United States
Environmental Protection
Agency
Office Of Water
(4503F)
EPA841-F-95-002
September 1 995
Number 1
Watershed Protection:
Clean Lakes Case Study
Phosphorus Inactivation and Wetland
Manipulation Improve Kezar Lake, NH
Key Feature:
Project Name:
Location:
Scope/Size:
Land Type:
Pollutant(s):
Pollutant Source:
Data Sources:
Data Mechanisms:
Monitoring Plan:
Control Measures:
A lake restoration effort using sediment
phosphorus inactivation and wetlands
management
Kezar Lake
USEPA Region I/Sutton, New Hampshire
Watershed area 2770 ha;
Lake area 73.5 ha
Ecoregion 58, Northeastern highlands
Sediment phosphorus
Historical POTW discharges
State and local
Modeling and sediment core analysis
Yes
Aluminum salts injection and wetlands
management
Figure 1. Location
Hampshire
of Kezar Lake in New
Summary: Kezar Lake, located in central New Hampshire (Figure
1), has had a long history of water quality problems. Following a
major tish kill and persistent algae blooms beginning in the early
1960s, a Diagnostic/Feasibility Study (Phase I of the Clean Lakes
Program) was initiated in 1980 under section 314 of the Clean Water
Act. The study established that the lake's problems were from internal loading of phosphorus, and outlined a management
strategy to restore the lake. Lake sediments, contaminated by years of effluent discharge from a nearby wastewater
treatment facility, were the source of this internal loading.
A Restoration/Protection Project (Phase II of the Clean Lakes Program) commenced in 1984 to implement the
recommended management strategy for Kezar Lake. Two main approaches were employed to reduce phosphorus
concentrations in the lake. First, aluminum salts were injected into the hypolimnion to inactivate sediment phosphorus.
The injections were performed using a modified barge system that was an efficient and cost-effective means of aluminum
salts application. Second, upstream riparian wetlands were manipulated by elevating water level and planting new species
to encourage phosphorus removal by sedimentation and vegetative uptake.
From 1984 to 1994, comprehensive water quality monitoring programs (including part of the Phase II project, a state-
assisted volunteer program, and an EPA Phase III Post-Restoration Monitoring Project) were conducted to assess the
effects of the restoration activities. Results from these efforts have generally indicated that water quality has improved
following aluminum salts injection, although some parameters did worsen during 1988 and 1993. Furthermore,
recreational use of Kezar Lake has increased substantially since restoration.
Contact: Jody Connor, New Hampshire Department of Environmental Services, Water Supply and Pollution
Control Division, 6 Hazen Drive, P.O. Box 95, Concord, NH 03301, phone (603) 271-3414.
-------
BACKGROUND
Kezar Lake is a "fairly shallow, north temperate,
dimictic, phosphorus-limited lake at 276 m above sea
level" that drains approximately 28 square kilometers of
land in central New Hampshire (Figures 1 and 2)
(Connor and Martin 1989a). Land use in the watershed
is comprised of forestland (approximately 70 percent),
urban/residential (25 percent), and agriculture (5
percent). The lake's volume is 1,975,500 m3 and its
shoreline measures 3400 m. Mean and maximum depths
are 2.7 m and 8.2 m, respectively. Flushing rate for the
lake is 44.5 days.
In addition to nonpoint sources of pollution (e.g., runoff
and erosion) associated with land use, one point source
of particular concern exists in the Kezar Lake watershed.
In 1931, the nearby Town of New London opened a
sewage treatment facility that began discharging effluent
into Lion Brook, the main tributary to Kezar Lake. The
New London treatment facility was upgraded in 1969
and decommissioned in 1981.
Water quality problems in Kezar Lake were first
documented in 1963, when blooms of algae
(Cyanophyceae) were observed. Five years later,
following continued blooms and a massive fish kill, lake-
shore property values around Kezar Lake dropped
significantly. Throughout the 1960s and early 1970s,
copper sulfate applications and mechanical
destratification were used to attempt to improve water
quality. The success of these efforts proved to be short-
lived, however, and eventually ineffective in preventing
algae blooms. Although New London's waste was
rerouted to a new treatment facility in the Town of
Sunapee in 1981, algae blooms persisted in Kezar Lake.
ASSESSING AND CHARACTERIZING
THE PROBLEM
The Clean Lakes Program, section 314 of the Clean
Water Act, provides assistance to states for identifying
and restoring lakes that are water-quality-impaired. In
1979, the biennial statewide assessment of lakes in New
Hampshire ranked Kezar Lake as having the highest
priority for restoration. A Diagnostic/Feasibility Study
(Phase I of the Clean Lakes Program) for Kezar Lake
was initiated in 1980. The purpose of a
Diagnostic/Feasibility Study is to determine the causes
and extent of pollution, evaluate potential solutions to
water quality problems, and recommend an effective and
feasible method for restoring and maintaining water
quality in a particular lake.
The Diagnostic/Feasibility Study for Kezar Lake, which
was completed in 1983, provided the following
information (Connor and Martin 1989a):
kilometers
OUT
CPB - Clark Pond Bk LBA - Lion Bk, Above
TRIE - Trib Bk LBB - Lion Bk, Below
IN - Lake Inlet OUT - Lake Outlet .
LAKE - !n-lake Station.
Figure 2. Kezar Lake and Chadwick Meadows
monitoring stations (from Connor and Martin 1989a)
• Examination of the existing water quality and
trophic state of the lake.
• Analysis of historical water quality trends.
• Determination of hydrologic and phosphorus
inputs and outputs (budgets) for Kezar Lake.
• Determination of the importance of the lake's
sediments in providing phosphorus to support
phytoplankton (algae) populations.
• Recommendations to improve the water quality
in Kezar Lake.
Water quality and quantity data for the study were
analyzed from the lake itself, tributaries, groundwater
seepage meters and shallow wells, rainfall gauges.
Sediments from the lake bottom were also collected and
analyzed. Nutrient budgets were developed using mass
balance equations.
Trophic state, a measure of a lake's level of biological
productivity and age, was assessed for Kezar Lake
during the Diagnostic/Feasibility Study. Three separate
classification models, from the State of New Hampshire,
EPA, and Dillon-Rigler, all confirmed that Kezar Lake
was eutrophic. Phosphorus, the limiting nutrient for
biological growth in the lake, existed in high
-------
concentrations (> 30 ju.g/1) at a depth of 6 m during
nearly the entire first year of study. Such high levels of
phosphorus translate into poor water quality because of
increased biological productivity. Water quality
parameters measured in Kezar Lake during the study
included high chlorophyll a concentrations (indicative of
algae blooms), low transparency, and low dissolved
oxygen levels, especially during summer months.
Another major determination made in the
Diagnostic/Feasibility Study was the source of the
phosphorus causing the water quality problems in Kezar
Lake. The main external source of phosphorus, the
New London Sewage Treatment Facility, had been
decommissioned in 1981, eliminating 71 percent of the
external phosphorus load. Blooms of algae persisted
after this date, however, forcing researchers to look
elsewhere for the source. Through sediment core
analysis, computer modeling, and mass balance, they
established that internal loading of phosphorus from lake
sediments was the controlling factor in determining the
trophic state of the lake (Snow and DiGiano 1976,
Connor and Martin 1989b). The models showed that
lake phosphorus concentrations were more sensitive to
changes in sediment loadings than to, morphological or
watershed loading changes. Lake sediments, which
often contain much higher concentrations of phosphorus
than does the lake water, can provide a net flux of
phosphorus into the water under anaerobic conditions
(Wetzel 1983).
The final part of the Diagnostic/Feasibility Study focused
on providing recommendations to restore and maintain
water quality in Kezar Lake. The main objective for
lake restoration was to prevent phosphorus in the
sediment from continuing to enter lake water. The
Diagnostic/Feasibility Study recommended that the most
feasible method to accomplish this objective was to inject
aluminum salts into the hypolimnion to inactivate the
sediment phosphorus.
Although the Diagnostic/Feasibility Study determined
that most of the phosphorus in Kezar Lake came from
the lake sediments, additional management measures
were also recommended to deal with external phosphorus
inputs from the watershed. The Study proposed
manipulating Chadwick Meadows, an upstream riparian
wetland area (Figure 2), to remove phosphorus that
would enter the lake from Lion Brook. According to the
hydrologic budget developed in the study, Lion Brook
contributes nearly 90 percent of the annual inflow to
Kezar Lake (Figure 3) and is therefore an appropriate
focal point for restoration. Specific activities proposed
in the wetland included increasing water level in the
Meadows and planting additional vegetation,
theoretically causing less phosphorus to enter the lake
because of sedimentation and vegetative uptake (Connor
and Martin 1989a).
Birch Brook
5%
Groundwater
2%
recipitation
4%
Runoff
2%
Lion Brook
87%
Figure 3. Kezar Lake inflow distribution, 1981-1982
(from Connor and Martin 1986)
IMPLEMENTATION AND MONITORING
EFFORTS
Based on recommendations from the 1983
Diagnostic/Feasibility Study, aluminum salts injection
and wetlands management projects were implemented to
reduce phosphorus concentrations in Kezar Lake. To
measure changes in the lake's status due to restoration
efforts, a water quality monitoring program was
instituted in 1984 and pursued through 1988 (Connor and
Martin 1989a). These activities were performed, in
part, through a section 314 EPA grant for a Restoration/
Protection Project (Phase II of the Clean Lakes
Program). Additional monitoring activities were also
performed from 1988 to 1994 through a state-assisted
volunteer program and an EPA Phase III Post-
Restoration Monitoring Project.
Phosphorus Inactivation
Aluminum salts injection was selected for Kezar Lake
partially because of the success this methodology has had
in reducing phosphorus concentrations in other thermally
stratified lakes (Connor and Martin 1989a). The
effectiveness of aluminum salts application rests on the
ability of aluminum to form complexes, chelates, and
insoluble precipitates with phosphorus, thereby removing
it from the water column and depositing it in the
sediment in forms unusable by phytoplankton.
Depending on pH, phosphorus concentration, aluminum
concentration, and the rate at which additional
phosphorus is supplied, aluminum salts can provide
-------
Table 1. A comparison of aluminum dose, cost, and productivity for phosphorus inactivation (from Connor and Martin
1989b)
Lake
Medical Lake,
Washington
Annabessacook
Lake, Maine1
Kezar Lake,
New Hampshire2
Lake Morey,
Vermont2
Cochnewagon
Lake, Maine2
Sluice Pond,
Massachusetts2
3 Mile Pond,
Maine3
'old barge system
2modified harvester
3new barge system
Year
Treated
1977
1978
1984
1986
1986
1987
1988
Area
Treated
(ha)
60
121
48
133
97
6
266
Aluminum Dose
8.0 g Al/m3
Aluminum Sulfate
25 g Al/m3
Aluminum Sulfate
Sodium Aluminate
40 g Al/m3
Aluminum Sulfate
Sodium Aluminate
45 g Al/m2
Aluminum Sulfate
Sodium Aluminate
18 g Al/m3
Aluminum Sulfate
Sodium Aluminate
20 g Al/m2
Aluminum Sulfate
Sodium Aluminate
20 g Al/m2
Aluminum Sulfate
Sodium Aluminate
Costs for
Chemicals,
Labor and
Equipment
$132,093
$234,000
$65,604
$165,640
$81,840
$13,196
$170,240
Personday/ha Cost/ha
No data $2,202
1.12 $1,934
0.50 $1,367
0.57 $1,245
0.41 $844
0.67 $2,199
0.06 $640
long-term inactivation of sediment phosphorus (Connor
and Martin 1989a). Furthermore, aluminum has been
shown to have no toxicity to aquatic life at the pH and
dose necessary for lake restoration (Cooke and Kennedy
1981). Although not all forms of phosphorus (e.g.,
dissolved organic phosphates) are removed by aluminum
salts application, this methodology has proven to be an
effective strategy for phosphorus inactivation in many
water-quality-impaired lakes.
The week prior to aluminum salts application, copper
sulfate was applied as an algicide to remove phosphorus
tied up in the phytoplankton. Theoretically, this
phosphorus could recycle in the lake system for many
years (Connor and Martin 1989a). Additionally,
bioassays were conducted to assess the impact of both
the copper sulfate and aluminum salts applications to
benthic macroinvertebrates in Kezar Lake. Results from
before and after the applications indicated no apparent
detrimental effects to the macroinvertebrate community
(Connor and Martin 1989a).
Pilot jar and tank studies were also performed before
aluminum salts application to determine the best ratio
and dosage of aluminum sulfate and sodium aluminate
for phosphorus inactivation. Based on results from these
studies, a 10-hectare portion of Kezar Lake was treated
using 30 mg Al/m2 at a 2:1 aluminum sulfate-to-sodium
aluminate ratio. Since no adverse impacts on aquatic
biota were observed following this application, an
additional 48-hectare area of Kezar Lake was treated at a
higher concentration (40 mg Al/m2 at the same ratio) to
improve flocculation.
A special method for applying aluminum salts on Kezar
Lake was developed to improve both efficiency and cost
(Connor and Smith 1986). Prior to the Kezar Lake
project, aluminum salts were applied using large barges
that were slow and imprecise. A weed harvester was
modified to simultaneously apply two aluminum salts and
carry a large payload. These alterations provided a less
cumbersome, more maneuverable means by which to
deliver aluminum salts accurately and quickly.
Table 1 summarizes cost-effectiveness information
associated with seven phosphorus inactivation projects.
Note the varying degrees of effectiveness based on the
application system used. Additional improvements (i.e.,
"new barge system" in Table 1) have further increased
the efficiency and cost-effectiveness of aluminum salts
application since the development of the modified barge
for Kezar Lake (Connor and Smith 1986).
As part of the Phase II Project for Kezar Lake, intensive
monitoring was conducted for 4 years to determine the
effectiveness of the aluminum salts applications. Water
quality parameters included in the monitoring program
were dissolved oxygen, pH, alkalinity, total dissolved
aluminum, total phosphorus, chlorophyll a,
transparency, phytoplankton, and zooplankton. A
qualitative summary of the response of each of these
parameters from 1984 to 1988 is given in Table 2.,
Initial success was realized following treatment, but
-------
Table 2. Water quality response to sediment phosphorus inactivation in Kezar Lake (from Connor and Martin 1989b)
Parameter
Dissolved Oxygen
pH
Alkalinity
Dissolved Aluminum
Total Phosphorus
Chlorophyll-a
Transparency
Phytoplankton
Zooplankton
Water Quality Response
reduced hypolimnetic DO
deficit
reduced variance; increased
hypolimnetic pH
reduced variance; reduced
concentration
no impact
reduced variance; reduced
concentration
reduced peak and mean
concentration
reduced variance; increased
transparency
reduced abundance; elimination
of noxious blue-greens
fewer cladocerans; elimination
of Daphnia as a co-dominant;
increased Keratella; decreased
Polyartha; increased ciliates
Duration of
Major
Response
treatment year only
4 years
3 years
n/a
3 years
3 years
2 years
3 years
7
Return to
Pre-Treatment
Conditions
1987
1988
1987
n/a
still better than
pre-treatment
conditions
1988
still better than
pre-treatment
conditions
1987, w/no major
blooms as of 1988
community still
altered as of
1988
Response Mechanisms
toxic effect on BOD-producing
microbes
decreased algal productivity;
reduced anoxia in hypolimnion
decreased algal productivity; reduced
anoxia in hypolimnion; direct effect
of treatment
none
immediate effect of alum; reduced
anoxia in hypolimnion; ongoing
effect of alum
reduced phosphorus supply
reduced phytoplankton abundance
reduced phosphorus supply
altered food chain by change in
phytoplankton community structure'
within 4 years many parameters returned to near
pretreatment levels, although this change may be due to
meteorologic variability. Most parameters did show
stabilization (i.e., less extreme variability), however, at
the end of the 4-year monitoring period (Connor and
Martin 1989b). Furthermore, and most significantly,
these levels were suitable for recreation, and average
attendance at Wadleigh State Park, which abuts the lake,
increased by almost 2000 people per summer in 1984
and 1986.
Additional monitoring from a state-assisted volunteer
program and an EPA Phase HI Post-Restoration
Monitoring Project was performed from 1988 to 1994 to
supplement the Phase II monitoring and provide a longer
time frame by which to evaluate water quality changes in
the lake. Results from these monitoring studies indicate
that water quality had, in fact, generally improved since
restoration and that the poor quality measured during the
last year of the Phase II project in 1988 (as well as in
1993) was not indicative of overall water quality trends.
A quantitative example of the concentrations of
chlorophyll a from 1980 to 1994, shown in Figure 4,
represents the improving water quality trend following
restoration.
Wetlands Management
The second management action taken to restore Kezar
Lake's water quality was manipulation of the 20-hectare
Chadwick Meadows (a seasonally flooded riparian area)
along Lion Brook. Research has shown that wetlands
attenuate phosphorus with distinct seasonal variation
(Connor and Martin 1989a). Although wetlands might
not attenuate or might even be a source of phosphorus in
the fall and spring during periods of high flow, several
studies have shown phosphorus removal in wetlands to
be greater than 80 percent during the summer growing
season, when algae growth is most common.
Macrophytic nutrient uptake and sedimentation of
suspended particulates are the primary mechanisms
responsible for phosphorus removal in wetlands.
4
To encourage sedimentation of phosphorus-laden
particles, the water level at Chadwick Meadows was
elevated in the fall of 1983 by installing flashboards
below the confluence of Lion Brook and Clark Brook
40
ir 35
t 30
I 25
f 20
e
10
5
Bars represent mean ± 1 standard error
1
I,.1
1 i S 8 1 8 1 i i
! I - i ' '
! i 1 s £ 8 I
Year
Figure 4. Kezar Lake summer mean chlorophyll a
values (from Connor and Martin 1989a)
-------
Pond (Connor and Martin 1986). The macrophyte
community in the wetland, composed primarily of blue-
joint grass (Calamagrostis canadensls), was also
supplemented with plantings of wild rice (Zinzania
aquatica) in 1985 and 1986, to aid in phosphorus
attenuation. It was anticipated that these manipulations
to Chadwick Meadows would decrease phosphorus
concentrations in Lion Brook, ultimately benefiting
Kezar Lake (Connor and Martin 1989a).
A monitoring program was established from 1984 to
1988 to calculate changes in the phosphorus budget and
measure the effects of the wetlands management
activities. Phosphorus concentrations and flow
measurements were taken monthly at the three tributaries
and at the outlet of Chadwick Meadows (Connor and
Martin 1989a). Results from the monitoring are shown
in Figure 5. Although there were a few months when the
wetland acted a sink, the overall effectiveness of
Chadwick Meadows in removing phosphorus from Lion
Brook was poor (Connor and Martin 1989a). The
restoration activities did, however, prove valuable in
increasing sedimentation and wildlife habitat.
Furthermore, costs associated with the wetland
manipulation were negligible, totaling $250.00 for the
purchase of wild rice.
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
I—I = 1985. im = 1986, ES3 = 1987, W = 1986
Figure 5. Monthly total phosphorus flux at Chadwick
Meadows, 1985-1988 (from Connor and Martin 1989a)
The conclusions of the Restoration/Protection Project in
the Phase II Final Report (Connor and Martin 1989a)
offered four main hypotheses for the water quality
response observed. First, the authors indicated the
possibility that the aluminum bonding sites provided by
the 1984 treatment eventually were all occupied,
preventing long-term phosphorus inactivation. Second,
the heavier aluminum salts, which initially created a
physical barrier between the sediment and water
interface, might have migrated vertically downward
through the sediment. This migration exposed some of
the sediment that might contribute additional internal
phosphorus loading. Third, additional phosphorus
entered the lake from the tributaries, perhaps as a result
of biological assimilation of phosphorus in Lion Brook
that occurred during effluent discharge from the New
London wastewater treatment facility. Fourth, historical
anoxic conditions that occur in the hypolimnion during
summer months in Kezar Lake increase the rate at which
sediment phosphorus is released into the hypolimnion.
A final hypothesis generated from more recent
monitoring data (collected from 1988 to 1994) suggests
that the water quality in Kezar Lake may be influenced
by the amount of annual precipitation (J. Connor, pers.
comm., May 1995). As Figure 4 indicates, chlorophyll
a levels following restoration (after 1984) fell below
those measured before restoration efforts, except during
1988 and 1993. During both of these years, annual
precipitation considerably exceeded normal amounts, as
did runoff. It is thought that nonpoint source loads from
the Kezar Lake watershed may contribute enough
additional phosphorus during periods of high
precipitation to noticeably decrease the water quality in
Kezar Lake. It appears now that the quality of Kezar
Lake is regulated by climatic conditions. High summer
precipitation produces high productivity, while drought
years, like 1995, produce record transparency and low
productivity.
LONG-TERM MONITORING STUDIES
As previously discussed, a state-assisted Volunteer Lake
Assessment Program was established to continue water
quality data collection for Kezar Lake and to provide a
means of public education following completion of the
Phase II Project in 1988. An ongoing 5-year EPA Phase
III Post-Restoration Monitoring Study is also assessing
specific longer-term effects of aluminum salts application
in Kezar Lake. Research in the Phase III Study
includes:
• An assessment of potential leaching of sediment
aluminum into overlying water.
A comparison of aluminum levels in horned
pout (Ictalurus nebulosus) and yellow perch
(Perca flavescens) between Kezar Lake and
several control lakes.
A comparison of macroinvertebrate diversity
and density between Kezar Lake and several
control lakes.
A comprehensive description of this research and the
results will be published in the near future in the Phase
III Final Report.
-------
REFERENCES
Connor, J.N. and M.R. Martin. 1986. Wetlands
management and first year response of a lake to
hypolimnetic aluminum salts injection. New Hampshire
Department of Environmental Services, Water Supply
and Pollution Control Commission, Staff Report Number
144. 76 pp.
Connor, J.N. and M.R. Martin. 1989a. An assessment of
wetlands management and sediment phosphorus
inactivation, Kezar Lake, New Hampshire. New
Hampshire Department of Environmental Services,
Water Supply and Pollution Control Division, Staff
Report Number 161. 109 pp.
Connor, J.N. and M.R. Martin. 1989b. An assessment
of sediment phosphorus inactivation, Kezar Lake, New
Hampshire. Water Resources Bulletin 25(4):845-853.
Connor, J.N. and G.N. Smith. 1986. An efficient
method of applying aluminum salts for sediment
phosphorus inactivation in lakes. Water Resources
Bulletin 22(4):661-664.
Cooke, G.D. and R.H. Kennedy. 1981. Precipitation
and imictivarion of phosphorus as a lake restoration
technique. U.S. EPA Ecological Research Series. EPA-
600/3-81-012. U.S. Environmental Protection Agency,
Washington, DC.
Snow, P.O. and F.A. DiGiano. 1976. Mathematical
Modeling of Phosphorus Exchange Between Sediments
and Overlying Water in Shallow Eutrophic Lakes.
Report ENVE.54-76-3 to the Massachusetts Division of
Water Resources, Department of Environmental Quality.
244pp.
Wetzel, R.G. 1983. Limnology. 2nd Edition. Harcourt
Brace Jovanich, Orlando, FL. 767 pp.
This ease study was prepared by Tetra Tech, Inc.,
Fairfax, VA, in conjunction with EPA's Office of
Wetlands, Oceans, and Watersheds, Watershed Branch.
To obtain copies, contact your EPA Regional Clean
Lakes Coordinator, or request a copy from:
NCEP1
11029 Kenwood Road, Building 5
Cincinnati, OH 45242
FAX 513-489-8695
-------
•£» TJ
CO CD
O 3
O SB
^
o"
-o
I
0
C/>
CD
DO
c
CO
=j'
<0
(«• 3" CO
5-J §
CO 3
sr 3
o 0
p a
o a.
O 13
10 a
a
o'
3
(Q
------- |