Biological Control of Eutrophication in Lakes - Environmental Science

29, 3, 784-786. Note: In lieu of an abstract, this is the article's first page. Click to increase image size Free first page. View: PDF | PDF w/ Links...
0 downloads 0 Views 339KB Size
Environ. Sci. Technol. 1995, 29, 784-786

Bioloyical Control of Eutrophication in lakes STEPHEN R . CARPENTER,*,' D A V I D L . CHRISTENSEN,+ JONATHAN J . C O L E , * KATHRYN L . C O T T I N G H A M , t X I H E , § J A M E S R. HODGSON," J A M E S F . KITCHELL,I SUSAN E . K N I G H T , + MICHAEL L. PACE,* DAVID M. POST,+ DANIEL E . S C H I N D L E R , + A N D N I CH 0 L A S V O I C H I C K t Center for Limnology, 680 North Park Street, University of Wisconsin, Madison, Wisconsin 53706, Institute of Ecosystem Studies, Box AB,Route 44A, Millbrook, New York 12545, National Marine Fisheries Service, 2580 Dole Street, Honolulu, Hawaii 96822, and Department of Biology, St. Norbert College, De Pere, Wisconsin 54115

Eutrophication of lakes, expressed as excess growth of planktonic algae, is caused by excessive inputs of phosphorus and can be mitigated by many mechanisms, including grazing. However, it has been hypothesized that grazing becomes ineffective with even modest increases in P input. W e tested this contention directly by fertilizing lakes that had contrasting food webs. A lake with zooplanktivorous fishes and small grazers accumulated algal biomass as predicted by Voll enwe ide r's mode I of eutrophic at ion. A lake with piscivorous fishes and large grazers accumulated about half the algal biomass predicted by the model. However, blue-green algae bloomed in both lakes. Grazing may effectively control total algal biomass over a relatively wide range of P input rates, but may not suppress irruptions of nuisance algae.

Introduction Eutrophication, which degrades water quality in many of the world's lakes (1, 2 ) , is caused by excessive inputs of nutrients, especially phosphorus, that stimulate nuisance growth of planktonic algae (3). P loading, the P input rate scaled by water volume (or area) and flushing rate, is a cornerstone of eutrophication management (1, 4). At a given P load, however, lakes respond differently, and this variation can be related to grazing by zooplankton ( I , 5 , 6 ) . Biomanipulation, a novel lake management technique, changes the food web to favor large zooplanktonic grazers that control algal biomass ( 4 , 7-9). Food web manipulation is accomplished by removing planktivorous fishes that consume grazers or by stocking piscivorous fishes that consume planktivores. Grazing by large zooplankton can be increased by either method. Numerous manipulations oflake food webs have demonstrated that large grazers can regulate algal biomass (9, 10-15). However, it is hypothesized that P loads above some upper limit or threshold may overwhelm the capacity of grazers to control algae (16, 20). This threshold has been variously estimated as '0.36-1.5 mg m-3 d-' (areal rates: '0.45-5.5 mg m-* d-l) (11, 13, 16, 18,21). Here, we report results of awholelake experiment designed to test the contention that grazer control becomes ineffective with increases in P loading. We added nutrients to lakes with contrasting food webs and measured responses of two indicators of eutrophication, chlorophyll and biomass of blue-green algae.

Methods Three small lakes of similar morphometry for which we have extensive background data (9) were compared. The reference lake (Paul Lake) was not manipulated and demonstrates the typical interyear variability known from lakes of the region (9). The top carnivores are piscivorous largemouth bass (Micropterm salmoides), which exclude planktivorous fishes from the pelagic zone of the lake. In the absence of planktivores, the herbivorous zooplankton is dominated by large cladocerans, mainly Daphnia Pulex. In the planktivore lake (Peter Lake), top carnivores are planktivorous minnows (predominantly golden shiner, Notemigonmcrysoleucas),whichwere stocked in 1991after removing all piscivorous fishes by angling, electrofishing, and ultimately rotenone treatment. Size-selective planktivory excludes large herbivores, and the zooplankton is dominated by rotifers and small-bodied crustaceans. The food web of the piscivore lake (West Long Lake) is similar to that of the reference lake. Dominant herbivores are large cladoceran grazers, especially Daphnia Pulex. Beginning in spring 1993,the planktivore and piscivore lakes were enriched using liquid fertilizer added continuously from mid-lake drip stations at a rate of 1 mg m-3 d-' for P in the epilimnia. The fertilizer had a N/P ratio of 25 by atoms. Areal loading rates were 3.06 mg m-2 d-l in the planktivore lake and 3.19 mg m-2 d-l in the piscivore lake. * Corresponding author: FAX: 608-265-2340; e-mail address: [email protected]. University of Wisconsin. 4 Institute of Ecosystem Studies. National Marine Fisheries Service. ' 1 St. Norbert College. +

*

784 1 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 29, NO. 3, 1995

0013-936W95/0929-0784$09.00/0

1995 American Chemical Society

0 Unenriched I

v) v)

V

I

T

I

,

,

,

I

I

I

,

T

g 4 A

58

40

Enriched 1

o Reference Planktivore

2L-dLdhA

v

Y

l

0

Vollenweider's model I

" -

91 92 93 Reference

91 92 93 Planktivore

91 92 93 Piscivore

FIGURE 1. Total herbivore biomass (A, g m-* dry biomass of zooplankton) and mean cladoceran length (B, mm) during 19911993 in the reference, planktivore, and piscivore lakes. Error bars are 95% confidence intervals for weekly observations from May to September.

These input rates exceed widely-accepted tolerance limits for eutrophication (1) and upper limits for grazer control of algae thought to apply to stratified lakes and reservoirs (areal rates of 1.3-2.7 mg m-* d-' or volumetric rates of 0.18-0.36 mg m-3 d-l) (16, 18). The expected in-lake phosphorus concentration [Pml resulting from measured inputs, flushing, and morphometry was calculated as

where Lp is specific surface loading of phosphorus (g m-2 y-l), qs is hydraulic load (m y r l ) , and 2, is mean depth (m) (1). Lp was calculated by mass balance (LP = change in standing stock of total P sedimentation outflow) for the reference lake in all years and for the planktivore and piscivore lakes prior to enrichment. P standing stock was measured in weekly profiles, and sedimentation was measured using hypolimnetic traps (9). Mass balance studies with LiBr tracer and rainfall data were used to estimate qs and outflow. Zm was calculated from hypsometry (9). Predictionsof chlorophyllfrom P loading using Vollenweider's model (1) were corrected for the retransformation bias of logarithmic regression (22). Water chemistry and plankton were monitored during summer stratification (approximatelyMay-September) at a central deep-water station in each lake (9). Chlorophyll a, corrected for pheopigments, was measured weekly at six depths. Phytoplankton were collected weekly from three depths in the epilimnion and counted and measured for biovolume by species. Zooplankton were sampled by vertical hauls and counted and measured by species. Hauls were corrected for net efficiencies. Animal dry masses were calculated from lengths using regressions.

+

+

Results Herbivore biomass in the reference lake (Figure 1A) was similar to those known from years of monitoring (9). Enrichment had no effect on herbivore biomass in the

I

I

I

I

90

1

Phosphorus Loading Rate FIGURE2 Chlorophyll econcentration (mg w 3 )versus mean annual P concentration (mg m 3 ) expected from loading, flushing, and morphometry. The solid line is the prediction of the Vollenweider model ( I ) . Means and 95% confidence intervals for weekly observations from May to September ara shown for the reference, planktivore, and piscivore lakes in 1991-1993.

planktivore lake, suggesting that herbivore biomass was effectively controlled by planktivory. Herbivore biomass increased after enrichment in the piscivore lake, although week-to-week variability was high. Length of cladoceran herbivores is among the best indicators of grazing intensity (5, 6, 9). Contrasts in the fish communities enforced substantial and consistent differences in cladoceran length among the lakes (Figure 1B). The planktivore lake had small cladocerans (Bosmina spp.). The reference and piscivore lakes had large cladocerans (Daphnia Pulex and Holopedium gibberum), Summer chlorophyll concentrations in the referencelake and the unenriched planktivore and piscivore lakes corresponded to predictions ofvollenweider's empiricalmodel (1)oflake eutrophication (Figure 2). Ordinate values above about 20 mg m-3 generallycause eutrophication (1).Upon enrichment, the chlorophyll concentration of the planktivore lake rose as predicted by the Vollenweider model (1). In contrast, the chlorophyll concentration of the piscivore lake was less than half the prediction of the Vollenweider model (1). Chlorophyll in the piscivore lake was significantly less than that of the planktivore lake, even though [P,] of the piscivore lake was almost twice as large as that of the planktivore lake. Although the lakes had very similar Lp, the piscivore lake had lower qsand therefore had a higher [Pm]. However, the measured post-enrichment total P concentrations in the lakes' epilimnions were very similiar (planktivore lake 25.9 mg m-? piscivore lake 24.3 mg m-3), Biomass of blue-green algae increased markedly after enrichment in both the planktivore and piscivore lakes (Figure3A). In both enriched lakes, mean biomass of bluegreen algae exceeded 2 g m-3, a level symptomatic of water quality problems (23). Biomass of blue-green algae was dominated by colonial taxa: Oscillatoriaspp. and Anabaena spiroides in the planktivore lake and A. spiroides and A. Jos-aquae in the piscivore lake. Blue-green algal concentrations in the reference lake remained low, as in past decades (9). VOL. 29, NO. 3,1995 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

785

0 Unenriched

Enriched

grazers from suppressing edible algae and, therefore, total chlorophyll. This finding does not support the hypothesis that a threshold for grazer control of chlorophyll results from interference by inedible algae. These findings suggest that thresholds of grazer regulation differ for edible and nuisance algae. Effective control of algal biomass and thereby water quality requires regulation of both types of algae. We need to further define the responses of algae across a range of P enrichments and food web structures in order to better predict benefits from biomanipulation.

%Jl ,&;,;,,AhJ Acknowledgments

We thank A. St. h a n d , S. Blumenshine, J. LeBouton, J. Reed, P. Troell, Y. Vadeboncoeur, and A. Wagner for assistance; D. M. Lodge and anonymous referees for comments on the manuscript; and the NSF for support.

w

0

91 92 93 91 92 93 Reference Planktivore

91 92 93 Piscivore

FIGURE 3. Biomass of blue-green algae (A, g ~ nwet - ~mass) and edible algae (9, g wet mass of algae less than 30pm in greatest axial linear dimension) in the phytoplankton of the reference, planktivore, and phcivore lakes during 1991-1993. Error bars are %YO confidence intervals for weekly observations from May to September.

Biomass of edible algae (