Clean Lake Study of Dillon Reservoir in Summit County, Colorado


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

List of Tables . ii

List of Figures vi

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

Design of the Study . 17

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

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

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

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

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

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

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

rt
[J
ru
I '
n>
SILVERTHORNE

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

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

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

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

Total Soluble 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

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

Spatial Survey Series

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

(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

(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

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

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

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.

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

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

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_ 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,

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

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

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

(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

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

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"

o
MI
o
Dillon Bay
Contour interval, 40ft
Tenmile

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

75 150 225 300
0

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

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

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

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.

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

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

£
0)
25
O
-J
60
30
O

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

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.

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

0.258
0.259
Measured
Acre-f ae

145,912
152,565

119.323
131,989
Outflow
•5
r j — '-'

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
**

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

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

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

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

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

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:

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

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

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

0
!C8
mix
stratified
mix
water,

soluble materials
particuiates

' I ! L
J L
JFMAMJJAS

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

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

168
152
7 .45
7.30
33.9
34.5
Decth, m
2C-25

167
153
1 . 36
7.59
34.4
TJi 1

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

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

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

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

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

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

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>

(X
n>
•d
rt
Q.
H-
P)
OP
i-<

-------
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

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

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

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

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

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

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_

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

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

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

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

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

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

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

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

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

rt
(X>
(M O
fl) t""tl
(-•
rt
("t D"
Cii fD

H- hj
a o
IX 1-1
H- O

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

N D
-------
00
c
CM
ft)
H-1
o'
C
a
Cu
o
(D
H-
p
H-

UJ
0
10
20
x 30
40
50 -
Astefionella, Lyngbya

Microflagellates

(Monoraphidium, Scenedesmus)
TTTT"
JULY PEAK
Synedra
(Lyngbya, Microflagellates)
Synechococcus

(Stephonodiscus,

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

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

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

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

0.73
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,

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

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

(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 -

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

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

Dillon 1931 402

Shadow Mountain 1963 459

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

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

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

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,

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

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

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

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

3.8
11.1
20.5
39.0
25.4
2.6
14.5
Dominant
Limitation

P
P
P
N
N
P/N
P
Ratio
P+N/C

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,
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

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

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

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

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

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

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

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

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

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
*

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
-------

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

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

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

0)
o»
a
0
t/>
5

-------
21
98!
JFMAMJJASO
1982
DlJFMAMJJASOND

0
-s_
.
— ,
1

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.

X.
o>
E
tf>
.£
"o
3
O
"•£
o
CL

— 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

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

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

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
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

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

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

* •
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

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

403
423
826
3366
4336
8^41
168,119
Thousands
Acre-ft/yr

32.54
40.45
47.02
120.0

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

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

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

.02
.37
.15
.55

.04
.04
.08
.005

2.86
8.49
9.20
20.55

.48
1.12
1.59
.18

1.36
.23
23.98
Nl trogen
»v»

.33
6.99
1.1 3
8.45

.84
.56
1.40
.08

1 .81
.03
11 .79
- kg/ha /y
Total
Soluble
N

5.49
20.61
15.49
41.59

2.69
2.15
4.84
.79

9.56
.44
57 . 54
r - 1981
Part.
N

.99
1.05
2.86
4.90

.05
.10
.21
.19

1.17*
.000
6.56

6.49
21.65
18.36
46.50

2.74
2.31
5.05
.99

10.72
.44
64.11

10.1
33.8
28.6
72.5

4.3
3.6
7.9
1.5

.09
.24
.86
1.17

.04
.09
.13
.02

.07
.01
1.41
t>
MVN

7.16
7.98
13.81
28.95

.42
.78
1.20
.32

1.140
.58
32.39
li trogen -
Nil, -N
4

.73
7.86
7.32
15.91

.98
.67
1.66
.16

.85
.06
18.71
- kg/ha/yr
Total
Soluble
N

14.33
28.06
37.05
79.44

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

.05*
.16*
.21*
1.30

.95*
.000
21.91

20.88
33.42
43.98
98.28

2.48
3.05
5.53
2.&9

4.07
1.16
113.39

18.4
29.5
38.8
06.7

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

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

DILLON VALLEY
STRAIGHT CREEK

PORCUPINE GULCH
WILDERNEST
(W2A %8
/ ~"—---_
WEST TENMILE (W5B)
KEYSTONE GULCH

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)

1982 541
1.55 0.34 17

3.52 0.43 32
75

35
Laskey Gulch (W7B)
1981
1982
Keystone Gulch (W6B)
1981
1982
Upper Snake (SR3)
1981
1982

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

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

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

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

34
1146
West Tenmile
(W5B)
1981
522
0.78

5.38
1.76
47
3.80
1.57*
200

66*
8426
1982
662
0.78

7.19
0.67
50
5.51
1.67*
213

-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.

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

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

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

0.17
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

5.84
0.64
58
4.09
1.75
103

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.

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.

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

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

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

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

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

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

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

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

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)

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

38
Climax 7
55
Package Plants and

16
3
81
Septic)

28
72
load, %
1982 Mean

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

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

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

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

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

-*- 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)

»- Minimum oxygen

*•• Carlson index

»- Plot % chango
-------
H-
OQ
Ln
CO
tn
(T)
OQ
0
c
en
(T>
O.

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.

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

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

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

2459
2618
47618
3190
2021

2596
Percent
WWTP*

17.4
16.2
1.7
12.8
22.3

30.0
28.3
1.6
23.2
36.6

28.5
Percent
Septic

8.2
14.7
1.5
31.9
0

6.6
12.0
0.7
27.2
0

16.7
Percent
Diversion

0
Percent
Background**
«
60.5
56.1
6.7
44.4
77.5

51.7
48.5
2.6
39.9
62.9

49.0
Percent
Other

13.9
13.0
0.2
10.9
0.2

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

6.1
6.4
74.
7.9
5.0

6.4
Chlorophyll
ug/1

7.1
8.0
141.
11.2
5.0

5.4
6.0
206.
7.9
4.1

5.9
Mean
Secchi , m

2.9
2.7
0.38
2.1
3.7

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

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

4.3
4.2
0.0
3.6
4.8

Mesotrophic
Mesotrophic
Eutrophic
Eutrophic
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

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

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

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

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

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

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 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

37
0
:24
o»
'21
O

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

10 0>
n>
H rt
< (D
ft) >-t
0- rt
H-
n.
c
M
H-

00
cr H
c a1
rt fD
p-

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

MT)
o
to o
o. H,
H-
a a
OP c
a
O D'
MI ro
to o
?i* Hi
m

M Co
M O
O 3
3 en

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

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-

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

-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.

Bailey-Watts, A.E. 1978. A nine year study of the phytoplankton of the

eutrophic and non-stratifying Loch Leven (Kinross, Scotland). J. Ecol.

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

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:

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:

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 -

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:

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:

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.

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

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:

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:

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:

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.

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

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:

Lund, J.W.G. 1970. Primary Production. Froc. Soc. Wat. Treat. Ex an. 19:

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:

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).

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.

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:

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:

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

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.

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)

-------
62.
63.
64 .

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.

10.
1 1 .
12.
13.
15.
1C.
17.
10.

19.
20.
21 .
22.
23.

24.
OOOOOOB
01241 IB

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.

41
42.
43.
44 .
4'j.
4G.
47
48.
40.
SO.
51
tj2.

53
b4
65.
«oo .
S7
5U
59
012722B

012750B
012751B

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

=(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

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

)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
-------