EPA910-R-06-003
United States
Environmental Protection
Agency
Region 10
1200 Sixth Avenue
Seattle WA 98101
Alaska
Idaho
Oregon
Washington
Office of Water & Watersheds
March 2006
A Classification of Pacific
Northwest Reservoirs with
Respect to Nutrient Processing
Western Forested
Mountains
-------
EPA910-R-06-003
March 2006
A Classification of Pacific Northwest Reservoirs with Respect to
Nutrient Processing
R.M. Vaga, US Environmental Protection Agency, Region 10, Office of Water and
Watersheds, 1200 Sixth Ave. Seattle, WA 98101
AT. Herlihy, Department of Fisheries and Wldlife, Oregon State University, 104 Nash
Hall, Corvallis, OR 97331
R. Miller, M.M. Sytsma, Environmental Sciences and Resources, Portland State University,
Portland Oregon, 97202
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency. It has been subjected to the Agency's peer and
administrative review and it has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
(Cover Photos: Top - Lake Billy Chinook, a Xeric West reservoir in eastern Oregon;
bottom - Bumping Lake, a Western Forested Mountains reservoir in Washington).
-------
TABLE OF CONTENTS
Page
Abstract iv
Introduction 1
Methods 1
Results 5
Discussion 14
Conclusions 16
References 18
Appendix I 20
Appendix II 21
-------
Abstract
To begin the process of developing nutrient criteria for Pacific Northwest reservoirs, 48
reservoirs were sampled in Oregon, Washington and Idaho during summer 2002. The
National Inventory of Dams lists 2419 dams in the Pacific Northwest. Over 75% of them
represented reservoirs that were less than 30 hectares and were not considered in this
study. After removing these small reservoirs and other non-target sites, there were 328
reservoirs that met our target criteria. The 48 sample reservoirs were selected from these
328 target reservoirs using a systematic, randomized design with Omernik Nutrient
Ecoregions (Xeric West or Western Forested Mountains) as a stratification variable.
Total phosphorus (TP) concentrations were higher and Secchi depths lower in the Xeric
West relative to the Western Forest Mountains. Total nitrogen (TN) and chlorophyll-a
concentrations were not significantly different between the two Ecoregions. All four nutrient
variables were strongly correlated with each other in both Ecoregions. However, reservoir
purpose class was not correlated with nutrient concentrations.
In the Western Forested Mountains, TP concentrations were very strongly related to
measures of ionic strength (e.g., base cations, alkalinity, conductivity); an indication that
there is a natural TP gradient related to watershed geology/soils. In the Xeric West TP was
less strongly related to potassium and sodium and not related at all to the other indicators
of ionic strength. The source of TP could be some specific geologic source that is only
related to potassium or sodium or it could be associated with fertilizers that are high in
potassium.
Principal components analysis and cluster analysis were performed on reservoirs in the
Western Forested Mountains and the Xeric West. Analysis did not reveal any clear
classification of reservoirs in the WFM. In the Xeric West 5 variables were (potassium,
alkalinity, absorbence at 326 nm, turbidity and chlorophyll) used to identify 4 clusters of
reservoirs. A reservoir typology resulting in 3 categories was developed from these four
clusters.
Category 1 - Low ionic strength, high transparency, low absorbence, low turbidity, low
chlorophyll.
Category 2 - Moderate ionic strength, moderate transparency, moderate absorbence,
moderate turbidity due either to algae or particulate matter.
Category 3 - High ionic strength, very high absorbence, very low transparency due to algal
and/or nonalgal turbidity.
These three classes differed in distributions of total P concentration and water transparency.
Total N and chlorophyll values were not different among categories.
The use of Nutrient Ecoregion provides a useful classification variable in categorizing
reservoirs in the Pacific Northwest. However, it appears that the Level III Ecoregion Eastern
Cascades should be placed into the Xeric West Nutrient Ecoregion when classifying
reservoirs. The reasons for this are unclear but may have to do with geology or
precipitation.
Future studies on reservoir nutrient dynamics in the Pacific Northwest should explicitly take
nonalgal turbidity into account to differentiate between the effects of nutrients and water
transparency on algal growth.
IV
-------
Introduction
According to the U.S. Environmental
Protection Agency=s (EPA=s) 1996 report
to Congress, a large percentage of U.S.
lakes and reservoirs are water quality
impaired, many due to excess nutrients.
In order to address this problem EPA has
devised a strategy for the development of
regional nutrient criteria (USEPA 1998).
Implicit in the strategy is the recognition
that excess nutrients are a major cause of
water quality impairment in the United
States. The strategy also acknowledges
that because of diverse geology, climate,
and morphometry, single national nutrient
criteria for lakes and reservoirs are not
appropriate. EPA's current approach for
developing nutrient criteria for lakes and
reservoirs involves the following steps
(Kennedy 2000):
1. Establish regional technical
assistance groups.
2. Delineate nutrient ecoregions.
3. Classify or group lakes/reservoirs
using non-nutrient parameters.
4. Establish an appropriate nutrient
related database.
5. Establish reference conditions for
each group or classification.
6. Develop nutrient criteria for each
group or classification.
Although the same basic physical,
chemical and biological processes
determine nutrient dynamics in lakes and
reservoirs, the age, morphometry,
location in the drainage basin and
hydrological characteristics of reservoirs
make them a unique ecosystem (Cooke
and Kennedy, 1989, Thornton et al.
1990). As such developing a
classification of reservoirs with respect to
nutrient processing must take the unique
nature of these systems into account
(Straskraba, 1993). This report describes
the methods and results of a study of
nutrient levels and classification of 48
randomly selected reservoirs in the
Pacific Northwest states of Oregon,
Washington and Idaho. In previous work
we have demonstrated that it is possible
to classify lakes using Level III
Ecoregions (Vaga and Herlihy 2004,
2005).
Methods
Site selection and population definition
Reservoirs range greatly in size and type
across the Pacific Northwest from small
farm ponds to giant hydropower/flood
control reservoirs on the Columbia River.
For this study, the target population of
interest was restricted to specific reservoir
types and size range. The population of
interest only includes reservoirs with a
surface area between 30 and 10,000
hectares. Target types include
hydropower, flood control, water supply,
irrigation, and recreation reservoirs. Mine
tailing, wastewater effluent, debris control,
run-of-the-river, and enhanced lake
reservoirs are not part of the target
population. Determination of type was
based on the National Inventory of Dams
purpose designations. Enhanced lake
reservoirs were defined as lakes in which
the dam height is less than 25% of the
maximum depth of the reservoir.
Reservoir location was expected to be an
important factor in explaining nutrient
concentrations so ecoregion was used as
a stratification variable in site selection.
To aid in developing nutrient criteria at the
national level, EPA has delineated the
conterminous U.S. into nutrient
ecoregions according to geologic and
climactic characteristics (Omernik 1998).
Portions of three nutrient ecoregions lie
within Region 10: Western Forested
Mountains (WFM), Xeric West (XW), and
Willamette and Central Valleys.
Reservoirs within Region 10's portion of
the Wllamette and Central Valley nutrient
ecoregion are not common and for this
project were considered part of the
Western Forested Mountains region.
Thus there were two ecoregion strata for
site selection.
The sample frame or explicit list of the
target population from which sites were
selected was the list of all reservoirs in
Page 1 of 21
-------
the National Inventory of Dams for
Oregon, Washington, and Idaho. Based
on this frame, there are 2,417 dams in
Oregon, Washington, and Idaho (Table
1). Reservoirs outside the type and size
cutoffs described above were removed
from the frame. Only 328 of the 2,417
reservoirs in the Pacific Northwest met
our target criteria. The majority of dams
in the database were non-target due to
small surface area (Table 1). To make
inferences to all 328 reservoirs in this
target population, sample reservoirs were
selected from this set using a stratified,
randomized design.
The desired sample size was 48
reservoirs, 24 from each of the two
ecoregion strata. To ensure a broad
geographic spread of reservoirs across
the study area, level III ecoregions
(Omernik, 1987) were used to define
spatial clusters. The Northern Basin and
Range ecoregion was split into Oregon
section and Idaho sections because the
region is considerably larger than other
Level III ecoregions in Region 10. The
328 reservoirs in the sample frame were
randomly sorted within level III ecoregion
and then level III ecoregion blocks
randomly sorted within each of the two
nutrient ecoregion strata. Sites were over
sampled so that there would be alternate
reservoirs. Therefore, 48 reservoirs were
selected from each strata, 24 primary and
24 secondary (alternates). After a
random start, reservoirs were picked an
interval of N/48 from the randomized
frame of all reservoirs in the strata (N is
the total number of reservoirs in the
strata).
Of the 48 primary randomly selected
reservoirs, 33 were successfully sampled
in the summer of 2002. Of the 15 lakes
not sampled, 6 were due to access
problems (private property, no launch site,
or remoteness), 2 were dry, 2 were
enhanced, 1 had no dam, and 3 were less
than 30 ha when visited in the field. In
addition, one reservoir wasn't sampled
due to time constraints. Thus 31%
(15/48) of the 328 reservoirs in the target
population were not really target in the
real world. If a primary reservoir could not
be sampled, one of the secondary
reservoirs from the same Level III
ecoregion was sampled. If that reservoir
could not be sampled another secondary
reservoir from the same Level III
ecoregion was sampled. If the pool of
secondary reservoirs within a Level III
ecoregion was exhausted, the original
sample frame was revisited and a tertiary
reservoir was selected from the same
Level III ecoregion.
In three cases this process was not
followed.
1. One Idaho Batholith reservoir: In
the Idaho Batholith Level III
ecoregion one primary reservoir
was not sampled because of time
constraints (Fish Creek
Reservoir). A secondary reservoir
with easier access was sampled
in its place (Brundage Reservoir).
2. Two Northern Rockies reservoirs:
In the Northern Rockies Level III
ecoregion the pool of secondary
reservoirs was exhausted and no
suitable tertiary reservoirs were
identified in the sample pool. One
reservoir located partially in the
Columbia Plateau and the
Northern Rockies ecoregions was
selected from the Columbia
Plateau's secondary list.
Laurence Lake was selected from
the Cascade Level III ecoregion to
replace the other missing
Northern Rockies reservoir.
3. One Puget Lowland reservoir:
The primary and secondary list
was exhausted, and due to time
constraints, a reservoir close to
Portland was sampled (Henry
Hagg Lake, tertiary, Wllamette
Valley ecoregion)
Inspection of the data, e.g. cations
suggested that East Cascade reservoirs
were far more similar to reservoirs in the
Xeric West that those in the Western
Forested Mountains. Under similar
considerations one reservoir (Thief Valley,
east end of the Blue Mountains
Ecoregion) which was originally in the
WFM was placed into the Xeric West.
Long Lake (WFM) was deleted from
analysis because it is currently 303(d)
listed for nutrients due to high nutrient
inputs from point sources. In addition,
Page 2 of 21
-------
after analysis of watershed versus dam
location it was determined that two
reservoirs (Conconully, Leader) are
actually in the WFM rather than the XW.
The resulting sample included 27 Xeric
west reservoirs and 20 Western Forested
Mountains (Fig. 1, Table 2).
Table 1. Sample Frame size and target
population size for Pacific Northwest
Reservoirs. Non-target reservoirs were those
reservoirs excluded from the Frame because
they did not meet criteria. Of the 2419
reservoirs in NID only 328 met the criteria.
About two-thirds of those were in the Western
Forested Mountains Ecoregion.
Table 2. List of sampled reservoirs with location and
surface area. Numbers refers to reservoir locations in
Figure 1. Most sensitive use: HM: Hydrologic
Management, RC: Recreation, WS: Water Supply.
Total Number in lnventory= 2419
A/on -Target Reservoirs
Too Small (< 30 ha)
Non-Target Types
Enhanced Lakes
Duplicate/Auxiliary Lakes
Inaccessible
Other
Total Non-Target
1828
122
71
53
7
10
2091
Target Reservoirs = 328
W. Forested Mt. Ecoregion
Xeric West Ecoregion
220
108
No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
WESTERN FORESTED MOUNTAINS
Lake Name
Brundage Res
Bumping Lake
Clear Lake
Conconully Res
Diablo Lake
Fern Ridge Res
Galesville Res
Henry Hagg Res
Judy Res
Lake Simtustus
Laurence Lake
Leader Lake
Mayfield Res
N Fork Clackamas
Ochoco Res
Prineville Res
Sage Hen Res
Skookumchuck Res
Toketee Res
Willow Creek
Area
87.0
527.3
56.7
133.6
400.7
3788.0
257.0
468.2
56.7
218.5
64.8
74.9
910.6
141.6
493.7
1449.0
96.3
218.5
41.3
130.3
LAT
45.0
46.8
46.6
48.5
48.7
44.1
42.8
45.4
48.4
44.6
45.4
48.3
46.5
45.2
44.3
44.1
44.3
46.7
43.2
42.4
LONG
-116.13
-121.30
-121.27
-119.75
-121.13
-123.29
-123.18
-123.21
-122.19
-121.23
-121.66
-119.70
-122.59
-122.28
-120.73
-120.78
-116.19
-122.72
-122.42
-122.44
Use
HM
RC
RC
RC
RC
RC
WS
WS
WS
RC
RC
RC
RC
RC
RC
HM
WS
HM
RC
WS
XERIC WEST
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Alexander Res
Barry Res
Ben Ross Res
Black Canyon Res
Blue Creek Res
Bray Lake
Bryant Mt. Res
Cedar Creek Res
Chickahominy
Daniels Res
Grasmere Res
JC Boyle Res
Kern Res
Little Blue Creek
Mann Creek Res
Mud Lake
Muddy Creek Res
Paddock Valley Res
Friday Res
Sand Creek Res
Scooteney Res
Stone Res
Thief Valley Res
Trail Storage Pond
Upper Rock Creek
Warmsprings Res
Weston Creek Res
494.9
60.7
142.9
445.2
76.1
82.6
48.6
424.9
214.1
151.8
31.6
170.0
43.7
53.8
127.5
68.0
78.9
542.3
85.8
31.6
617.2
100.0
373.5
104.8
155.4
1862.0
45.3
42.6
42.1
44.5
43.9
42.3
43.0
42.1
42.2
43.5
42.3
42.3
42.1
42.9
42.2
44.3
42.2
42.2
44.2
42.3
44.2
46.6
42.0
45.0
43.0
42.6
43.5
42.1
-111.70
-119.53
-116.44
-116.44
-116.18
-114.88
-121.33
-114.88
-119.61
-112.44
-115.90
-122.04
-118.78
-116.12
-116.89
-119.72
-120.52
-116.60
-119.90
-111.61
-119.03
-112.69
-117.78
-115.32
-119.31
-118.21
-112.13
HM
HM
WS
HM
WS
HM
HM
HM
HM
HM
WS
RC
HM
WS
RC
HM
HM
HM
HM
RC
HM
WS
RC
HM
HM
RC
WS
Page 3 of 21
-------
Field and laboratory methods
Reservoirs were visited during the
summer growing season, May through
September, 2002. Reservoirs were
sampled by boat and those from different
nutrient ecoregions were sampled on a
rotating schedule as much as Ipgistically
possible to avoid seasonal bias. The
deepest location in each reservoir was
found by starting at the dam and using an
electronic depth finder. A water sample
and in situ measurements were collected
at the deepest spot in each reservoir. In
situ profiles of temperature, dissolved
oxygen, pH, Turbidity, redox potential,
light extinction coefficient, and
conductivity were collected using a
Hydrolab Sonde 4a meter. Secchi depth
was determined as the average of the
disappearance and reappearance depths
of a 20 cm Secchi disk. Two, 1 L water
samples were collected from 1 m depth
using a 2.5 L, Model 1010X Niskin bottle.
One sample was preserved by
acidification to pH < 2 for analysis of total
nitrogen and phosphorus. The other
sample was filtered with a Whatcom glass
fiber filter that had a nominal pore size of
0.7 urn. The filter paper was frozen, kept
in the dark and analyzed for chlorophyll-a.
The filtrate was kept and used for analysis
of anions, base cations, DOC, alkalinity,
and light absorbence. All water samples
were kept cold and in the dark until
analysis.
n the lab, chlorophyll-a was determined
fluorometrically. Base cations were
analyzed by flame atomic absorption
spectrometry and sulfate and chloride by
ion chromatography. Alkalinity was
determined by Gran titration, and DOC by
carbon analyzer. Total phosphorus and
nitrogen were analyzed by persulfate
digestion and colorimetry. Light
absorbence at 325 and 440 nm was
measured with a spectrophotometer.
Field measurement of reservoir physical
habitat and substrate composition were
made using a condensed form of the U.S.
EPA EMAP lake protocols (Baker et al.,
1997). Sampling locations were set up
following a random start at 10 equal
interval shoreline stations around the
reservoir. Percentage category estimates
Figure 1. Location of the 47 reservoirs used in
this study. Numbers refer to reservoir number
in Table 2.
were made at each station for substrate
classes (clay, sand, cobble/gravel,
boulder, bedrock), vegetation cover
classes (barren, herbs, shrubs, trees),
and macrophyte cover classes
(submergent, emergent, floating) as well
as presence/absence of human influence
categories. These data were analyzed by
averaging the 10 measurements together
to calculate shoreline cover estimates for
the entire reservoir. In addition, depth
below maximum pool estimates were
made by comparing lake level at the time
of sampling to shoreline evidence of
maximum reservoir conditions.
Data Analysis
Data were entered into Excel
spreadsheets and converted into SAS
databases for analysis. As is typically
observed for water chemistry data, the
data are not normally distributed. Most
water chemistry variables are highly
skewed. For all statistical analyses,
chemical variables were Iog10
transformed as were reservoir surface
area and maximum depth. Shoreline
percentage data and Secchi depth were
not transformed. All correlation analyses
were done using Spearman rank
correlations in SAS.
A two-way analysis of variance (ANOVA)
was used to test for differences in nutrient
concentration due to ecoregion (Xeric
West versus Western Forested Mountain)
and reservoir purpose (Hydrologic
management, Drinking water, or
Recreation). Reservoir purpose was
taken from the National Inventory of
Dams database. Many reservoirs were
Page 4 of 21
-------
coded with multiple purposes. The
hydrologic management class includes
hydropower, irrigation, or navigation
categories. A single purpose class was
assigned to each reservoir as a "most-
sensitive" purpose rating sensitivity as
Water supply > Recreation > Hydrologic
management. In other words if there
were any one purpose listed as water
supply, the reservoir was coded as water
supply even though it may have been also
listed for other purposes. ANOVAs were
run in SAS using PROC GLM.
To get at an initial indication of possible
reservoir clusters or types, a principal
components analysis (PCA) was
performed on the major reservoir typology
variables. Axis or factor scores were then
clustered via cluster analysis to look for
groups of reservoirs. PCA was done
using SAS PROC FACTOR,
method=PRINCOMP. Cluster analysis
were
done using PC-ORD using the Flexible
Beta method (beta=-0.5).
Results
Ecoregional Patterns
Of the total number of reservoirs included
in this analysis, 27 were in the Xeric West
and 20 were in the Western Forested
Mountains Ecoregion. By most sensitive
purpose class, 11 were water supply, 17
were recreation, and 19 were hydrologic
management reservoirs. Most of the
hydrologic management reservoirs were
in the Xeric West and most of the
recreation reservoirs were in the Western
Forested Mountains (Table 2).
The use of Nutrient Ecoregion as a
classification variable for reservoirs
proved to be useful in characterizing
reservoirs (Table 3). Reservoirs in the
WFM tended to be larger, deeper and of
greater volume that those in the XW.
However, distributions of reservoir surface
areas within each of the two ecoregions
were virtually identical. Similarly, there
was little difference in surface area across
purpose class. In a two-way ANOVA,
mean surface areas across both
ecoregion and purpose class were not
significantly different (p > 0.25).
Maximum reservoir depths, however,
were significantly (F=7.0, p=0.01) higher
in the Western Forested Mountains than
the Xeric West (Tables 3, 4). Depths
across purpose class were not
significantly different.
Water transparency as measured by
extinction coefficient and Secchi depth
were greater in the WFM (Table 3).
Reservoirs in the XW were more turbid
that in the WFM. Surface water
temperatures tended to be greater in the
XW. Reservoirs in the WFM tended to be
more dilute that in the XW. Median
conductivity, alkalinity and ANC were all
much lower in the WFM than in the XW
(Table 3). Chlorophyll concentrations
were higher in the XW as was DOC.
Reservoirs in the XWwere higher in DOC
(Abs325) and humics (Abs440) than in
the WFM (Table 3).
Cations (K, Na, Ca, Mg) were all higher in
the XW reservoirs as compared with
WFM reservoirs. Anions (CL, SO4) were
also higher in the XW (Table 3).
The primary nutrient criteria variables,
total phosphorus (TP), total nitrogen (TN),
chlorophyll-a, and Secchi depth, and their
distributions within each nutrient
ecoregion are shown in Figures 4 - 7.
Mean TP concentrations were
significantly higher in the Xeric West
Ecoregion and in Hydrologically managed
reservoirs (Tables 3, 4, Fig. 4). TN
concentrations showed no difference
across either ecoregion or purpose class
(Table 4, Fig. 5). Chlorophyll-a
concentrations were higher in
hydrologically managed reservoirs but
showed no ecoregion effect whereas
Secchi depth was higher in the Western
Forested Mountains but showed no
reservoir purpose effect (Table 4, Figs, 6,
7). For all four variables, the ecoregion-
purpose interaction terms in the ANOVA
were not significant.
Turbidity, light extinction coefficient, and
absorbence at 325 nm (DOC) and 440 nm
(fulvic acids) were also analyzed. The
correlation matrix of the relationships
among these variables shows that in both
Ecoregions the variables are highly
correlated with each other (Table 5). In
the Xeric West there were very high
Page 5 of 21
-------
correlation coefficients (r>0.8) for the
relationships among the transparency
variables (Secchi, extinction, absorbence)
and between transparency and turbidity.
The relationship between extinction
coefficient and chlorophyll-a was
significant but much lower (r~0.5)
indicating that transparency in the Xeric
West is affected by abiotic turbidity in
addition to algal production (Fig. 8).
^ BQOOj
1000=
100 =
10-
Westem Forested Mountains
Xerte West
Nutrient Ecoreglon
Figure 2. Box and whisker plot of reservoir
surface area by nutrient ecoregion. Boxes
show the interquartile range and median,
whiskers the minimum/maximum values.
I 40-
30-
20-
10-
o-
8 Western Forested Mountains Xeric West
Nutrient Ecoregion
Figure 3. Box and whisker plot of maximum
depth by nutrient ecoregion. Reservoirs in the
Western Forested Mountains are significantly
deeper than those in the Xeric West.
On the other hand, in the Western
Forested Mountains, transparency is
more strongly related to chlorophyll-a
(r~0.8), indicating a more important role
for algal production in transparency. This
hypothesis is also supported by the
scatterplot of Secchi depth versus
turbidity (Fig. 9). With one exception, the
Xeric West reservoirs have almost a
Table 3. Median (± 1 SE) values for physical
and chemical characteristics of reservoirs
sampled in the Western Forested Mountains
and Xeric West. Volume is (m3 x 10s), depth
is maximum depth and ions are (uEq/L).
Variable
Area (ha)
Volume
Depth(m)
Ext(rrr1)
Secchi (m)
Turb(NTU)
Temp (°C)
PH
Eh (mV)
Dosat (%)
TP (ug/L
TN (ug/L)
K
Na
Ca
Mg
SO4
Cl
Alk (mg/L)
Cond(uS)
ANC(uEq/L
Chla (ug/L)
DOC(mg/L)
Abs325
Abs440
WFM XW
Median
218.5
36.7
11.2
0.5
3.4
0.9
17.4
7.9
340.0
97.0
23.0
213.2
22.8
187.0
289.4
157.7
43.6
64.1
28.2
60.0
563.6
4.6
2.2
50.0
9.3
se
180.0
57.5
2.7
0.1
0.5
1.1
0.8
0.2
18.5
3.3
13.6
54.3
6.9
77.7
75.5
48.8
49.1
14.7
7.8
18.3
155.5
3.1
1.6
15.2
2.5
Median
104.8
6.3
6.0
2.1
0.7
14.7
22.0
8.4
318.0
96.9
109.0
430.9
88.2
357.6
524.0
244.1
77.2
101.8
57.6
113.4
1151.5
7.5
11.7
267.3
55.8
se
75.8
27.7
0.9
0.4
0.3
8.7
0.6
0.1
6.6
5.1
43.7
46.3
13.7
140.2
116.2
213.0
63.4
147.7
11.5
41.4
230.0
27.8
2.8
87.4
26.8
Western Forested Mountains
Nutrient Ecoregion
Figure 4. Box and whisker plot of total
phosphorus by nutrient ecoregion. Total
phosphorus is significantly higher in the Xeric
West reservoirs.
Page 6 of 21
-------
I
z
"ffl
S
Western Forested Mountains
Nutrient Ecoregion
Figure 5. Box and whisker plot of total
nitrogen by nutrient ecoregion. Total
nitrogen concentrations are not
significantly different between ecoregions.
straight-line relationship between turbidity
and Secchi depth (the one outlier is a
reservoir with chlorophyll-a=519 ug/L
which is responsible for the 0.15 m Secchi
depth). In the Western Forested
Mountains, the plot isn't nearly as linear,
indicating non-turbidity factors influencing
Secchi depth.
Reservoirs with high TP tend to have high
TN (Fig. 10). Reservoirs in the Xeric
West tend to have higher concentrations
of both TP and TN. However, the slope of
the log(TN) - log(TP) regression was
significantly (p < 0.05) higher in the WFM.
Chlorophyll-a concentrations increased
with TP in both ecoregions (Fig. 11) but
again the slopes were significantly lower
in the XW.
Chlorophyll also increased with TN in both
nutrient ecoregions (Fig. 12). Secchi
transparency decreased with increasing
TP in both ecoregions (Fig. 13).
However, the lowest transparencies
observed in the Xeric West were not due
to chlorophyll but non-algal turbidity.
TP is often used a measure of trophic
status. A commonly used scheme is to
use TP concentrations of 10, 30 and 50
ug/L as the cutoffs between oligotrophic-
mesotrophic, mesotrophic-eutrophic, and
eutrophic-hypereutrophic lakes. Using
this scheme, none of the reservoirs in the
Xeric West are oligotrophic, and 62.5% of
the reservoirs are hypereutrophic. In
contrast, in the Western Forested
Mountains, one-third of the reservoirs are
oligotrophic and 25% are hypereutrophic.
As these data are from a random sample
of all Pacific Northwest reservoirs, these
percentages should reflect conditions in
all 328 reservoirs that met our target
criteria. Extrapolating the values in Table
3 to all reservoirs in the PNW using the
two Nutrient Ecoregion estimates and the
total number of reservoirs in each
ecoregion strata (Table 1), 22.3% are
oligotrophic, 25% are mesotrophic, 3%
are eutrophic, and 37.3% are
hypereutrophic.
§» 100:
2.
o
Western Forested Mountains Xeric West
Nutrient Ecoregion
Figure 6. Box and whisker plot of
chlorophyll-a by nutrient ecoregion.
Chlorophyll concentrations were not
significantly different between ecoregions.
Watershed Chemistry and Nutrients
Base cation concentrations are typically
controlled by watershed geology/soils and
weathering rates. There were clear
differences between ion relationships
between Nutrient Ecoregions. In the
WFM correlations among most cations
and anions were high and significant
(Table 6). In the Xeric West, the divalent
base cations (calcium and magnesium)
are strongly related to each other and the
monovalent base cations (sodium and
potassium) are strongly related to each
other but the two groups are not strongly
related to each other.
These results suggest that in the Western
Forested Mountains all of the base
cations come from a similar source. In
contrast, in the Xeric West there appears
to be independent sources for these two
groups of base cations .
Page 7 of 21
-------
6-
4-
2
T
01
Western Forested Mountains
g
Xerfc West
Nutrient Ecoregion
Figure 7. Box and whisker plot of Secchi depth
by nutrient ecoregion. Reservoirs in the
Western Forested Mountains are significantly
more transparent that those in the Xeric West.
* * Western Forested Mountains
Xerlc West
Total Phosphorus (ug/L)
Figure 10. Relationship between total nitrogen
and total phosphorus in Pacific Northwest
reservoirs. Total nitrogen increases in both
ecoregions with total phosphorus.
' Western Forested Mountains
" " - Xerlc West
-s -.
Chlorophyll (ug/L)
Figure 8. Relationship between Secchi depth
and chlorophyll in Pacific Northwest
reservoirs. The dependence of water
transparency on chlorophyll is much stronger
in the WFM as compared with the XW.
tt,0:
Western Forested Mountains
" Xeric West
H ."
Total Phosphorus (ug/L)
Figure 11. Relationship between
chlorophyll and total phosphorus.
Chlorophyll increases with increasing total
phosphorus in both ecoregions.
E 6
* Vfestern Forested Mountains
3 Xerlc Wast
Turbidity (NTU)
Figure 9. Relationship between Secchi depth
and turbidity in Pacific Northwest reservoirs.
Highest turbidities in the Xeric west are due to
non-algal turbidity.
I "
e
£
o
Western Forested Mountains
Xerlc West
200
Total Nitrogen (ug/L)
Figure 12. Scatterplot of chlorophyll versus
total nitrogen. Chlorophyll increases with total
nitrogen in both ecoregions.
Page 8 of 21
-------
Table 4. Two-way analysis of variance results
on the effects of ecoregion and reservoir
purpose class on nutrient criteria variables.
Bold values were significant at p < 0.05.
Variable
Total
Phospho
rus
Total
Nitrogen
Chlorop
hyll-a
Secchi
depth
Class
Ecoregion
Reservoir
Purpose
Ecox
Purpose
Interaction
Ecoregion
Reservoir
Purpose
Ecox
Purpose
Interaction
Ecoregion
Reservoir
Purpose
Eco x
Purpose
Interaction
Ecoregion
Reservoir
Purpose
Eco x
Purpose
Interaction
F
5.8
(0.021)
4.8
(0.013)
0.64
(0.53)
3.5
(0.071)
2.0
(0.20)
0.50
(0.61)
0.23
(0.63)
3.5
(0.039)
0.16
(0.85)
5.4
(0.026)
1.0
(0.37)
0.02
(0.98)
Means
Xeric > W.
For. Mt.
Mng > Rec
= Water
n.s.
n.s.
n.s.
n.s.
n.s.
Mng >
Water = Rec
n.s.
W. For. Mt >
Xeric
n.s.
n.s.
Nutrient Ecoregion: Xeric=Xeric West, W. For. Mt.
= Western Forested Mountains. Use:
Mng=Hydrologic Management, Rec=Recreation,
Water=Water Supply, n.s. = Not significant
In the WFM total P is highly correlated with
all cations but not correlated with Sulfate
and weakly correlated with Chloride (Table
6). In the XW total P is weakly correlated
with Potassium and Sodium not correlated
with any other ions. Total N is weakly
correlated with Sodium, Calcium and
Magnesium in the WFM but not correlated
with any ion in the Xeric West.
Furthermore, total P and total N are not
correlated in the WFM but loosely
correlated in the Xeric West.
The strong correlation among cations and
total P concentrations in the WFM are an
indication that there is a natural total P
gradient related to watershed
geology/soils. In the XW, since total P is
related to the monovalent group but not
the divalent group, the total P source could
be a specific geologic source that is only
related to Potassium or Sodium or it could
be associated with land use, e.g. fertilizers
that are high in the monovalent cations.
This latter interpretation is supported by
the fact that total P and total N were
correlated in the Xeric West but not in the
Western Forested Mountains, i.e. due to
exogenous inputs of both nutrients into
these watersheds.
Table 5. Spearman correlation coefficient
matrix for nutrient related variables in 24
reservoirs in the Western Forested Mountain
ecoregion and 24 reservoirs in the Xeric West
ecoregion. Correlations significant at p<0.05
are shown in bold, correlations significant at
p<0.001 are starred (*).
WESTERN FORESTED MOUNTAINS
Vari
TP
TN
Chl-
Sec
Turb
Ext.
Abs
Abs
rp
0.6
0.5
-0.5
0.0
0.5
0.2
0.2
FN
0.4
-0.5
0.3
0.6
0.6
0.6
;hl-a
-
1.
0.7
0.
0.
5ec
....
0.0
-
-0.5
-0.4
furb
0.7
0.5
0.6
Ext.
0.7
0.7
tos3
0.9
XERIC WEST
Vari
TP
TN
Chl-
Secc
Turb
Ext.
Abs3
Abs4
TP
....
0.68*
0.60
-0.86*
0.80*
0.84*
0.73*
0.68*
TN
0.55
-
0.49
0.56
0.45
0.35
;m-a
-0.55
0.48
0.50
0.23
0.16
Seech
-0.88*
-0.89*
-0.80*
-0.77*
Turb
0.96
0.83
0.87
Ext.
0.83
0.85
i\bs3
*
* 0.96
)OC
0.9
0.8
0.7
DOC
....
....
TP = Total Phosphorus, TN = Total Nitrogen, Chl-a
= Chlorophyll-a, Secchi = Secchi Depth, Turb =
Turbidity, Ext. = Light Extinction Coefficient, Abs325
= Absorbence at 325 nm, Abs440 = Absorbence at
440 nm
Page 9 of 21
-------
f
°
1
ffi
ID
Western Forested Mountains
"""Xeric West
1 10 BO 1000
Total Phosphorus (ug/L)
Figure 13. Relationship between Secchi depth
and total phosphorus. Water transparency
decreases with increasing total phosphorus.
* Western Forested Mountains
Xeric West
Potassium (ueq/L)
Figure 14. Scatter plot of total phosphorus
versus potassium concentration in Pacific
Northwest reservoirs. Total P has a higher
correlation with Potassium in the WFM than
the XW.
Table 6. Spearman rank correlation
coefficients among base cations, anions, total
phosphorus (TP) and total nitrogen (TN) in 24
Xeric West and 24 Western Forested Mountain
reservoirs. Bold values are significant at p <
0.0001. Values with significance below 0.1 are
shown as n.s.
K
Na
Ca
Mg
S04
CL
TP
Western Forested Mountains
K
A/a
Ca
Mg
S04
CL
TP
TN
-
0.90
0.72
0.77
0.56
0.55
0.84
ns
-
0.73
0.86
0.59
0.64
0.82
0.53
-
0.86
0.79
0.58
0.66
0.57
-
0.65
0.71
0.73
0.66
-
0.67
ns
ns
0.58
ns
-
ns
Xeric West
K
A/a
Ca
Mg
S04
CL
TP
TN
-
0.75
ns
ns
ns
0.80
0.61
ns
-
0.56
0.53
0.63
0.89
0.48
ns
-
0.91
0.76
ns
ns
ns
0.66
ns
ns
ns
-
0.65
ns
ns
-
ns
ns
-
0.63
As the nutrient variables are highly
correlated with each other, it's not
necessary to consider how other variables
relate to all of them. TP and chlorophyll-a
were selected as nutrient indicator
variables and analyzed to see how other
reservoir variables related to them (Table
6). In examining the data it became
-J IWtl:
I
We Seen Forested Mountains
Xeric West
Potassium (ueq/L)
Figure 15. Scatter plot of total
phosphorus versus potassium
concentration in Pacific Northwest
reservoirs. Calcium and Potassium are
more highly correlated in the WFM than in
the XW.
apparent that the two ecoregions behaved
differently so the results are presented
separately for each ecoregion. In the Xeric
West, no variables were highly correlated
(p < 0.001) with either TP or chlorophyll-a.
TP was significantly related (p < .05) to
potassium, dissolved oxygen, shoreline
disturbance, and maximum reservoir
depth. In the Western Forested
Mountains, many variables were
significantly related to both TP and
chlorophyll-a. Indicators of water ionic
strength (base cations, alkalinity,
conductivity were highly correlated with
both nutrient variables. Reservoir surface
area, depth below full pool, sulfate, redox
potential, and shoreline ground cover w
Page 10 of 21
-------
were not related to either nutrient variable
in either ecoregion (Table 6).
Potassium was most strongly related to TP
in both ecoregions (Fig. 15). The log-log
plots have similar slopes in both
ecoregions though the relationship is less
variable in the Western Forested
Mountains. On the other hand, calcium is
strongly related to TP in the Western
Forested Mountains but is not related to
TP in the Xeric West. Plots of calcium
versus potassium show that the two base
cations are strongly related in the Western
Forested Mountains (r=0.74) but unrelated
\in the Xeric West (r=0.28; Fig. 15). In the
Western Forested Mountains, all four base
cations are correlated with each other with
an r>0.75.
In the WFM DOC was strongly correlated
with absorbence at both 325 and 440 nm
(Table 5). In contrast there was no
correlation among these variables in the
Xeric West. Considering that median DOC
was five times higher in the XWthan in the
WFM, the absence of a correlation
between these parameters in the XW is
problematic. This may have to do with the
high turbidities in the XW, which were
highly correlated with absorbence and thus
masked any relationship to DOC.
Shoreline structure indices were not
strongly correlated with any water
chemistry variable in the data set.
Although there appeared to be a rough
negative correlation between total
shoreline vegetation ground cover and
total P concentrations, correlations were
insignificant. This result has not been
controlled for season, i.e. reservoir
draw down.
Cluster Analysis
One of the objectives of this
research was to develop classes or a
typology for Pacific Northwest reservoirs to
aid in developing nutrient criteria. Two
approaches were tried; visual inspection of
anion/cation chemical composition, and an
overall cluster analysis. To investigate
anion/cation composition, trilinear
diagrams were used that show the relative
% composition of various ions on a
triangular plot. For both anions and base
cations, there were no apparent clusters of
lake types and this doesn't appear to be a
useful approach for developing a reservoir
typology
To develop clusters from the reservoir data
for each Nutrient Ecoregion, a principal
components analysis (PCA) was used to
identify the major components in the data.
These components or factor scores for
each site were then clustered. Results of
the PCA are very dependent on what
variables are included. It's also desirable
to have non-redundant variables. It was
also desirable to select a set of variables
that coyer a range of reservoir
characteristics that are not correlated with
one another. Five variables were chosen
for the PCA analysis for the Xeric West
(Table 7). Potassium and alkalinity were
selected to represent ionic strength. Both
were used as they represent different
processes in the Xeric West and were not
correlated with each other. Water clarity
was represented by three separate factors:
Abs325, turbidity and chlorophyll.
The resulting PCA had 4 factors or axes
that each explained more than 10% of the
variance and these 4 site factor scores
were used for the subsequent cluster
analysis. All together, these 4 components
explained 98% of the variance in the 5
variables. In looking at the correlation of
the original 5 variables to the axis scores,
axis 1 can be considered an ionic
strength/transparency axis, axis 2 is
related to alkalinity, axis 3 is related to
chlorophyll, and axis 4 is a turbidity axis.
We included chlorophyll as a variable in an
attempt to separate algal and nonalgal
turbidity.
From the cluster analysis of the four axes
scores, four clusters were identified from
the resulting dendrogram (Fig. 16). Box
and whisker plots showing how the
clusters differ on reservoir variables are
useful in assigning cluster attributes .
Cluster 1 - Low ionic strength, high
transparency, low absorbence, low
turbidity.
Cluster 2 - Moderate ionic strength,
moderate transparency, moderate
Page 11 of 21
-------
absorbence, moderate turbidity due either
to algae or particulate matter.
Cluster 3 - Moderate ionic strength, low
transparency due to non-algal turbidity.
Cluster 4 - High ionic strength, very high
absorbence, very low transparency due to
algal and/or nonalgal turbidity.
We found no components for reservoirs in
the WFM. The magnitude of total P in
these reservoirs ranged from 4 to 61 ug/L
(Appendix 1). Base cations and total P
were all correlated in this Nutrient
Ecoregion, suggesting that a common
mechanism explains the concentrations of
total P in these reservoirs (Table 6). The
highest total P concentrations were found
in reservoirs (e.g. Lake Simtustus,
Toketee, Willow Creek) that have very
four PCA axes, and the amount of variance
explained by each PCA axis. Cumulatively, the
four axes explained 98% of the variance in the
5 variables.
Variable
Potassium
Alkalinity
Abs325
Turbidity
Chlorophyll
% variance
Axis
1
0.52
0.03
0.50
0.55
0.41
58%
Axis
2
0.31
0.89
-0.28
-0.15
0.11
24%
Axis
3
0.18
0.20
0.45
0.05
-0.85
14%
Axis
4
-0.32
0.15
-0.41
0.81
-0.20
3%
high natural sources of phosphorus.
Excluding these reservoirs, the other
reservoirs in the WFM all had total P
Table 7. Variables used in principal
Reservoir Nome
Alexander Reservoir
Ch icknhoniiny
Barry Reservoir
Mud Late
Grastne re Reservoir
Upper Rock Creek Reservoi
Pridov Reservoir
Ben Ross Reservoir
Snnd Creek Reservoir
Bryant Ml. Reservoir
Block Canyon Reservoir
Monn Creek Reservoir
Daniels Reservoir
Scoo tenev Reservoir
Nes ton Creek Reservoir
Stone Reservoir
Blue Creek Reservoir
Kern Reservoir
Little Blue Creek Rese rvo
Trail Storage Pond
Bray Lake
Cedar Creek Reservoir
JC Boyle Reservoir
Thief Valley Reservoir
(No r risp r i ngs Reservoir
Muddy Creek Reservoir
Paddock Valley Reservoir
concentrations less that 40 ug/L. If there
2
6
10
12
H
16
18
FI ex i bIe-B eta Distance
Figure 16. Dendrogram of the cluster analysis of the four PCA axes for the 27 reservoirs in the
Xeric West showing the 4 cluster used in subsequent analysis.
components analysis, their correlation with the
are significant differences among these
Page 12 of 21
-------
reservoirs with respect to landscape, we
weren't able to discern them with in our
data set.
Reservoir Categories
Developing a typology or classes is an
essential first step for developing nutrient
criteria for reservoirs. Based on water
quality data for each of the clusters
described in section 5, we propose a
reservoir typology for the Xeric West with
three categories.
Category 1 - Low ionic strength, high
transparency, low absorbence, low
turbidity, low chlorophyll.
Category 2 - Moderate ionic strength,
moderate transparency, moderate
absorbence, moderate turbidity due either
to algae or particulate matter.
Category 3 - High ionic strength, very high
absorbence, very low transparency due to
algal and/or nonalgal turbidity.
Categories 1 and 2 are the same as
clusters 3 and 4, respectively. Category 3
is a combination of clusters 1 and 2 (Fig.
21).
Cutoff criteria for these types were
selected to preserve the original clustering
as much as possible. Thus, this typology
preserves the segregation observed in the
original clustering but has very simple
parameters used to define the types.
Reservoir typing could be done with a
simple field visit and measurements of
conductivity, turbidity, Secchi and
chlorophyll. These measurements can all
be made fairly inexpensively.
Characteristics of Reservoir Categories in
the Xeric West
There was no apparent relationship
between reservoir category and
geographic location (Fig. 17). However, all
but two of the reservoirs in the Xeric West
(Scooteney in Columbia Basin and Thief
Valley in Blue Mountains) were in the Level
III Northern Basin and Range Ecoregion.
Therefore if additional reservoirs are
sampled in the Columbia Plateau this
absence of apparent geographic effect
may not hold. Conversely, classification of
reservoirs by Nutrient Ecoregion may
prove to be effective as an initial method to
stratify reservoirs by location for purposes
of developing nutrient criteria.
The three types of reservoir thus
determined show differences in terms of
total P, total N Secchi but not chlorophyll
concentrations (Fig. 18-21).
For natural water bodies, once a typology
has been established, nutrient criteria
would be established by looking at the
distribution of nutrient levels in a set of
undisturbed of reference systems of a
given category. That would provide a
scientifically defensible picture of what
natural nutrient concentrations ought to be
for setting criteria. For reservoirs,
however, there is no way to have natural
reference sites. Reservoirs by their very
nature are artificial human built systems so
the reference site concept doesn't apply.
Thus, setting nutrient criteria for reservoirs
will be especially problematic.
One method to define nutrient criteria is
based upon a percentile distribution of
scores of all systems in each category
(EPA, 2000). Recent data suggest that in
the Pacific Northwest the median-75%
interquartile is an appropriate range for
total P values in lakes and reservoirs
(Vaga and Herlihy 2004). Percentile
scores and medians for nutrient criteria
variables for each reservoir category are
listed in Table 8. Using the percentile
approach, total P nutrient criteria for
Category 1 reservoirs would be <33 ug/L ,
<238 ug/L in Category 2 reservoirs and
<686 ug/L in category 3 reservoirs.
Similarly, criteria for Secchi depth would
be 2.1,0.70 and 0.15 meters, respectively.
It is evident that the high variances of total
N and chlorophyll preclude any
determination regarding meaningful
assignments of criteria using this method.
Whether a larger sample size would
provide more precise determinations of
these parameters among reservoir
categories is problematic. Category 1
reservoirs are clearly relatively oligotrophic
compared with the other categories.
Page 13 of 21
-------
Chlorophyll values in Category 2 reservoirs
indicate a variability based upon season
but water transparency is sufficient to allow
for development of significant algal
biomass. Category 3 reservoirs evidently
have the potential for significant
accumulation of algal biomass but that
potential can be limited by nonalgal
turbidity. In this study we attempted to
differentiate among the effects of color,
absorbence, algal turbidity and nonalgal
turbidity. It is evident that in Category 3
reservoirs we were unable to accomplish
this differentiation due to the fact that we
did not have data on Total Suspended
Solids. In future studies it is imperative to
collect Total Suspended Solids data to
explicitly differentiate the effects between
algal and nonalgal effects on water
transparency. At all events, the results
presented here provide a useful guide to
future investigations regarding reservoir
nutrient dynamics in the Pacific Northwest.
Discussion
The magnitude of the median values for
total P, total N, Secchi and chlorophyll
compare favorably with published values
(Table 9). Total P values were roughly two
times those reported earlier (Vaga and
Herlihy 2004). This is reasonable, since
Vaga & Herlihy included a preponderance
of natural lakes versus reservoirs and thus
their values would be expected to be lower
than for reservoirs alone. A similar
relationship exists for Secchi and
chlorophyll, i.e. the reservoirs appearto be
more eutrophic. Interestingly total N
values reported here appearto be one-half
of the lake+reservoir study cited above.
These data suggest that reservoirs are
relatively nitrogen poor as compared with
natural lakes but the reason for this is
unclear. However, these values for total P
and total N suggest that natural lakes in
the WFM and XW have N:P ratios of 38.7
and 13.2, respectively. In contrast
reservoirs in the WFM and XWhave ratios
of 9.7 and 3.9, respectively. Thus
irrespective of Nutrient Ecoregion, nutrient
criteria for reservoirs may have to take
total N into account.
Reservoirs typically are relatively nutrient
rich during the first years after construction
due to leaching of nutrients from the
recently flooded soils (Thornton et al.,
1990). Reservoir age did not appear to be
a factor in this study. Reservoirs in the
WFM were constructed between 1910 and
1989 (average age 56 years) and those in
the XW between 1895 and 1970 (average
age 64 years). In neither Nutrient
Ecoregion were any water quality/nutrient
parameters correlated with reservoir age,
suggesting that any effects due to
construction are no longer factors in the
nutrient dynamics of these reservoirs.
Reservoirs tend to be long and narrow
resulting in a gradients in physical,
chemical and biological parameters from
the upstream end to the dam. This
gradient can be divided into three general
zones, e.g. riverine, transition and
lacustrine (Thornton, et al., 1990).
Typically the transition zone exhibit the
highest productivity and the lacustrine
portion of reservoirs exhibit relatively low
productivity, since nutrients are removed
by algal settling in the transition zone. In
this study we limited sampling to the
lacustrine areas of the reservoirs and more
specifically to the forebay. Thus our
estimates of nutrient concentrations,
Secchi and chlorophyll are only
representative of the most nutrient poor
region of the reservoirs. Other zones in
these reservoirs would in all probability
exhibit different nutrient dynamics.
In addition to sampling location, season
also plays an important part in the be be
expected to have higher nutrients and
chlorophyll concentrations, nutrient
dynamics of reservoirs. As in lakes,
reservoirs often exhibit a spring bloom of
phytoplankton followed by a clear water
phase in midsummer and a secondary
bloom in the fall (Wetzel 2001). In seven
WFM reservoirs total P was < 10 ug/L
(Appendix 1,2). These low levels are
probably indicative of nutrient conditions in
these reservoirs but without further study
we cannot rule out the possibility of higher
concentrations early in the year. The
same holds for the six reservoirs in the XW
with concentrations < 30 ug/L.
In the Xeric West high chlorophyll
concentrations were not always observed
Page 14 of 21
-------
in the presence of high total P
concentrations (Appendix 1, 2). For
example, in Category 3 reservoirs
Chickahominy and Alexander had over 600
ug/L of total P and chlorophyll above 400
ug/L (Fig. 11, Appendix 2). However,
several reservoirs had over 100 ug/L of
total P (e.g. Upper Rock, Priday, Barry,
Grasmere, Warmsprings) but less than 10
ug/L of chlorophyll. In several cases the
presence of nonalgal turbidity (Upper
Rock) or absorbence due to DOC (e.g.
Barry, Priday, Grasmere) may have
prevented development of high algal
biomass. However, we cannot discount
the fact that we may have missed an algal
bloom in these reservoirs.
Reservoir
Categories
1
2
A 3
Figure 17. Location of the three reservoir
categories in the Xeric West. There is no
apparent relationship between category and
geographic location.
Table 8. Median and 75th percentile (Q3) values
for nutrient criteria variables in the three
proposed nutrient criteria reservoir categories, n
= number of reservoirs in category.
c
1
2
3
n
9
11
7
TP
Med(Q3)
25(33)
137(238)
344( 686)
TN
Med(Q3)
258(399)
483(578)
385(1930)
Chl-a
Med(Q3)
2.1(4.6)
27.9(76.9)
7.9( 460.6)
Secchi
Med(Q3)
2.10(4.4)
0.70(0.9)
0.15(0.3)
Table 9. Comparison of median (± 1 se) values
for total P, total N, Secchi and chlorophyll
observed in this study with those calculated from
Vaga and Herlihy (2004) for lakes and reservoirs
in corresponding Nutrient Ecoregions.
Variable/
Source
TP (ug/L)
This study
Vaga/Herl
TN (ug/L)
This study
Vaga/Herl
Secchi (m)
This study
Vsga/Herl
Chla (ug/L)
This study
Vaga/Herl
WFM XW
Mean
23.0
11.3
213.2
437.5
3.4
4.5
4.6
2.0
se
13.6
2.4
54.3
63.6
0.5
0.2
3.1
0.5
Mean
109.0
62.5
430.9
825.0
0.7
1.6
7.5
4.0
se
43.7
14.4
46.3
156.4
0.3
0.4
27.8
1.8
The differences observed in reservoir total
P and Secchi depth in the two different
Nutrient Ecoregions reflects differences in
landscape. Land use in the WFM is roughly
0,12 and 80% agricultural, range land and
forest, respectively. In the Great Basin
and Range Ecoregion, where all but one of
the XW reservoirs is located, land use is
agricultural = 0, Range Land = 83 and
Forest = 10%. These differences as well
as differences in geology and weathering
probably account for the observed
differences in total P and Secchi depth.
The high correlation of Potassium and total
P in the WFM (Fig. 14, Table 6) suggests
a single relatively homogenous source for
phosphorus in these reservoirs, with
relatively low effect of land use differences.
The much weaker correlation observed in
the XW is likely due to a more
heterogeneous source. Whether this
heterogeneity is due to land use e.g.
fertilizers or geophysical processes, or a
combination of both, is unknown.
3- :
1
100 =
« -
= :
o
10-
a E
1 !
i-
i *
*?
6
Reservoir Category
Figure 18. Distribution of total P values in the
three reservoir categories in the Xeric West.
There was a clear difference among the
categories as defined by cluster analysis.
Boxes show the interquartile range and
median, whiskers the minimum/maximum
Page15of^lues-
-------
IB
."5
1 2 3
Reservoir Category
Figure 19. Distribution of total N values in the
three reservoir categories in the Xeric West.
Distributions of total N overlapped more than for
total P among categories. Boxes show the
interquartile range and median, whiskers the
minimum/maximum values.
Conclusions
This study suggests that Nutrient
Ecoregions are useful as an initial
stratification variable for developing numeric
nutrient criteria for reservoirs. There
appears to be several lines of fruitful future
investigation to further the development of
nutrient criteria:
1) The minimum size of reservoir should be
determined that will be subject to criteria.
Small reservoirs, particularly in the Xeric
West, are most likely to be subject to draw
down during the summer to the level that
makes criteria development unnecessary.
2) Conversely, the maximum size of
reservoir should be determined where the
relationship between measurable landscape
parameters and nutrient concentrations in
the reservoir can be empirically determined.
For example, Columbia River reservoirs
likely have no need for development of such
criteria.
3) Greater resolution of nutrient dynamics in
reservoirs in the Western Forested
Mountains should be undertaken to
ascertain to what extent these reservoirs
can be considered as a single category.
Such a determination may be made simply
by accounting for local anomalies, e.g. high
total P groundwater. It is also possible that
reservoirs in the WFM are all similar
enough to warrant placement into a single
category.
4) Greater resolution of nutrient dynamics
in reservoirs in the Xeric West should be
undertaken to determine whether the three
categories suggested in this study are
sufficient to account for all reservoirs in
this Ecoregion. In this study there was
only one reservoir in the Columbia Plateau
(Level III Ecoregion). Vaga and Herlihy
(2004) found that lakes/reservoirs had a
median and 75%percentile total P
concentrations in the Columbia Plateau of
30 and 71 ug/L. In the Northern Basin and
Range (Level III Ecoregion) similar values
were reported and 90 and 204 ug/L. Thus
it may not be appropriate to combine these
two Level III Ecoregions into one
population. In addition TSS rather than
chlorophyll should be used in the
classification process.
5) Downstream effects. The obvious high
concentrations of total P in the Xeric West
reservoirs should be evaluated with
respect to tailwater effects on downstream
habitats.
6) The value of N:P ratios, e.g. 3.9 N:P in
the Xeric West, should be evaluated with
respect to the need for nitrogen criteria
and the concomitant effects on
downstream systems.
7) Shoreline characteristics seem to be
of less importance in determining nutrient
concentrations than overall watershed
characteristics. We surmise that at least
for reservoirs that experience draw down,
the effects of season are more important
than shoreline characteristics at full pool.
Acknowledgments. We would like to thank Peter
Leinenbach and George Gibson for reviewing an
earlier draft of this report.
Page 16 of 21
-------
i
Reservoir Category
Figure 20. Distribution of Secchi depth in the
three reservoir categories in the Xeric West.
There was a clear difference among the
categories as defined by cluster analysis.
Boxes show the interquartile range and
median, whiskers the minimum/maximum
values.
p
§> 100,0 =
f
o.
o 10.0=
5 :
1,0-
6
L
r
Reservoir Category
Figure 21. Distribution of chlorophyll values in
the three reservoir categories in the Xeric West.
The low variability in Category 1 reservoirs is
due to relatively low nutrients. Category 2
reservoirs exhibit a higher median and variance
due to seasonal effects, i.e. low nonalgal
turbidity but higher total P. Category 3
reservoirs have low median but high potential
for algal growth. Boxes show the interquartile
range and median, whiskers the
minimum/maximum values.
Page 17 of 21
-------
References
Baker, J.R., D.V. Peck, and D.W. Sutton (eds.). 1997. Environmental Monitoring and
Assessment Program Surface Waters: Field Operations Manual for Lakes. EPA/620/R-
97/001. U.S. Environmental Protection Agency, Washington DC.
Cooke, G.D. and R.H. Kennedy. 1989. Water Quality Management for Reservoirs and
Tailwaters. In: Reservoir Water Water Quality and Management Techniques. Tech. Rep.
E-89. U.S. Army Corps of Engineers, Washington, D.C.
Emmons, E.E., Jennings, M.J., and C. Edwards. 1999. An alternative classification method
for northern Wsconsin lakes. Can. J. Fish. Aquat. Sci. 56: 661-669.
Environmental Protection Agency. 2000. Nutrient Criteria Technical Guidance Manual Lakes
and Reservoirs, EPA-822-BOO-001, U.S. Environmental Protection Agency:
Washington, DC.
Kennedy, R.H. 1999. Reservoir design and operation: limnological implications and
management opportunities. In: Tundisi, J.G. and Straskraba, M., eds. Theoretical
reservoir ecology and its applications. Brazil and the Netherlands: International Institute
of Ecology, Brazilian Academy of Sciences, and Backhuys Publ.. pp. 1-28.
Kennedy, R.H. 2000. Nutrient criteria: considerations for Corps of Engineers Reservoirs.
Water Quality Technical Note. U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Omernik, J.M. 1987. Ecoregions of the conterminous United States. Annals Assoc. Amer.
Geogr. 77(1):118-125.
Perkins, R.G. and G.J Underwood. 1999. Gradients of Chlorophyll A and Water Chemistry
Along an Eutrophic Reservoir with Determination of the Limiting Nutrient by In Situ Nutrient
Addition. Wat. Res. Vol. 34 No.3, 713-724.
Ryder, R.A. 1978. Ecological heterogeneity between north-temperate reservoirs and glacial
lake systems due to differing succession rates and cultural uses. Verh. Internat. Verein.
Limnol. 20: 1568-1574.
Straskraba, M. J.G. Tundisi and A. Duncan 1993. Comparative Reservoir Limnology and
Water Quality Mangement.
Thornton, K.W., B.L. Kimmerland F.E. Payne. Reservoir Limnology: Ecological Perspectives.
Wley-lnterscience Publication, New York. 246pp.
USEPA. 1996. National Nutrient Assessment Workshop Proceedings. USEPA Office of
Water. EPA-822-96-004.
USEPA. 1998. National strategy for the development of regional nutrient criteria. USEPA
Office of Water. EPA-8221-R-98-002.
USEPA. 2000. Nutrient Criteria Technical Guidance Manual: Lakes and Reservoirs. USEPA
Office of Water. EPA-822-BOO-001.
Vaga, R.M. and AT. Herlihy. 2004. A GIS inventory Of Pacific Northwest Lakds/Reservoirs
and an analysis of historical nutrient and water quality data. EPA 910-R-04-009.
Page 18 of 21
-------
Vaga, R.M. 2005. Analysis of hydrologic and Water Quality Data for American Falls Reservoir,
ID. EPA910-R-05-001. EPA 910-R-05-002.
Vaga, R.M. and AT. Herlihy. 2005. A Classification of lakes in the Coast Range Ecoregion
with respect to nutrient processing. EPA 910-R-06-003.
Willen, E., Hajdu, S., and Pejler, Y. 1990. Summer phytoplankton in 73 nutrient poor Swedish
lakes: classification, ordination, and choice of long term monitoring objects. Limnologica,
20: 217-228.
Wetzel, R.G. 1990. Reservoir ecosystems: conclusions and speculations, pp. 227-238. In:
K.W. Thornton, B.L. Kimmel, and F.E. Payne (eds.), Reservoir Limnology: Ecological
Perspectives. John Wiley and Sons, Inc., New York, NY.
Page 19 of 21
-------
Appendix 1. Nutrient and site data for Western Forested Mountain reservoirs.
NIDID
ID00337
WA00263
WA00264
WA00291
WA00170
OR00016
OR00748
OR10020
WA00183
OR00548
OR83030
WA00223
WA00021
WA00152
OR00550
OR00098
OR00579
ID00115
WA00153
OR00554
OR00212
Lake Name
Brundage
Bumping Lake
Clear Lake
Conconully
Diablo Lake
Fern Ridge
Galesville
Henry Hagg
Judy
Lake Simtustus
Laurence Lake
Leader Lake
Long Lake
Mayfield
N. Fork Clackamas
Ochoco
Prineville
Sage Hen
Skookumchuck
Toketee
Willow Creek
Latitud
45.0418
46.8733
46.6333
48.5583
48.7133
44.1150
42.8492
45.4917
48.4717
44.6933
45.4583
48.3617
47.8367
46.5033
45.2417
44.2983
44.1133
44.3255
46.7850
43.2647
42.4800
Longitud
116.1306
121.3000
121.2667
119.7450
121.1300
123.2917
123.1778
123.2139
122.1883
121.2300
121.6583
119.6967
117.8383
122.5900
122.2817
120.7250
120.7800
116.1941
122.7167
122.4194
122.4433
Chl-
2.4
0.1
1.1
7.5
1.8
36.9
5.7
6.0
1.3
34.8
7.5
9.8
3.5
4.6
0.1
2.1
21.2
1.4
2.8
0.4
18.8
Secc
4.2
7.8
6.9
3.7
7.3
0.3
4.1
3.3
3.0
1.5
3.4
3.3
5.2
3.4
4.6
3.2
3.8
4.7
4.2
0.7
Max
8.7
11.0
17.4
6.8
48.5
6.5
31.7
11.2
6.8
42.0
12.5
8.0
16.7
27.5
33.2
23.6
34.0
9.4
29.5
7.6
7.9
TP
10
8
21
26
4
36
10
17
9
61
17
37
110
9
14
30
39
9.5
440
58
36
TN
82
279
55
182
116
93
192
164
61
403
864
56
227
177
99
173
58
167
NIDID = National Inventory of Dams ID code
Chl-a = Chlorophyll-a (ug/L)
Secchi = Secchi Depth (m)
Max Depth = Maximum Reservoir depth at time of sampling (m)
TP = Total Phosphorus (ug/L)
TN = Total Nitrogen (ug/L)
Page 20 of 21
-------
Appendix 2. Nutrient and site data forXeric ecoregion reservoirs.
NIDID
ID00060
OR00108
ID00136
ID00282
ID00194
ID00042
OR00344
ID00045
OR00228
ID00006
ID00190
OR00559
OR00181
ID00193
ID00285
OR00569
OR00377
ID00250
OR00369
ID00010
WA00566
ID00007
OR00592
ID00239
OR00157
OR00082
ID00074
Reservoir Name
Alexander
Barry
Ben Ross
Black Canyon
Blue Creek
Bray Lake
Bryant Mt.
Cedar Creek
Chickahominy
Daniels
Grasmere
JC Boyle
Kern
Little Blue Creek
Mann Creek
Mud Lake
Muddy Creek
Paddock Valley
Friday
Sand Creek
Scooteney
Stone
Thief Valley
Trail Storage Pond
Upper Rock Creek
Warmsprings
Weston Creek
Latitude
42.6446
42.0950
44.5235
43.9302
42.3097
43.0344
42.1050
42.2237
43.5450
42.3455
42.3555
42.1231
42.9183
42.2917
44.3918
42.2117
42.1950
44.1983
42.3400
44.2025
46.6667
42.0680
45.0150
43.0553
42.6867
43.5850
42.1302
Longitud
1 1 1 .6960
119.5317
116.4446
116.4357
116.1807
114.8760
121.3283
114.8786
119.6133
112.4424
115.9039
122.0439
118.7817
116.1203
116.8928
119.7150
120.5200
116.5980
119.9000
111.6144
119.0333
112.6942
117.7783
115.3196
119.3050
118.2083
112.1260
Chl-
519.
2.9
2.2
2.6
2.2
76.8
10.2
64.0
460.
4.8
4.4
58.3
27.9
18.8
7.5
29.3
176.
222.
4.9
0.9
12.7
4.6
14.8
5.5
7.9
4.4
2.8
Secc
0.2
0.2
2.5
4.2
0.9
0.7
0.8
0.7
0.2
4.5
0.3
0.2
0.7
0.8
1.7
0.2
0.5
0.3
0.4
1.1
1.3
1.1
0.6
0.2
1.4
4.7
Max
15.7
4.0
12.0
21.0
5.8
2.8
1.1
7.8
3.6
9.8
5.0
8.9
5.8
6.3
14.4
2.7
6.4
4.5
3.8
2.9
11.0
6.7
9.1
6.0
3.7
11.3
4.7
TP
686
344
22
22
108
137
23
151
887
42
201
252
73
109
26.5
456
325
110
201
25
46
23
225
104
340
238
33
TN
243
385
118
93
509
371
598
468
193
258
367
123
274
292
123
453
579
929
365
268
792
151
483
514
289
251
399
NIDID = National Inventory of Dams ID code
Chl-a = Chlorophyll-a (ug/L)
Secchi = Secchi Depth (m)
Max Depth = Maximum Reservoir depth at time of sampling (m)
TP = Total Phosphorus (ug/L)
TN = Total Nitrogen (ug/L)
Page 21 of 21
------- |