Ecological Research Series
NUTRIENT DIVERSION: Resulting Lake Trophic
State and Phosphorus Dynamics
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal
species, and materials, Problems are assessed for their long- and short-term
influences. Investigations include formation, transport, and pathway studies to
determine the fate of pollutants and their effects. This work provides the technical
basis for setting standards to minimize undesirable changes in living organisms
in the aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-77-003
January 1977
NUTRIENT DIVERSION: RESULTING LAKE TROPHIC
STATE AND PHOSPHORUS DYNAMICS
by
Eugene B. Welch
University of Washington
Seattle, Washington 98195
Research Grant R 800512
Project Officer
Kenneth W. Malueg
Corvallis Environmental Research Laboratory
Corvallis, Oregon 97330
CORVALLIS ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
-------
DISCLAIMER
This report has been reviewed by the Corvallis Environmental
Research Laboratory, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
FOREWORD
Effective regulatory and enforcement actions by the Environmental
Protection Agency would be virtually impossible without sound scientific
data on pollutants and their impact on environmental stability and human
health. Responsibility for building this data base has been assigned to
EPA's Office of Research and Development and its 15 major field installa-
tions, one of which is the Corvallis Environmental Research Laboratory
(CERL).
The primary mission of the Corvallis Laboratory is research on the effects
of environmental pollutants on terrestrial, freshwater, and marine eco-
systems; the behavior, effects and control of pollutants in lake systems;
and the development of predictive models on the movement of pollutants in
the biosphere.
This report describes the effects of wastewater diversion to a mesotrophic
lake in Washington.
A.F. Bartsch
Director, CERL
iii
-------
ABSTRACT
Lake Sammamish, Washington, was studied during 1970-75 to determine
the response to wastewater diversion in 1968. The results were compared
with a pre-diversion study by Seattle METRO in 1964-65. Diversion re-
duced the phosphorus loading by about one third - from 1.02 to 0.67 g
P/m2.yr and about 119 to 68 yg/1 in the inflow. Winter total P remained
constant at about 30 +_ 2 jjg/1 and no trend was shown in chlorophyll a_ in
spring-summer with a year-to-year average of about 7 +_ 2 vg/1. Water trans-
parency remained the same - a summer mean of 3.3 in Secchi disk depth.
Paleolimnological evidence suggests that the lake has probably been
near its present mesotrophic state for possibly 100 years with some
alteration during the late 19th and early 20th century logging activity
in its watershed. A thorough survey in 1913 showed trophic state in-
dices at similar levels as today. The lake has a history of relative
stability in trophic state even though P loading was decreased by about
one third and also increased by at least that much.
The lake's biological state did not respond to the P loading change
largely because the water column P content did not change. Its stability
is in turn thought to be controlled by the anaerobic-aerobic release
and sedimentation of iron and its complexes. Plastic column experiments
In situ showed rates of anaerobic release to range from 3-5 mg P/m2-day
and measured sedimentation rates to exceed the release rates, with the
greatest sedimentation occurring after turnover and the reestablishment
of an aerobic water column, although as much as three-fourths of the
water column P content could have come from sediments by the end of
summer stagnation it was largely unavailable for spring phytoplankton
because of subsequent sedimentation with iron.
The increase or excess in P loading during the greatest urbaniza-
tion impact in the 1960's may well have gone to the sediment as was
suggested by mass balance considerations if P sedimentation varies with
P loading. Also the introduction of waste P into the major inflow
stream 2 km above the mouth, rather than directly into the lake, argues
for its sedimentation upon entering the lake. Without diversion, how-
ever, loading would now be at 1.36 g P/m2.day and probably some increase
in water column P content in winter would be apparent.
Seasonal dynamic changes in P and chl a_ during spring were studied
with a mathematical model that simulated those and two additional state
variables. The model provided an interesting procedure for studying
the mechanisms controlling the timing and maximum for the s-pring diatom
bloom. The model should be extended to study the dynamics of P and Fe
in the fall-winter period.
IV
-------
CONTENTS
Foreword iii
Abstract iv
List of Figures vi
List of Tables ix
Acknowledgments x
Sections
1. Conclusions 1
2. Recommendations 3
3. Introduction 4
4. Geographic Description of Lake Sammamish 7
5. Morphometric and Hydro!ogic Description of Lake Sammamish 15
6. Limnological Characterization 17
7. Methods and Materials 20
8. Results 32
9. Discussion 79
10. References 88
-------
LIST OF FIGURES
FIGURES PAGE
1. Land Use Map of the Lake Sammamish Watershed, 11
2. Vertical distribution of percent water, percent
organic matter, and phosphorus as a fraction of
the dry weight and organic matter of the sediment.
Open circles represent values from Lake Washington
cores (Shapiro ejtal_., 1971). 12
3. Location of sampling stations and bathymetric 2_
map for Lake Sammamish.
4. State variables, transfer processes, and functional
equations in the phosphorus model for Lake Sammamish
See Tables 8 and 9 for further explanation of process
functions.
5. Mean concentrations in the photic zone (usually top
8 m) of growing season chl a^ (Mar-Aug) and Winter
(Dec-Feb) total phosphorus and nitrate nitrogen
relative to pre-diversion 1965 levels. The 1965
levels were: chl a_ 6.5 yg/£ (actually a mean of 1964
and 1965 data), total P 31 ug/£ and N03-N 390 yg/£
The % blue green algae of the total phytoplankton
volume was compared against the pre-diversion mean _ft
for June-October in 1965 and July-Oct in 1964 (67.5%) -30
6. Total and ortho phosphorus concentration (mean of .,
epilimnion) at Lake Sammamish central station in 1974.
7. Oxygen isopleths in Lake Sammamish for 1973.
8. Dissolved oxygen, total phosphorus, and total iron
concentrations in Lake Sammamish for the 1972 turnover .
period (from Rock, 1974).
9. Dissolved oxygen, total phosphorus, and total iron
concentrations in Lake Sammamish for the 1973 turnover ..
period (from Rock, 1974).
10. Water column phosphorus content for turnover period
(After Krull, 1973).
11. Temporal variation in chlorophyll a (integrated means for
photic zone) in Lake Sammamish, 1970 through 1973 Aft
(1970-71 data from Emery 1972).
12. Mean seasonal variation in Secchi disk readings for
Lake Sammamish. Dashed line represents mean for the Rn
plotted data (1970-71 data from Emery (1972)). ou
vi
-------
LIST OF FIGURES - (Cont,)
FIGURES PAGE
13. Total Zooplankton in Lake Sammamish, 1965 vs 1972. 52
14. Percent Composition of Zooplankton Categories, 1972. 54
15. Phosphorus sedimentation rates in Lake Sammamish
determined at Stations 612 (28 m depth, 26 m Trap)
and 2A (30 m depth, 27 m Trap). 57
16. Total sedimentation rates in Lake Sammamish at Station
612 (28 m depth, 25 m Trap) and Station 2A (30 m depth,
27 m Trap). 58
17. Changes in Total Iron and Total Phosphorus in an Opaque
Column from 8 August to 16 December 1974 (The column 6-|
was partially reoxygenated by nitrogen aeration).
18. Changes in Total Iron and Phosphorus for 'an Opaque
Column from 27 July to 8 January 1975. (KHLPOy, - fi?
0.236 mg/1 was added on day 120). ^
19. Changes in Total Iron and Phosphorus in a Transparent ^
Column from 17 July 1974 to 8 January 1975.
20. Changes in Total and Soluble Phosphorus in the Lake
Hypolimnion, Station 6.2, from 29 May to 27 November
1974. (Day 140 illustrates the commencement of the 64
annual fall overturn).
21. Release of Total and Soluble Iron in the Lake Hypo-
limnion, Station 612, from 29 May to 27 November 1975. 65
22. Simulated and Observed Temperature Profiles for Lake 68
Sammamish (1970).
22-b. Simulated and Observed Temperature Profiles for Lake
Sammamish (1971). 69
22-c. Simulated and Observed Temperature Profiles for Lake 7Q
Sammamish (1972).
22-d. Simulated and Observed Temperature Profiles for Lake 71
Sammamish (1973).
23. Calculated daily mixing depth in Lake Sammamish during 7o
March-July, 1972-72 (from Tang, 1975).
24. Simulated and observed Chi a_ content in the epilimnion
of Lake Sammamish during the springs of 1970-73. 73
VII
-------
LIST OF FIGURES - (Cont.)
FIGURES
25. Simulated and observed ortho-P concentrations in
the epilimnion of Lake Sammamish during the springs
of 1970-73.
26. Simulated and observed chl a_ content in the epi-
limnion of Lake Sammamish during the spring of 1972.
Note relatively better simulation of the spring
increase in chl ^ with the modified growth model.
27. Phosphorus loading graph comparing the relative.
position of three manipulated lakes. Relationship
after Vollenweider (1974).
PAGE
75
78
86
vi ii
-------
LIST OF TABLES
TABLE PAGE
1. Summary of Basin Geography. 7
2. Historical Evens of Interest in the Lake Sammamish
Bain (After Fish, 1967). 9
3. Diatom Ratios for Lake Sammamish Core 08. 14
4. Summary of Pertinent Hydrologic and Morphometric -ic
Characteristics for Lake Sammamish.
5. Hypolimnetic (depth greater than 15 m) Oxygen Content
and "Actual" Oxygen Deficit Rate. 18
6. Trace Metal Survey on October 24, 1971 for Lake
Sammamish. 19
7. Summary of Routinely Monitored Physical, Biological 2s
and Chemical Parameters.
8. Mathematical Expressions of Involved Subprocesses
in the Phosphorus Cycling Model. 30
9. Notations and Units Used Along With Values for
Constants for the Phosphorus Cycling Model. 31
10. Phosphorus Loading (Kg/Yr) to Lake Sammamish before
and after Wastewater Diversion. P Estimates Are
Based on a Normal Water Year. 35
11. Nitrogen Loading to Lake Sammamish in Kg/Yr;
11 tributaries plus Issaquah Creek. 36
12. Summary of Present Limnological Characteristics in
the Photic Zone (except oxygen deficit). 39
13. Annual and Growing Season Means of Phytoplankton
Chlorophyll a. (weighted means for the euphotic zone)
and Daily Rate of Primary Productivity in 1970-1974. 49
14. Comparison ofJUigust Net Hauls 1913 vs 1972
(Number per m ). 51
15. Zooplankton Species in Lake Sammamish. 53
16. A Comparison of P Release Rates From in situ Column
Experiments and Observed Hypolimnetic Changes in
Concentration During 1972-75. 66
ix
-------
ACKNOWLEDGMENTS
This report is a compilation of the thinking, work, discussions
and writing of faculty and a large number of students in the Civil
Engineering Department, University of Washington. Although there is
not room to list them as authors, while in fact many of them really are,
they should be given special mention here. Special mention should go to
Dr. D. E. Spyridakis and Dr. B. W. Mar for their active participation in
the direction of the sediment and modelling effort, respectively. The
students who were active in research in Lake Sammamish and who completed
theses contributing to this report are in chronological order: Dr. R. M.
Emery, Mr. C. E. Moon, Mr. M. Morton, Mr. S. Lanish, Mr. J. Krull,
Mr. F. W. Monahan, Dr. C. A. Rock, Dr. 6. Pederson, Ms. S. Guttormsen,
Dr. C. H. Tang, Mr. J. C. McDonnell, Ms. C. M. Noah, Dr. P. Birch,
Mr. R. L. Barnes and Mr. J. Mock. Special thanks is also due laboratory
technicians Ms. S. Hamel and Mr. R. Tarn.
Partial support for this work was also provided by the IBP-Coniferous
Forest Biome, National Science Foundation Grant No. GB-20963. This is con-
tribution no. 263 from the Coniferous Forest Biome.
-------
SECTION 1
CONCLUSIONS
1. About 7,000 kg/yr of phosphorus was diverted from Lake Sammamish
with waste water in 1968. This amounted to an aerial loading reduc-
2
tion of about one-third - from 1.02 to 0.67 g P/m -yr or a reduction
in mean flow weighted inflow concentration from 119 to 68 ug/1.
2. Continuous monitoring of the major inflow stream showed that daily
observations were necessary to avoid missing 25-30% of the annual
input of P. High concentrations of P (200-300 ug/1) and high flows
occurred together on four separate days in 1973-74 and four in 1974-
75 which amounted to such a high fraction.
3. The reduction in loading had no measurable impact on lake water con-
centration of total P and chlorophyll a_t water transparency or
hypolimnetic oxygen deficit rate. The blue green fraction of the
phytoplankton did seem to have declined significantly - by over 40%.
4. Although most trophic state indicators did not change, the lake
nevertheless remains mesotrophic with a mean summer transparency
(Secchi disk depth) of 3.3 m and a maximum of near 6 m. This has
been the state of the lake since early in this century and possibly
since the settlement in the watershed of European man although log-
ging had an impact in the late 19th and early 20th centuries.
5. Maintenance of lake trophic stability is most probably due in large
part to the constancy of the winter P concentration, which in turn
is controlled by rather high anaerobic release rates from the sedi-
ment (3-4 mg P/m -day) in summer, but more importantly high sedi-
mentation rates following autumn turnover of usually about 4 mg P.m
2
day, but initially up to 14 mg P/m -day. Much of the P fluctuation
was correlated with iron, which is no doubt controlling the P cycle
in the lake - anaerobic release and aerobic sedimentation. Much of
the higher pre-diversion P load evidently did not occur in the water
column, was thus unavailable to algae, and only served to increase
the sedimentation of P through the above and other processes.
1
-------
Although the effect of a one-third reduction, as well as a similar
increase, in P load to Lake Sammamish was not apparent, the diver-
sion surely had a preventative value. Without diversion loading
2
would now be 1.36 g P/m *yr and could well have exceeded the
stability mechanism(s) for water-column P.
A mathematical model with four state variables simulated P and
chl ^concentration reasonably well, but the model served primarily
to study seasonal dynamics of the system. Although the work had
only begun it has much promise for better understanding of the
lake's cycling processes.
-------
SECTION 2
RECOMMENDATIONS
Work should continue on the seasonal dynamics phase and associated
modeling activity. With more continuously monitored key variables
the carefully measured rates of P interchange with the sediments
could be effectively utilized in the modeling process to determine
the causative processes involved in maintaining such a steady lake
water concentration of P in winter.
Effort should also include the nearshore area where changes in
attached algal biomass should first appear as a result of the
increasing urban development and its associated greater storm water
input to the lake.
-------
SECTION 3
INTRODUCTION
Since the turn of the century, cultural eutrophication of the
Nation's waters has become a problem of ever-increasing magnitude. The
rapid, sprawling growth of our communities has perturbed surrounding
watersheds, augmenting the point and non-point sources of nutrient sup-
ply to many lakes. These artificial sources have stimulated primary
production, often culminating in massive, foul-smelling plankton blooms
that cloud the water. Typically, such events arouse adverse public
reaction, with a resultant out-cry for corrective action.
The strategies and schemes for remedial action run the gamut from
nutrient diversion to sealing the lake bottom with plastic sheeting.
Perhaps one of the most frequently applied solutions has been to divert
incoming nutrients away from an affected lake. The results can be very
spectacular, as in the Lake Washington case (Edmondson, 1970, 1972),
but also extremely expensive. The cost of the Lake Washington sewage
diversion project exceeded $125 million (Gibbs, et al., 1972).
Subsequent to the Lake Washington project, the Municipality of
Metropolitan Seattle (METRO) initiated a similar program for Lake
Sammamish, which seemed to show early signs of movement to a eutrophic
state (Isaac, et al., 1966). The secondary effluent from the city of
Issaquah, Washington, and waste from a milk processing plant were
diverted in September of 1968 at a cost of $3 million. Located only
3 kilometers east of Lake Washington, Lake Sammamish was expected to
display a rapid improvement like its sister lake. To determine if the
lake would respond to nutrient diversion and the extent and rate of such a
response, monitoring of the lake has been carried out from 1969 to 1975.
The Pre-diversion data on the lake consists of a two-year study by
METRO (Isaac, et al., 1966) and a one-day survey by Kemmerer, et al.
(1923) in 1913.
-------
In addition to continuous monitoring of limnological characteris-
tics, special studies of secondary production (zooplankton and fish),
nutrient exchange rates between sediment and water, phytoplankton up-
take of nutrients, feeding rates of zooplankton, profundal bottom
fauna, and dynamic modelling of the phosphorus and nitrogen cycles have
been conducted, as well as a careful evaluation of the nutrient (partic-
ularly P) income (Birch, 1974; Emery, 1972; Hendrey, 1973; Horton,
1972; Monahan, 1974; Moon, 1973; Pederson, 1974; Sturtevant, 1974;
Rock, 1974; Welch, et al., 1973; and Welch, et al., 1975).
Most of this effort has been for the purpose of estimating the
change in the nutrient income, and defining observed changes in trophic
state indicators and the processes that have permitted the lake to re-
main mesotrophic in spite of alteration of the P loading through the
diversion project. In fact, core analyses have shown that several in-
dicators of trophy have remained unchanged for over 100 years.
The diversion has subsequently been shown to have amounted to one-
third of the lake's P loading. Apparently the lake's internal sediment-
water interchange mechanism controlled by iron has resisted P loading
2
changes over a range of at least 0.7 - 1.1 g P/m -yr. and allows the
available water column P content to remain remarkably stable. This may
well be the principal cause for the lake's lack of response to diver-
sion. However, stability may result for other reasons and could not be
expected to persist over a much greater range in loading and when viewed
over the range of trophic states and loading that exist in the world's
lakes, the range examined in Sammamish appears rather small.
For this project, the specific goals were to determine changes in
water column concentrations of nutrients, phytoplankton chlorophyll a_
and species composition, and transparency following diversion of
nutrients and determine, through measurement of external loading, the
resultant fractional decrease in nutrient income. Further, the sedi-
ments were to be analyzed for N, P and C content and exchange rates be"
tween sediment and overlaying water determined in order to evaluate the relative
-------
importance of the sediments as an internal source of P. Lastly, the
understanding of the lake's behavior in response to changing nutrient
input was to be extended by incorporating knowledge of the principal
process rates into a mathematical model. Models of the simple input-
output, steady state type described by Vollenweider (1969) and modified
by Lorenzen (1973), as well as the multiparametric type (Tang, 1975),
have been used. The steady state model application has verified the
potential stability of the lake P content in response to changing P
input while the multiparametric model has elucidated a more dynamic
accounting of the P (and N) supply during the spring phytoplankton
outburst.
A principal advantage in defining the response of Lake Sammamish
to nutrient diversion that was cited earlier, is its proximity and com-
parability to that of Lake Washington. Although many characteristics
such as geologic age, water quality type, meteorologic conditions and
flushing rate are relatively similar, some are distinctly different.
Lake Washington is essentially twice the depth as Sammamish, does not
completely lose its hypolimnetic oxygen, while Sammamish does, and
apparently is more subject to wind mixing. If not for those differing
factors, the response of the two lakes to two fractions of diverted P
could be considered with more validity. The diversions removed nearly
three-fourths of the entering P from Washington, while only about one-
third was removed from Sammamish. Whether the principal reason for no
response in Sammamish is the control on water column P by the sediments
or the smaller fraction of P diverted can be debated. The input-output
model suggests that it is the former, while Vollenweider and Dillon
(1974) indicate that >50% diversion of P should be necessary for im-
provement. To be sure,the year-to-year variation in the P input to
Sammamish has been about equal to the fraction diverted. Probably a
combination of the two is involved as will be emphasized in the report.
-------
SECTION 4
GEOGRAPHIC DESCRIPTION OF LAKE SAMMAMISH
The waning of the Wisconsin glaciation (14,000 BP) left the Puget
Sound lowlands dominated by striated hills, rolling uplands, and deeply
cut troughs. Today one trough is occupied by Lake Sarranamish, a second
by Lake Washington with the meandering Sammamish River connecting the
two. A mild, maritime climate now prevails, annually producing 90
centimeters of precipitation and a mean monthly temperature of 11.5°C
(52.7°F). Direct sunshine is present 45 percent of the daylight hours.
The lake is monomictic with stratification beginning in May and building
to a maximum in August. The thermocline is completely eroded by late
November and the lake remains homothermal until the following May.
Table 1 provides a summary of the pertinent geographic conditions.
Table 1.
SUMMARY OF BASIN GEOGRAPHY
Parameter
Lake Sammamish
Location
Altitude (meters above mean sea level)
Longitude
Latitude
p
Size of Drainage Basin (km )
Duration of ice cover
Evapotranspiration (cm)
Evaporation (cm)
Precipitation (cm)
Maximum monthly precipitation (cm)
Minimum monthly precipitation (cm)
12
122°05'W
47°36'N
253
none
23.7
5.1
90
39
0
-------
The predominant surface stratum of the drainage basin is a light,
gray till. This till is a hard unsorted mixture about 46 meters thick,
consisting of clay, sand, silt, and gravel. Although the till is rela-
tively impermeable, thin beds of sand and gravel commonly yield small
quantities of perched water. Aquifers transect the basin, with several
artesian wells Surfacing within the basin (Liesch, et al., 1963). Coal
seams are located in the southern half of the watershed, while high
quality sand and gravel, refractory grade clay, quarry basalt and cin-
nebar deposits are scattered throughout the basin (Livingston, 1971).
A geologic cross-section cutting through Issaquah in an east-west
direction shows base rock consisting of marine sedimentary rocks on the
west side of the Lake Sammamish valley. On the east side is volcanic
rock with overlying layers of clay, advanced stratified drift, till and
sedimentary deposits (Liesch, et al., 1963).
Prior to the arrival of European settlers in 1862, the Lake Samma-
mish basin was covered in a climax formation of Western Red Cedar
(Thuja plicata), Western Hemlock (Tsuga heterophylla), and Douglas Fir
(Pseudotsuga taxifolia) (Hansen, 1938). Heavy logging around the turn
of the century left the basin in second growth forest. Today 80% of
the watershed remains in second growth, primarily red alder (Alnus
oregona) with scattered maple (Acer, sp.) and willow (Salix sp.).
The significant historical events in the basin are listed in Table 2.
The population of the basin has grown from three families in 1862
to the present 40,000, the majority of the growth coming in the last 10
years. The only sizeable concentration is located in the town of
Issaquah, population 4,500. The town is comprised of the small busi-
nesses required to support a residential community. The only industrial
development is a dairy processing plant and a state salmon hatchery.
Within the watershed are several gravel operations and a county sanitary
landfill. Large residential developments have been built throughout the
entire west side of the lake. The east side is dotted with small farms,
but the major portion of the land remains in second-growth. A narrow
8
-------
Table 2 HISTORICAL EVENTS OF INTEREST IN THE LAKE SAMMAMISH BASIN
(AFTER FISH, 1967).
Date Event
1862 Three families settled along lakeshore
1864 Three settlers and four Indians killed in uprising
1868 Hops started and become principal agricultural crop
1870 Census listed 28 men; 24 were farmers
1880-1920 Logging very important, lake often covered with logs
1887 Large scale coal mining (1913-14 boom years), dies out
in 1920's
1888 Railroad built around eastern shore
1889 Donnelly mill closed (large sawdust pile on lakeshore),
large mill located at Monohan (pop. 200)
1900 Hops no longer farmed
1912 Sammamish Slough dredged, shortened from 19 km to
10.5 km
1913 Issaquah Sewerage Agency formed
1925 Town of Monohan burned down
1940 Secondary treatment (trickling filter) built for
0.15 MGD capacity
1960 Population of Issaquah is 1,870
1968 Treated effluent diverted from Issaquah Creek
1970 Population of Issaquah is 4,314
-------
strip of land along the east shore of the lake has been subdivided into
residential tracts. The upper valley drained by Issaquah and Tibbets
Creeks is primarily forested with scattered farms and small clusters of
houses (Figure 1).
Lake Sanmamish State Park, located at the south end of the lake,
draws over 600,000 visitors each summer. Two commercial and an un-
developed county park are also located on the lakeshore. In addition,
a large county park touches the northern end of the lake and extends
along the Samnamish River. Both Pine and Phantom Lakes have county
Parks located on their shores. Presently the waters of Lake Sammamish
support a variety of water sports, including water skiing, pleasure
boating, boat racing, swimming and fishing, besides providing an
aesthetically pleasing background for picnics and related activities
carried out at the several lakeside parks.
The primary point sources of wastewater within the basin were the
town of Issaquah, the milk processing plant, and the fish hatchery.
Since 1968, the effluent from the town's trickling filter plant (568
3 3
m /d) and the milk plant (284 m /d) have been diverted out of the
3
basin. Today only the milk plant cooling water (227 m /d from ground-
water) and the hatchery passthrough water, which originally comes from
Issaquah Creek, are discharged to Issaquah Creek, and to Lake Sammamish.
Only the sparsely settled east side and upper valley sections of the
watershed remain on septic tanks. Hence the percentage of the basin
population on septic tanks is small.
PALEOLIMNOLOGY
Paleolimnological evidence suggests that Lake Sammamish has maintain-
ed a trophic state within the mesotrophic range for at least the past 100
years in spite of settlement in the watershed by European man and an in-
crease and decrease in the phosphorus income of at least one-third. The
only shift in the 150 years of paleolimnologic history shown in the organic
matter profile (Fig. 2) from Lake Sammamish is from 1880 to 1910, using a
sedimentation rate of 3.5 mm/yr determined from a recent historical event. This
10
-------
N
A
LAKE SAMMAMISH
Beaver Lake
Phantom ~=
Lake S^^
North Fork
Issaquah
Tibbetts
Creek
N
Drainage
Boundary
East Fork
Issaquah
Creek
::: INDUSTRIAL
COMMERCIAL
RESIDENTIAL
x *r
Figure 1. Land Use Map of the Lake Sammamish Watershed.
11
-------
Percent Water
v-
10-
*E
£
120-
A
lAl
U.
0
530-
0.
Uf
40-
50-
"N
60
o
o
(
o
0
\
o
o
o
Figure 2. Vertical distrib
Percent Organic Mat.
0 10 15 20 °
o
rag P/gm dry wt,
5
mg P/gm organic wt.
o
- o
0 10 20 30 40
Vertical distribution of percent waiter, percent organic matter, and phosphorus as a fraction
of the dry weight and organic matter of the sediment. Open circles represent values from Lake
Washington cores (Shapiro et al., 1971).
-------
bulge corresponds to the period of intense logging in the Pacific
Northwest and a shift in pollen from confers to alders. The organic
matter bulge probably resulted from sawdust and bark deposits from
mill and log rafting activities. If a change from aerobic to anaerobic
conditions had occurred during this period it surely would have been
evident in the organic matter content.
Comparison with Lake Washington results (Shapiro, et al., 1971)
shows that both phosphorus and organic matter content were higher than
levels in Lake Washington in early years, lower than Lake Washington
in recent years, but remained relatively constant^ over the past century
(Fig. 2). The sediments in Sammamish clearly show less impact of man
than in Lake Washington and are characterized most strongly by the con-
stancy of their contents.
In a more detailed analysis of historical sedimentation using Pb210
dating, four distinctly different rates were observed with the highest
rate having occurred between 1932 and 1944 following logging activity.
Although the concentration of P in the sediment actually decreased, the
loading rate to the lake during that period may have been four times the
present loading if the increased rate of sedimentation is considered.
The increased sedimentation rates are thought to be a result of increased
erosion following deforestation, which was at a peak during 1910-1930
(Birch, 1976).
A one-day study of the lake by Kemmerer, et al. (1923) in August,
1913, strongly suggests that the lake was anaerobic and mesotrophic
then. Phytoplankton species, Secchi disk depth, and except for the ab-
sence of Diaptomus, zooplankton species were all similar then to those
found in the lake today.
The pennate/centrate ratio of diatoms in the lake's sediments have
also remained relatively stable over the past 100 years (Table 3).
According to Stockner (1971) these values are within the 1.0 - 2.0 range
for mesotrophy. For more details of the paleolimnological results see
Rock (1974). Also, Wiederholm (1976) has found chironomid species that
indicate mesotrophy throughout a 50 cm deep core and a 100 year history.
13
-------
Table 3. DIATOM RATIOS FOR LAKE SAMMAMISH CORE 08.
Date
1972
1955
1940
1925
1910
1895
1880
1865
1850
Pennate/Centrate
1.2
1.0
1.5
1.1
1.2
1.1
1.5
0.9
0.9
Thus, it appears that the trophic state of Lake Sammamish has re-
mained relatively unaffected by activities in the watershed since the
settlement of European man. The associated increases in P income during
this period have been at least on the order of 30 to 40 percent and no
doubt much greater considering all potential point and non-point sources,
14
-------
SECTION 5
MORPHOMETRIC AND HYDROLOGIC DESCRIPTION OF LAKE SAMMAMISH
Lake Sammamish occupied a 13 km section of the Sammamish River
Valley after the Wisconsin glaciation when the retreating Vashon glacier
left a terminal moraine blocking the valley. Today the lake level is
controlled by a wier at the head of the Sammamish River. The deepest
section of the lake is located less than 3 km from the south end. If
15 m is considered as the division between deep and shallow, the surface
area ratio of deep to shallow water is
12km2
7.8 km2 (shallow)
while the ratio of epilimnion to hypolimnion volume is 0.99.
The study of water currents has been limited to the movement of
Issaquah Creek water in the lake (Moon, 1972). During the period of
winter mixing the creek water dispersal is primarily influenced by wind
direction and velocity. The water was sufficiently dispersed at a dis-
tance of 500 m to make the tracer undetectable. Similar studies made
during thermal stratification showed the creek water plunging into the
metalimnion and dispersing in a fan-like pattern. Additional morpho-
metric and hydrologic data are summarized in Table 4.
Waste-water effluent from Issaquah entered Lake Sammamish through
Issaquah Creek. In winter, during high flow waste water P entered
along with high levels of suspended sediment. In summer the waste P
followed the Issaquah Creek inflow water which probably entered the
open water area at depth.
15
-------
Table 4. SUMMARY OF PERTINENT HYDROL06IC AND MORPHOMETRIC
CHARACTERISTICS FOR LAKE SAMMAMISH.
Parameter Lake Sammamish
.
Surface Area of Lake (km ) 19.8
2
Drainage Area (km ) 253
Lake Volume (km }* 0.35
Depth
Mean (m) 17.7
Maximum (m) 32.0
Epilimnion (m) 8.8
Euphotic (m) 7.3
Width
Mean (km) 1.5
Maximum (km) 2.4
Length of Lake (km) 13.0
Length of Shoreline (km) 34.0
Water Retention Time (yrs) 1.8
3
Stream Inflow (km /yr} 0.198
o
Stream Outflow (km /yr) 0.203
o
Groundwater Infiltration (km /yr) 0.0
Groundwater Exfiltration (km3/yr) 0.01
Duration of Stratification (mos.) 7
*
influenced by wier
16
-------
SECTION 6
LIMNOLOGICAL CHARACTERIZATION
TEMPERATURE
Lake Sammamish is a monomictic lake that begins thermal stratifica-
tion in May. Maximum water column stability occurs by late August and
destruction of the thermocline is complete by late November. The tem-
perature range is from a minimum 5.5°C to a maximum 25.5°C.
LIGHT
The depth of visibility has been determined by means of the Secchi
disk. The annual mean for the six years of data is 3.3 m. The lowest
seasonal mean (3.0m ) occurs in the winter due to turbidity from the
winter mixing and runoff. The springtime mean is only slightly higher
(3.1 m), but these low values are due to the diatom pulse. The light
penetration increases during the summer (3.5 m) and reaches its highest
values in the autumn (3.6 m). The maximum Secchi disc measurements
usually occur in September, the deepest recorded being 6.1 m.
Light extinction was determined by a submarine photometer. The
bottom of the euphotic zone was considered to be at a depth receiving
1% of the surface light intensity. The mean depth of the euphotic zone
is 7.0 m, while the range is from 5.0 to 12.5 m.
ALKALINITY AND pH
Lake Sammamish has a pH range of 6.3 to 9.6 due to biological
activity. Correspondingly the alkalinity as CaCOg ranges from 26 mg/1
(0.52 meq/1) to 42 mg/1 (0.84 meq/1), while the mean is 33.3 mg/1
(0.67 meq/1).
DISSOLVED OXYGEN
During the winter, the oxygen content essentially remains at an
air saturation level, approximately 12 mg 02/1, due to continual
17
-------
circulation. The development of thermal stratification in early May
results in a clinograde CL curve that approaches zero oxygen levels
(0.1 mg Op/1) in the bottom waters by late July or early August. The
hypolimnetic oxygen deficit continues to increase until early October.
By this time the entire hypolimnion (below 15 meters) has less than
1 mg 02/1. Oxygen levels start to increase with the coming of the
autumnal circulation.
The depletion of oxygen in the hypolimnion is dramatically shown
in the calculation of hypolimnetic oxygen content in early October
(Table 5). Also shown are "actual" deficit rates. The "actual" rates
are based on oxygen saturation of 12 mg 02/1 at 7.5°C occurring on
April 15th and the oxygen present at the end of stagnation.
Table 5 HYPOLIMNETIC (DEPTH GREATER THAN 15 m) OXYGEN CONTENT AND
"ACTUAL" OXYGEN DEFICIT RATE.
Total Hypolimnetic Total Hypolimnetic Number of
v 02 present at 0? present at end days of
Tear start of stratifi- of stratification stratifi-
cation (metric tons) (metric tons) cation
1970
1971
1972
1973
1120
1120
1120
1120
21.3
12.5
45.7
21.3
178
192
176
185
02 Deficit
Rate
(mg Og/
o
cm- day)
0.051
0.047
0.050
0.049
18
-------
MAJOR AND MINOR IONS
The results of a single survey are shown in Table 6. Neither S04"
nor Cl" have been measured, while the only trace metals measured have
been Mn++, Zn++ and Pb++
Table 6. Trace Metal Survey on October 24, 1971 for Lake Sammamish
Location
Lk. Sam. 612
Surface
8 m
16 m
25 m
Islaquah Ck.
Tibbetts Ck.
Ca
12.
8.
8.
12.
24.
i
80
40
95
40
70
V.
3.
3.
3.
3.
8.
- Pi
ig
42
44
68
70
15
N
8.
8.
8.
8.
9.
14.
la
43
47
15
17
31
59
K
1.01
0.98
0.94
1.00
0.94
1.52
Fe
40
63
280
1020
450
110
ypi
Mn
19
40
600
1660
35
20
Zn
376
300
35
34
318
150
Pb
0.5
0.5
0.8
0.8
0.9
4.2
Outflow
Sammami sh
River 6.05 3.00 8.16 1.13 25 9 7 0.6
19
-------
SECTION 7
METHODS AND MATERIALS
LAKE MONITORING
The lake water column has been monitored at a central location from
1970 through 1975 by students and faculty in the Civil Engineering
Department, University of Washington. Prior to the 1968 wastewater
diversion the Municipality of Metropolitan Seattle (METRO) monitored
the water column at the same location for 1.5 years from July 1964
through 1965. Thus, the effect of diversion is evaluated primarily
from the changes in trophic state indicators determined at that central
station designated as 612 in Figure 3.
Sampling frequency has varied from monthly to weekly throughout the
period, with frequency generally increasing in recent years. Twice
monthly was generally considered necessary particularly during the grow-
ing season. After examining the 1970-71 data, Swayne (1973) showed that
a frequency greater than 20 days between sample collections would result
in an artificial smoothing in the temporal dynamics of many of the vari-
ables.
One station was considered representative. Emery (1972)
showed that there was no statistically significant difference between
chemical and biological variables at station 612 compared with a nor-
therly located station indicated by METRO as 611 (Fig. 3).
During 1970-73 the photic zone was sampled at four depths based on
the amount of light received. The depths corresponded to 95, 60, 30 and
1 percent of incident light. Secchi disk depths were used to locate
depths for given light levels based on a nomograph determined by Emery
(1972). This was largely for the purpose of insuring that maximum
photosynthetic rates were measured regardless of incident light varia-
tion. The maximum in Lake Sammamish was found to vary from 30-60% of
incident light. Samples for chemical analysis were also collected in
20
-------
Sommamish River
OLYMPIC PENINSULA
LAKE SAMMAMISH STUDY AREA
Sediment
Trap-1
Issaquah Creek
Figure 3.
Tibbetts Creek
Location of sampling stations and bathymetric map for
Lake Sammamish.
21
-------
the hypolimnion, one half the distance between the 1% depth and the
bottom and 1 meter above the bottom - a depth of 28 m. This station
was not located at the lake's maximum depth, which is 31 m, but rather
at 28 m (Fig. 3). Procedures were slightly modified in 1974-75 when
specific depths were regularly sampled throughout the water column -
surface, 5 m, 10 m, 15 m, 20 and 22 m. Also, in situ oxygen and tem-
perature determinations were made at more frequent depth intervals than
other chemical measurements.
The variables determined, the methods used and respective sources
for the methods are shown in Table 7. The only method not explained is
phytoplankton volume, which has varied with investigations; METRO used
an inverted scope sedimentation along with specific cell measurements,
Emery (1972) used an upright scope, centrifugation and specifically
measured each observed alga cell, while subsequent work has been with
an upright scope, but with a sedimentation method in 1973 and a milli-
pore filter method in 1974-75. Both later sets were developed with the
use of average volumes for a given species. Although direct comparisons
of volumes are not comparable the percent volumes occupied by various
groups, for example blue green algae, is considered valid.
IRON AND PHOSPHORUS AT OVERTURN
The relationship between iron and phosphorus was studied at over-
turns in 1972 and 1973. Samples for total P total Fe and oxygen were
determined at weekly intervals and every 2 meters in the water column.
Total Fe was determined by atomic absorption (APHA, 1971).
NUTRIENT INCOME ESTIMATION
Phosphorus and nitrogen loading into Lake Sanunairrish has been esti-
mated by measuring the concentration of total P and total N (only N03 in
water year 71) in several minor tributary streams. METRO monitored 9
in 1964-65 (Isaac, et_ al_., 1966) and 12 in 1970-71 (Moon, 1972). The
major surface inflowIssaquah Creekwas also monitored during those
periods and again in 1972.
22
-------
Table 7. SUMMARY OF ROUTINELY MONITORED PHYSICAL, BIOLOGICAL AND
CHEMICAL PARAMETERS
Parameter
Method
Reference
Physical
Temperature Profile
Light Penetration
Euphotic Zone
YSI and Hydrolab
Temperature Probe
Seechi Disk
Submarine Photometer
Biological
Primary Production
Phytoplankton Volume
Chi a
14C Light-Dark
Bottles
Variable Sample
Extraction with 90%
Acetone
Strickland and Parsons
(1968)
Strickland and Parsons
(1968)
Chemical
Dissolved Oxygen
Profile
Total, Ortho-
phosphate P
Nitrate, Nitrite N
Silicate
PH, Alkalinity
Total N
Azide Modification-
Winkler
Molybdate Complex-
ing Reaction
Cadmium-Copper
Column
Molybdate Complex-
ing Reaction
pH Meter
UV Light Oxidation
American Public Health
Assoc. (APHA) et al.,
(1971)
Strickland and Parsons
(1968)
Strickland and Parsons
(1968)
Strickland and Parsons
(1968)
APHA et al., (1971)
Strickland and Parsons
(1968)
23
-------
sampled the streams once and Moon twice per month for nutrient content
and flow, except that Issaquah Creek was continuously gauged by the
U.S.G.S. at a point about 1,9 km from the lake. Moon collected 37
samples from Issaquah Creek in 1970-71 and 30 in 1971-72 (WY 71 and 72).
Issaquah Creek was shown by Moon (1972) to carry about 70% of the inflow
of P and 57% of the N (N03).
Because of the variability in nutrient budget estimates from person
to person and by different means of data handling (see Rock, 1974), con-
tinuous monitoring of the Issaquah Creek nutrient content was begun in
the 1973 water year. An automatic sample collector was installed in
the U.S.G.S. gauging hut and samples were collected every eight hours
for nutrient analysis. Over 1000 samples were provided for analysis
from Issaquah Creek in the 73 WY. This insured that periods of short
term peak runoff, in which total P was known to vary greatly, could be
accurately determined. Daily collections were also made in the 1974
water year for total P and total N content.
In Moon's (1972) water budget for the lake, subsurface was assumed
to contribute relatively little water to the lake in comparison to sur-
face stream flow. Thus, the groundwater contribution of nutrient was
also ignored in the budgets. Atmospheric input with rain to the lake
surface was included, however, by monitoring the rainfall, albeit
rather infrequently, during the 71 WY. The final significant source is
the Washington Department of Fisheries salmon hatchery, which was
briefly evaluated and included in the below mentioned special study.
To help resolve the problem of how much nutrient was actually
diverted in 1968, a joint study was undertaken with METRO to reevaluate
the contribution of the Darigold milk processing plant and the Issaquah
sewage treatment plant effluents to the pre-diversion loading to the
lake.
24
-------
SEDIMENTATION
Sediment raining down from the epilimion was collected in traps
located near the bottom at three additional stations besides the control
612 station. Sediment was collected by four 10 cm diameter plastic
funnels which channeled the material into 50 ml centrifuge tubes. These
four units were secured to a 30 x 30 cm polyvinyl chloride platform.
Traps were suspended at various depths in the lake, with emphasis here on
results from the bottom traps, 2 m above the bottom at all four stations.
The traps were located after collection intervals ranging from one
week to several months by means of buoys submerged 1-2 m below the sur-
face. This technique successfully avoided interference by the boating
public. To minimize the decomposition of material, 5 ml of chloroform
was added to each centrifuge tube as a preservative. A disadvantage of
this technique is that congregating zooplankton are trapped and killed
and then must be removed before analysis by filtration (0.5 mm) and then
detritus particles > 0.5 mm must be removed from the net and added back
to the sediment samples.
Sediment analysis consisted of centrifugation at 9000 rpm for seven
minutes, decanted, dried at 60° C for 24-36 hours and weighed. This was
begun immediately upon returning from the field. Prior to this, samples
were examined microscopically to evaluate the qualitative nature of the
material as to allochthonous or autochthonous origin.
Subsamples of dried sediment (.20-100 mg) were analyzed for C using a
Leco semi-automatic carbon analyzer and for N using the semi-microkjeldahl
method (Bremner, 1960). Total P was determined according to the ascorbic
acid-molybdenum blue method in Standard Methods (1971) on digested sediment.
Sediment (20-100 mg) was digested in 10 ml Teflon crucibles by incubating
for 12 hours at 120° C with 5 ml of 40% HF. Then 5 ml of concentrated
HN03 was added and treated at 240° C for 1-2 hours. Oxidation was com-
pleted by further addition of 5 ml HN03 and cautious addition of 5 ml of
70% HCL03. Just prior to evaporation samples were cooled, removed with
1 ml concentrated HCL, filtered through #4 prerinsed Whatman paper and
diluted with 50 ml demineralized, deonized water.
25
-------
SEDIMENT-WATER INTERCHANGE
Phosphorus release from the sediment in Lake Sammamish was deter-
mined in situ by using six plastic columns inserted into the sediment in
the vicinity of sediment trap 1 (Fig. 3). Two opaque columns and one
transparent column were placed at depths from 10-12 m and two transparent
and one opaque column were placed at 6-8 m depths.
Each column was 3.7 m long with a weighted base plate attached 0.6 m
above the bottom. Volumes varied from 42 to 62 £ depending upon diameters.
Sampling ports (1.3 cm dia) were located at 0.3, 1.9 and 2.8 m above the
sediment. Tygon tubes 15 m long were extended from each port to a sur-
face float. Each column was equipped with a plexiglass piston that
allowed for the extraction of sample water and prevented the influx of
oxygenated water and avoided the formation of a vacuum Inside,
After samples were extracted deoxygenated water was forced back into
the columns. About 5 I or 10% of the column volume was removed with
each set of samples.
Analytical procedures for P, Fe, DO, pH and alkalinity were the same
as previously indicated.
Experiments were conducted at various time intervals over a three-
year period. Usually, columns were sampled every 10-14 days. Effort was
made to maintain the column samples under anaerobic conditions by collect-
ing the water in chambers previously purged wtth N«. However, when filters
clogged and were changed Q? was introduced thus biasing the soluble P
results.
The columns similated two conditions in the lake. The opaque columns
represented the dark hypolimnetic water while the transparent ones repre-
sented the processes in the euphotic zone.
The experimental periods in 1973-74 were May to July and July or
August to December or January. For the early experiments the duration
ranged from 63-76 days for three columns. For the later experiments the
periods ranged from 121-171 days. In some instances the columns were
purged with N2 to immediately exhaust the DO while in others DO was
allowed to exhaust naturally via respiration.
Z6
-------
DYNAMIC MODEL
The model developed for Lake Sammamish is made up of two major sub-
models, a mixing model, which computes the daily mixing depth and tem-
perature profile, and a phsophorus model, which utilizes the physical,
chemical and biological parameters pertinent to the lake. The model
determines the magnitude of the pertinent state variables.
Mixing model
The mixing model computes the temperature profile by first calculat-
ing the thermal energy input to the surface waters considering insolation,
conduction, evaporation and back radiation. The temperature throughout
the water column is then determined by summing the temperature value
from the previous day at each depth with the effect of the diffusivity
of energy within the water column at that depth.
The mixing depth is then computed by determining the point in the
water column where the kinetic energy provided by the wind shear on the
lake's surface coupled with the convective mixing energy due to insta-
bility caused by temperature differential, is equal to the potential
energy of the thermal gradient integrated from the surface.
The last operation the mixing model performs is to average the
temperature from the surface to the mixing depth and adjust the temper-
ature at each depth within this zone to an average value.
Phosphorus
To facilitate modeling, the lake was divided into three zones, the
epilimnion, defined as the layer between the surface and the mixing
depth, the hypolimnion, defined as the layer between the mixing depth
and one meter above the bottom, and the sediment-water interface zone,
defined as the layer containing the bottom sediments and the water one
meter above the bottom. From these definitions it is observed that on
days when complete mixing occurs the hypolimnion does not exist.
The phosphorus model has four state variables: phytoplankton (XI);
ortho-phosphorus (X2); zooplankton (X3); and detritus (X4). These state
27
-------
variables are represented by the pools (ovals) in Figure 4. The equa-
tions for the changes in those state variables over time are also shown
in Figure 4.
Table 8 lists the subprocesses which make up the functional equa-
tions for the state variables. The notation used for system constants
and variables is listed in Table 9 along with their definitions, dimen-
sions and the values for the constants. As may be readily observed from
Table 8, all of the subprocesses, with the exception of photosynthesis
(Photo), are either Michaelis-Menton type or logistic functions.
The arrows in Figure 4 show the model processes in graphic form
including sources, sinks and the exogenous variables (driving functions)
which are mixing depth (H), temperature profile (T[I]), incident solar
radiation (RO) and inflow {Q. ). Outflow is assumed equal to inflow in
in
this model and is also notated Q. . The sediment acts as a source and/or
sink for the phosphorus depending upon physical conditions.
The phosphorus model was developed as a differential equation model
using a fourth-order Runge-Kutta technique to approximate the time
dependent changes in the state variables on a daily basis. The model is
presently implemented in both its original form and as a difference equ-
tion model using a variable time step to insure that no state variable
changes more than five per cent (5%) in any one time period.
28
-------
^jf = -UPTAKE + PREG - DILUT
- P/l/day)
PHOTO - PRESP - GRAZ - PDEA - SINK - PFLUSH
- ch-/l/day)
PRESP
Figure 4. State variables, transfer
processes, and functional equations
in the phosphorus model for Lake
Sammamish. See Tables 8 and 9 for
further explanation of process
functions.
dX4
~dT
ZDEA-PREG+DEXC-SED-DFLUSH
+PDEA+DSR
- P/I/day)
ZRESP
ZRESP - ZDEA +ZRESP
C/l/day)
-------
Table 8 MATHEMATICAL EXPRESSIONS OF INVOLVED SUBPROCESSES IN THE
PHOSPHORUS CYCLING MODEL.
PHOTO = kl.e(T/10).(2R/RO-dz)e(1"2R/RO'dz).(X2/(k2+X2))Xl
PRESP = k3-T-Xl
GRAZ = k4-X3(Xl/(k9+Xl))
SINK = k5-dXl/dz
PDEA = klO.Xl
PFLUSH = Q1n-Xl/V
UPTAKE = CO.COT(PHOTO)
PREG = k7-X4
DILUT = (.25.k6-X4)Qin/V
ZGROW = C01-C02-GRAZ
ZRESP = ZRES-T-X3
ZDEA = k8-X3
DZD = CO-ZDEA
DEXC = (1-C02)-CO-C01-GRAZ
SED = kll-dX4/dz
DFLUSH = (.3-k6-X4)Qin/V
DPZ = COCO!PDEA
DSR = kl2-Ase(j-l/V(zn)
where: A H = surface area of the bottom or sides in contact
sea with the layer
V(zn)= volume of the layer
30
-------
Table 9. NOTATIONS AND UNITS USED ALONG WITH VALUES FOR CONSTANTS
FOR THE PHOSPHORUS CYCLING MODEL.
Notation Definition
XI
X2
X3
X4
X5
RO
R
T
H
kl
k2
k3
k4
k5
k6
k7
k8
k9
klO
kll
k!2
IRES
CO
CO!
C02
Values
PHYTOPLANKTON BIOMASS
PHOSPHORUS CONCENTRATION
ZOOPLANKTON BIOMASS
DETRITUS MASS
TOTAL PHOSPHORUS
INCIDENT SOLAR RADIATION
AVAILABLE RADIATION
TEMPERATURE
MIXING DEPTH
PHYTOPLANKTON MAXIMUM UPTAKE RATE
PHOSPHORUS HALF SATURATION CONSTANT
PHYTOPLANKTON RESPIRATION RATE
ZOOPLANKTON MAXIMUM GRAZING RATE
PHYTOPLANKTON SINKING RATE
PHOSPHORUS CONCENTRATION OF INFLOW
DETRITUS REGENERATION RATE
ZOOPLANKTON DEATH RATE
chl a HALF SATURATION CONSTANT
PHYTOPLANKTON DEATH RATE
SETTLING VELOCITY OF DETRITUS
NET SEDIMENT RELEASE RATE (ANAEROBIC)
ZOOPLANKTON MAXIMUM RESPIRATION RATE
PHOSPHORUS TO CARBON RATIO
CARBON TO chl a RATIO
FRACTION OF ZOOPLANKTON DIGESTION
Dimensions
yg-chl a/1
yg-P/1
yg-c/i
yg-P/1
yg-P/1
cal/cm /sec
2
cal/cm /sec
°C
meter
day"1
yg-P/1
day"1 "C"1
yg-chl a/1/
yg-C/1/day
m-day"
yg-P/1
day"1
day"1
yg-Chl a/1
day"1
m-day
2
yg-P/m /day
day"1 °C"]
mg-P/mg-C
Value
-
-
_
X2+X4
-
_
0.22*
0.62*
0.0005
0.015*
0.2
0.035*
0.3
1.0*
0.05
0.1*
2300.0*
0.0015
0.02
mg-C/mg-chl a_ 50.0
mg-C/mg-C
determined either in situ or in laboratory experiments
0.5
for
Lake Sammamish.
31
-------
SECTION 8
RESULTS
EXTERNAL NUTRIENT LOADING
Rock (1974) reviewed problems in calculating the pre-diversion
nutrient loading in Issaquah Creek, particularly dealing with the choice
of either calendar or water year data. By weighting the data on a
monthly time basis the short lived high peaks in P content and flow that
would bias the annual input estimate were damped out. Thus, while
several different approaches (by different individuals) to calculating
nutrient load had produced values ranging from 23,000 to 11,000 Kg P/yr,
Rock established that about 16,000 Kg P/yr was the most reasonable
estimate for WY 1965 and 12,000 Kg P for the calendar year. Total N
input was 170,000 Kg N for WY 1965; 122,000 of that was inorganic N.
These were determined by multiplying mean monthly flows by mean monthly
concentrations.
Post-diversion loading estimates in Issaquah Creek showed consider-
able year-to-year variation. Moon estimated 8,500 Kg P and 100,700 Kg
N (inorganic) in Issaquah Creek for WY 1971. Guttormsen (1974) esti-
mated 198,000 Kg inorganic N in 1972-73 and 258,000 total N. Rock
calculated 15,000 Kg P from the 30 samples collected in WY 72 and
7,126 Kg P for WY 73 when a complete data base on P concentration was
available. Much of the year-to-year variation was due to flow since the
mean annual flow in WY 1972 was 5.58 m3/sec and 2.91 m3/sec in WY 1973.
The input from June 1974 to June 1975 was 7,128 Kg P, almost identical to
the 1973 WY value even though the mean flow was higher3.98 M3/sec.
Rock found that most monthly or twice monthly sampling procedures
would result in underestimates of the true Issaquah Creek load by 13 to
26%. However, if the 10 year annual flow and the annual mean concentra-
tion of 65 yg/1 P (12-312 yg/1) was used there was almost perfect agree-
ment with the true load. The difficulty with estimating annual nutrient
input involves the simultaneously large, but short lived increases in
32
-------
flow and concentration. Missing such important points will lead to great-
ly underestimated loads. For example, 25% of the annual P load came into
Sammamish in the 1973 WY during two days in December and two days in
January. The mean flow for those four days was 22.6 m /sec and the mean
concentration was 230 yg/1 P. In 1974-75 there were four similar days
in January with a mean flow of 20.7 m /sec and a mean concentration of
303 yg/1 P. Those four days contributed 30% of the 1974-75 load. Thus,
it is clear that daily observations of concentration as well as flow are
necessary, particularly during periods of high flow, if serious under-
estimates of loading are to be avoided.
Thus, it was important to attempt to normalize flow during pre- and
post-diversion years to obtain an accurate estimate of the quantity of P
diverted. For flows below 2.8 m /sec., flow and P concentration in 1973
were inversely correlated, but above that flow they were directly corre-
lated. Both correlations were statistically significant. The process
operating that seems to produce such variation is a seasonal one. As the
flows increase in the fall and winter the material accumulated in the
streams and on the land during the summer is flushed out. During that
period concentration may increase dramatically with flow. Once the mate-
rial has ,been flushed out in late winter and spring, increased flow may
carry very little particulate matter with it and concentrations may de-
crease in spite of high flows. For example, for flows over 2.8 m3/sec
the November-December-January mean was 88 yg/1 while the February-June
mean was 39 yg/1. The annual mean was 65 yg/1.
To accurately estimate the fraction of P diverted, years of equal
flow must be compared. The 1973 WY was a very low flow year compared to
the 1965 WY, however, when all post-diversion data, were used the result-
ing correlation showed a relatively good fit (r = 0.41) and it seemed
reasonable to let the equation describe an average post-diversion year.
The mean flows for all post-diversion data through the 1973 WY is 4.3 m3/
sec., essentially the same as the mean pre-diversion flow of 4.25 m3/sec.
33
-------
The estimated P loading calculated by the above post-diversion equa-
tion is 9000 Kg. Thus, comparison of this value with the most reasonable
pre-diversion estimate of 16,000 Kg P leaves 7000 Kg P diverted by the
METRO project in 1968. Assuming Issaquah Creek is now 70% (Moon, 1972)
of the surface income to the lake this means that about 35% of the total
surface input P was diverted. Including an estimate for atmospheric in-
put to the lake surface lowers the diverted percentage to 34. If around
7200 Kg P is considered as the post-diversion Issaquah Creek input, be-
cause of the close agreement in 1973 and 1974, then the diversion would
be 8800 Kg P or 43%, which is probably high.
Another approach was used to estimate the fraction diverted and
that was to back calculate the quantity of waste-water P that was prob-
ably entering before diversion. As part of this, METRO conducted a mass
balance of the inputs and outputs to and from the Creek within the town
of Issaquah during high flow and low flow. In addition the
Issaquah treatment plant flow before diversion was used with literature
values of dairy waste and sewage P to estimate the wastewater P load
before diversion (see Rock, 1974,for details). The P loads measured
in the Creek did not balance and thus failed to provide usable data
because of diurnal variations, but the indirect approach provided an esti-
mate of 4,400 Kg P diverted, which would have been 33% of the Issaquah
load (normal flow year), but only 255t of the lake's total load. Using
this approach would mean that Issaquah Creek carried a pre-diversion
load of 13,400 Kg P.
Because of the many uncertainties in the latter approach it seemed
most reasonable to stick with the previously mentioned approach; that
of correcting the post-divers ion data to a normal flow year, which the
1965 WY apparently was,comparing that to the 16,000 Kg best estimate for
Issaquah Creek before diversion. Therefore, the most probable estimates
of pre- and post-diversion phosphorus loads are shown in Table 10. The
remaining wastewater P input comes from the salmon hatchery in Issaquah,
which is a rather crude estimate based on the two mass balances done on
the various inputs to Issaquah Creek by METRO.
34
-------
Table 10. PHOSPHORUS LOADING (Kg/Yr)TO LAKE SAMMAMISH BEFORE AND AFTER
WASTEWATER DIVERSION. P ESTIMATES ARE BASED ON A NORMAL
WATER YEAR.
Pre-diversion (%) Post-diversion (%)
Issaquah Creek
sewage and dairy waste 7,000 (34) 0 (0)
land runoff 8,500 (42) 8,500 (64)
salmon hatchery 500 (3) 500 (4)
11 Minor Tributaries 3,900 (19) 3,900 (29)
Atmosphere to Lake Surf. 400 (2) 400 (3)
Total external loading 20,300 (100) 13,300 (100)
Loading in g/m2 yr 1.02 0.67
In terms of normal year-to-year variation it was noted that for
the 1972 WY, which was wet, the Issaquah Creek loading was 15,000 Kg P.
For the 1973 WY, which was dry, the loading was 7,110 Kg P, with 9000
Kg P as the estimate for an average or normal post-diversion year.
The respective flow weighted mean concentrations for the wet 1972 and
dry 1973 WY's were 85 and 78 ug/1. The surprising thing is that 1974-
75 flow was about normal, 3.98 m3/sec versus the 11-year mean of 4.25
m3/sec, and the flow-weighted mean concentration was 57 jjg/1. Thus,
there may be more variation in flow from year to year than in concen-
tration.
If these post-diversion flow-weighted mean concentrations for 1973-
75 are compared with the pre-diversion concentration (68 versus 119 g/1)
then that represents a 43% reduction in concentration in Issaquah Creek
or 34% for all inflows if Issaquah Creek is 70% of the P input. That
estimated reduction in concentration is in agreement with the reduction
in loading.
35
-------
Nevertheless, the P load can vary annually due largely to flow on
an order similar to the fraction directed. This again emphasizes the
necessity of comparing relatively similar hydrologic years to estimate
manipulated loading changes.
The nitrogen balance was treated less intensively than phosphorus
because the lake is primarily P limited and the difference in analytical
methods for NCK and organic N make the comparison of before and after
diversion data rather questionable. However, the various estimates of
the total N input for two post-diversion years compared to METRO'S
estimate of pre-diversion input (NOg corrected to results with the
Cd-Cu reduction method - see Emery, 1972) are shown in Table 11. To ex-
trapolate to total lake input, Issaquah Creek loading was assumed to
be 57% of the total surface water input. See Appendix B for 1974 N
loading to Issaquah Creekit was nearly the same as that observed in
1972-73, the basis for Table 11.
Table 11. NITROGEN LOADING TO LAKE SAMMAMISH IN Kg/Yr; 11 TRIBUTARIES
PLUS ISSAQUAH CREEK.
Pre-diversion
(1965 WY)
Post-diversion Post-diversion
(3/72 - 2/73; (6/74 - 6/75
Guttormsen, 1974
Organic + NH3-N
N02 + N03 - N
Total N
Loading in g/m -yr
69,000
174,000
243,000
12.3
60,000
198,000
258,000
13;0
135,600
179,470
314,730
15.9
The increase noted in Table 11 cannot be considered significant in
the face of the analytical differences between pre- and post-diversion
period investigators. However, the ratio of N:P loading is 12 before
36
-------
and 20 after diversion, suggesting that P should be most limiting under
most circumstances. Nitrogen fixation has been shown to be rather
insignificant to the total N supply (J. Staley, personal communication1)
Atmospheric input has not been measured, but should not account for
more than a few percent of surface water input.
RESPONSE TO DIVERSION
Common trophic state indicators in Lake Sammanrish have remained
relatively stable, or at least no significant trend is apparent, in the
six post-diversion years the lake has been monitored. Figure 5 shows
the mean concentrations of total P and N03~N measured at station 612
during December through February (the non-growth months) when the lake
is well mixed. These values are weighted means in the photic zone,
which corresponds reasonably well with the epilimnetic depth in this
lake, and are compared with the pre-diversfon values assigned as 100%.
P remained very stable* well within 10% of the pre-diversion level.
Although NOo decreased*the correction of prediversion data for a
methods difference (a factor of 2) makes the significance of that
decrease rather doubtful. In any event levels are very near 100% in
the last two years.
Chi a^ concentration was determined from March through August and
weighted for time, similar to the analysis of stream nutrient data men-
tioned earlier, so that months with more frequent data points would not
be weighted disproportionately heavy. Although chl a_ has varied con-
siderably from year to year no trend of either increasing or decreasing
algal biomass in the lake is apparent. The cause for such variations
in spite of stable pregrowth P levels is not related to year-to-year
variation in inflow P, but probably to internal mechanisms of P supply
as will be discussed in later sections. In association with algal
biomass,Secchi disk depth measurements have also not changed. The mean
summer season Secchi measurement remains at 3.3 m, in fact the greatest
Secchi disk value recorded is 6.1 m - a pre-diversion value.
^J. Staley, Dept. of Microbiology, Univ. of Washington, Seattle,
37
-------
150 -
0)
c
o
r-
vt
0)
o
(U
Ou
0>
U
d)
O.
v
\ Blue Green /
x Algae '
Nutrient
Diversion
25
\/
O
1964-
1965
Figure 5.
1970
1971
1972
1973
1974
1975
Mean concentrations in the photic zone (usually top 3 m) of
growing season chl a_ (Mar-Aug) and winter (Dec-Feb) total
phosphorus and nitrate nitrogen relative to pre-diversion
1965 levels. The 1965 levels were: chl a_ 6.5 yg/2, (actually
a mean of 1964 and 1 65 data), total P 31 ugA and NOo-N
390 yg A. The % blue green algae of the total phytoplankton
volume was compared against the pre-diversion mean for June-
October in 1965 and July-Oct in 1964 (67.5%).
38
-------
2
Oxygen deficit rate has remained right at 0.05 mg 02/cm -day since
diversion, a little more than the pre-diversion rate, but with year-to-
2
year variation no more than a few yg/cm -day.
The one variable that seems to have changed significantly is the
fraction of the growing season (in this case June through October)
Phytoplankton that is composed of blue green algae. This value, deter-
mined from surface and/or 1 m depth samples, has been consistently below
that of the 1964-65 pre-diversion period, and except for one year has
been markedly less. Including all years the average is 46% below the
pre-diversion level.
The lake can clearly be considered as mesotrophic with respect to
algal biomass and productivity. However, nutrient content and oxygen
Deficit would suggest a lake at least on the border between mesotrophy
and eutrophy (Table 12). From the standpoint of Carlson's (1974) tech-
nique it falls into the 40-50 numerical range based on total P, chl a.
and Secchi disk depth.
Table 12. SUMMARY OF PRESENT LIMNOLOGICAL CHARACTERISTICS IN THE PHOTIC
ZONE (EXCEPT OCYGEN DEFICIT).
Parameter
Chl a
Primary Production
Total P
Ortho P (photic zone)
NO.+NO.-N
«J c.
Oxygen Deficit (15-30 m)
Annual Mean
3
4.0 mg/m
494 mg C/m2/d
24.4 yg P/l
7.9 yg P/l
180 yg N/l
Growing season* or
Winter Means**
6.0 mg/m
700 mg C/m2/d*
31 yg P/l**
12 yg P/l**
275 yg P/l**
0.049 mg 02/cm -day
*Growing season is March to August
**Winter is December to February
1
Carlson, R., McGill Univ., Montreal, Canada, unpublished manuscript.
39
-------
WATER COLUMN PHOSPHORUS
The cycle of P in the epilimnion of Lake Sammamish is somewhat
variable from year to year, but generally concentrations of both total
and ortho P are maximized at water column turnover during November -
December (Fig. 6). Levels usually less than the maximum persist during
winter (Jan. - Feb.). Ortho P is depleted to a slightly greater extent
than total P in spring and summer due to the conversion of P into algal
biomass. The concentrations of ortho P reached near undetectable levels
in the summer during some years, e.g. 1974.
The post-diversion annual mean epilimnetic content of total P is
about 24 yg/1 and ortho P 8 yg/1. The winter (Dec. - Feb.) mean for
total P is 31 yg/1 and 12 yg/1 for ortho P. One can see from Fig. 6
that 1974 represents a rather typical year with respect to mean values
of the two forms of P. The annual mean total P content for the entire
water column, which includes the anaerobic hypolimnion in summer, is
36 yg/1.
The question of the stable 30 yg/1 winter concentration in Lake
Sammamish in spite of changed P loading is of extreme interest and has
been a major thrust of this research. Intensive monitoring (weekly) of the
turnover period during October through December was undertaken by Rock
(1974) to approach the problem.
The lake is stratified from mid May to mid November. Portions of
the hypolimnion are anoxic from late July or early August until turnover
(Fig. 7).
The pattern of D.O., total P and total Fe are shown in Figures 8
and 9 during 1972 and 1973. Note the inverse relationship between D.O.
and Fe and D.O. and total P. Although concentrations in the hypolimnion
seem potentially capable of at least doubling the epilimnetic concentra-
tion at overturn it does not happen presumably because once the hypolim-
nion becomes aerated and Fe is oxidized much of the hypolimnetic P is
resedimented. The high levels of Fe and P were not observed in 1973 as
40
-------
o?
*»x
CD
\Sl
o
o.
to
o
o.
10 -
M
M
N
Figure 6. Total and ortho phosphorus concentration (mean of epilimnion) at Lake Sammamish
central station in 1974.
-------
ro
0.
UJ
30
I f M A M J
Figure 7. Oxyqen isopleths in Lake Sammamish for 1973.
I ' A ' S "
-------
D
wo-\
10-
o-J
D
100 H
0J
10-
EPfLIMNION
1
11
:::
HYPOLIMNION
0-
Figure 8.
..-.-.
m
:':
I
m
m
1
m
1
I
:
e»
«SS
?:?}
;-
1
9-30 10-8 10-22 11-2 11-11 11-18 11-25 12-2 12-9
Dissolved oxygen, total phosphorus, and total iron concentrations in Lake
Sammamish for the 1972 turnover period (from Rock, 1974)
t.o
1
p
X,
M
1.0
-0.5
-------
n
100
£50
10
0J
D
100
50-
10
Ss-
8-1
EPILIMNION
'*':
1
1
HPYOLIMNION
IvM
1
s
;;$
9-28 10-17 10-26 11-2 11-9 11-17 11-25 12-5
1.0
0.5 |5
1.0
0
-0.5 E
Figure 9. Dissolved oxygen, total phosphorus, and total iron concentrations in Lake
Sammamish for the 1973 turnover period (from Rock, 1974)
-------
in 1972 because the 0,0. content began to increase earlier (compare
early November values) than in 1972, Thus, earlier aeration of the
hypolimnion initiated earlier oxidation and sedimentation rate increase
in 1973.
This efficient sedimentation of hypolimnetic P as a result of
aerating a previously anaerobic water mass may be the stabilizing force
for the water column P concentration. Thus, added P to the lake, par-
ticularly if added to the hypolimnion during the warm stratified period,
could simply be removed through the sedimentation process and never be
available for spring-summer algal growth. In fact, the inflow from
Issaquah Creek probably does not enter the epilimnion, but rather dives
to the lower part of the metalimnion as shown by a dye study (Moon,
1972).
There has been an apparent change in the peak water column P con-
O
tent after diversion, that is, the total amount of P in g/m . Figure 10
shows how these peak values have substantially decreased in the post-
diversion years 1972-73. Krull (1973) suggested that the high pre-diver-
sion peaks were related to the rate of water column breakup at turnover,
but Rock (1974) showed that was probably not the total cause for the
differences in Figure 9 because turnover or temperature decline rates
for 1964, 1965 and 1973 were similar. However, he did hypothesize that
O
the difference in calculated P retention before (0.848 g/m -yr) and
after (0.489 g/m2-yr) diversion was the result of sedimentation differ-
ences, because all other evidence points to the release rate from sedi-
ment remaining constant. He argued that the excess pre-diversion sedi-
mentation could have resulted in a build-up of loose particulate matter
on the lake bottom and thereby was supplying the amount of P necessary
to explain the higher pre-diversion peaks (Fig.10) through mixing.
PHYTOPLANKTON
Phytoplankton biomass shows a peak in the spring composed primarily
of diatoms. The dominating genera during winter and spring are Melosira
and Ste^hanodisous. During the summer and fall, Fragilaria, Synedra.
45
-------
6
0
o o 1964
D o 1965
1972
«1973
D
I
\!\
I \
I
ii
;;
11
II
D
Figure 10. Water column phosphorus content for turnover
period (after Krull, 1973}
46
-------
MelosIra, Rhizosalenia and Asterionella are the major diatoms. The
blue-green algae are comprised predominantly of Aphanocapsa, Microcystis,
Coelosphaerium. Anabaena and Gomphosphaerium. In 1973-74 the appearance
of Aphanizomenon has been less pronounced than earlier while the abund-
ance of Lyngbya has increased. Predominant chlorophyseans are Oocystis,
Spjiaerocystis. Closteriopsis, Chlamydomonas and Staurastrum. Also pre-
dominant in the phytoplankton of the lake is the chrysomonad Mallomonas.
The pattern of phytoplankton biomass is one of typically the
largest concentrations occurring in the spring - the diatoms (Fig. 11).
During some years biomass increase begins early (March) while in other
years the increase is delayed until April. However, in all years the
chl a_ content has subsequently declined to 5 yg/1 or less. Thus,
although the composition of blue green algae comprised from 15 to 69%
of the June - October biomass in post-diversion years (1970-75) the
abundance of blue greens is not great.
The maximum chl ai content shown in Fig. 11 was 28 yg/1 in 1971.
That equals the pre-diversion maximum in 1965. However, the highest
value was in 1975 - 37 jig/1 in mid April.
The growing season mean chl a content is 7 yg/1 and that value has
remained rather stable during the 6 year study. However, growing season
Productivity Can vary as much asr+ 100% from year to year (Table 13).
Curiously, the years with low mean productivity are also the years when
tne diatom bloom has started in early March. Productivity was not
measured by Metro by similar methods so comparison of normalized (for
47
-------
-P.
00
ro 30
E
£ 25
oi
_J
jJ 20
X
Q.
O 1^
CC I0
O
-J
o 10
5
0
J
Figure 11.
M
M
0
N
D
Temporal variation in chlorophyll a_ (integrated means for photic zone) in Lake Sammamish,
1970 through 1973 (1970-71 data from Emery 1972)
-------
Table 13 ANNUAL AND GROWING SEASON MEANS OF PHYTOPLANKTON CHLOROPHYLL
a. (WEIGHTED MEANS FOR THE EUPHOTIC ZONE) AND DAILY RATE OF
PRIMARY PRODUCTIVITY IN 1970-1974.
Lake Sammamish
Year
1970
1971
1972
1973
1974
Average
Chlorophy
(mg/m3)
Yearly
5.7
6.6
4.3
4.0
6.0
5.3
11 a.
Growing
Season
7.7
10.9
4.8
4.7
6.8
7.0
Primary
(mg
Yearly
711
467
799
496
789
652
Productivity
C/nr-day)
Growing
Season
899
575
952
545
904
775
'ightjvalues before and after diversion is not possible.
Transparency of the water column has been even more stable and less
altered after diversion than chl a.. As Fig. 12 shows the mean summer
value has remained very constant around 3.5 ± 0.2 m. Interestingly
enough an August value taken in 1913 by Kemmerer, et al. (1923) showed
3.3 m.
ZOOPLANKTON
Vertical net hauls were collected at station 612 (see Fig. 3) at
frequencies varying from twice weekly to once per month with the least
frequency at periods of low reproductive activity. Relatively little
2°oplankton work had been done in Lake Sammamish prior to 1970. The
first recorded zooplankton haul in Lake Sammamish was made by Kemmerer,
et al. (1923) in 1913. Table 14 shows the results of this vertical net haul
(18 m) from August 13, 1913, versus the results from a net haul (25 m)
on August 14, 1972.
49
-------
0
I
0
I
-- 1
h-
a.
S 4
64 65 70 71 72 73
WINTER
64 65 70 71 72 73
SPRING
u
1
E 2
x
f 3
O.
LU
0 4
5
-
-
-
"
-
i it ii i i ii ii i
u
1
2
3
4
5
-
-
^^M
^^i
-
-
-
U U 1
, ,
,
64 65 70 71 72 73
SUMMER.
64 65 70 71 72 73
AUTUMN
Figure 12. Mean seasonal variation in Secchi disk readings for Lake
Sammamish. Dashed line represents mean for the plotted
data (1970-71 data from Emery, 1972)
50
-------
Table 14. COMPARISON OF AUGUST NET HAULS 1913 vs 1972 (NUMBER PER m3).
1913
Bosmina
Piaphanasoma
Cyclops
l£jschura
PJjPtomus
Mastjgocerea
Nothpjka (=Kell
P^lyarthra
Cladocerans
393
905
Copepods
5780
9805
-
Rotifers
393
Icottia) 113
113
1972
Bosmina
Diaphanasoma
Cyclops
Epischura
Pi apterous
51
331
3268
151
6573
0
0
0
The two samples are quite similar except for Diaptomus being absent
in the sample collected in Lake Sammamish in 1913, and it is the most
common zooplankter now in the lake. The difference in numbers in the
other species with only 1 sample as a comparison is not significant as
Population peaks may be shifted several weeks from year to year.
The first major zooplankton study of Lake Sammamish was a portion
of the METRO study of June, 1964 to December, 1965 (Isaac, et al . , 1966).
pigure 13 is a graph of total zooplankton (number per m3) in 1965 versus
19?2. The total numbers are quite similar and any difference probably
can be explained on year-to-year or sampling variation. For example,
the major peak in both years was due to a rotifer bloom, which reached
Teo.OOO/m3 in 1965 and only about 90,000/m3 in 1972. This could be
explained by postulating that the 1965 sample was taken at the height of
the bloom while the 1972 sample was a week on either side of the peak.
Species recorded during the 64-65 study are the same as those found in
tne lake at the present time.
51
-------
to
'
ro
en
ro
OC
LJ
Q.
UJ
CD
80-
70
^ 60
50
40-
=5 30
20
10
0
1972
/
/
/
'1965
M
M
N
Fiqure 13. Total zooplankton in Lake Sammamish, 1965 vs 1972
-------
Table 15 is a list of the zooplankton species that have been col
lected in Lake Sammamish during this study. Figure 14 shows the per
cent numerical occurrence of the three major groups of zooplankton
throughout the year and the most abundant is the copepod Diaptomus
ash 1 andi.
Table 15. ZOOPLANKTON SPECIES IN LAKE SAMMAMISH.
Copepods
*Diaptomus ashlandi
*Epischura nevadensis
*Cyclops blcuspidatus
CJadocerans
*Daphnia thorata
*D. schgilderi
*Bosmi na longirostris
*Diaphanasoma leuchtenbergianum
Leptodora kindtii
Scapholeberis kingi
Rotifers
*Kellicottia longispina
j<. bos ton iens is
*Polyarthra sp.
Kerate11 a cochlearis
J<. Quadrata
*Conochilus unicorm's
Collotheca mutabilis
C^. pelagica
Notholca squamula
Ploesoma hudsoni
Gastropus sp.
Synchaeta sp.
Trichocerca sp.
Filinia sp.
v
indicates most common species.
53
-------
en
CLADOCERANS
Figure 14. Percent composition of zooolankton categories, 1972
-------
SEDIMENTATION
Sediment trap data are summarized in Table 16 while seasonal
events are illustrated in Figures 14 and 15. The trend for increasing
sedimentation with depth was due to relatively coarse and dense alloch-
thonous material transported into the hypolimnion which settled rapidly
and tended to be intercepted by deeper traps. Also in late autumn sedi-
ment released P continuously coprecipitates with Fe-hydroxy floes as
the thermocline erodes and upper hypolimnetic waters are reaerated.
The Fe floc-P material was collected by deeper traps although at over-
turn some of it was also collected by shallow traps. The littoral
(Station 3) and sub-littoral (Station 1) traps collected only about 70%
as much as traps suspended at the same depths in pelagic regions (Sta-
tions 2a and 612) mainly because they were somewhat sheltered from aloch-
thonous inputs during times of high stream flow.
With regard to seasonal events (Figures 15 and 16), high sedimen-
tation rates occur during spring, late autumn and winter. Microscopic
examination of freshly collected material indicated that high sedimen-
tation rates in the spring resulted from settling out of the diatom
bloom which occurred in March-April. During summer the settled material
was dominated by blue-green algae detritus. In situ experiments (Birch,
1976) indicated that this material decomposes much more rapidly than dia-
tom detritus, a factor believed responsible for the relatively small
collections in traps at that time.
In late summer an unusual organism tentatively identified as a
bacterium Metal!ogeniurn Personatum appeared in abundance in the sample
from 26 m and in the autumn in the 15 m samples as well. This organism
has also been found in hypolimnotic water samples from Lake Washington
(Shapiro et aj_., 1971). The radiating tubules of this organism are
coated with brown Fe and Mn containing compounds (Libby, 1972) to which
P may be adsorbed. The significance of this mechanism in Fe, Mn and P
cycling in Lake Sammamish has not yet been clearly identified.
Immediately after thermal destratification (12/4/74) there was a
55
-------
spectacular increase in the rate of sedimentation due to the chemical
effect of reaerated bottom waters. Most of the Fe which had been
released following reduction of surface sediments during the anaerobic
period was reoxidized and precipitated as a floe. Sediment released P
is sorbed on to these Fe (III) hydroxy compounds and co-precipitated.
Since sediment released P, until overturn, was contained in lower hypo-
limnetic waters and below the photic zone its contribution to algal
nutrition in Lake Sammamish is considered minimal.
During winter high sedimentation rates of total material and of P
were due to increased input of allochthonous material resulting from
high stream flows. More material, particularly total sediment, was
deposited at Station 2A becuase it is closer to the major inlet, Issa-
quah Creek. High sedimentation rates of P during winter means that a
significant proportion of the external P income, most of which is in-
troduced during winter, is rapidly deposited and effectively removed
as a source for spring-summer algal nutrition.
A comparison of primary production and sedimentation into the 7 m
trap (Station 612) during the productive season (March-October) indicat-
ed that only a very small proportion (%15%) of this produced material
settled out of the photic zone as detritus. This indicates that nutri-
ents are efficiently recycled in the trophogenic zone during the stra-
tified period. The growing season rates of sedimentation were 43,
4.8 and 0.67 mg/m day for C, N and P respectively.
56
-------
14
12
1O
O)
O
H
<
z
LU
O
UJ
(A
OVERTURN
r
I
I
I
STN 612
Stn 2A
MAMJ JASONDJ FMAMJJ A
1974 1975
TIME IN MONTHS
Figure 15. Phosphorus sedimentation rates in Lake Sammamish determined
at Stations 612 (23 m depth, 26 M Trap) and 2A (30 m depth,
27 m Trap).
57
-------
o
-------
SEDIMENT-WATER INTERCHANGE OF P
During the course of the column experiments there was a doubling in
ambient lake temperature from 8° C in early spring to 20° C in summer and
back down to 10° C in autumn. The column DO content decreased rapidly;
from 8-11 mg/£ to 2-3 mg/5, in 30 days, with a subsequent slower decrease
to as low as 0.1 mg/H in some columns, DO depletion in transparent
columns was much slower due to photosynthesis,
In all cases the loss of DO was accompanied by increases in P and
Fe content. Alkalinity and pH responded consistently to the onset of
anaerobic conditions. The opaque column's pH decreased to 6.5 and alka-
linity increased from 32 to 39 mg/£ as CaCOg. Subsequently pH increased
to 6.8-7.0. There was an initial increase in pH in the transparent
Columns due to photosynthesis, but then they attained levels similar to
the opaque columns.
Rates of release of Fe and P were determined by considering only
positive slopes on the concentration time curves and only after DO had
decreased to levels of 2-3 mg/£. Values were excluded where problems of
^2 intrusion were known to have occurred.
As representative results from column experiments Figures 17-19
are presented. Figure 17 shows changes in total P, total Fe and DO in a
column initially purged with N2 to effect early DO depletion. That pro-
cedure showed clearly that the time of incubation was indeed short before
Fe and P began to release. After day 20 total and soluble Fe increased
at rates of 25 and 12 mg/m2 day, respectively, and total and soluble P
also increased at 3.9 and 4.3 mg/m2 day, respectively.
Figure 18 shows results of an opaque column allowed to lose its DO
naturally, which appeared nearly as fast as the previously described
experiment in which the column was purged. The rates in this experiment
Were 3.15 and 3.22 mg/m2 day for total and soluble P, respectively. Total
and soluble Fe increased at 10 and 5 mg/m2 day. Note the only slight
increase in P following the addition of P at 3 times the lake loading rate
On day 120.
59
-------
Although the DO decreased in the transparent column more slowly than
in the opaque columns, release rates were as great as in the opaque
columns once DO became low (fig. 19}. It took 90 days to reach a concen-
tration of 0.5 tng/5, in two such transparent columns, one at 7 m and the
2
other at 10 m depths. Total and soluble P released at 4,4 and 4.1 mg/m
day once DO was depleted to the low value and total and soluble Fe at 36
2
and 32 mg/m day, respectively.
For comparison the observed changes in P and Fe in the hypolimnion
are shown in Figures 20 and 21. During 1974, P, Fe and DO were determined
at depths of 23-25 and 27 m at Station 612 - the 28 m Station, The re-
sults in Fig. 20 and 21 were calculated for a 5 m water column at that
deep station. As shown P increased gradually after day 13 (July 8} but
not until day 102 (Sept. 10) did the increase appear marked. The rates
2
of release for the period were 5.2 and 4.9 mg/m day for total P and
2
soluble P, respectively, and 25 and 12 mg/m day for total and soluble
o
Fe. The DO loss rate was 0.69 mg/m day which was the same as in one
opaque column - the other two opaque columns showed rates of 0.54 and
2
0.58 mg/m day.
Interestingly enough the maximum P value reached, 290 yg/8, (1799
2
mg/m ) by day 165 (Nov. 13);was a concentration comparable to most maxima
in opaque columns during summer, The maximum Fe content occurred the same
time as that of P - 3.4 mg/a (2.97 g/m2).
There was rather close agreement in the P release rates in eight of
n
nine experiments in 1974. The values ranged from 2.6 to 4.4 mg/m day
for total P with a mean of 3.4. The one high value of 12.6 mg/m2 day
was not included because it was thought excessive mixing during sampling
caused that high value. There was no difference in release rates between
transparent or opaque columns once DO had been reduced to 2 mg/SL or below.
Table 16 shows a comparison of all values of release of P by various
investigators in Lake Samrnamish. The only values that seem excessively
out of line are those of Monahan (1974) in the spring which were later
shown to be accounted for by excessive mixing induced during sampling.
60
-------
UJ
>
0
20
40
60 80 100
TIME IN OflYS
120
140
160
DflYS
DflYS
VERSUS DO
VERSUS TP
OflYS VERSUS TFE
10 VflLUES
10 VflLUES
10 VflLUES
Figure 17. Changes in Total Iron and Total Phosphorus in an Opague
Column from 8 August to 16 December 1974 (The column was partially de-
oxygenated by nitrogen aeration). From McDonnel (1976).
61
-------
3.5
20
40
60 80 100 120
TIME IN DRYS
140 160 180
DfiYS
DRYS
VERSUS 00
VERSUS TP
^ DflYS VERSUS TFe
Figure 18. Changes in Total Iron and Phosphorus for an Opague
Column from 27 July 1974 to 8 January 1975. (KH-PO. - 0.236 mg/1 was
added on day 120). From McDonnell (1976). ^
-------
20
40
60 80 100
TIME IN DflYS
120
140
160
DflYS
DflYS
OflYS
VERSUS DO
VERSUS TP
VERSUS TFe
9 VflLUES
9 VflLUES
9 VflLUES
Figure 19. Changes in Total Iron and Phosphorus in a Transparent
Column from 17 July 1974 to 8 January 1975. From McDonnell (1976),
63
-------
LflKE SflMMflMISH
**STflTION 612«*
DEPTHS [2O to 2 7m]
0 20
29 MAY J
40 60 80 100
TIME IN DflYS
120 140
160 180
27 NOV 7^
DflYS VERSUS DO
DflYS VERSUS TP
DRYS VERSUS TOP
14 VflLUES
14 VflLUES
14 VflLUES
Figure 20. Changes in Total and Soluble Phosphorus in the Lake Hypolimnion,
Station 612, from 29 May to 27 November 1974. (Day 140 illustrates the
commencement of the annual fall overturn). From McDonnell (1976).
64
-------
LflKE SfWIRMISH **STRTIQN 612**
0 20
2k Jul 74
60
TIME IN DflYS
80
100 120
12 Nov 74
DflYS VERSUS DO
DflYS VERSUS TFe
DflYS VERSUS DFe
8 VflLUES
8 VflLUES
8 VflLUES
Figure 21. Release of Total and Soluble Iron in the Lake Hypolimnion,
Station 612, from 29 May to 27 November 1975. From McDonnell (1976).
65
-------
Table 16. A COMPARISON OF P RELEASE RATES FROM in situ COLUMN EXPERIMENTS
AND OBSERVED HYPOLIMNETIC CHANGES IN CONCENTRATION" DURING 1972-1975.
lotal P 2 Soluble P
mg/m day
In situ columns
McDonnell (1975)
hypolimnion Station 612
1974
2.6-4.4
5.3
1.9-4.3
4.9
in situ columns 4. 0-6.0 (fall)
Monahan (1974) 17.0-26.0 (spring)
In sjtu columns 2.8 (summer) 2.6
Horton (1972) 6.2
The effect of adding P to some columns (see Fig. 18) with respect to
the maximum concentrations attained without addition suggests a situation
of equilibrium and sediment buffering in the lake. P attained a maximum
of 270 yg/ji following P addition to one column, while two other columns
and the deep lake showed maxima between 260-280 yg/Ji. Overall maxima
ranged between 100-300 yg/£. Near 300 ug/£ may represent an equilibrium
value for the anaerobic bottom water assuming adequate contact with bottom
sediments occurs.
66
-------
MODELLING
Temperature profile output from the model is presented in Figures 22a
through 22d for days when corresponding observed data were available
for the years 1970-1973. These figures show that the simulated data
give a close fit to the observed during this period.
Figure 23 gives the simulated mixing depth for March-July of 1971
and 1972. When this is compared to the epilimnion chl ^ values given
in Figure 24 it may be noted that the earlier peak for the spring diatom
pulse related closely to the consistently greater mixing depths the
model had simulated. This output has caused us to reevaluate the hypoth-
esis that the years of early blooms also have more light available to
mixing diatoms because of a shallower mixing depth. The model suggests
that mixing, which effects continued nutrient entrainment, may be more
important than light availability. Another question raised about the
1971 spring diatom bloom concerned the source of P04-P being utilized.
The P04-P pool, supply thought to be available was not sufficient to
create a bloom of the magnitude of that which actually occurred. The
model was able to simulate a bloom of similar magnitude only when fed
P04-P from an external source. Diffusion from the lake bottom was
eliminated as the major source because of the enormous diffusion coeffi-
cient which the model indicated would be required to effect the transfer.
Littoral release and surface runoff are other possible sources to the
P04-P pool to consider. The source from surface runoff is quite possi-
bly asolution in the 1971 case since there was prolonged rainfall and
subsequent runoff shortly before the spring bloom began.
Another example of information generated by the model is the fact
that a sensitivity analysis of the model parameters showed that if the
light intensity required for maximum phytoplankton production was held
constant, as first hypothesized (Tang, 1975), the sensitivity of the
model to that parameter was much greater than could be expected in the
lake. Accordingly, it was felt that the function in that form was not
an accurate representation and was therefore changed to its present
67
-------
6 8 10 6 8 10
(°C)
6 8 10 6 8 10 12
M
0)
0)
°
20
25
30
«
f
-» - \J
5
10
15
20
25
30
«
r
» \j
5
10
15
20
25
30-
-. u
5
10
15
20
25
30
<
/
March ..27
April 11 April 25 May 9
10 is
0
10 15 20 25
10'
w
20
25
30
5
10
15
20
25'
30.
June 12
July 14
observed
simulated
, Figure 22. Simulated and Observed Temperature Profiles for Lake
Sammamish (1970).
68
-------
57-9 .579
5 7 9
10 15
0
y-
10
en
S is
0)
O
5 20
P.
0)
Q
25
30
*
.
*
, £
0
5
10
15-
20
25
30
t
0
*
t
5
10
15
20
25
30
.,...».
«
*
i ;
10'
15
20
25
30
s
^r
'
*
March 20
April 10 April 24
May 21
10 15 18
10 15 20
0
10
.'s
CO
Vi
01
i_)
0* , _
S 15
& 20
25
30
15
20
25
30
'June 30
July 20
observed
simulated
Figure 22-b. Simulated and Observed Temperature Profiles for Lake
Sammamish (1971).
69
-------
0
5
10
to 15'
0)
5 20.
fi
4J
a
0)
Q
25
30
6
8 10
...<_ Q
5
10
15
20
*
0
25
i
t
30
i
6 !
.
3 10
-TT ' ' 0
5
7
f 10
15
20
25
j 30
T
6 8
i
.
( C)
10
-* ~ o
5
10
15'
20
25
30'
6 8
*
10
.r*- 0
7
/ 5
[
10-
15
20
25
30
5
9
10 15_
r
*
*
J
/
March 11 April 4 April 18
April 27
May 25
10 15 20
15 20
0
10
m
M
Q)
20
25
30
10
15
20
25
30
10 15 20
10
15
20
25
30
June 2
June 13
July 10
observed
simulated
Figure 22-c. Simulated and Observed Temperature Profiles for Lake
Sammamlsh. (1972)
70
-------
689
0 ( ' f
10 10
(0
M
(X.
0)
p
0
6 8 9 11 13
5 10 15
5 15 15
20 20
25 25
30 30
March 31 April 27
10
15
20J
25
30
5 10 15 18
0
5
10-
15
201
25'
30-
May 11
May 28
0
5 . 10 15 20 5 10 . 15. 20 22 5- 10 15 20
" " - i " ' i i > ' '." ' ' n . < t ' - t - i '- i I, I « r . i i r r
10
15
5 20
-------
30
a.
20
CL.
UU
O
»,»
» «%
MARCH APRIL MAY JUNE JULY
Figure 23. Calculated daily mixing depth in Lake Sammamish during
March-July, 1971-72 (from Tang, 1975).
72
-------
20
10
30
20
10
^ 0
J 15
«" 10
0
15
10
1970
1971
1972
1973
MARCH APRIL
MAY
JUNE
Figure 24. simulated and observed chl a.
content in the epilimnion of
Lake Sammamish during the
springs of 1970-73.
JULY
simulated
observed
73
-------
form in which maximum production occurs at fifty per cent (50%) of the
incident light level (see PHOTO, Table 8 ).
Figure 25 shows a comparison of simulated to observed concentra-
tions in the PO.-P pool in the epilimnion. It can be seen that,
although the model does not simulate the PO^-P pool exactly from year to
year, some of the simulations are reasonably accurate and well within
experimental error. It should be noted that no attempt was made to
calibrate the model to provide the best fit possible.
MODEL OF THE NITROGEN CYCLE
The above-described multiparametric lake model considers both light
and phosphorus as potentially growth-limiting factors to Lake Sammamish
phytoplankton. There is some suspicion, however, that nitrogen may well
have some effect on growth rates, particularly in the event of increased
loading of P. In order to determine when nitrogen might be the limiting
nutrient and to what extent its role would be in controlling productivi-
ty, and biomass, a nitrogen model has been developed which approximates
the concentrations of the two major inorganic forms, ammonia and
nitrate(Noah, 1976). These forms of N are not only the most abundant
forms of N, but also the most significant in terms of plankton algal
growth rate.
The following major processes can be evaluated by the model on a
daily basis: (1) assimilation of inorganic N by phytoplankton, (2)
denitrification of nitrate, (3) fixation of ammonia from molecular
nitrogen, (4) lake inflow and outflow concentrations, (5) mineraliza-
tion of ammonia from the detritus pool, (6) nitrification of ammonia to
nitrate, (7) reduction of nitrate to ammonia, (8) input of N in rain,
and (9) sediment release of ammonia.
In order to evaluate some of these rate processes, it was necessary
to determine dissolved oxygen concentrations as a controlling factor.
Rather than entering a complex array of concentrations over time and
depth as exogenous input, an oxygen model was developed. The reactions
74
-------
10
0
10
o?
3.
£ 0
a. 20
o
a.
o
I
o
10
0
10
0
1970
*
1971
1972
1973
MARCH APRIL MAY JUNE JULY
simulated
observed
Figure 25. Simulated and observed ortho-P concentrations in the
epilimnion of Lake Sammamish during the springs of 1970-73,
75
-------
on which the system depends are (1) production of oxygen via photosyn-
thesis, (2) consumption according to a biochemical oxygen demand func-
tion, (3) reaeration on the lake surface, and (4) benthal oxygen demand
at the sediment layer.
Coordination and information transfer during the development of
the nitrogen model and the sensitivity analysis of the phosphorus model
revealed an interesting problem. Both models used a Michael is-Menton
function to limit phytoplankton growth. It was discovered that the
Michael is-Menton model has an apparent anomaly at limiting concentra-
tions when used in a difference equation model.
As described by Mar (1976)' the anomaly occurs at low concentrations
of nutrient and large biomass when the equation used is
C
y = us ^ K + S ^ (equation 1)
q
where v = growth rate = Productivity/unit time = AB
3 biomass B
S = concentration of nutrient
PS = growth rate of infinite concentration
K = 1/2 saturation constant
If S » K then y approaches yg and when S « K y approaches zero. It
is at these latter concentrations that the problem occurs. If S«K
then equation 1 reduces to
or AB = u {j
This equation says that the change in biomass (AB) for a given concen-
tration of substrate is a function of how much biomass in present.
This means that no matter how little substrate is available the amount of
Manuscript in review, ASCE.
76
-------
growth it can support can be increased by increasing the amount of
biomass assimilating it. This is clearly incorrect and in a difference
equation model causes an overestimation of growth if the nutrient con-
centration approaches zero.
A new form of the equation has been proposed by Mar (1976) which
is of the form
As S approaches zero this equation reduces to
AB , / S
T + ys (T
S
ar > B
This means that growth is dependent only upon the concentration of sub-
strate at these values while at values where S » oB. The equation
reduces to the same form as the Michaelis-Menton function.
This new equation has been incorporated in the nitrogen and phos-
phorus models with the result output of chl a_ for 1972 as shown in
Fig. 26. The maximum growth rate was also reduced by 75% to delay the
ch/s increase.
Comparison of this information to that provided by the earlier
phosphorus model (Fig. 24) shows a much closer fit of the simulated
data to the observed data. Because of its more accurate simulation
the new equation is considered a better approach to practically deal
with modelling of nitrogen and phosphorus dynamics in lakes.
The model has not shown nitrogen to limit phytoplankton growth
during the March-to-July period and under the existing P loading regime.
Nitrogen may indeed prove to be temporarily limiting growth rate, or at
least have an interactive role, in late summer or early fall and/or under
higher P loading rates. Such extended application of the N model was
beyond the scope of this project, however. The formulation, parameter
evaluation and ^validation of this model was described by Noah (1976).
77
-------
15
e*
"g 10
C
Observed
Simulated
Simulated with
modified growth model
MARCH
APRIL
MAY
JUNE
JULY
Fig. 26. Simulated and observed chl a_ content in the epilimnion
of Lake Sammamish during the spring of 1972. Note
relatively better simulation of the spring increase
in chl a_with the modified growth model.
78
-------
SECTION 9
DISCUSSION
HYPOTHESIS FOR STABILITY
Lake Sammamish has shown no detectable changes in most of the per-
tinent lake quality variables in the six years since one third of the
phosphorus income was diverted. Water clarity has remained unchanged
since diversion - the mean summer Secchi disk depth is 3.3 m. Mean
winter total phosphorus concentration has remained at about 30 yg/1 +
2 yg/1 before and after diversion. Chlorophyll a_, averaged over the
growing season and throughout the photic zone, has remained at about
7 yg/1 +_ 2 yg/1. The only variable that may have changed is the frac-
tion of blue green algae, which shows an average decrease of 46 percent
compared to pre-diversion (1964-65) values.
The apparent stability in the trophic state of Lake Sammamish, in
spite of the watershed settlement by European man beginning over 100
years ago and the diversion of about one-third of its phosphorus income
in 1968, can be largely attributed to the sediment-water interchange
mechanism for phosphorus. As much as three-fourths of the phosphorus
increase into the water column during the stratified period in May-
November could potentially be released as soluble P from the anaerobic
sediments as shown in plastic column experiments. The moderately high
and constant phosphorus content in the sediment profile along with the
indication that the lake's hypolimnion has been anoxic in late summer
for a long time suggests that the P release rate has always been large
in Lake Sammamish. Water column balances of P for the turnover period
show a greater maximum before diversion (5-9 g/m ) than after diver-
sion (<2 g/m2). Since there has been no change in the winter water
column content of P (31 yg/1) since diversion, the excess at turnover
in pre-diversion years 1964 and 1965 was most probably lost to sedi-
mentation (Rock, 1974).
79
-------
The release and sedimentation mechanism is thought to be largely
controlled by iron. Phosphate is released in proportion to iron as
oxygen content reaches low levels and is eventually exhausted in late
summer. Just prior to lake turnover the water column content of dis-
solved P is at the annual maximum, most of which is in the hypolimnion.
Instead of these high concentrations being maintained throughout the
winter, much of that P is resedimented when turnover restores the aerob-
ic conditions and P decreases in proportion to Fe. Presumably much of
this P is complexed with iron. The relatively constant winter concentra-
tion of 30 yg/1 following the fall turnover appears to be residual, the
excess having been largely removed by the complexation with iron. Pre-
sumably the additional phosphorus income before diversion had not in-
creased the pre-growth period residual P concentration because of the
ample quantity of iron available for P complexation and sedimentation -
up to 0.9 mg/1 Fe at maximum anaerobiosis. The large prediversion
quantities of P that appeared at fall turnover then could have come from
accumulations of flocculent sedimented P, that is, the quantity of
annual P income removed by Fe over and above the residual P level. The
anaerobic nature of the hypolimnion in this lake then seems to behave
as a greater effective sink for P than if the hypolimm'on were aerobic.
That is, the large quantities of Fe released under anaerobic conditions
would not have been so released in the aerobic state (Fe in the Lake
Sammamish hypolimnion is 10 times that in Lake Washington) and these
would not be available for the precipitation of biologically mineralized
P.
While complexation of hypolimnetic P with iron and sedimentation
subsequent to turnover (Nov. - Dec.) is probably the principal mechanism
controlling the residual level of P, another contributing mechanism
operates during Jan - Feb. To account for the continual decrease in
water column P content during those later winter months, the relatively
large amount of P input during that period (runoff is usually high) is
believed to be rapidly sedimented along with the large quantity of P
reactive colloidal material brought in through the high flows. Secchi
80
-------
disk measurements are lower during that period than any other time
during the year and a large fraction of the material caught in sediment
traps is clay, which together reflect the large contribution from land
runoff.
Additionally, early spring blooms of diatoms remove P from the
epilimnion more effectively than non-diatoms. Years in which early
diatom blooms occur also have lower spring and summer mean P concentra-
tions in the photic zone. This mechanism is not contributing to mainte-
nance of the wintertime 30 yg P/l residual, but may contribute to resis-
tance to increased P income of lakes with low mixing potential.
Interestingly, the years with early spring diatom blooms show one-half
the March - August productivity compared to non-early-bloom years.
Thus, it appears that in the range of P income to Lake Sammamish,
1.0 g P/m-yr before to 0.67 g P/m -yr after diversion, and assuming no
change in the sediment release of P, the P content could have remained
relatively unchanged with only a change in the sedimentation rate. In
further support of this relative stability in that range of P income,
a modified version of a mass balance model (Vollenweider, 1969) can be
used to predict only a slight change in water column content of P (Krull,
1973; Welch, et al., 1973). Phosphorus input, outflow, and exchange with
the sediments were included in the model. The lake was assumed to be two
layered during the stratified period and uniformly mixed during the
remainder of the year, and the water column phosphorus content was con-
sidered to be influenced only by the above mentioned gains and losses;
complex cycling within the water column was ignored. The model con-
sidered sedimentation and release rate separately. The release rate was
assumed constant, while the sedimentation rate was not held constant.
This was the key departure from the Vollenweider model (Vollenweider,
1969) which earlier had shown that Lake Sammamish P should exhibit a
prompt and significant recovery (Emery, et al., 1973). The modified
model indicated only a slight change (7 percent) in phosphorus concen-
tration to a one-third reduction in phosphorus loading.
81
-------
Although the above model allowed the sedimentation rate to vary,
the sedimentation rate varied proportional to the phosphorus concen-
tration in the water column. Therefore, as an another approach to
realistically reflect the lake dynamics the sedimentation rate Rock
(1974) allowed to vary with the phosphorus loading to the lake. The
second modified model was;
dP = M_|yp.nc
dt A V R S
where,
M
S = sedimentation rate = K, ^
P = total phosphorus in the water column
R = release rate
M = input loading rate
A = surface area of lake
KQ = output rate coefficient
K, = sedimentation rate coefficient
The release rate was assumed constant. Data showed that the winter
column concentration of total phosphorus was unchanged from pre-diver-
sion to post-diversion. Under these conditions the model could be
calibrated. At pre-diversion steady state;
where,
D = retention rate of phosphorus in lake,
thus, equation (1) becomes;
D = S - R (2)
and for pre-diversion lake conditions equation (2) becomes;
20,000 - 3,300 kg = 20.000 kg . K _ R
19,8 km2 19.8 km2 ]
82
-------
at post-diversion steady state, equation (2) becomes;
13.000 - 3,300 kg = 13.000 kg . K _ R (4)
19.8 km2 19.8 km2 ]
equations (3) and (4) were solved by simultaneous equations, hence
K = 0.96.
Solution of equation (1) yields the change in steady state phos-
phorus concentration (AP) for a given step change in loading (AM)
P=f (1 -K,) (5)
The solution of equation (5) for the 1968 sewage diversion (AM =
7000 kg P) with K] = 0.96 gives a P of only 0.014 g/m2 in the water
column. This represents a concentration change in the water column of
only 0.5 yg P/l (^2%). Thus, the model indicates that if a decrease
in loading was followed by a decrease in sedimentation rate, the water
column phosphorus concentration should not have been significantly
reduced.
IMPLICATION OF NON-RECOVERY
Although non-recovery would suggest that the sewage diversion pro-
ject was a failure, such a conclusion is completely unwarranted. The
primary objective of the project was to stop the cultural eutrophication
of Lake Sammamish. Although the trophic state in terms of lake concen-
trations was not reversed, the present trophic state does not represent
deteriorating nor intolerable conditions. In fact, the mean summer chl
a level in Lake Sammamish is now about equal to that found in Lake
Washington following diversion (Edmondson, 1972). Further, summer water
clarity, as measured by the Secchi disk has been greater than 3 meters
on the average during all the years investigated, before and after
diversion. Lake Washington on the other hand showed average summer
Secchi disk depths of only 1 m before diversion (Edmondson, 1972) and
now has similar clarity as Sammamish.
83
-------
Although Lake Sammamish did not noticeably improve following diver-
sion of wastewater, the action was no doubt a preventative measure. The
lake may have in fact advanced in trophic status without the 1968 diver-
sion. The lack of response over the past 100 years to the range of P
2
income of 1.0 to 0.67 g P/m .yr tempts one to speculate that future in-
crease may also not affect trophic state. That is probably not tenable
because although most of the above increase in P probably came in the last
10 years. Issaquah, the main source of sewage, has nearly tripled in
population since 1960 and the wastewater from the Issaquah area has doubled
since diversion in 1968. This suggests that the increase in P income to
Lake Sammamish over the past century has been in the "lag phase" and only
very recently could the increase be described as entering the "growth
phase11 of an exponential function. Without diversion, the P income would
not be 0.67 g P/m .yr, or 1.0 g P/m2.yr (the pre-diversion rate), but
2
rather 1.36 g P/m .yr, which would reflect the recent growth in the
Issaquah area since 1968. Such an increase would most probably not all
go into the sediments and the present winter mean content of 31 yg/1 would
surely have increased. This seems especially reasonable since the iron
supply for complexation would not have increased proportionately.
Although Lake Sammamish has remained stable over a range in a P
income of 0.67 to 1.0 g P/m .yr that probably has little implication for
future increases tn Income. The important question now relates to the
expected increase from suburban and urban runoff that can be expected as
the forested areas around the lake, particularly the east side, are reduced
to residential areas. At first approximation the yield of P from such
developed areas in the Sammamtsh watershed do not appear great, but in
light of what has been demonstrated by careful analysis of the P trans-
port in Issaquah Creek (high runoff rates yield high P concentrations)
the question of yield increase from residential development must receive
more careful study if a totally reliable prediction is to be obtained.
The implication of this lake's response to P income manipulation
84
-------
to the management of other lakes is not entirely clear. Certainly one
is left with the overriding impression that lakes with anaerobic hypo-
limnia are more unpredictable than those that remain aerobic. If the
degree of P manipulation in Sammamish is compared with that of
Washington on the Vollenweider graph (Figure 27) it becomes of course
clear that the degree of manipulation was greater in Washington. Yet
from that graph one would have expected a proportional change in lake
trophic state, other factors being equal, since the loading in Sammamish
was decreased to a point near the "excessive" line similar to that in
Washington. Obviously such a proportional response did not occur, thus
one must agree with Vollenweider and Dillon (1974), for anaerobic lakes
at least, that P removal in excess of 50% of the input is necessary to
observe a significant recovery. In support of that Larson, et al. (1975)
have observed an improvement in the trophic state in Shagawa Lake, which
has an anaerobic hypolimnion, as a result of an interception of 70% of
the P income (Fig. 27). However, Shagawa was very eutrophic at a
similar loading as Sammamish (Malueg, et al., 1973). Clearly more of
the incoming P in Lakes Washington and Shagawa ended up in the water
column and available to algae than was the case in Lake Sammamish. The
waste water P entered Lakes Washington and Shagawa directly while
in Sammamish the waste water entered through Issaquah Creek. This
could be at least part of the reason why much of the loaded P to
Sammamish does not end up in the water column but rather in the sedi-
ment. Issaquah Creek carries a rather high load of iron so possibly
much more of the P added to Issaquah Creek was destined for the sedi-
ments of Lake Sammamish than was the case for the waste water P added
to the other two example lakes.
85
-------
oo
1.0 -
\J
2:
OJ
c
(Q
a
_i
to
i o.,-
CO
o
C"
-_*-_
Q.
2
n.m -
Eutrophic ^
^Washington 1964 ^
Mesotrophic i ^^
Sammamish 1965 £ i Mesotrophic ^s'
| 0Washj*x^ comparing the relative
position of three manipu-
lated lakes. Relationship
after Vollenweider (1974).
0.1
1.0
10.0
100
Z/TW (Meters/year)
-------
CONTRIBUTION OF THE MODEL TO UNDERSTANDING
As the results have shown, the model has provided some interesting
information as well as raising questions which have provided impetus and
direction for new or more thorough research in certain critical areas.
As with any modeling effort its major contributions are as a test vehi-
cle for new hypotheses and as a source of direction for research which
will expand understanding.
This model has helped to increase understanding in such areas as
the importance of mixing of Lake Sammamish waters to algal growth, the
effect of light intensity on algal growth, the relative importance of
diffusion in the water column versus external inputs of nutrients and
effects of wind and insolation upon themixed depth of the lake.
Interesting questions have been raised as to littoral and bottom
release as contributors to the PO^-P pool as well as the effects of
surface runoff and atmospheric fallout. The model is available for use
in testing any new theories which are generated in these or other areas
of interest using Lake Samamish as the study site.
87
-------
SECTION 10
REFERENCES
American Public Health Association, and others, "Standard methods
for the examination of water and wastewater," 13th ed., APHA, New
York (1971).
Birch, P., "Sedimentation in lakes of the Lake Washington Drainage
Basin," M. S. Thesis. Univ. of Wash. (1974).
Birch, P. B., "The Relationship of Sedimentation and Nutrient Cycling
to the Trophic Status of Four Lakes in the Lake Washington Drainage
Basin," Ph.D. dissertation, Univ. of Wash. (1976).
Bremner, J. M., "Determination of Nitrogen in Soil by the Kjeldahl
Method," J. Agri. Sci. 52_, pp. 137-146 (I960).
Edmondson, W. T., "Nutrients and Phytoplankton in Lake Washington,"
In Nutrients and_Eutrgphication - Special Symposium, Limnol. and
Oceanog., 1, pp. 172-193 (1972).
Emery, R. M., "Initial responses of phytoplankton and related factors
in Lake Sammamish following nutrient diversion," Ph.D. dissertation,
Univ. of Wash. (1972).
Emery, R. M., Welch, E. B., and Moon, C. E., "Delayed Recovery of a
Mesotrophic Lake after Nutrient Diversion," J. Water Pollut. Cont.
Fed., 45, pp. 913-925 (1973).
Fish, E., "The past and present in Issaquah, Washington," (1967).
Gibbs, C,, Henry, C., and Kersnar, F., "How Seattle beat pollution,"
Water and Wastes Engineering, 9^ pp. 30-40 (1972).
Guttormsen, S., "A comprehensive nitrogen study of Lake Sammamish,"
M. S. Thesis, Univ. of Wash. (1974).
Hansen, H., "Postglacial forest succession and climate in the Puget
Sound region," Ecology, 19^ pp. 528-542 (1938).
Hendrey, G., "Productivity and growth kinetics of natural phyto-
plankton communities in four lakes of contrasting trophic state,"
Ph.D. dissertation, Univ. of Wash. (1973).
Horton, M., "The chemistry of P in Lake Sammamish," M. S. Thesis.
Univ. of Wash. (1972).
Isaac, G. W., Matsuda, R. I., and Walker, J. R., "A limnological
investigation of water quality conditions in Lake Sammamish," Water
Quality Series No. 2., Municipality of METRO Seattle (1966).
-------
Kemmerer, G., Bovard, J., and Boorman, W., "Northwestern lakes of
the U.S.: Biological and chemical studies with reference to pos-
sibilities in production of fish," Bull. U.S. Bur. Fish., 39,
pp. 51-140 (1923).
Krull, J., "Phosphorus response model for Lake Sammamish," Unpub.
manuscript, Univ. of Wash. (1973).
Larson, D. P., ejt al_., "Response of Eutrophic Shagawa Lake, Minne-
sota, U.S.A., to Point-source, Phosphorus Reduction," Verh. Intl.
Ver. Llmnol. 19., pp. 884-892 (1975).
Libby, R. A., "The application of Newton Activation Analysis to
Water Samples," M. S. Thesis, Dept. of Nuclear Engineering, Univ.
of Wash. (1972).
Liesch, Price and Walters, "Geology and groundwater resources of
Northwest King County, Washington," Water Supply Bulletin No. 20.
USGS (1963).
Livingston, Jr., V., "Geology and mineral resources of King County,
Washington," Wash. Dept. of Nat. Resources Bull. No. 63 (1971).
Lorenzen, M. W., "Predicting the effects of nutrient diversion on
lake recovery," In Modeling the Eutrophication Process. Middlebrooks,
J. E., et al_. (Eds.) Utah State Univ., pp. 205-210 (1973).
Malueg, K. W., et aK, "The Shagawa Lake Project; Lake Restoration
by Nutrient Removal from Wastewater Effluent," Environmental Protec-
tion Agency, Corvallis, Ore., 49 pp. (1973).
McDonnell, J. C., "In Situ Phosphorus Release Rates from Anaerobic
Lake Sediments," M. S. Thesis, Univ. of Wash. (1975).
Monahan, F., "An In Situ study of sediment nutrient release in Lake
Sammamish," M. S. Thesis". Univ. of Wash. (1974).
Moon, C. E., "Nutrient budget following waste diversion from a meso-
trophic lake," M.S. Thesis. Univ. of Wash. (1973).
Noah, C. M., "A Multiparametric Model of the Nitrogen System in
Lake Sammamish," M.S. Thesis, Univ. of Wash. (1976).
Pederson, G. L., "Zooplankton Population Dynamics and Production
in Three Lakes of Contrasting Trophic Status," Ph.D. dissertation,
Univ. of Wash. (1974).
89
-------
28. Rock, C. A., "The trophic status of Lake Sammamish and its relation-
ship to nutrient income," Ph.D. dissertation, Univ. of Wash., (1974).
29. Shapiro, J., Edmondson, W. T. and Allison, D. E., "Changes in the
chemical composition of sediments of Lake Washington, 1958-1970,"
Limnol. Oceanog. Ijj, pp. 437-452 (1971).
30. Stockner, J., "Preliminary characterization of lakes in the Experi-
mental Lakes Area, Northwestern Ontario, using diatom occurrences
in sediments," J. Fish. Res. Bd. Canada, 28, pp. 265-272 (1971).
31. Strickland, J. D., and Parsons, T. R., "A practical handbook of
seawater analysis," Bull. Fish. Res. Bd. Canada. No. 167 (1968).
32. Strutevant, P., "Growth Rate Parameters and Biomass Measurement
Ratio for Natural Algal Populations," M. S. Thesis, Univ. of Wash.
(1974).
33. Swayne, M., "Enviornmental monitoring from a communication engineer-
ing point of view," M. S. Thesis. Univ. of Wash. (1973).
34. Tang, C. H., "A multiparametric lake model," Ph.D. Dissertation,
Univ. of Wash. (1975).
35. Vollenweider, R. A., "Possibilities and limits of elementary models
concerning the budget of substances in lakes," Arch. Hvdro1obiology_»
66, pp. 1-36 (1969). "
36. Vollenweider, R. A. and Dillon, P. J., "The application of phos-
phorus loading concept to eutrophication research," N.R.C. Tech.
Rep. 13690, 42 pp. (1974).
37. Welch, E. B., Rock, C. A. and Krull, J. D., "Long-term lake recovery
related to available phosphorus," In Modelling the Eutrophication
Process. Middlebrooks, J. E., et al. (Eds.), Utah State Univ.,
pp. 5-14 (1973).
38. Welch, E. B., Hendrey, G. R., and Stoll, R. K., "Nutrient Supply
and the Production and Biomass of Algae in Four Washington Lakes,"
Oikos, 26., pp. 47-54 (1975).
39. Wiederholm, T., "A survey of the bottom fauna of Lake Sammamish,11
Northwest Science. 50, pp. 23-31 (1976).
90
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/3-77-003
4. TITLE AND SUBTITLE
Nutrient Diversion: Resulting Lake Trophic State
and Phosphorus Dynamics
7. AUTHOR(S)
Eugene B. Welch
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Washington
Seattle, Washington 98195
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Con/all is Environmental Research Laboratory
200 S.W. 35th Street
Con/all is, Oregon 97330
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
January 1977
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1 BA031
11. CONTRACT/GRANT NO.
EPA R-800512
13- Bar -E?ffi m% i WTRED
14. SPONSORING AGENCY CODE
EPA - ORD
15. SUPPLEMENTARY NOTES
lb. ABSTRACT
Lake Sammamish, Washington, was studied during 1970-75 to determine its
LaKe bammamisn, wasmngton, was suuuicu uU( ius «"< - :~.
response to wastewater diversion in 1968. The results were compared with a
pre-dtversTon study in 1964-65. Diversion Reduced the phosphorus loading
by about one-third (from 1.02 to 0.67 g P/nT. yr and about 119 to 68 p g/1 in
the inflow). Winter total phosphorus remained constant and no trend was shown
tn chlorophyll a in spring-summer. Water transparency remained the same.
PaleolimnoTogical evidence suggests that the lake has been near its present
mesotrophic state for about 100 years. This stability is thought to be due to
the constancy of the water phosphorus concentration which is in turn controlled
by the anaerobic-aerobic release and sedimentation of iron and its complexes.
Although the effect of a one-third reduction in phosphorus Joading.as well
as a similar increase resulting from the urbanization impact of the 1960 s, was
not apparent in the lake, without diversion loading would now be increased to
1.36 g P/m2. yr.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Lakes *
Limnology
Phosphorus *
Algae
Eutrophicatton *
Trophic Load .
(*denotes major descriptors;
b.lDENTIFIERS/OPEN ENDED TERMS
:. COSATI Field/Group
02H
04A
o5C
IS. DISTRIBUTION STATEMENT
Release unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
100
20. SECURITY CLASS (Thispage)
Unclassified
22, PRICE
EPA Form 2220-1 (0-73)
91
ft U. S. GOVERNMENT PRINTING OFFICE: 1976-796-587 / 27 REGION 10
------- |