United States
                    Environmental Protection
                    Agency
Environmental Research
Laboratory
Corvallis OR 97330
                    Research and Development
EPA-600/S3-83-057 Dec. 1983
SERA         Project  Summary
                    Response of  Skinner  Lake
                    (Indiana)  to Agricultural Drainage

                    C. D. McNabb, B. J. Premo, F. C. Payne, T. R. Batterson, and J. R Craig
                      Agricultural runoff has been shown
                    to be relatively rich in plant nutrients
                    that promote high densities in algae
                    populations in lakes. Decomposition
                    of these algae stresses the oxygen
                    system, which in very productive lakes
                    results in release of water quality de-
                    grading substances from sediments,
                    and causes shifts in animal and plant
                    populations toward less desirable spe-
                    cies. Turbidity in the water is increased
                    in the process. For lakes in the north
                    central States, phosphorus is most
                    commonly the key nutrient responsi-
                    ble for development of excessive algae
                    densities. Various agricultural  land
                    management practices are known to
                    promote conservation of this and other
                    nutrients on the land.
                      From measurements over an annual
                    cycle in 1978-79, a phosphorus runoff-
                    algae density model was developed for
                    a lake at the bottom of an agricultural
                    watershed in northern Indiana.  Like
                    other models in  the literature,  this
                    model  stated that mean total phos-
                    phorus and chlorophyll a concentra-
                    tions in the epilimnion of the lake in
                    the ice-free period of the year were
                    predictable from mean total phospho-
                    rus in inflowing streams and residence
                    time of water in thejake  (t^,). For this
                    particular system, [TP]epLand [chl  a]epj
                    were predictable from [TP]streams and
                    tyy measured during spring and summer,
                    rather than over a longer interval of the
                    year. Measurement of the last  two
                    variables during the period of spring
                    runoff just prior to thermal stratifica-
                    tion of the lake  when the flushing
                    coefficient (p) equaled 1, and continued
                    through the duration of stratification,
                    was sufficient
                      Land management practices were
                    implemented on erosion-sensitive por-
tions of the watershed between 1978
and  1981. These included conserva-
tion  tillage, terraces, livestock exclu-
sion, grassed waterways, diversions.
group tile mains, settling basins, and
stabilizing stream banks with vegeta-
tion. The model was tested in 1982 on
the assumption that this work would
change [~n*lrtmam& and alter total phos-
phorus and chlorophyll a concentra-
tions in the lake. It proved to be area-
sonable predictor of UPlapj and [chl a]^
for the ice-free period of the year. In the
absence of unusual rain events and by
design of the model, tw in 1979 and
1982 were nearly the same; 0.60 and
0.65 respectively. [TFJ^r^,,,,, was 127
mg rtv3 in 1979 and 100 mg rtv3 in
1982. Total  phosphorus and chloro-
phyll a concentrations in the lake were
reduced between years from 63 to 54
mg nv3 and 15 to 10 mg nr3 respec-
tively. The model predicted 55 mg nr3
UPlepi and  13 mg nv3 [chl a]^,  for

  When a lake model can be calibrated
to fit specific characteristics of a water-
shed, it becomes a powerful manage-
ment tool. It focuses attention of plan-
ners on management alternatives most
likely to achieve intended goals. This
study showed that proper manage-
ment of an agricultural watershed can
improve water quality of a receiving
lake. Maintenance of established land-
use  practices and  implementation of
others at key sites on the watershed
will likely result in continued improve-
ment of this lake.

   This Project Summary was developed
by EPA's Environmental Research Lab-
oratory, Corvallis, OR, to announce key
findings of the research project that is
fully documented in a separate report

-------
 of the same title (see Project Report
 ordering information at back).

 Introduction
   During 1978-1981, various land man-
 agement practices were put in place on the
 agricultural watershed of Skinner Lake in
 Noble County, Indiana. The objective was
 to reverse the process of eutrophication in
 the lake. Studies reported here were con-
 ducted to evaluate the effectiveness of the
 program.
   Skinner Lake had an average  depth  of
 4.4  m, a maximum depth of 10 m and
 surface  area of 49.4 ha.  One  hundred
 twenty-five permanent residences line the
 shore of the lake. Lake level was controlled
 by a concrete sill  dam that was con-
 structed in 1962. The lake discharged  to
 the  Croft Drain that  entered the South
 Branch of the Elkhart River some 8.3 km
 downstream.
   Stable thermal stratification during sum-
 mer was characteristic of Skinner Lake. It
 existed from  mid-May through mid-Octo-
 ber.  Figure 1  shows a typical temperature
 regime for the ice-free season. The hy-
 polimnion approached anaerobic  condi-
 tions rapidly  as stratification became es-
 tablished (Figure 2). Relative areal oxygen
 deficits in the hypolimnion calculated for
 the  interval 1 May through 15 May  in
 1978 and 1979 were in the range of 700-
 900 mg m"2  day"1.  Lakes with areal oxy-
 gen deficits greater than 550 mg  m"2 day"1
 have been considered  highly  fertile  or
 eutrophic. Other expressions of  eutrophy
 for Skinner Lake included extensive growth
 of aquatic macrophytes such as Nymphaea
 odorata. Ceratophyllum demersum, and
 Myriophyllum spicatum, and blooms  of
 blue-green algae (Aphanizomenon  sp.).
 The dominant  fish in Skinner  Lake in-
 cluded small sunfish (Lepomis spp.) and
 crapp\es(Pomoxisspp.), and a large rough
 fish population composed of the white
 sucker (Catastomus commersoni) and the
 golden shiner (Notemigonus crysoleucas).
 The lake's turbidity, algal blooms, macro-
 phytes and low oxygen conditions were
 blamed  for a general decline in fishing
 quality over the last two decades.

 The Watershed
  The area of Skinner Lake watershed was
 3649 hectares, 68%  of which was used
• for agriculture. The remaining 32% was in
 woodlands and wetlands. The  Rimmell
 inlet was the most important continuously
 discharging surface flow to Skinner Lake
 (Figure  3).   Eighty-two percent of the
 watershed (3091   ha)  was  within the
 Rimmell drainage system. Land there was
 put to varied  crop and livestock use. Sub-
     April    May     June     July     August  September October  November Dec.

Figure 1.    Temperature isopleths (°C) for Skinner Lake during the ice-free season of 1979.
   7

   8

   9

  10
                                                                      10 11 12
     April   May     June      July     August  September  October November Dec.
Figure 2.    Dissolved oxygen isopleths (mg I -') for Skinner Lake during the
          ice-free season of 1979.
 Figure 3.   The Skinner Lake watershed in Noble County, Indiana. Section boundaries
            form the grid shown.

-------
surface tiles  were widely employed  for
drainage. Three smaller streams carrying
tile and surface runoff entered the lake
from  agricultural land: the Hardendorff
system drained 1 80 ha, the Riddle stream
107 ha, and the Weimer stream 42 ha of
the watershed. The overflow from nearby
Sweet Lake ran along a 0.8 km channel
through a woodland to Skinner Lake. The
Croft-Sweet system drained 229 ha of
watershed.
  The particular land treatment practices
implemented on the watershed of Skinner
Lake  from  1978 to 1981  were chosen
from the data and experience generated by
the nonpoint-source pollution study of
Black Creek in nearby Allen County, Indiana
These included settling basins,  conser-
vation tillage, group tile mains, terraces,
livestock exclusion, diversions, grassed
waterways, and planting vegetation  on
critical erosion sites. The work accom-
plished is shown in Figure 4.

Phosphorus - Algae Model
  The product of the  evaluation studies
was a model  which described  the quan-
titative relationship between phosphorus
concentration in inputs  to  the  lake and
algal  biomass in the lake in the ice-free
period of the  year. This model was based
on eutrophication models that were devel-
oped  in the literature over the past several
decades. More specifically, it resulted from
a modification of a  model developed by
Vollenweider and Kerekes  from data of
200 lakes in 50 countries participating in
the Organization for Economic Coopera-
tion and  Development (OECD) Coopera-
tive Program on  Lake Eutrophication.
Skinner Lake was considered phosphorus
limited, ratherthan nitrogen limited, on the
strength of high ratios of [TN]/[TP]. These
ranged from  19  to 220 in  Skinner Lake
during 1979.
  Modifications were made  to the Vollen-
weider and Kerekes model by considering
(1)  major sources contributing to [TP] in
loading to Skinner Lake,  (2) seasonal dif-
ferences in [TP]  in input and water resi-
dence time of the lake (tw), and (3) dif-
ferences in epilimnetic and  hypolimnetic
[TP] during summer, and their relation to
[cnta].
  Atmospheric  bulk loading to Skinner
Lake over a 12-month interval in 1978-
1979 amounted to 0.04 g TP nr2. While
this measurement fell within the range
predicted for the  geographical region, it
constituted only 1 A% of the annual load-
ing. Septic tank loading was estimated
using a constant, 0.08 kgTPcapuf1 yr1.
There were 125  cottages  and approxi-
mately 375 people served by septic tanks
         Minimum tillage


         Pasture & hay seedings


         Settling & lake basins
         Tile


     I    Terrace


   MM  Tree plantings


 ••«••••••  Diversion


 ----- Watershed boundaries


 sssss=5t  Grassed waterway


    •    Water control structure


    •    Animal waste pit

 Figure 4 (a).    Land management practices applied to the west portion of the Skinner Lake
              watershed during 1979-1981.
around  Skinner  Lake. By this constant,
they contributed 30 kg TP yr"1 . This septic
tank loading comprised 2% of the 1 978-
79 annual TP load.
  Estimates of internal phosphorus loading
were calculated for intervals of the ice- free
period of 1 979 from the following:
Tpinternai = A 
-------
Figure 4 (b).    Land management practices applied to the east portion of the Skinner Lake
              watershed during 1979-1981.
streams and t^ These are illustrated in
Table 1  and in  Figure 5 for the principal
stream discharging to Skinner Lake.  Be-
cause of the high flushing rate during the
spring overturn period (Table 1), TP inputs
from streams during the preceding winter
and period ofjjpring snow-melt had little
influence on [TP] in the lake during spring
overturn and the following summer.

Table 1.   Flushing Coefficients (p = Volume
          Outflow/Volume Lake) of Skinner
          Lake during Periods of
           1978-1979

        Period              (p)

Fall Overturn and Winter       0.02
9/6/78 - 2/6/79
Snow Melt                 1.30
2/6 - 3/20/79
Spring Overturn              1.06
3/20 - 5/22/79
Summer Stratification         0.17
5/22 - 9/6/79
Whole Year                 2.55
9/6/78 - 9/6/79
  The following model fit the 1978-1979
data for Skinner Lake:

       (1)[TPep,]ss=[TPstrearnl^s
                     1 + ^ss 2

In the  model, ss is the period of spring
overturn and summer stratification where
the length of the spring period is defined
as the  interval prior to the onset of strati-
fication during  which the  flushing coef-
ficient (p) of the laj
-------
      1000-
       ioo -4
   IS:
           0.1
                                  1
                                                       '\
                                                       10
                                   100
                                   Residence Time
£  represents plot of Skinner Lake coordinates of 1979.
•  represents plot of Skinner Lake coordinates of 1982.
Aa probable case for Skinner Lake with diversion of Rimmell stream around lake to outlet stream;
&[> probable case for Skinner Lake if settling basin on Rimmell stream removed 100% TP in
    stream discharge.

Figure 6.    Diagramatic representation of the Skinner Lake model.
                                          Skinner Lake, diverting the stream would
                                          have increased tw   from 0.63 to 2.35.
                                          The net  effect  of diverting the Rimmell
                                          would have been to reduce the [TP.DJSS
                                          from 63 mg nr3 to 45 mg m'3, and [chla]ss
                                          from 1 5  mg  nr3  to 11  mg rrr3. Another
                                          alternative for reducing the effect of runoff
                                          and nutrients on  the lake was a settling
                                          basin on the Rimmell stream just above its
                                          inlet to Skinner Lake. On the basis of 1979
                                          data, if the basin removed 100% of panic-
                                          ulate phosphorus from the water,  [TPstream]ss
                                          would have been reduced from 127 mg
                                          nr3to52 mgm-^lTPepJss fram63 mg nr3
                                          to 20 mg rrr3, and[chla]ssfrom  15 mg nr3
                                          to 7.5 mg m"3. As shown in Figure 6, this
                                          would represent a considerable improve-
                                          ment in water quality at Skinner Lake.
                                           Expectations in these examples provide
                                          a focus for considering costs and benefits
                                          of various strategies for land management
                                          and lake  improvement.  Results of this
                                          study call attention to the need for site-
                                          specific information on  which to  base
                                          cost-benefit judgements.
year variability in the systems. It should be
noted that by definition of variables in the
Skinner Lake model, comparable portions
of the water-year in 1979 and 1982 were
used  in this study. Between-year differ-
ences in rain events on the watershed, for
example, differences in the intensity of
individual events and associated differ-
ences in erosional runoff, were minimized
by the model. The time interval over which
tWs  was calculated was nearly the same
for  both years,  as  was the  value to t^,
(0.63  in 1979  and 0.65 in  1982]S
Unusually intense storms did not occur in
those intervals in either year. It is probable
that land management was the  primary
cause for the improvement observed in
Skinner Lake during this study. However,
ordinary between-year variability was not
measured in the system during years prior
to land  treatment  and only a continuing
trend  of improvement with  maintenance
of existing practices and implementation
of others  on the watershed in the years
ahead will confirm this.

Applications
  A calibrated model fit to a specific lake's
characteristics becomes a useful manage-
ment tool. One treatment considered by
planners of the Skinner Lake project was
diversion of the Rimmell stream directly to
the lake's  outlet Contours of the existing
landscape  favored such an approach. Com-
pensation  for disrupting intervening land
use would have been costly. The Skinner
Lake model could predict the effect of this
strategy if, for example, it had been imple-
mented in 1979.  Figure 6 shows that
diversion_pf the Rimmell would not have
lowered [TPstream]ss appreciably. Other in-
lets were  important m determining  total
[TpstreaJss-  Because  the Rimmell  pro-
vided a large percentage of inflow water to

-------
     C. D. McNabb, B. J. Premo. F. C. Payne. T. R. Batterson, and J. R. Craig are with
       Department of Fisheries and Wildlife, East Lansing. Ml 48824.
     Spencer A. Peterson is the EPA Project Officer (see below).
     The complete report,  entitled "Response of Skinner Lake (Indiana) to Agricultural
       Drainage," (Order  No. PB 83-233 676; Cost: $10.00, subject to change) will be
       available only from:
             National Technical Information Service
             5285 Port Royal Road
             Springfield,  VA 22161
             Telephone: 703-487-4650
     The EPA Project Officer can be contacted at:
             Environmental Research Laboratory
             U.S. Environmental Protection Agency
             Corvallis, OR 97333
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
FFICIALMAH.
                                                                                                 p"jMTy
                                                                                                 ™"  1  = 0291
                                                                                                   f*SOCj
                                                              ^^L^TTrsi  .  !
                                                                •*-   *, [• r:'-i T t\ C    '

                                                                                 * U.S. GOVERNMENT PRINTING OFFICE: 1984-759-102/808

-------