




     lakes of the Experimental Lakes Area, northwestern Ontario:




     implications for primary production estimates. Limnol.  Oceanogr.  21:




     767-783.




Fogg, G.E., W.D.P. Stewart, P- Fay, and A.E. Walsby.  1973. The  Bluegreen




     Algae. Academic Press, N.Y. 459 pp.




Fogg, G.E. 1975. Algal Cultures and Phytoplankton Ecology. Univ. of




     Wisconsin Press, Madison. 175 pp.




Ganf, G.G. 1975. Photosynthetic production and irradiance-photosynthesis




     relationships of the photoplankton from a shallow equatorial lake  (Lake




     George, Uganda). Oecologia 18: 165-183.




Gibson, C.E. 1978. Field and laboratory observations on the  temporal and




     spatial variation of carbohydrate content in planktonic blue-green




     algae in Lough Neagh, Northern Ireland. J. Ecol. 66: 97-115.




Glooschenko, W.A., J.E. Moore, and R.A. Vollenweider. 1973.  Chlorophyll a




     distribution in Lake Huron and its relationship to primary




     productivity. Great Lakes Res. 16: 40-49.




Goldman, C.R. 1974. Eutrophication of Lake Tahoe emphasizing water quality.




     EPA Report 660/3-74-034. 408 pp.




Golterman, H.L. (ed.). 1976. Interactions between Sediments  and Fresh Water.




     Junk, The Hague. 473 pp.




Grant,  M.C. and W.M. Lewis, Jr. 1982. Precipitation chemistry in the




     Colorado Rockies. Tellus 34: 74-88.




Grasshoff, K. 1976. Methods of Seawater Analysis. Verlag Chenie, Weinhei=.




     317 pp.

-------
                                                                       249




Harris, G.P. 1978. Photosynthesis, productivity, and  growth:  the




     physiological ecology phytoplankton. Arch. Hydrobiol.  Beih.  Ergebn.




     Limnol. 10: 1-171.




Harris, G.P. 1980. Temporal and spatial  scales  in  phytoplankton ecology.




     Can. J. Fish. Aquat. Sci. 37: 877-900.




Heaney, S.I., and J.F. Tailing. 1980.  Dynamic aspects  of  dinoflagellate




     distribution patterns in a small  productive lake. J. Ecol. 68:  75-94.




Higgins, J.M. and B.R. Kim. 1981. Phosphorus retention models for Tennessee




     Valley Authority reservoirs. Water  Resour. Res.  17:  571-576.




Robbie, J.E. and J.L. Tiwari. 1977. Ecosystem models vs.  biological  reality:




     Experiences in the systems analysis of an arctic pond. Verh. Internat.




     Verein. Limnol. 20: 105-109.




Hutchinson, G.F. 1938. On the relation between oxygen deficit and the




     productivity and typology of lakes. Int. Rev. ges. Hydrobiol. Hydrogr -




     36: 336-355.




Hutchinson, G.E. 1957. A Treatise on Limnology. Vol.  I. Geography, Physics,




     and Chemistry. Wiley, New York.   1015 pp.




Hutchinson, G.E. 1967. A Treatise on Limnology, Vol.  II.  Introduction to




     Lake Biology and the Limnoplankton. Wiley; New York. 1115 pp.




Jones, A.R. and G.F- Lee. 1982. Recent advances in assessing  impact  of




     phosphorus loads on eutrophic-related water quality. Water Res. 16:




     503-515.




Jones, J.R. and R.W. Bachmann. 1976. Prediction of phosphorus and




     chlorophyll levels in lakes. J. Water. Pollut. Cont. Fed. 48:




     2176-2182.

-------
                                                                       350





Kirk, J.T.O. 1975. A theoretical analysis of the  contribution  of  algal  cells




     to the alteration of light within natural waters.  New  Phytol.  75:




     11-20.




Komarek, J. 1974. The morphology and taxonomy of  crucigenoid algae




     (Scenedesmaceae Chlorococcales). Arch. Protistenk.  116: 1-75.




Latnbou, V.W., S.C. Hern, W.D. Taylor, and L.R. Williams.  1982.  Chlorophyll,




     phosphorus, secchi disk and trophic state.   Water  Resources  Bulletin.




     18: 807-813.




Larson, D.P. and H.T. Mercier. 1976. Phosphorus retention capacity  of lakes.




     J. Fish. Res. Bd. Canada 33: 1742-50.




Lasenby, D.C. 1975. Development of oxygen deficits in 14 southern Ontario




     lakes. Limnol. Oceanogr. 20: 993-999.




Lean, D.R.S., and F.R. Pick. 1981. Photosynthetic response  of  lake  plankton




     to nutrient enrichment: a test of nutrient limitation. Linmol.  Oceanogr




     26: 1001-1019.




Leopold, L.3., M.G. Wolman, and J.p. Miller- 1964. Fluvial  Processes in




     Geomorphology.  W.H. Freeman, San Francisco. 522 pp.




Levine, S.N., and D.W. Schindler. 1980. Radiochetaical analysis  of




     orthophosphate concentrations and seasonal changes in  the  flux  of




     orthophosphate to seston in Canadian Shield  lakes. Can. J. Fish. Aquat.




     Sci. 37: 479-437.




Lewis,  W.M., Jr. 1974. Primary Production in the  plankton community  of a




     tropical lake. Ecol. Konogr. 44:  377-409.




Lewis,  W.M., Jr., 1980. Comparison of spatial and temporal  variations in  the




     zoo plankton of a lake by means of variance components. Ecology  59:




     666-671.

-------
                                                                       351






Lewis, W.M., Jr., and D. Canfield. 1977- Dissolved organic  carbon  in  some




     dark Venezuelan waters and a revised equation for  spectrophototnetric




     determination of dissolved organic carbon. Arch. Hydrobiol. 79:




     441-445.




Lewis, W.M., Jr. and F.H. Weibezahn.  1976.  Chemistry, energy  flow, and




     community structure in some Venezuelan fresh waters. Arch. Hydrobiol.




     Suppl. 50(2/3): 145-207.




Lewis, W.M., Jr. and M.C. Grant. 1978.  Sampling and  chemical  interpretations




     of precipitation for mass balance  studies. Water Resour. Res. 14:




     1098-1104.




Lewis, W.M., Jr. and M.C. Grant. 1979.  Relationships between  stream




     discharge and yield of dissolved substances from a Colorado mountain




     watershed. Soil Science 128: 253-363.




Lewis, W.M., Jr. and M.C. Grant. 1980a. Change in the output  of ions  from a




     watershed as a result of acidification of precipitation. Ecology 60:




     1093-1097-




Lewis, W.M., Jr. and M.C. Grant. 1980b. Relationship between  snow  cover and




     winter losses of dissolved substances  from a mountain  watershed. Arct.




     Alp. Res. 12(1):  11-17.




Lewis, W.M., Jr., and J.F. Saunders,  III. 1979. Two  new integrating samplers




     for zooplankton, phytoplankton,  and water chemistry. Arch. Hydrobiol.




     85: 245-249.




Likens, G.E. 1975. Primary production of inland aquatic ecosystens. l£ H-




     Lieth and R.H. Whittaker (eds.)  The Primary Productivity of the




     Biosphere. Springer-Verlag, New  York.

-------
                                                                       352






Likens, G.E. F.H. Bormann, R.S. Pierce, J.S. Eaton, and N.M. Johnson.  1977.




     Biogeochenistry of a Forested Ecosystem. Springer-Verlag, New York.  146




     pp.




Lund, J.W.G. 1961. The periodicity of U algae in three English lakes.  Verh.




     Internat. Verein. Limnol. 14: 147-154.




Lund, J.W.G. 1965. The ecology of freshwater phytoplankton. Biol. Rev. 40:




     231-293.




Lund, J.W.G. 1970. Primary Production. Froc. Soc. Wat. Treat. Ex an. 19:




     332-358.




Lund, J.W.G. 1971. The seasonal periodicity of three planktonic desmids in




     Windermere. Mitt. Internat. Verein. Liamol. 19: 3-25.




Lund, J.W.G. 1979. Changes in the phytoplankton of an English lake,




     1945-1977. Hydrobiol. Jour. 14:  6-21.




Margalef, R. 1978. Life-forms of phytoplanktcn as survival alternatives in




     an unstable environment. Oceanol. Acta 1:  493-509.




Mortimer, C.H. 1941. The exchange of  dissolved substances between mud  and




     water. J. Ecol. 29: 280-329.




Murphy. J. and J.P. Riley. 1962. A modified single solution method for the




     determination of phosphate in natural waters. Anal. Chim. Acta 27:




     31-36.




Moss, B. 1972. Studies on Gull Lake,  Michigan. Seasonal and depth




     distribution of phytoplankton. Freshwat. Biol. 2: 289-307-




Nalewajko, C., and D.R.S. Lean. 1980. Phosphorus, pp. 235-258 In I. Morris




     fed.) The Physiological Ecology  of Phytoplankton. Univ. Calif. Press,




     Rerkelev.

-------
                                                                       353




Nauwerck, A. 1963. Die Beziehungen  zwischen  Zooplankton  und  Phytoplankton




     im See Erken. Symb. Bot. Upsal.  17:  1-163.




Nelson, W.C. 1971. Comparative  limnology  of  Colorado-Big Thompson  project




     reservoirs and lakes. Colorado Division of Wildlife Report  (mineo).




OECD. 1982.




Park, A.R. 1979. Modifications  to the model  CLEANER  requiring  further




     research, pp. 87-108 D.  Scavia and A. Robertson (eds.)  Perspectives on




     Lake Ecosystem Modeling. Ann Arbor Science, Ann Arbor.  326 pp.




Parsons, T.R., M. Takahashi and  B.  Hargrave.  1977. Biological  Oceanographic




     Processes. Pergamon, New York. 332 pp.




Rai, Hakumat (ed.) 1980. The measurement  of  photosynthetic pigments in




     freshwaters and standardization  of methods. Arch. Hydrobiol.  Beih.




     Ergeb. Limnol. 14: 1-106.




Ragotzkie, R.A. 1978. Heat budgets  of lakes,  pp. 1-19. ^n_ A. Lerman (ed.)




     Lakes: Chemistry, Geology,  Physics.  Springer-Verlag, New  York. 363 pp.




Ramberg, L. 1976. Relations between phytoplankton and environment  in two




     Swedish forest lakes. Scripta  Limnol. Upsal. 426: 1-97.




Ramberg, L. 1979. Relations between phytoplankton and light  climate in two




     Swedish forest lakes. Internat.  Rev. ges. Hydrobiol. 64:  749-782.




Rieman, B. 1980. A note on the use  of methanol as an extraction  solvent for




     chlorophyll a determination. Arch. Hydrobiol. Beih.  Ergebn. Linnol.




     14: 70-78.




Rigler, F.H. 1968. Further observations inconsistent with the  hypothesis




     that molybdenum blue method measures orthophosphate  in  lake water.




     Limnol. Cceanogr. 13: 7-13.

-------
                                                                       354





Rohlf, J.F. and R. Sokal. 1969. Statistical Tables. W.H. Freeman,  San




     Francisco. 253 pp.




Saunders, G.W., F.B. Trama, and R.W. Sachmann. 1962. Evaluation  of  a




     modified C-14 technique for shipboard estimation of photosynthesis  in




     large lakes. Great Lakes Res. Div. Publ. No. 8. 61 p.




Sawyer, C.N. 1947- Fertilization of lakes by agriculture and urban  drainage.




     J. New England Water Works. Ass. 61: 109-127-




Schindler, D.W., R.W. Newbury, K.G. Beaty, and P. Campbell. 1976.  Natural




     water and chemical budgets for a small Precambrian lake basin  in




     Central Canada. J. Fish. Res. Bd. Canada. 33:  2526-2543.




Smith, R.C. 1968. The optical characterization of natural waters by means of




     an "extinction coefficient." Limnol. Oceanogr. 13: 423-429.




Smith, V.H. 1982. The nitrogen and phosphorus dependence of algal  bioinass in




     lakes: an empirical and theoretical analysis. Limnol. Oceanogr. 27:




     1101-1112.




Solorzano, L. 1969. Determination of ammonia in natural waters by  the




     phenol-hypochlorite method. Limnol. Oceanogr. 14: 799-801.




Solorzano, L., and J. Sharp. 1980a. Determination of total dissolved




     phosphorus and particulate phosphorus in natural waters. Limnol.




     Oceanogr. 25: 754-758.




Solorzano, L., and J. Sharp. 198Cb. Determination of total dissolved




     nitrogen in natural waters. Limnol. Oceanogr. 25: 751-754.




Strom, K.M. 1931. Fetovatn: A physiographic and biological study of a




     mountain lake. Arch. Eydrobiol. 22: 491-536.

-------
                                                                       355




Stumm, W. and P. Baccini. 1978. Man-Made  Chemical  Perturbations  of  lakes.




     pp. 91-123. In_ A. Lerinan  (ed.)  Lakes -  Chemistry,  Geology,  Physics.




     Springer-Verlag, New York. 363  pp.




Stumm, W. and J.J. Morgan. 1981. Aquatic  Chemistry.  Wiley,  New York.  780  pp.




Taguchi, S. 1976. Relationship between photosynthesis  and  cell size of




     marine diatoms. J. Phycol. 12:  185-189.




Tailing, J.F. 1969. Measurement of photosynthesis  using illuminated constant




     temperature baths, pp.  108-117  J[n_ R.A.  Vollenweider (ed.) A Manual on




     Methods for Measuring Primary Production  in Aquatic Environments.




     Blackwell, London.




Tailing, J.F. 1971. The underwater light  climate as  a  controlling factor  in




     the production ecology  of freshwater phytoplankton. Mitt. Internat.




     Verein. Limnol. 19: 214-243.




Tailing, J.F. 1976. The depletion of carbon  dioxide  from lake water by




     phytoplankton. J. Ecol. 64: 79-121.




Tailing, J.F. 1979. Factor interactions and  implications for the prediction




     of lake metabolism. Arch. Hydrobiol.  Beih. 13:  96-109.




Tailing, J.F. 1982. Utilization of solar  radiation by  phytoplankton. pp.




     619-631. ^n_ C. Helene, M. Charlier Th.  Montenay-Garestier and  G.




     Laustriat (eds.) Trends in Photobiology.  Plenum,  New York.




Vollenweider, R.A. 1968. Scientific  fundamentals of  the eutrophication of




     lakes  and flowing waters, with  particular reference to nitrogen and




     phosphorus as factors in eutrophication.  Organization  for Economic




     Cooperation and Development Report.  DAS/CSI/68.27, 182 pp.

-------
                                                                       356





Vollenweider, R.A. 1969. Possibilities and limits of  the  elementary  models




     concerning the budget of substance in lakes. Arch. Hydrobiol. 66:




     1-36.




Vollenweider, R.A., W. Rast and J. Kerakes. 1980. The  phosphorus  loading




     concept and Great Lakes eutrophication. pp. 207-234  _In_ Loehr, R.C.,




     C.S. Martin, and W. Rast (eds.) Phosphorus Management  Strategies  for




     Lakes. Ann Arbor Science, Ann Arbor- 490 pp.




Wahlstrom, E.E. and D.Q. Eornback. 1962. Geology of the H.D. Roberts Tunnel,




     Colorado. Geological Society of America Bulletin  73:1477-1498.




Welch, E.B. 1980. Ecological Effects of Wastewater. Cambridge University




     Press, New York. 337 pp.




Wolk, C.P. 1973. Physiology and cytological chemistry  of  blue-green  algae.




     Bacterioi. Rev. 37: 32-101.




Wright, R.F. 1974. Forest Fire: Impact on the hydrobiology, chemistry and




     sediments of small lakes in northeastern Minnesota.  Interim Rept. No.




     10, Limnology Research Center, University of Minnesota, Minneapolis,




     MN. 129 pp.

-------
                                                                       357

                     Summary of Public  Participation

     The Steering Committee of the Lake  Dillon  Clean  Lakes  Study was

composed of the following members:

     AMAX Inc.
     Keystone Corporation
     Copper Mountain Water and Sanitation  District
     Breckenridge Sanitation District
     Frisco Sanitation District
     Dillon/SiIverthorne Joint Sewer Authority
     Denver Water Department
     Town of Breckenridge
     Town of Frisco
     Town of Dillon
     Summit County
     Northwest Colorado Council  of Governments
     Colorado Department of Health
     Region VIII, U.S. Environmental Protection Agency
     During the progress of  the  study one or more representatives of

the contractor, Western Environmental Analysts,  Inc., attended  12

meetings at which discussions or  formal  presentations were sade

concerning the Study-  The date,  place and  nature of each meeting is

listed below.

Jan. 6, 1981 - Frisco - Steering  Committee  meeting  to discuss work
     plan.
Feb. 10, 1981 - Frisco - Steering Committee meeting to discuss addendun
     to work plan.
Apr. 8, 1981 - Frisco - Steering  Committee  meeting  to discuss sampling
     plan and parameter measurements.
Nov. 23, 1981 - Frisco - Steering Committee meeting to discuss  first
     year results and second year plans.
Apr- 13, 1982 - Breckenridge - presentation of  first year results at
     public meeting.
June 18, 1982 - Frisco - description of  Study and first  year  results  to
     joint meeting of members of  Colorado River  and Middle Park Water
     Conservancy Districts.
July 7, 1982 - Frisco - discussion of comments  on first  annual  report
     with Steering Committee and  representatives from two cor.s-j-Ltir.g
     firns, Engineering Science  and Gulp Wessner Gulp.

-------
                                                                       353
Feb. 16, 1983 - Frisco - status report on analysis of  first  and  second
     year results to meeting of Billon watershed discharge permit
     holders and others.
Mar- 1, 1983 - Denver - status report on analysis of first and  second
     year results to meeting of American Waterworks Association  members
     and guests.
May 19, 1983 - Frisco - Steering Committee meeting to  discuss draft
     final report.
July 12, 1983 - Frisco - presentation of first and second year  results
     to Dillon Reservoir Clean Lakes Policy Committee  (policy-making
     representatives from Steering Committee membership as opposed  to
     technical representatives described previously as the "Steering
     Committee").
July 26, 1983 - Breckenridge - presentation of final (first and  second
     year) results at public meeting.

-------
                                                                       359
                              Appendix I
                       DILLON CLEAN LAKES MODEL
                        Input, Code, and Output

  For a model run (Scenario 1A: Low growth, no diversions, wet year)
Input  - Pages 1-1  - 1-2
Code   - Pages 1-3  - 1-15
Output - Pages 1-16 - 1-21

-------
 21 1
3572 42-1
2322  83
    0
    0
    0
BR13572.
BR2 424.
BR4 231
BRG 253
BR7 235.
SRI 319.
SR2 310.
SR3 500.
SR4 364.
TH1 433.
TM4 406.
IM7 513.
MC2 325.
SC2  77.
AAA 325.
BBB 325.
CCC 325.
ODD 325.
EEE 325.
       235
       192
         0
         0
         O

       63.
231 253
213 165
0    0
0    0
0    0
   25.
11003.
 9719.1410
 6133.   11
 5165.  178
 1041   909
 2G78.1305
 9789.
 4 1 22.
 G329.
11422.
 6447
 1000.
 2212.
 1  143.
 1226.
  619.
     319 310
     120 160
        O
        0
        0
   5OO
   240
   O
   0
   0
364
227
 0
 0
 0
433
259
  0
  0
  0
406 513
151 435
   0
   0
   0
325
126
0
0
0
          4263.
  33.
1817.
     367.
     440.

      86.
294.
252.2388.
536.

    1258.
317.
697
647.
843.

 64.
133.

135.
                 129.
                 45.
                            43.  28.
                              6700.
                           386.
                           177.
                                 28.
                                178.
                                 10.
                            78.   90.
                       42.
                       67
          302.
                                    2500.
77
21
O
0
0
325
126
  0
  0
  0
325
126
325
126
325
126
325
126
                                         2650.
                                    3230.
                                      3300.

-------
1
2
3
4
5
6
7
8
9
10
1 1
12
15
45
74
105
135
1 66
19S
227
2S8
288
319
349
8901
8977
8974
8974
8980
9OOO
9018
90)8
901 8
9017
9016
9015
4
3,
5,
7
33,
92,
56,
27,
14.
9.
6,
6,
. 78
.65
, 19
73
, 71
,04
, 30
,48
, 71
93
,23
, 19
12.
1 1
9
4 ,
3,
6,
30,
2G
15,
10,
13.
5,
, 54
13
.07
. 68
, 82
. 15
. 38
53
35
17
, 12
.79
0
0
o
0
0
0
0
0
0
0
0
0

-------
        rv  or  MINNESOTA FORTRAN COMPILER AVERSION  5.4  -  79/03/01)  ON THE 6-400 UNDER KRONOS 2.1.0  ON  eo/O4/oi  AT 16 03
NIVEFtSITY  COMPUT I NO CENTER - UNIVERSITY OF COLORADO
         MliF, e=
                            =DILEXP, L =
         OOOOOOD
         O10335E
          OI0335B
          010335B
     v> .
     0.
Ol 1C! IB
Cll IG^bU
          Ol 1 G.-'Ul
          O I 1 b-l'-
          01 I G.l'j
                  c
                  c
                  c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
                   c
   PROGRAM  DILEXP (INPUT,OUTPUT,TAPE1,TAPE2,TAPE3)
THIS  IS  THE LAND USE ELEMENT OF THE DILLON  CLEAN LAKES MODEL.
IT PREDICTS NUTRIENT YIELD TO THE LAKE GIVEN LAND USE DATA
AS INPUT AND AMOUT OF RUNOFF
   DIMENSI ON NAME(19),WATR(19).AREA(19),PTOT(19),NTOT(19),
  1PBKG(19).NBKG(19),LUSE(19,8).PYLD(19,6),NYLD(19,8)
  2,USTOP(8),USTON(8),USPCP(8),USPCN(8),HI WAT(8),
  3LOWAT(8),DI V(12),DVCONP(12),DVCONN(12)
TAPE1 ONE IS THE DATA SOURCE,STORED ON DISK AS DILEX
THE FIRST RECORD ON TAPE1  IS THE RUNOFF(EXCLUDING ANY PUMPED
WATER).  THIS IS  VARIABLE USGS AND IS  GIVEN  AS THOUSANDS
OF ACRE  FEET PER YEAR.  IT IS THE SUM  OF THE 4 USGS
GAUGES  (SNAKE,KEYSTONE GULCH,BLUE,TENMILE).   THE
SECOND RECORD IS A LIST OF THE WATER  YIELDS FOR  THE
19 SEGMENTS OF THE WATERSHED FOR 1982 IN MM PER  YEAR.
THE THIRD RECORD IS THE SAME FOR 1981.  THE FOURTH
RECORD  IS A LIST OF THE AMOUNT OF WATER DIVERTED
OR PUMPED FROM OUTSIDE THE WATERSHED  IN EACH OF  THE
12 MONTHS OF THE YEAR,  AS THOUSANDS OF ACRE FEET
PER MONTH.   THE NEXT RECORD GIVES THE P CONCENTRATION
OF THE  PUMPED WATER MONTH BY MONTH AND  THE  NEXT  GIVE
THE N CONCENTRATION. THE LAST  19 RECORDS GIVE THE
AREA, WATER YIELD, AND LAND USE  INFORMATION OF EACH OF
THE  19  WATERSHED SEGMENTS.
   REAL  LUSE,PYLD,NYLD,NTOT,PTOT,PGRAND,NORAND,NBKG,PBKG
   1,PRECP,PRECN,PRPCTP,PRPCTN,PCTSTOP,PCTSTON,NETYLDP,NETYLDN
   2,HtWAT.LOWAT
 I I SETS THE NUMBER OF RUNS
   DO 901  11=1,1
THE AREA OF EACH SEGMENT  IS ENTERED ON  TAPE1 AS  IS THE
WATER YIELD.  IF THE WATER YIELDS ARE SPECIFIED  AH?NO W!TH
THE LAND USE DATA, THE SPECIFIED VALUES ARE USED.IF
THEY  ARE NOT SPEC IFlED,THE DISTRIBUTION OF  WATER YIELDS
 IS SET  PROPORTIONAL TO THE  1981 YIELDS  FOR  A YEAR
BELOW AVERAGE RUNOFF AND PROPORTIONAL TO  THE 1982 YIELDS
FOR  AVERAGE OR AtiOVE AVERAGE RUNOFF.
 THE  RUNOFF  IS VARIABLE USGS AND  IS THE  SUM  OF THE 4 GAUGES
 IN THOUSANDS OF  ACRE FEtT PER  YEAR.   THE  MATRIX  LUSE HAS 19 ROWS
 THAT  ARE 19 SEGMENTS OF THE  CATCHMENT AND  8 COLUMNS THAT
 ARE  THE LAND USLS AS FOLLOWS:  1 )RES I DENT IAL SEWER NONPOINT.AS
 NUMBER OF PERSONS. 2)URBAN SEWER AS PERSONS,3)SEPT IC AS
 PERSONS.4)INTERSTATE HIGHWAY AS  HECTARES  ROAD PLUS RIGHT OF WAY,
 t>)SKI AREA AS HECTARES OF CLEARED AREA, 6)PACKAGE PLANT OUTFALLS,
 AS NUMBER OF PERSONS SERVED , 7 ) TE.RT 1 ARY  OUTFALLS  SAME.8)AMAX
 MINING AS WORKFORCE SIZE
 THE  GROSS YIELDS ARE IINCORRECTED FOR  INVENTORY CHANGE IN THE
 VALLEY BOTTOMS,  THE NET YIELDS ARE CORRECTED FOR THIS.
 READ IN THE AMOUNTS OF RUNOFF.
    READ (1,260)  USGS, ( III WAT(K),K=1 , 19>, (LOWAT(K),K=1 , 19)
2GO FORMAT  ( F5 . O , / . 1 9F-1 . 2 , / . 1 9F-1 . 2 )
 READ IN THE  INFORMATION ON  PUMPING AMD  DIVERSIONS
    READM.297)   , ( DVCONP(K ) , K= 1 , 1 2) ,
   1 ( DVCONN < K ) , « -- 1 . 1 2 )
297 FORMAT  ( 1 2f-b .(),/, 1 21-'o  C),/  I^Fb.O)
    DO 201   1-1.1 9

-------
10.
1 1
12.
01 1G4GB
01 1670B
01 1G70B
READ (1.202) NAMEC I ) , WATR( I ) , AREA( 1 ) , (LUSE( 1 ,K),K=1 , 8)
202 FORMAT ( A3, F5 . 0, F7 . 0, 8F5 . O)
201 CONTINUE

C COMPUTE THE LOADING BY DIVERSION AND PUMPING AND CONVERT
C DIV TO MILLIONS OF M3 PER MONTH
1 3.
14.
15.
16.
17
is.
19.

20.
21 .
O ft
23.
24.
25.
26.
27
28.
29.
30.

31 .
312.
33.
3-1.
35.
36.
37
38.
39.
40.
41 .
42.
43.
44 .
45.
46.
47,
48.
43.
50.
O1 1672B
01 1672B
01 1G73B
01 1675B
01 1677B
01 1702B
01 1705B

01 17068
01 1710B
01 1713B
0 1 1 7 1 5B
01 1717B
01 1717B
01 1721B
01 1723B
01 1727B
01 1730B
011 73 IB

01 1731B
O1 1733B
01 1740B
01 17468
01 1754B
01 1762B
01 1 770B
01 1774B
0120036
O12012B
012014B
012017D
012021B
O12024B
012026B
01 203 IB
O12033B
01203613
012045B
012054B
DVSUMP=0. 0
DVSUMN--O.O
D8 298 1=1,12



DIV(I)=DIV(I) *1000. *1233. 5* .000001
DVSUMP=DVSUMP+DI V( 1 )*DVCQNP( I )
DVSUMN=DVSUMN+DIV( I )*DVCONN( I )
298 CONTINUE
C SET THE WATER YIELDS
DO 280 1=1,19



IF (WATR(2) .GT. 10. )GO TO 280
IF (US6S.GT. 180. )GO TO 281
WATR( I )=LOWAT( 1 >*(USGS/1 16.5)
GO TO 280
201 CONTINUE


WATR( I )=HIWAT( I ) * ( USGS/21 0. 9)
280 CONTINUE
WATR( 1 )=WATR(2)
PGRAND=0.0
NGRAND=O.O
C APPLY LAND USE EQUATIONS
DO 203 1=1,19
PYLD( I , 1 ) = . 000001 *LUSE<
NYLDC I , 1 )=. 000001 *LUSE(
PYLDC I , 2)=.000001*LUSE(
NYLD( I ,2)=. 000001 *LUSE(
PYLD( I , 3)=. 000001 *LUSE(
NYLD( I , 3) = .000001 *LUSE(
PYLD( I . 4 ) = . 000001 *LUSE(
NYLDC I ,4)=. 000001 *LUSE(
PYLDf I , 5)=. 000001 *LUSE(
NYLD( I ,5)-.O00001*LUSE(
PYLD( I ,6)=. 000001 *LUSE(
NYLD( I ,6)=. 000001 *LUSE(
PYLD( I , 7)=. 000001 *LUSE(
NYLD( I ,7)= . 000001 *LUSE(
PYLD( I , 8) = . 000001 *LUSE(
NYLDC I , 8)= . 000001 *LUSE(
PBKG ( I ) =AREA ( I ) * 1 OOOO . *
NBKG( I )=AREAf I )* I 0000 . *
PTOT( I )=PYLD( I , 1 )+PYLD(






, 1 )*2. 95*WATR( I )**1 . 372
, 1 )*536. *WATR( I )**1 . 154
(2)*1 .50*WATR( I )*»1 . 372
,2)*410. *WATR( 1 )**1 . 154
, 3)*3440. *WATR( I )**0.759
,3)* (44300. *WATR( I )-5205. J
,4)*10000. * , 00209*WATR( 1 )**1 . 799
,45*10000. *1 .71*WATR( I )**! . 13
,5)*6. 50*10000.
, 5)*42. *10000.
,6)*270000.
,6)*1 197000.
, 7)*47000.
, 7)*1 197000.
, 8)*98000.
, 8)*19930000.
000001*. 000782*WATR( I )**1 .372
OOOOO1 * . 0842*WATR( I >**1 . 154
, 2)+PYLD( I , 3)+PYLD( I ,4) +
1PVLD( 1 , 5)+PYLD( I ,6)+PYLD( 1 , 7>+PYLD( I , 8>+PBKG( I )
51
0120650
NTOT( I )=MYLD( I , 1 )+NYLD(
, 2 ) +NYLD ( I , 3 ) +NYLD ( I , 4 ) +
1NYLD( I . b)+NYLD( I , 6)+NYLDf 1 , 7>+NYLD( 1 , 8>+NBKG( I )
52.
S3.
54.
012076B
0121 OOB
012102B
PGRAND=PGRANn+PTQT( I )
NGRAND = NGRANC1*NTOT< 1 )
203 CONTINUE



C ADD 3.5 PERCENT FOR CONSTRUCTION ACTIVITIES P.2.2 PERCENT N
55.
56.
57
50.

59.
CO.
61 .
012103B
01 2104B
01 21 OOB
01 211 OB

O121 12B
0 1 2 1 1 CD
0121 16B
CGNSTP=PGRAND*0. 035
CONSTN=NGRAND*0. 022
PGRAND=PGRAND< CONSTP
NGRAND=MGRAMD i CONSTN
C WFUTE OUT PREDICTED YIELDS
WRITE (2, 190)
19O FORMAT (///////)
WRITE (2,226)




BY SEGMENTS




-------
62.
63.
64 .



65
GO.

67.
G6.
69.
70.
T\
72.
73.
74.

75.
76
77
78.
79.
80.
81 .
82.
83.
84.
85.
86.


87
08.
69.
90.
91
92.
93
94.
95
9G.
97
98.
99.
10O.
101
102.
103.
1.1-1 .
105.
lOf>.
IO/
1OO
109.
110.



1 1 1
012I21B
012I21B
012124B



O1 2124B
012125B

012151B
012151B
O12153B
012157B
012162B
012162B
012165B
012166B

012212B
012214B
012220B
012223B
012223B
012226B
O1222GB
012227B
012241B
012241B
OI2243B
012251B


01 2251B
01 2215 IB
0122'o3B
O12254B
OI2256B
012260B
0122GOB
01 2260B
01226.7B
0122G2B
0122G3B
0122G5B
01226GB
012270B
012271B
01 2273B
0122750
Ol 230 IB
Ol 23CKII)
Ol ?30 IB
01 I-'Jij/h
Ol :>.> I :)B
Ol r.Jl :m
U 1 231 L.Li



oi :'.•> i (.M
22O FORMAT (49H GROSS P YIELD FOR WATERSHED SEGMENTS, KG PER YR )
WRITE (2.215)
215 FORMAT ( 1 5H STATION NAME ,11H MM RUNOFF, 11H AREA HA
111H RES SEW ,11H URB SEW .11H SEPTIC ,11H HI WAY
2 1 1 H SK 1 SLOPE , 1 1 H PKG PLT , 1 1 H TERT PLT , 1 1 H AMAX
31 1H BKGRND )
DO 204 1=1,19
WRITE (2,205) NAME( 1 ) , WATRC I ) . AREA( I ) , (PYLD( 1 , K) , K=l . 6) ,
1PBKG( I )
205 FORMAT ( / , 4X. A3, 5X . 1 1 F1 1 . 1 )
204 CONTINUE
WRITE (2, 190)
WRITE (2,227)
227 FORMAT (49H GROSS N YIELD FOR WATERSHED SEGMENTS, KG PER YR)
WRITE (2,215)
DO 206 1=1,19
WRITE (2,205) NAME( 1 ) , WATR( 1 ) , AREA( 1 ) , (NYLD( 1 , K) , K=1 , 8) .
1NBKG( I )
206 CONTINUE
WRITE (2. 190)
WRITE (2,256)
258 FORMAT (42H GROSS P AND N YIELD BY SEGMENT. KG PER YR)
WRITE (2,257)
257 FORMAT ( 6H NAME , 8H PTOTAL , 8H NTOTAL )
DO 2O7 1=1,19
WRITE (2,208) NAME( I ) , PTOT( I ) , NTOT( 1 )
208 FORMAT ( / , IX, A3, 2F8 . O)
207 CONTINUE
WRITE (2.209) PGRAND, NGRAND
209 FORMAT (/////. 32H GROSS KG PER YR P IN RUNOFF IS.F6.0,/,
132H GROSS KG PER YR N I N RUNOFF IS.F0.O,/)
C COMPUTE THE TOTAL P AND N YIELDS TOTAL FOR EACH LAND USE, GROSS
PRECP=72G.
PRECN=9672.
GRNDP=10.
GRNDN=1066.
OO 210 J = 1 , 8
USTGP( J)=0.0
USTON( J)=0. 0
210 CONTINUE
TBKGf =0.0 .
TBKGN-0.0
DO 250 1=1.19
TBKGP=TBKGP+PBKG( 1 )
TBKGN=TBKGN+NBKG( 1 )
2bO CONTINUE
DO 211 J = 1 . 8
DO 212 1=1.19
USTOP( J)=PYLD( 1 . J)+USTOP(J>
USTON( J)=tlYLO( 1 . J) +USTONI J)
212 CONTINUE
•J; 1 1 CONT 1 NUE
WRITE (2,270)
27O FORMAT ( 5OH GROSS YIELD BY SOURCE KG PER YR P ( ABOVE ), N( BELOW ))
WRITE (2,271)
271 FORMAT f/,1311 RES 1 D . SEWLK'.IOII URB SEW , 1 OH SEPTIC
11 OH III WAY . IOH SKI SLOPE, 1011 PKG PLT ,1011 TERT PLT,
21011 AMAX 1011 BKGRND , 1 OH PRECIP , 1 0H GRNDUAT ,
310II DIVERSION, 1 OH COIIMKIIOT)
UKI re .2,^13) ( us Torn- > , K- i , o) , IBKGP, PUECP, GRNOP, ovsurip, CUNSTP
M
 I

-------
i 12.
13.
114.
1 1t>.
lie.
1 17
1 18.
119.
120.
121 .
22.
23.
24 .
25.
26,
127
128.
129.
130.
131
132.
133.
134.
13S.
136.
137.
133.
139.
140.
141 .
142.
143.
144.
145.
146.
147.
148.
149.
150.
151 .
152.
153.
154.
155.
156.
157.
158.
159.
O12343B
01234GB
01 234GB
012350B
012353B
012355B
012360B
O12362B
0123G48
012367B
O12372B
01237GB
012402B
01240GB
012412B
012422B
0124318
012443B
012454B
012455B
012457B
012461B
012-4G3B
0124658
O12467B
012471B
012474B
012476B
012500B
012502B
012504B
012507B
012511B
012513B
012315B
012521B
012524B
012524B
01252VB
012527B
012532B
012557B
012562B
0125G2B
O125G5B
012612B
O12615B
Ol 2G22H
160.
       O12G22B
     1(USTON(K).K=1,8),TBKGN,PRECN,ORNDN,DVSUMN,CONSTM
      WRITE(2,190)
  213 FORMAT  ( 13F10.1,/,13F1O.1)
C  COMPUTE INVENTORY  CHANGE  AND ADJUST YIELD   ACCORDINGLY
      PCTSTGP=USGS*(-.40)+87.6
      PCTSTON=USGS*(-.57)+112.
      DO 290  1=1,6
      USTOP(I)=USTOP(I)*(100.-PCTSTOP)*.01
      USTON(I)=USTONfI)*(1OO. -PCTSTON)*.Ol
  290 CONTINUE
      USTOP(8)=USTOP(8)*(100.-PCTSTOP)*.01
      USTGN(8)=USTGN(8)*(100.-PCTSTON)*.01
      CONSTP=CONSTP*(100.-PCTSTOP)*.Ol
      CONSTN=CGNSTN*(100.-PCTSTON)*.Ol
      TBKGP=TBKGP*<1OO.-PCTSTOP)*.01
      TBKGN=TBKGN*(1OO.-PCTSTON)*.01
      WWTPP=PYLD(1,7>+PYLD(16J7)+PYLD(18,7)
     1+PYLD(11.7)*(100.-PCTSTOP)*.01
      WWTPN=NYLD(1,7)+NYLD(16,7)+NYLD(18,7)
     1+NYLD(11,7)*(100.-PCTSTON)*.01
      N£TYLDP=(PGRAND-USTOP(7))*(100.-PCTSTGP)*.01+WWTPP+PRECP+GRNDP
     1 +DVSUMP
      NETYLDN=(NGRAND-USTGN(7))*(100.-PCTSTON)*.Ol+WWTPN+PRECN+GRNDN
     1*DVSUMN
      USTOPC7)=WWTPP
      USTON(7)=WWTPN
C  COMPUTE THE PERCENTAGES OF EACH CONTRIBUTION
      DO 295 J=O ,8
      USPCP(J) = (USTOP(J)/(NETYLDP))* 100.
      USPCNC J) = (USTON(J)/NETYLDN)* 100.
  235 CONTINUE
      PRPCTP=CPRECP/(NETYLDP))*100.
      GRPCTP=(GRNDP/(NETYLDP))* 100.
      TBKGPPC=(TBKGP/(NETYLDP))* 100.
      DVPCP=(DVSUMP/NETYLDP)* 100.
      CONSPCP=(CONSTP/NETYLDP)* 100.
      TBKGNPC=(TBKGN/NETYLDN)* 100.
      PRPCTN=(PRECN/NETYLDN)* 1 00.
      GRPCTN=(GRNDN/NETYLDN)* 100.
      DVPCN=(DVSUMN/NETYLDN)* 100 .
      CONSPCN=(C0NSTN/NETYLDN)* 100.
      WRITE (2,190)
      WRITE (2,296)
  296 FSRMAT(42H  VALUES  UP  TO  THIS POINT UNCORRECTED FOR  ,/,
     )55H   INVENTORY  CHANGE IN RIVER BOTTOMS,  NOW CORRECT THIS.,//)
      WRITE (2,370)
  370 FORMAT (5OH NET  LOADING BY SOURCE KG PER YR P(ABOVE),N(BELOW))
      WRITE (2,271)
      WRITE (2,213)  (USTOP(K),K=1,8),TBKGP,PRECP,GRNDP,DVSUHP,CGNSTP,
     1(USTON(K),«=1,8),TBKGN,PRECN,GRNDN.DVSUMN,CONSTN
      WRITE (2,275)
  275 FORMAT (//,4BH THE  NET YIELDS BY PERCENT. ARE P(ABOVE),N(BELOW))
      WRITE (2,271)
      WRITE (2,213)  (USPCP(K),K=1,8),TBKGPPC,PRPCTP,GRPCTP,DVPCP,
     1CONSPCP,(USPCN(K),K=1,8),TBKGNPC,PRPCTN,GRPCTN,DVPCN,CONSPCN
     WRITE (2,190)
      WR1TE(2,265)  NETYLDP.NETYLDN
  205 FORMAT(48H  THE  NET  P  YIELD  TO THE LAKE IN KG PER YEAR  IS,F6.O,
     1./.48H  THE NET  N YIELD TO  THE LAKE IN KG PER YEAR  IS.Fa.O)
      WRITE (3,902) NETYLDP,PRECP,USGS,(DIV(K),K=1,12),
                                                                                                                          M
                                                                                                                           I
                                                                                                                          CTi

-------
                      1(DVCONP(K),K=1 , 12)
161     012640B    902  FORMAT  (3F7.0,/,12F6.2,/,12F6.2)
162.    012G40B    901  CONTINUE
163.    O12642B         STOP
164.    012643B         END
                                                                                                                           t-t
                                                                                                                           I

-------
UNIVERSITY OF MINNESOTA FGRTKAN  COMPILER (VERSION 5.4 - 79/03/01) ON  THE  6400 UNDER KROMOS 2.1.O ON 83/04/O1
UNIVERSITY COMPUTING CENTER  -  UNIVERSITY OF COLORADO
                                                                                                      AT 16.03
     3.

     4.
     6.
     7.

     8.
     9.

    10.
    1 1 .
    12.
    13.
    15.
    1C.
    17.
    10.

    19.
    20.
    21 .
    22.
    23.

    24.
          OOOOOOB
01241 IB


01241 IB

01241 1 B


0126238

012G44B
0126440

012662Q
012662B

012662B
012664B
0126&4B
012G71B
012G73B
012674B
012676B
OI2677B
012702B

012703B
012704CS
012705B
0127I2B
O12714B

O12715Q
                   MNF,E = 4,l=OILTP,L=LL.
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
                   C
      PROGRAM  DILTP  (INPUT.OUTPUT,TAPE11, TAPE2,TAPE3,TAPE4)
   THIS  IS THE TROPHIC STATUS ELEMENT OF THE DILLON
   CLEAN LAKES MODEL.   THIS VERSION IS SET UP TO
   INTERFACE WITH  LAND USE ON THE INPUT SIDE AND
   EFFECTS ON  THE  OUTPUT SIDE.   THE PROGRAM TAKES
   USER SUPPLIED  INPUT ON TAPE11  AND INPUT SUPPLIED
   BY D1LEXP ON TAPE3.  KRONOS COMMANDS FOR INTERFACE
   ARE ON DILPROC.   VARIABLES COMING IN BY MONTH
   ON TAPE 11 INCLUDE MONTH NUMBER (M<3N). DAY NUMBER
   (DAY), LAKE LEVEL IN FEET (LEV),  TOTAL SURFACE
   WATER IN MILLIONS OF M3 PER MQNTH (VIN),TOTAL
   OUTFLOW IN  MILLIONS OF M3 PER MONTH (VUT). VARIABLES
   COMING IN ON TAPE3 FROM DILEXP ARE TOTAL P LOAD  IN
   KG PER YEAR (NETYL.DP),  P LOAD DUE TO PRECIPITATION
   IN KG PER YR (PRECLO),  AMOUNT OF DIVERTED OR
   PUMPED WATER IN MILLIONS OF M3 PER MONTH FOR  ALL
   12 MONTHS,  AND  P  CONCENTRATION MONTHLY  IN PUMPED
   OR DIVERTED WATER,  AS uo PER L.
   THE PROGRAM WRITES TO TAPE2 AND OUTPUT  IS
   PACKED BY DILPROC.
      DIMENSION MGN(12>. DAY(12),LEV<12),VIN(12),
     1VUT (12), PLD  (12),  ARE (12),  VOL (12),  RLDU2)
     2,DVCONP(12)
      REAL MON, LEV,  PLD,  PLOTS,  PLAKE, LANDLO,
   SET NUMBER  OF  DATA SETS TO BE TREATED
      DO 107 11=1,1
   READ  INPUT  DATA FOR EACH OF 12 MONTHS
                                                                             DIV(12)
PRECLO,NETYLDP
      READ  (11,3)  (MON(K),  DAY (K), LEV  (K), VIN  (K),  VUT(K)
      1  . K=1,12)
    3 FORMAT  (3F5.0.2F7.0)
      READ  (3,101)  NETYLDP,PRECLO,USGS,(DIV(K),K-1,12),
      1(DVCOMP(K),K=1,12)
  101 FORMAT  (2F7.0,F7.0,/,12F6.2,/,12FC.2)
      DVSUMP=0.0
C  COMPUTE  LOADING  FRACTIONS
      DO 150  1=1,12
      DVSUMP=DVSUMP+DIV(I)*DVCONP(I)
  ISO CONTINUE
      LANDLO = NETYLDP - PRECLO-DVSLIMP
C  TOTAL UP THE WATER  EXCLUDING DIVERSION  INPUT AND ADD GROUNDWATER
C  VTOT WILL  BE TOTAL  NONDI VERSION WATER  INPUT PER YR IN M3
      VTOT=0.0
      DO 103  1=1,12
      VIN( I )=VIN( I ) t-,575
      VTOT=VTOT+VIN(1)*100OOOO.
  103 CONTINUE
C  COMPUTE  MONTHLY  P  INPUTS IN KG PER MONTH
      DO 102  1=1,12
      RLD(I)=O.OG3*PRECLO
      PLD(I1=LANDLO*(VIN(I)*10OOOOO./VTOT)+DVCONP(I)*DIV(I)
  102 CONTINUE
      RLD(G)=0.3*PRECLO
C  WHITE PRELIMINARY  DATA DATA ONTO TAPE  2
      WRITE (2,2)
                                                          I
                                                         oa

-------
O I 2722B
                  2  FORMAT  I//////, '    »MONTH*
                    I1   »LEVEL  FT*  ',   *MIL  M3MO
                   2'   *MIL M3MO OUT*  '  '   »KG P
                                                                     ,'      «DAY«
                                                                  ROFF*  ',•  *M!L  M3MO  DIVER*
                                                                  NONPREC*  ','  »KG  P PREC*  ')
20.

27

26.
29.

30.

31

32

33.

34.

35.

36.

37.

38.
39.
40.

41
42.
43.
44 .
4'j.
4G.
47
48.
40.
SO.
51
tj2.



53
b4
65.
«oo .
S7
5U
59
012722B

01 2750B

012750B
012751B

01 27G1B

0127G7B

012775B

013003B

Ol 301 IB

013017B

013025B

01 3O33B

01 304 IB
O13O43B
01 3044B

01 3047B
01 304 VB
013050B
013O51B
01305 IB
01 30t>2B
Ol 3054B
Ol 3O5GB
Ol 3OG1B
01 3OG3B
0130G5B
Ol 3O70B



01 307 IB
O 1 3O72B
Ol 3O74B
Ol 3O76B
Ol 31 03B
Ol 3 1 UUJ
O 1 3 1 1 1 B
WRITE (2,6) (MON(K> , OAY(K> ,LEV(K) , VIN(K),DI V(K)
1VUTCK) . PLD(K) , RLD(K) , K=1 , 12)
6 FORMAT (F7.0, F14.0, F12.0,F12.2, 2F18.2, 2F17.
C COMPUTE LAKE VOLUME At-40 AREA FROM LEVEL
DO 7 I = 1,12
IF (LEV ( 1 ) .GT. 8000. . AND. LEV( I ) . LE . 8950. )ARE( 1 )
1(M -6900. )+1000.
IF (LEV( 1 ) .GT.8900. .AND. LEV( 1 ) .LE. 8925. )VOL( 1 )=
1(1) -8900. )+40.
IF (LEV( 1 ) .GT. 8925. .AND. LEV ( 1 ) . LE. 8950. )VOL( 1 )=
1 ( 1 1-8923. )+67.
IF (LEV ( I ) .GT.6950. . AND. LEV( I ) . LE . 8975. )ARE( 1 )
1(1) - 8950. )+1600.
IF (LEV ( 1 ) .GT. 8950. .AND. LEV ( I ) . LE . 8975. ) VOL
1(1) - 8950. ) +103.
IF (LEV ( 1 ) .GT. 6975. .AND. LEV ( 1 ) . LE . 90OO. ) ARE( 1
1(1) - 8975. ) +2000.
IF (LEV( 1 ) .GT. 6975. .AND. LEV ( 1 ) . LE. 900O. )VOL( I )
1(1) - 8975. )+150.
IF (LEV ( I ) .GT. 9000. . AMD. LEV ( I ) . LE . 9025 . ) ARE( I
1(1) - 9000. )+2700.
IF (LEV ( 1 ) .GT. 9000. . AND. LEV ( I ) . LE . 9025. )VOL( 1 )
1(1) - 900O. )+210.
VOL ( I ) = VOL ( 1 ) * 1 000 . « 1 233 .
ARE( 1 ) = ARE( I )* .4047
7 CONTINUE
C SUM VARIABLES OVER YEAR VOLS IN M3
VUTTd = 0.0
PLDTd =0.0
AR£TO = 0.0
VOLTO = 0.0
VINTd=VTOT
DO 8 1 = 1,12
VUTTO = VUTTO + VUT (I)*1000OOO.
PLDTO = PLDTO + PLD (I) + RLD(I)
ARETO = ARE TO + ARE(I)
VOLTO = VOLTO + VOL ( 1 )
VI NTO = VINTO + DI V( I ) * 1 000000.
8 CONTINUE
C COMPUTE P LOAD PER UNIT AREA IN MG PER M2 PER YR,
C RESIDENCE TIME IN YR,
C VOLUME LOAD I NO IN MG PER M3 PER YR, AMD MEAN DEPTH
PLDTO = K-LOTO/36S.
AREMN = ARETO/ 12.0
VOLMN = VOLTO/ 12.0
ARELD - (PLDTO* 1000. * 1000. *3G5 .) /(AKEMN* 1000O. )
VOLLD = ( PLD1O* 1000. * 1 OOO . *3Gl> . ) /VOLMN
DEPMN = VOLMN/ ( AREMN Jt 1 OOOO. )
RESI D = VOLMN --V INTO
t

1 )


=(580. /50. )*(LEV

(28. 15/25. )*(LEV

( 37./2S)* (LEV

=(400. /25. )*(LEV

( I )-(47. 725. )*(LEV

)=(700. 725. )*(LEV

=(60. 725. )*(LEV

)=( 1050. 725. )*(LEV

=(75.725. )*(LEV

















HYDRAULIC

IN M







bl
             C   COMPUTE THE  EFFECT I VL  SEDIMENTATION  COE1 F 1 Cl EMT , S 1 GMA
0131 I3B           SIGMA = USGS*0 0036' . G3
             C    COMPUTE THE PREDICTED IN-I.AKE  P  CONCENTRATION
             C   US I NO VOLLENWEIDER MODI I  I LI)  INDEX
OlJllbU           PLAKE= (ARELD/dlEFMN* (  1 /KF.blD  'SIGMA)))
             C   COHPIME THE  CHI (li.'OI'HYl.l  (IVI'R   MIL  U 1>M  L.AYER,
             C   STRATI PICA II OI-J  SLASON  111 AH
                     CHLA= 1 . -I-I'J »Al OL'. I O( I I M' h ) - . :)'J13

-------
03.    O13126B         CHLA=10**CHLA
               C  WRITE  OUT RESULTS
C4.    013131B         WRITE (2,1O) AREMN, V8LMN,  VINTO
C5.    013140B     10  FORMAT (/////,2X,'MEAN LAKE AREA OVER THE YEAR IN HECTARES  IS'
                    1F6.0./.2X,
                    2'AVERAGE LAKE VOLUME OVER THE  YEAR IN M3 IS ',
                    3F11.0,/,2X,
                    4'TOTAL ANNUAL INFLOW IN M3  INCLUDING DIVERSION IS'. F11.0)
66.    013140B         WRITE (2,11) VUTTO, PLDTO
67    O13145B     11  FORMAT J/.2X,'TOTAL ANNUAL  OUTFLOW IN M3 IS',F11.0,/,2X,
                    1'TOTAL LOADING IN KG PER DAY IS',  F7.3)
68.    013145B         WRITE (2,12) ARELD, VOLLD,  DEPMN
09.    013153B     12  FORMAT (/,2X,'AREAL LOAD  IN M6 PER M2 PER YR  IS  ',
                    1F6.1,/,2X,'VOLUME LOAD IN MG PER M3 PER YR IS',
                    2F6.1,/,2X,'MEAN DEPTH IN M  IS',
                    3F5.1 )
70.    O13153B         WRITE (2,13)RES ID,PLAKE
71.    013160B     13  FORMAT (/,2X,'TAU  IN YR  IS', F5.2,/,2X,
                    1'PREDICTED  P CONCENTRATION IN TOP 15 M IS',F5.1,//)
72.    013i60B         WRITE (2,14)CHLA
73.    013164B     14  FORMAT (/,2X,'MEAN CHLOROPHYLL A 0-5M DURING STRATIFICATON  IS'
                    1F7.2,//)
               C  WRITE  TO TAPE4
74.    013164B         WRITE (4,60) USGS,PLAKE,CHLA,DEPMN
75.    013173B     GO  FORMAT (F8.O,2F8.2,F8.1)
76.    013173B     107  CONTINUE
77.    013175B         STOP
78.    013176B         END
                                                                                                                           H
                                                                                                                            I

-------
UNIVERSITY OF  MINNESOTA FORTRAN COMPILER  (VERSION  5.4 - 79/O3/O1> ON THE 64OO UNDER  KRONOS 2.1.O ON 63/O4/O1 AT  16.O3
UNIVERSITY COMPUTING CENTER - UNIVERSITY  OF COLORADO

                    MNF,E=4,I=DILEFP,L=LLL.

                          PROGRAM DILEFP (INPUT,OUTPUT.TAPE4,TAPE2,TAPES)
                          REAL PCON
                          DO 61  11=1.1
                       THIS IS THE EFFECTS ELEMENT  OF THE DILLON CLEAN
                       LAKES MODEL.   IT TAKES  INPUT FROM THE TROPHIC
                       STATUS ELEMENT ON TAPE4  AND  WRITES TO TAPE2.
                       THE PURPOSE OF THE  PROGRAM  IS TO CONVERT
                       THE PREDICTED CHANGES IN P CONCENTRATON TO
                       PREDICTED CHANGES IN OTHER VARIABLES OF INTEREST.
                          READ (4,1) USGS,PCON,CHLA,2BAR
                        1  FORMAT (F8.O,2F8.2,F8.1)
                       PREDICT THE MEAN AND MIN SECCHI DEPTH,  JULY-OCTOBER
                          SD=2.4-0.68*ALOG(CHLA)
                          SD=EXP(SD)
                          SDMIN=SD».7O
                          WRITE (2,2) SD.SDMIN
                        2 FORMAT :t 01O4O1I1
D-l OIO.IU.'U
23H
33H
43H
S3H
63H
f. 1 OONT
STOP
KND
3
4
5
G
7
1


t
,
,
(
,
3H
311
3H
3H1
3(11
3,
.1 ,
0,
8,
. 8,
F5
F5
511
5H
5H
.0,
. o,

1
-59
/ , 3H
/ , 3(1
0, /,
3. , / ,
. O, / ,
3
4
311
311
311
NUt














                                                        3, 5H
                                                        4, 5H
1 ,
0,
0,
0,
                                                                  511
                                                                         O,/,
             ,/.
             , / .
             /,
5,3H  5,5H     0,,
0, 3112.2, 3H1 3. , / ,
7, 3112. 2, 5H-59. O)

-------
          UNIVERSITY COMPUTING CENTER
          UNIVERSITY OF COLORADO

          SPSS- - STATISTICAL  PACKAGE FOR THE SOCIAL SCIENCES

          VERSION 8.3 (NOS)  -- MAY 04,  1982

          070000 CM MAXIMUM  FIELD  LENGTH REQUEST
                                                                              63/04/01.
                                                                                           16.03.43.
                                                                                     PAGE
                     RUN NAME
                     VARIABLE  LIST
                     N OF CASES
                     INPUT MEDIUM
                     INPUT FORMAT
                   PLOT
                    VAR001 TO VAR003
                    14
                    TAPES
                    FIXED(F3.0,F3.0,F5.0)
                    ACCORDING  TQ  YOUR INPUT FORMAT, VARIABLES ARE  TO  BE READ AS FOLLOWS

                    VARIABLE   FORMAT  RECORD    COLUMNS
                    VAR001
                    VAROO2
                    VAR003
          F 3.  0
          F 3.  0
          F 5.  0
1 -
4-
7-
 3
 6
11
 THE  INPUT FORMAT PROVIDES FOR    3 VARIABLES.
 IT PROVIDES F8R  1 RECORDS  (*CARDS*>  PER CASE.
 A MAXIMUM OF    11 *COLUHNS* ARE USED ON A RECORD.
                             3 WILL BE  READ.
CPU  TIME REQUIRED..
                            .115  SECONDS
              GIVEN
PLOT                PLOTS=VAR003(-100.100)  WITH VAR002(0,5) BY VAR001(1,7)/
                   TITLE=AS APPROPRIATE/                  :
                 TITLEX=   TOTAL P        CHLA        SECCHI      BOTTOM 02/
                 TITLEY=BETTER,PERCENT                     WORSE,PERCENT/
                 SIZE=25,25/
                   SYMBOLS=0,O,0,0,0,-3,-3/
                  PLOTS = VAR003(-100, 100)WITH VAR002(0,5) BY VAROO1M,?)/
                      TITLE=AS APPROPRIATE/
                 T!TLEX=   TOTAL P      CHI.A     SECCHI
                 TITLEY^BETTER,PERCENT
                 SIZE=9,9/
                  XDIV=10/
                  YD1V=10/
                 SYMBOLS=0,0,0,0,0,-3,-3/
READ INPUT DATA

 3 VARIABLES,  INITIAL CM ALLOWS FOR    271  CASES
               MAXIMUM CM ALLOWS FOR    949  CASES
                                                                               BOTTOM O2/
                                                                               WORSE,PERCENT/
                                                                             H

                                                                             M
                                                                             t-o
- - - WARNING  -  -  -

MO PLOTS WILL  BE PRODUCED UNLESS THE  PLOT FILE (TAPE99) IS DISPOSED.
THE FOLLOWING  JOB  SETUP WILL YIELD  MEDIUM L1WF ON PLAIN PAPER-

-------
    SPSS.

    DISPOSE,TAP£G9,GP=CP3.


EE THE CALCOMP MANUAL FOR  OTHER PAPER/PEN COMBINATIONS.
                                                                                                                                  t -\
                                                                                                                                  I

-------
PLOT                                                                         63/04/01.     16.03.43.     PAGE


             OPT!OH  -  1
             IGNORE  MISSING  VALUE INDICATORS
             (NO MISSING VALUES  DEFI NED, . .OPT I ON 1  MAY HAVE BEEN FORCED)
PLOTS WILL NEED APPROXIMATELY    53 INCHES OF PAPER

-------
PLOT

FILE
N.JNAME
         (CREATION  DATE  =   03/04/01  )
                                                                            83/O4/O1.
                                                                                          16.03.43.
                                                                                                        PAGE
'LOT   1
3N THE X-AXIS.

.IN THE Y-AXIS.
         VAROO2
                VAROO3
                       1
                       2
                       3
                       4
                       5
                       6
                       7
                       VALUES
                       VALUES
                       VALUES
                       VALUES
                       VALUES
                       VALUES
                       VALUES
PLOTTED
PLOTTED
PLOTTED
PLOTTED
PLOTTED
PLOTTED
PLOTTED
2
2
2
2
2
2
2
PI OT   ;.
pi  THE  \-AXIS..

O i THE Y-AXIS. .
          VAROO2
                 VAR003
1
2
3
4
5
6
7
VALUES
VALUES
VALUES
VALUES
VALUES
VALUES
VALUES
PLOTTED
PLOTTED
PLOTTED
PLOTTED
PLOTTED
PLOTTED
PLOTTED
                                                      2
                                                      2
                                                      2
                                                      2
                                                      2
                                                      2
                                                      2

-------
PLOT
                                                                             83/04/01.
16.03.43.
                                                                                                        PASE
CPU TIME REQUIRED..
                           .835  SECONDS
                    FINISH

TOTAL CPU TIME USED..        .965 SECONDS
RUM COMPLETED

NUMBER OF CONTROL CARDS  READ   21
NUMBER OF ERRORS DETECTED       O
STATION NAME

   BR1

   BR2

   BR4

   BRG
   SRI

   SR2

   SR3

   SR4

   TMI

   TM4

   TM7

   MC2

   SC2

   AAA

   BOB

   CCC

   ODD





)R WATERSHED SEGMENTS, KG PER YR
1M RUNOFF AREA HA RES SEW URB SEW
424.0
424.0
231 .0
253.0
235.0
313.0
310.0
500. 0
3G4.0
433.0
406. O
513.0
325.0
77.0
325 . 0
325. 0
325. 0
325. O
25.
1 1003.
9719
6133,
S165.
1041 ,
2678.
9789,
4122.
6329.
1 1422.
6447.
1800.
2212.
1 143.
122G.
619.
33,
,0
, 0
, 0
0
0
,0
, 0
0
0
0
0
0
0
. 0
0
0
0
O
0
, 7
7.3
. 1
.9
7.3
10. 1
0
0
0
0
O
0
. 3
2. 1
4. 4
0
a
0
0
11.2
0
O
0
O
0
0
2.3
2. 5
0
.4
0
1O.O
O
0
5.3

SEPTIC
0
107.6
149.2
148.4
182.8
0
17. 1
51 .2
0
46.6
0
0
35.8
O
1 .9
0
12. 5
0

H I WAY
0
0
0
0
0
0
0
0
0
48.6
69.0
0
0
0
0
0
O
O

SKI SLOPE
O
0
25. 1
0
0
0
1 1 .5
0
5. J
0
19.6
O
0
0
0
0
O
O

PKG PLT
O
7.6
O
0
48.1
2.7
0
0
24.3
0
0
0
0
0
0
0
0
0

TERT PLT
314. 9
0
0
0
0
0
0
0
0
0
117.5
0
0
0
0
151 .8
0
155. 1

AMAX
0
0
0
0
0
0
0
0
0
0
0
259.7
0
0
0
0
0
0
1-1
1
BKGRND
.8
346. 3
133.0
S5. 1
72.3
22.2
54.8
386. 3
105.2
205.0
338. 7
263.5
39.3
6.7
25.0
26.8
13.5
.7

-------
                                         i a i 7 . o
                                                                                                              19.3
                                                                                                                                                                                                 39. 7
GROSS N YIELD  FOR  WATERSHED SEGMENTS,   KG PER YR
iTION NAME
BR1
BR2
BR4
BR6
BR7
SRI
SR2
SR3
SR4
TMJ
TM4
TM7
MC2
SC2
AAA
BOB
CCC
ODD
EEl"
MM RUNOFF
424 . 0
424. O
231 . O
253.0
235.0
319.0
310.0
500. O
364. 0
433. 0
4O6. O
513.0
325. O
//. 0
325. 0
325 0
325. 0
325. 0
325 . 0
AREA HA
25.0
1 1O03.O
9719.0
6133.0
5165.0
1 04 1 . 0
2678. 0
9789. 0
4122. 0
6329. 0
1 1422.0
6447.0
18OO. 0
2212. 0
1 143. O
1226.0
619. 0
33. 0
1817.0
RES SEW
O
36.3
403.6
3.3
52.0
377.7
524.6
0
O
0
O
0
0
23. 7
107. 0
227. 5
0
O
O
URB SEW
0
O
933. 5
0
O
0
0
0
0
165. 9
184.7
0
27. 9
0
775. 4
0
0
408. 5
0
SEPTIC
O
5952.6
7129.0
7248. 1
8771 .7
0
878.6
2945. 3
O
2588. 9
0
0
1856.6
O
100. 7
0
647 7
0
618. 9
HI WAY
0
0
0
0
0
0
0
0
0
664.7
1015.6
0
0
0
0
0
0
0
330. 1
SKI SLOPE
0
O
162. 1
0
0
0
74. 3
0
32.8
0
126. 8
O
O
0
0
0
0
0
O
PKG PLT
0
33.5
O
0
213.1
12.0
0
O
107. 7
0
0
0
O
0
0
0
O
0
0
TERT PLT
8019. 9
0
0
0
0
0
0
0
0
0
2992. 5
0
0
0
0
3066. 3
0
3950. 1
0
AMAX
0
0
0
0
0
0
0
0
0
0
0
52814. 5
O
0
O
0
0
0
0
BKORNO
22.7
9972. 5
4370.6
3063. 3
2369. 2
679.4
1691 .O
10731 . 5
3132. 8
5877. 0
9846. 8
720O 2
1200. 3
200. 0
762. 2
8)7.6
412.8
22. 0
1211 7
OKOb:> P Alii")  N  YIMD  |4Y  .MMIIIIT   KG  ITK  Yl?
NAfIL"   P10IAI-    (IKJTAL

           •-11 r,       .in i <

-------
BR2
BFM
BR6
BR7
SRI
SR2
SR3
SR4
TH1
TM4
TM7
MC2
SC2
AAA
DBB
CCC
ODD
EEE
-162.
32G .
244.
304 .
32.
94.
437
135.
302.
547
523.
75.
7.
39.
183.
26.
161 .
71 .
15995.
12999.
10315.
1 1406.
1069.
3169.
13677.
3273.
9316.
14166.
60O95.
3005.
304.
1745.
491 1
1 OCO .
4381 .
2161
JRflSS KG PER YR P  IN RUNOFF IS,  4435.
JROSS KG PER YR N  IN FHJN5FF IS,  185155.


>RCSS YIELD BY SOURCE KG  PER YR f(ABOVE),N(BELOW)


:eSIO.  SEWER  URB SEW   SEPTIC    HI WAY     SKI SLOPE  PKG PLT    TERT  PLT  AMAX      BKGRND   PRECIP     GRNDWAT  DIVERSION  CONSTRUCT

     33.2      31.G     765.0     136.9       61.3       82.6      739.3     259.7    2175.1      726.0       10.0         0      150.0

   1755.9    2496.0   38738.0    2030.3     396.1      366.3    18828.0   52814.5   63743.4     9872.0     1068.O         0     3985.7





                                                                                                                               IH
                                                                                                                                I
                                                                                                                               M
                                                                                                                               00

-------
VALUES UP TO THIS POINT UNCORRECTEO FOR
  INVENTORY CHANGE IN  RIVER BOTTOMS, NOW  CORRECT THIS.
ET LOADING  BY SOURCE KG PER YR P(ABOVE>,N(BELOW>

•.SID. SEWER  URB SEW   SEPTIC    HI WAY      SKI SLOPE  PKG PLT
    32.2       30.6     740.5     132.5       59.3      80.O
  19O1.1     2702.4   41941.6    2198.2      428.8     396.6
                                                     TERT PLT   AMAX
                                                     735.5      251.4
                                                   19076.3   57182.3
                                                                     BKGRND   PREC1P      GRNDWAT
                                                                    2105.5     726.0       10.0
                                                                   69015.0    9872.0     1068.0
                                                                                   DIVERSION CONSTRUCT
                                                                                        0      145.2
                                                                                        0    4315.3
IE NET YIELDS BY PERCENT ARE P(ABOVE),N(BELOW)
.SID.  SEWER
      .6
      . 9
URB SEW
    .6
  1 .3
     SEPTIC
      14.7
      20. O
H1 WAY
  2.6
  1 .0
SKI SLOPE
  1 .2
   .2
PKG PLT    TERT PLT  AMAX       BKGRND   PRECIP     GRNDWAT
  1.6       14.6       5.0       41.7      14.4         .2
   .2        9.1      27.2       32.8       4.7         .5
                       DIVERSION CONSTRUCT
                            0       2.9
                            0       2. 1
THE  NET  P YIELD TO  THE LAKE IN KG PER  YEAR IS, 5049.
THE  NET  N YIELD TO  THE LAKE IN KG PER  YEAR IS, 210098.
 •MONTH*
    1 .
    2.
    3.
    4 .
    5.
    6.
    7
    8.
    9.
    1 O.
    1 1
    12.
  »DAY*
    15.
    45.
    74.
   105.
   1 35.
   1GG.
   190.
   ?£7
   25B.
   200
   319.
   3-1U.
      *LEVEL FT*   *MIL  M3MO ROFF*  »MIL M3MO DIVER*  *MIL M3MO  OUT*   *KG P NONPREC*   *KG P PREC*
         8981 .
         8977.
         8974.
         8974.
         8980.
         9OOO.
         9018.
         901 6.
         9010.
         9017
         9O16.
         9O15.
       5.35
       4.42
       5.76
       8.30
      34. 29
      92.61
      56. 88
      28.06
      15. 28
      10.50
       6. 80
       6. 76
                  0
                  0
                  0
                  0
                  0
                  0
                  0
                  0
                  0
                  0
                  0
                  0
                      12.54
                      11 13
                       9.07
                       4.68
                       3.82
                       6.15
                      30. 38
                      26.53
                      15.35
                      10. 17
                      13. 12
                       5. 79
  84.2
  69.6
  90.6
 130. 5
 538.9
1455.7
 893. 9
 441 . 0
 240.2
 165. 1
 107.0
 106. 3
 45. 7
 45.7
 45. 7
 45. 7
 45. 7
217.8
 45.7
 45. 7
 45.
 45.
 45.
7
7
7
                                                                                                                   45.7
 MEAN  LAKL AREA OVL l< THE YEAR  IN HECTARES IS  1130.
 AVCKAOL"  LAKE VOI.U1L OVER  THE  YEAR IN M3 IS   2623659GO.
 TOTAL ANNUAL UiriOW III M3 INCLUDING DIVERSION  IS  l'750-IOOOO.
                                                                                                                                M
                                                                                                                                I
 TOTAL  ANNUAL OUHIOU  IN M3  IS 148730OOO.
 TOTAl  lOADIUii IN M) I'EK DAY  IS 13.819
 AKLAI  LOAD IN flu
 701 Ill-It I Oft II  IN 1-K
t'LK
 I-LU
           2 Pl-R YR l
           in PER YR
   44G. 2
    19.2

-------
  tAN DEPTH  IN H  IS  23.2


  AU III YR  IS   .95

  REDICTFD   P CONCENTRATION  IN TOP 15 H IS  7.9
 MEAN CHLOROPHYLL A  0-5M  DURING STRATIFICATON IS   7.97
MEAN SECCHI DEPTH  IN  H  JULY-OCT IS,  2.69


MIN SECCHI DEPTH  IN M JULY-OCT IS,  1.38




CARLSON TROPHIC STATE INDEX IS,   21.




THE AREAL HYPOLIM  O2  DEFICIENCY IS,  631


THE SUMMER DEPLETION  IN PPM IS,   A. 4


THE MINIMUM 02 5M  OVER  THE BOTTOM IN PPM IS,  4.6
                                                                                                                              I
                                                                                                                              i-o
                                                                                                                              O

-------
DATE .

JOB
ACCT
1 983/CM/O1

18GG9 F253CCV
                             REEL.   140652    START..1641:30    STOP..1641:50
                                       PAGES.  .OO0022    STATUS.  .8000800O
END-OF-REPORT
ENO-OF-REPORT
END-OF-REPORT
END-OF-REPORT
END-OF-REPORT
END-OK-REPORT
ENO-OF-REPORT
ENO-OF-REPORf
END-OF-REPORT
ENO-OF-REPORT
END-OF-REPORT
END-OF REPORT
ENO-OF-REPORT
END-GF-REPORT
END-OF-liEfJOF
-------