The Condor97:141-153 0 The CooperOrnithological Society 1995
WATERBIRD PREDATION ON FISH IN WESTERN LAKE ERIE: A BIOENERGETICS MODEL APPLICATION ’
CHARLES P. MADENJLW National Biological Survey, Great Lakes ScienceCenter, Lake Erie BiologicalStation, 6100 ColumbusAvenue, Sand&y, OH 44870
STEVEN w. GABREY U.S. Department of Agriculture,Denver Wildlfe ResearchCenter, 6100 ColumbusAvenue, Sandusky, OH 44870
Abstract. To better understand the role of piscivorouswaterbirds in the food web of westernlake Erie, we applied a bioenergeticsmodel to determine their total fish consump- tion. The important nesting speciesincluded the Herring Gull (Laru.r argentatus),Ring- billed Gull (L. delawarensis).Double-crested Cormorant (Phalacrocoraxauritus). Great Blue Heron (Ardea herodias), Black-crowned Night-Heron (kycticorax nycticorax);and Great Egret (Casmerodiusalbus). The impact of migrant waterbirds, including the Red-breasted Merganser(Merges serrator),on westernLake Erie fish biomasswas also consideredin the analysis.According to the modeling results,during the early 1990s piscivorouswaterbirds consumed13,368 tonnesof fish from westernLake Erie eachyear. This tonnagewas equiv- alent to 15.2% of the prey fish biomassneeded to supportthe walleye (Stizostedionvitreum) population in westernLake Erie during a singlegrowing season. The model application was useful in quantifying energyflow between birds and fish in a large lake ecosystem. Key words: Bioenergetics;ecological impact; fish consumption;Lake Erie; large lakes; modeling:predation; waterbirds.
INTRODUCTION There appears to be substantial overlap in the Large populations of several piscivorous water- diets of walleye and piscivorous waterbirds in bird species,including the Herring Gull (Larus western Lake Erie. The most important prey fish argentatus), Ring-billed Gull (L. delawarensis), for western Lake Erie walleye included the giz- Double-crested Cormorant (Phalacrocorax au- zard shad (Dorosoma cepedianum), alewife (Alo- rims), Great Blue Heron (Ardeu herodius), Black- sa pseudoharengus), white perch (Morone amer- crowned Night-Heron (Nycticorax nycticorax), icana), emerald shiner (Notropis atherinoides), and Great Egret(Casmerodius albus), residealong and spottail shiner (Notropis hudsonius) (Hart- the shores of western Lake Erie from spring man and Margraf 1992). Other prey fish included through autumn (Blokpoel and Tessier 199 1, Pe- the rainbow smelt (Osmerus mordax), freshwater tejohn and Rice 199 1, Scharfand Trapp 1994). drum (Aplodinotus grunniens), and yellow perch In addition, some piscivorous waterbird species, (Percaflavescens) (Hartman and Margraf 1992). suchas the Red-breasted Merganser (Mergus ser- Belant et al. (1993) and Hoffman (1977) have ratus), use the western basin of Lake Erie as a shown that gizzard shad, shiners, and alewives staging area during their spring and fall migra- were important diet constituents of the Herring tions (Southern 1974, Petejohn and Rice 199 1). Gull, Great Egret, Great Blue Heron, and Black- The walleye (Stizostedion vitreum) is the pre- crowned Night-Heron from western Lake Erie. dominant piscivorous fish in western Lake Erie Petejohn and Rice (199 1) reported large flocks (Knight et al. 1984, Hartman and Margraf 1992, of Red-breasted Mergansers feeding on gizzard Knight and Vondracek 1993). Although the wall- shad in Lake Erie during autumn. The diet of eye is assumed to be the dominant piscivore in Double-crested Cormorants from western Lake this ecosystem(Knight and Vondracek 1993) a Erie has not been examined (J. Ludwig, pers. rigorouscomparison of fish consumption by birds comm.), however the diet of Lake Ontario Dou- with that by walleyes has not been attempted. ble-crested Cormorants included alewife, yellow perch, white perch, and rainbow smelt (Weseloh and Casselman 1992). In summary, there is po- ’ Received20 June 1994. Accepted24 August 1994. tential for feeding competition between pisciv-
11411 142 CHARLES P. MADENJIAN AND STEVEN W. GABREY orous waterbirds and walleyes in western Lake a basic ecology standpoint, waterbird predation Erie. on fish is one link in the western Lake Erie food If prey fish populations are sufficiently re- web. Quantification of that link represents one duced, walleye growth and recruitment would be steptoward building an overall ecosystemmodel reduced. When prey fish are unavailable, wall- for the western basin of Lake Erie. Such a model eyes resort to a diet of invertebrates, and growth could be used to identify the important energy is substantially slower on an invertebrate diet pathways within the ecosystem. than on a fish diet (Forney 1966, Priegel 1970). Furthermore, recruitment of juvenile walleyes MODEL into the adult population may be greatly reduced if juvenile walleye diet is chiefly comprised of MODEL STRUCTURE FOR NESTING invertebrates rather than fish (Madenjian and BIRD POPULATIONS Carpenter 199 1). We estimated fish consumption by waterbird Bioenergeticsmodels can be used to estimate populations nesting within the perimeter of the the amount of food consumption by a particular western basin of Lake Erie. For purposesof com- animal population. Hartman and Margraf (1992) parison with the Hartman and Margraf (1992) fitted a bioenergetics model to the walleye pop- study, we defined the western basin of Lake Erie ulation (approximately 50 million age-one year as that portion of the lake west of Point Pelee on or older fish) of western Lake Erie, and calculated the Canadian side and west of the mouth of the that walleyes consumed, on average, 88,200 Huron River (in Ohio) on the U.S. side. Nesting tonnes of prey fish during each growing season populations of the following specieswere mod- (May through November) from 1986 through eled: Herring Gull, Ring-billed Gull, Double- 1988. Bioenergetic simulation models have been crested Cormorant, Great Blue Heron, Black- applied to seabird populations on the Oregon crowned Night-Heron, and Great Egret. SeeData coast (Wiens and Scott 1975) and in the Gulf of for Nesting Populations section for information St. Lawrence (Cairns et al. 199 1) to estimate fish on the bird nest surveys. Specieswith fewer than consumption. For freshwater systems, Weseloh 100 nests, such as the Common Tern (Sterna and Casselman (1992) used a simple bioener- hirundo), were not included in the modeling geticsmodel to estimate fish consumption by the analyses. Only nesting colonies within the west- Double-crested Cormorant population of Lake em basin perimeter were included in this anal- Ontario. ysis. Our objectives were to: (1) apply a bioener- For the nesting populations, we used a bio- getics model to the waterbird populations of energetics approach similar to the one used by western Lake Erie to estimate annual fish con- Cairns et al. (199 1). This approach, based on the sumption by birds, and then (2) compare fish work of Fumess (1978) and Wiens (1984) in- consumption by birds with that by walleyes in corporated the allometric relationship between western Lake Erie. field metabolism and body mass developed by Results of our modeling exerciseshould be of Bitt-Friesen et al. (1989). Birt-Friesen et al. (1989) interest to applied and basic ecologistsalike. The observed a strong correlation between the loga- walleye population in western Lake Erie supports rithm of field metabolism, as measured by dou- valuable commercial and recreational fisheries bly labeled water and activity timers, and the (Knight and Vondracek 1993). Due to a decrease logarithm of bird body mass. Thus, we modeled in environmental contaminant levels, the Dou- daily energy expenditure (DEE) as a function of ble-crested Cormorant nesting population in bird body mass. We assumed an assimilation western Lake Erie has increased dramatically efficiency of 0.80 for all speciesconsidered in the from the mid- 1980s to the presenttime (Weseloh analysis (Cairns et al. 1991). DEE divided by et al., in press).As has already happened on Lake assimilation efficiency equalled the total daily Ontario (Weseloh and Casselman 1992) anglers energy consumption by an adult. Total daily en- and commercial fish harvesterson western Lake ergy consumption divided by the average energy Erie may become concerned about the effectsof density of the diet yielded the daily food mass an increasing cormorant population on the fish- per adult. The product of daily food mass per ery. The results of our study can be used to help adult bird and the proportion of fish in the diet guide decisionsmade by fishery managers.From representedthe daily fish consumption per adult BIRD PREDATION ON FISH IN LAKE ERIE 143 bird. Refer to Appendix 1 for more details on in any breedingactivities during a particular year. adult bird bioenergetics. We assumedthat the ratio of non-breeders to the Concerning the bioenergetics of egg produc- number of nests was 0.609 (Cairns et al. 199 1). tion and the food intake needs of chicks, we pro- Clutch size, hatching success,fledging success, ceededas outlined in Kendeigh et al. (1977). Re- hatch-year mortality, non-breeder mortality, fer to Appendix 1 for more information on the breeder mortality, and diet composition were de- bioenergetics modeling of hatching-year birds. rived from values reported in the literature and Each nesting population was groupedinto three from personal observations from researchers life stages: hatching-year, non-breeder, and conducting studies in the western Lake Erie vi- breeder. Within each life stage, the number of cinity. Refer to Table 1 and Figure 1 for more individuals was assumed to decline exponen- details on the life history characteristicsof the tially with time (Perrins and Birkhead 1983). See nesting bird populations employed in the com- Appendix 1 for referenceson estimating instan- puter simulations. Becausethe bulk of the non- taneous mortality rates. breeding population was believed to be com- prised of immature birds, we assumed that the DATA FOR NESTING BIRD POPULATIONS non-breeding bird mortality was equal to the Number of nests were obtained from surveys subadult mortality reported in the literature. En- conducted by the Canadian Wildlife Service ergy content of the diet items was taken from the (Blokpoel and Tessier 199 1) for gull nests on the literature (Cummins and Wuycheck 197 1, Stein Canadian side of Lake Erie and by the U.S. Fish and Murphy 1976, Stewart et al. 1983, Hewett and Wildlife Service (Scharf et al. 1994, Scharf and Johnson 1992). and Trapp 1994) for nestsofall waterbird species The undocumented values (i.e., those without on the U.S. side of Lake Erie during 1989 through footnotes) in Table 1 represent values of life his- 199 1. We assumed that the number of breeding tory characteristicsthat were unavailable from birds was equal to double the number of nests. the literature for the particular speciesof interest, This assumption was supported by gull colony and therefore the value from a related species studies, which showed that polygyny rates were was used. For example, hatch-year mortality re- extremely low (
PRESENCEOF BREEDING ADULTS
_____, PRESENCE OF NON-BREEDERS
1 EGG-LAYING
HERRING I I GULL ......
I RING-BILLED I GULL ...... ~
DOUBLE-CRESTED m I CORMORANT ...... ~
GREAT BLUE I I HERON ...... ~
GREAT I n EGRET ......
I BLACK-CROWNED I NIGHT-HERON ......
J I I I I I I I I I I I I JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC MONTH
FIGURE 1. Typical arrival, egg-laying, and departure dates for gulls, cormorants, herons, and egrets, nesting on western Lake Erie. The nesting population of Herring Gulls was present during the entire calendar year. Data were taken from Vermeer (1970) No1 and Blokpoel(l983), Petejohn and Rice (199 l), We&oh and Casselman ’ (1992) and Gabrey (unpubl. data), or provided by B. Buckingham (Ohio Department of Natural Resources, -j- Vickery, Ohio, pers. comm.), M. Shieldcastle (pers. comm.), and D. Weseloh (pers. comm.). All female breeders of a given species were assumed to lay their eggs on one day. def’hiu), Ring-billed Gull, Herring Gull, and the multiplied by the estimated population size to Double-crested Cormorant in the migrant bird yield the number of bird-days. The product of analysis. bird-days and daily fish consumption per adult We employed a simple algorithm to estimate bird yielded the total fish consumption during fish predation by migratory bird populations the autumn migration. Daily fish consumption during autumn. The duration of the fall migra- per adult bird was determined as described in tion was assumed to be eight weeks (Peterjohn the Model Structure for Nesting Bird Popula- and Rice 199 1). The estimated sizesof migrating tions section. Merganser and Bonaparte’s Gull populations were 2 10,000 Red-breasted Mer- diets were assumedto be 100% fish. Adult weights gansers(Peterjohn and Rice 1991), 25,000 Com- of mergansersand Bonaparte’s Gulls were taken mon Mergansers (adjusted average from Christ- from Dunning (1993). mas bird counts 1983-1992), 20,000 Bonaparte’s Estimates of the sizes of the migratory bird Gulls (extrapolation from fall surveysfrom 1986 populations were taken from a variety of sources. to 1993 by R. Dolbeer, unpubl. data), 200,000 Red-breasted Merganser and Double-crested Ring-billed Gulls (based on fall surveys by R. Cormorant numbers were based on aerial ob- Dolbeer), 10,000 Herring Gulls (based on fall servations. The Common Merganser migratory surveys by R. Dolbeer), and 6,500 Double-crest- population size was estimated by adjusting the ed Cormorants (M. Shieldcastle, pers. comm.). Christmas bird counts. The Common Merganser For each migratory population, the 56 days was population size was expected to peak at Christ-
146 CHARLES P. MADENJIAN AND STEVEN W. GABREY mas time (Peterjohn and Rice 199 1); but because in total annual fish consumption resulting from of incomplete coverageof the entire western ba- the input perturbation was determined. sin by Christmas bird counts, the averageChrist- mas bird count between 1983 and 1993 was in- RESULTS creased by 50%. Fall surveys of gulls along the Gull nestscomprised the bulk of the 27,411 sur- southern shore of western Lake Erie have been veyed nests,with 10,274 Herring Gull nests(37% conducted by the personnel at the Ohio Field of the total nest number) and 6,676 Ring-billed Station of the Denver Wildlife Research Center Gull nests(24%) (Table 1). The Great Egret pop- from 1986 to the present time (R. Dolbeer, un- ulation was the smallest, with 1,446 nests (5%). publ. data). This survey covered roughly one- The largest adult bird masseswere for the Great sixth of the western basin shoreline, and there- Blue Heron (2.4 kg) and the Double-crested Cor- fore the average count over years 1986 through morant (1.9 kg), whereas the lowest adult bird 1993 was multiplied by 6 to arrive at the pop- mass of the nesting birds was for the Ring-billed ulation size estimate. For Herring Gulls and Ring- Gull (0.5 kg). Of the nesting birds, fish comprised billed Gulls, this estimate was adjusted by sub- the highest proportion in the Double-crested tracting the estimated size of the nesting popu- Cormorant diet (99% fish), while fish was least lation. important in the diet of the Ring-billed Gull (35% Spring migration of waterbirds acrosswestern fish) (Table 1). Lake Erie is not nearly as spectacularas the fall According to the modeling results, waterbirds migration. Spring flock sizeswere roughly an or- consumed a total of 13,368 tonnes of fish from der of magnitude lessthan those observed in the western Lake Erie each year (Table 2). The nest- fall (Peterjohn and Rice 199 l), and the duration ing bird populations, including breeders, non- of spring migration was typically only six weeks breeders, and hatching-year birds, consumed a rather than eight weeks (Trautman 1940). We total of 5,999 tonnes of fish, whereasthe migrant therefore multiplied the fish consumption during bird populations consumed an estimated 7,369 the fall migration by 1.075 to arrive at the total tonnes. Of all of the individual categories of annual fish consumption by each migrant bird waterbirds that were considered in the analyses, population during spring and fall combined. the migrant Red-breasted Mergansers were re- sponsible for the greatest amount of fish con- SENSITIVITY ANALYSIS sumption at 4,914 tonnes (Table 2) or about A sensitivity analysis was performed to identify 67% of the fish consumption by migrant birds. the most important model inputs affecting the Migrant Ring-billed Gulls accounted for an ad- model output of annual fish consumption by the ditional 16% of the fish consumption by migrant nesting waterbird community in western Lake birds. Erie. The individual parameter perturbation Of the nesting bird populations, the Herring method was used (Bartell et al. 1986). Fifteen Gull population was the leading consumer of fish model inputs were examined. A total of 30 sim- (Table 2). The Double-crested Cormorant and ulation trials was conducted for each of the six Great Blue Heron nesting populations were in- nesting waterbird species.In any simulation trial, termediate consumers of fish, while the nesting only one model input was changedfrom its nom- populations of the Ring-billed Gull, Black- inal value; all the other model inputs were re- crowned Night-Heron, and Great Egret were rel- tained at their nominal value. A simulation was atively minor fish predators. performed with a particular model input in- Examining the nesting bird populations by life creased 10% from its nominal value; and then history stage,the breeding birds (all speciescom- another simulation was performed with that same bined) were the major fish predators, consuming model input decreased 10% from its nominal a total of 3,518 tonnes (Table 2). The hatching- value. For each input perturbation, the annual year group accounted for a total of 1,509 tonnes, fish consumption by the composite nesting pop- and the non-breeders consumed a total of 972 ulation of waterbirds was calculated; the com- tonnes. posite nesting population of waterbirds included The sensitivity analysis revealed that number nesting populations of all six speciesconsidered of nests,daily energy expenditure of post-fledged in the modeling analysis.Then, the percentchange birds, proportion of fish in diet, and assimilation BIRD PREDATION ON FISH IN LAKE ERIE 147
TABLE 2. Estimated fish consumption (tonnes/year)by gulls, cormorants,herons, and mergansersresiding on or migrating through western Lake Erie.
Hatcl&$year Non-breeding Species birds Breeding birds Migrant birds Total
Herring Gull 441 429 1,515 130 2,521 Ring-billed Gull 100 101 356 1,178 1,735 Double-crested Cormorant 365 159 720 236 1,480 Great Blue Heron 489 182 591 1,268 Great Egret 35 40 130 205 Black-crowned Night-Heron 73 61 200 334 Red-breasted Merganser 4,914 4,914 Common Merganser 762 762 Bonaparte’s Gull 149 149 Total 1,509 972 3,518 1,369 13,368
efficiency were the most important model inputs fish consumption by the walleye population in affecting the estimate of total annual fish con- western Lake Erie would be 117,600 tonnes. sumption by the composite waterbird population Therefore, on an annual basis, piscivorous water- nesting on western Lake Erie (Table 3). Total birds would take 11.4% of the quantity of fish annual fish consumption behaved linearly in re- eaten by the walleye population in western Lake sponse to changes in the number of nests. Fish Erie. energy density was of intermediate importance To fully evaluate the impact of fish consump- in determining total fish consumption. Non- tion by birds on the western Lake Erie ecosystem, breeder to nest ratio, the various mortality rates, joules per egg, duration of incubation, time be- tween hatching and fledging, and mass at time TABLE 3. Sensitivity of the estimate of fish con- of hatching had relatively minor impact on the sumption, by the composite population of nesting pi- scivorous waterbirds in western Lake Erie, to pertur- estimate of total fish consumption (Table 3). bations of various inputs to the bioenergeticsmodel. Overall, the model output was robust to uncer- Each input was separatelyperturbed + 10% and then tainty in most of the model inputs. - 10% from its nominal value. Sensitivity was mea- sured as the percent changein total fish consumption DISCUSSION resultingfrom the perturbation to the model input (see Sensitivity Analysis section for more details on the We estimated that waterbirds annually con- procedures).A sensitivity value of magnitudeless than sumed 13,368 tons of fish from western Lake 0.05 was reported as zero. Erie, based on nest counts made from 1989 through 1993. Hartman and Margraf (1992) cal- Input perturbation error culated that the walleye population in western Model input +10% -10% Lake Erie consumed, on the average, 88,200 Number of nests +10.0 - 10.0 tonnes of fish between May and November each Non-breeder to nest ratio +I.6 -1.6 year during the late 1980s. Thus, the amount of Hatch rate +2.5 -2.5 -1.1 fish biomass removed by waterbirds is relatively Fledge rate +1.1 Hatch-year mortality -0.8 +0.8 small (15.2% ofwalleye ingestion) comparedwith Non-breeder mortality -0.3 +0.3 fish consumption by walleyes in western Lake Breeder mortality -0.6 +0.6 Erie during a single growing season. According Daily energy expenditure to a bioenergeticsmodel application to Lake Erie (for post-fledgedbirds) +9.5 -9.5 Proportion of fish in diet +8.1 -9.4 yellow perch (Kitchell et al. 1977) approxi- Caloric density (of fish) -6.1 +7.1 mately 75% of the annual food consumption by Calories per egg 0.0 0.0 yellow perch in Lake Erie occurred during the Assimilation efficiency -9.1 fll.1 growing season (May through November). If a Incubation -0.5 +0.5 similar pattern in annual consumption was as- Fledging -0.3 +0.3 Mass at hatch 0.0 0.0 sumed for walleye, then an estimate of annual 148 CHARLES P. MADENJIAN AND STEVEN W. GABREY the biomass of fish consumed by birds should be to nest ratio is difficult to estimate (D. Weseloh, compared with the standing stock of prey fish. pers. comm.), and there were no available esti- Unfortunately, there are no reliable estimates of mates of this ratio for the western Lake Erie wa- total prey fish biomass in western Lake Erie cur- terbirds. Fortunately, the estimate of total fish rently available. Bottom trawls used by several consumption was relatively insensitive to this natural resourceagencies to survey prey fish pop- model input. Furthermore, the total fish con- ulations do not sample the entire water column, sumption estimate was very robust to uncertain- and therefore may lead to underestimation of fish ties in the various mortality estimates. biomass. An effort to estimate prey fish biomass According to the modeling results,the Double- via acoustical apparatus combined with bottom crestedCormorant nesting population in western trawls has been initiated for western Lake Erie Lake Erie annually consumed 1,244 tonnes of (D. Stewart, pers. comm.). fish, which was 9.3% of the total fish predation For nesting waterbirds, the model output (i.e., by birds in the western basin. The cormorant fish consumption) was most sensitive to the fol- population has expanded rapidly sincethe 1970s lowing model inputs: number of nests, daily en- when lessthan 60 cormorant nestswere recorded ergy expenditure (DEE) of post-fledged birds, throughout all of Lake Erie (Weseloh et al., in proportion of fish in the diet, and assimilation press). In 1990 a total of 1,954 nests was ob- efficiency. Number of nests was taken from ex- served, and a total of 3,600 nests was estimated tensive surveys of the western Lake Erie basin, for 1993. Presently, there may be some concern and relatively little error was expected in these that if the cormorant population were to con- nest surveys (H. Blokpoel, pers. comm.; W. tinue to increase, the prey fish base for walleye Scharf, pers. comm.; D. Weseloh, pers. comm.). in western Lake Erie would be in jeopardy. Yet, DEE was estimated from an allometric relation- if the nesting cormorant population doubled in ship developed by Bitt-Friesen et al. (1989); this size while populations of the other waterbirds relationship was based on field measurements of remained constant, total fish consumption by pi- bird metabolism using the doubly labeled water scivorous waterbirds from western Lake Erie technique, which allows for accurate estimation would only show a modest increase from 13,368 of population energy requirements (Bitt-Friesen to 14,612 or just 9.3%. et al. 1989). Data used to build the allometric Blokpoel and Tessier (199 1) reported a total relationship were from observations on cold wa- of 34,02 1 Ring-billed Gull nests on Fighting Is- ter seabirds, such as the Northern Gannet (Zulu land in the Detroit River during 1990. Using the bassanus), but the relationship should be appli- bioenergeticsmodel, we estimated that the Fight- cable to other waterbirds with flapping flight from ing Island nesting population would annually cool water regions (Birt-Friesen et al. 1989). consume 2,838 tonnes of fish. However, this Doubly labeled water measurements of the field nesting population was not included in our anal- metabolic rate of waterbirds in western Lake Erie ysis for the following reasons.The Detroit River, would be a valuable researchendeavor, because rather than Lake Erie or Lake St. Clair, should such an exercise would help verify or refine the be the source of the vast majority of the fish above-mentioned allometric relationship. The consumed by thesebirds during the breeding and proportion of fish in diet used in our modeling fledging periods, because the Ring-billed Gull exerciserepresented an average (temporally and typically foragesfor fish within 24 km of its nest spatially for a particular species)diet proportion at this time (W. Southern, pers. comm.). Fighting based on North American studies.There was no Island is approximately 24 km from the mouth reason to believe that the proportion of fish in of the Detroit River in western Lake Erie, and the diets of western Lake Erie waterbirds would about 30 km from Lake St. Clair. After the fledg- substantially deviate from these average num- ing period, adult and juvenile Ring-billed Gulls bers. Nevertheless, more detailed diet informa- may dispersefrom the nesting site (W. Southern, tion specifically for western Lake Erie waterbirds pers. comm.). The autumn feeding of the Fight- would improve the accuracy of the model. The ing Island gulls on fish in the wcstem basin would assimilation efficiency value of 0.8 used in our be taken into account by our migrant bird anal- modeling study was considered an accurate av- ysis. We have estimated that a large migrant pop- erageestimate of assimilation efficiency for most ulation (200,000 birds) of Ring-billed Gulls use waterbirds (Cairns et al. 199 1). The non-breeder western Lake Erie as a staging area for their fall BIRD PREDATION ON FISH IN LAKE ERIE 149 migration, and we have included these birds in 198 1, Burger 1986). Thus, reliable data on the arriving at our total fish consumption estimate. seasonal changes in bird metabolism and diet, Overall, it appearsthat our estimate of total con- as well as determination of activity budgets and sumption of fish in western Lake Erie by water- subsequent incorporation of behavior- and sex- birds needs little adjustment to account for the specific energy expenditures, could increase the Fighting Island gull population. accuracy of the model. From an ecologicalperspective, our modeling exercise quantifies a link (i.e., the consumption SUMMARY of fish by piscivorous waterbirds) in the western This bioenergetics model application not only Lake Erie food web. The zebra mussel (Dreissena estimated the magnitude of overall predation on polymorpha), a filter feeder on phytoplankton, fish by waterbirds in western Lake Erie, but also invaded the western basin during the late 1980s. identified the major consumersof fish within the Water clarity has increased (Leach 1993), how- western Lake Erie waterbird community. The ever researchershave not yet fully evaluated the model application has indicated that, indeed, impact of the mussels on primary production, walleye is the dominant piscivore in the western on the walleye population, and on recycling of Lake Erie ecosystem. Furthermore, the model microcontaminants. To understand the flow of allowed for prediction of total fish consumption energy and microcontaminants within this in- by waterbirds with varying cormorant popula- teresting ecosystem,a Lake Erie ecosystemmod- tions. The total fish consumption estimate by eling project has been proposed by the Intema- nesting waterbirds was strongly dependent on the tional Joint Commission (V. Cairns, pers. number of nests, and there was a high degree of comm.). Our bioenergeticsmodeling work would confidence in nest number estimates. The model contribute to proposed effort by supplying in- application to the western Lake Erie ecosystem formation on the trophic link between piscivo- revealed gapsin knowledge, particularly with re- rous waterbirds and prey fish. gard to seasonalchanges in bird metabolism and The bioenergeticsapproach that we have used diet and to further refinement and verification to determine fish consumption employed sim- of the allometric relationship for estimation of plifying assumptions concerning bird metabo- daily energy expenditure. The model may be im- lism and diet. For example, seasonalchanges in proved as more data become available. the daily energy expenditure and diet were ig- nored. However, diets of some birds (especially ACKNOWLEDGMENTS gulls) may change over time, particularly at the W. Scharf,J. Trapp, M. Shieldcastle,J. Ludwig,and end of the breeding period (Vermeer 1970, D. Bestprovided us with nest countsfrom the U.S. side of Lake Erie, and H. Blokpoeland D. Weseloh Haymes and Blokpoel 1978, Fox et al. 1990, furnishedus with the complementarydata from the Belant et al. 1993). Additionally, digestibility may Canadianside of the lake.We are gratefulto B. Buck- vary from one diet item to another (Brugger1993). ingham for sharinghis personalobservations on the For lack of detailed information on diet through- life historycharacteristics of the breedingpopulations out the year in western Lake Erie, we used “av- of waterbirdsin westernLake Erie. K. Bruggeradvised us on modelingwaterbird bioenergetics. We thank H. erage” proportions of fish and other diet items Blokpoel,R. Dolbeer,M. Hansen,K. Muth, D. We- acrossall seasons,based on literature values, in seloh,and an anonymousreviewer for reviewingthis our simulations. Furthermore, energy expendi- manuscript.T. Girard supplieda criticalreview of the ture may be affected by a bird’s behavior, prox- nroiectthroughout its development.Thanks to M. Bur, imity to a food source, or foraging strategy. The B. ickes, and R. Dolbeer for suggestingthis research nroiect. This article is contribution 884 of the National metabolic rate of an individual bird varies de- Biological Survey, Great Lakes Science Center, 1451 pending on its activity (Birt-Friesen et al. 1989), Green Road, Ann Arbor, Michigan. with flying the most energetically-demanding ac- tivity. Hunt (1972) found that Herring Gulls LITERATURE CITED nesting nearer to a food source (a landfill) spent BARTELL,S. M.,J.E.Bana,R.H. GARDNER,ANDA. more time on territory, and therefore less time L. BRENKERT.1986. Individual parameter per- flying, than those nesting further from the food turbation and error analysis of fish bioenergetics models. Can. J. Fish. Aquat. Sci. 43:160-168. source. Also, males and females within a popu- BELANT,J. L., T. W. SEAMANS,S. W. GABREY,AND S. lation may exhibit different foraging strategies K. Ictcns. 1993. Importance oflandfills to nesting (Fox et al. 1990) and different behaviors (Pierotti Herring Gulls. Condor 95:8 17-830. 150 CHARLES P. MADENJIAN AND STEVEN W. GABREY
BENT, A. C. 1921. Life histories of North American HARRISON,H. H. 1975. A field guide to birds’ nests. gulls and terns. U.S. Nat. Mus. Bull. 113. Houghton Mifflin, Boston. BENT, A. C. 1922. Life histories of North American HARTMAN,K. J., AND F. J. MARGRAF. 1992. Effects petrelsand pelicansand their allies. U.S. Nat. Mus. of prey and predator abundances on prey con- Bull. 121. sumption and growth of walleyesin western Lake BENT, A. C. 1926. Life histories of North American Erie. Trans. Am. Fish. Sot. 121:245-260. marsh birds. U.S. Nat. Mus. Bull. 135. HAYMES,G. T., AND H. BLOKPOEL.1978. Food of BIRT-FRIESEN.V. L.. W. A. MONTEVECCHI.D. K. Ring-billed Gull chicks at the easternheadland of CAIRNS,AND S. A. MACKEX 1989. Activity-spe- the Toronto Outer Harbor in 1977. Can. Field cific metabolic rates of free-living Northern-G&- Nat. 92:392-395. nets and other seabirds.Ecology 70:357-367. HENNY, C. J. 1972. An analysis of the population BU~KPOEL,H., AND G. B. M&EATING. 1978. Fish- dynamics of selected avian species,with special eating birds in Canadian Lake Erie and adjacent reference to changesduring the modem pesticide waters. Can. Wildl. Serv. Pron. Notes No. 87. era. U.S. Fish Wild. Serv. Wildl. Res. Reo. No. 1. BU)KPOEL,H., AND G. D. TESSIER~1986. The Ring- HEWETT, S. W., AND B. L. JOHNSON. 1992._ An up- billed Gull in Ontario: a review of a new problem grade of a generalized bioenergeticsmodel of fish species.Can. Wildl. Serv. Occas. Pap. No. 57. growth for microcomputers. Univ. of Wisconsin, BLOKPOEL,H., ANDG. D. TESSIER.199 1. Distribution Sea Grant Institute, Tech. Rep. WIS-SG-92-250, and abundance of colonial waterbirds nesting in Madison, WI. the Canadian portions of the lower Great Lakes HOFFMAN,R. D. 1977. The diets of heronsand egrets system in 1990. Can. Wildl. Serv. Tech. Rep. Ser. in southwesternLake Erie, p. 365-369. In A. Sprunt No. 117. IV, J. C. Ogden, and S. Winckler [eds.], Wading BLOMME,C. 1981. Status and breeding successof birds. National Audobon Society, New York. Double-crestedCormorant in two North Channel HUNT, G. L., JR. 1972. Influence of food distribution (Lake Huron) colonies in 1979. Ont. Field Biol. and human disturbanceon the reproductive suc- 35:70-78. cessof Herring Gulls. Ecology 53:1051-1061. BRUGGER,K. E. 1993. Digestibility of three fish spe- JEHL,J. R., JR., J. FRANC~NE,AND S. I. BOND. 1990. cies by Double-crested Cormorants. Condor 95: Growth patterns of two races of California Gull 25-32. raisedin a common environment. Condor 92:732- BURGER,J. 1986. Selection for equitability in some 738. aspectsof reproductiveinvestment in Herring Gulls KADLEC, J. A., W. H. DRURY, JR., AND D. K. ONION. Larus argentatus. Omis. Stand. 18:17-23. 1969. Growth and mortality of Herring Gull BUTLER,R. W. 1992. Great Blue Heron. Birds of chicks. Bird-banding 40~222-233. North America. No. 25, The Academy of Natural KENDEIGH,S. C., V. R. DOL’NIK, ANDV. M. GAVRILOV. Sciences,Philadelphia. 1977. Avian energetics, p. 127-205. In J. Pi- CAIRNS.D. K., G. CHAPDEWNE,AND W. A. MONTE- nowski and S. C. Kendeiah Ieds.1. Granivorous VE&HI. 1991. Prey exploitation by seabirds in birds in ecosystems.Camb>dge Univ. Press,Lon- the Gulf of St. Lawrence. D. 277-291. In J.-C. don. Theniault [ed.], The Gulfof St. Lawrence: small K~XHELL, J. F., D. J. STEWART,AND D. WEININGER. oceanor big estuary?Can. Spec.Publ. Fish. Aquat. 1977. Applications of a bioenergeticsmodel to Sci. 113. yellow perch (Percaflavescens) and walleye (Stizo- CAREY,C., H. RAHN, AND P. PARISI. 1980. Calories, stedion vitreum vitreum). J. Fish. Res. Board Can. water, lipid and yolk in avian eggs. Condor 82: 34:1922-1935. 335-343. KNIGHT, R. L., F. J. MARGRAF,AND R. F. CARLINE. CUMMINS,K. W., ANDJ. C. WUYCHECK.197 1. Caloric 1984. Piscivory by walleyes and yellow perch in equivalents for investigations in ecological ener- westernLake Erie. Trans. Am. Fish. Sot. 113:677- getics.Mitt. Int. Theor. Angew. Limnol. 18:l-l 5 1. 693. DUNNING,J. B., JR. 1993. CRC handbook of avian KNIGHT, R. L., AND B. VONDRACEK.1993. Changes body masses.CRC Press, London. in prey fish populationsin westernLake Erie, 1969- FE~EROLF,P. M. 1983. Effectsof investigatoractiv- 1988, as related to walleye, Stizostedion vitreum, ity on Ring-billed Gull behavior and reproductive predation. Can. J. Fish. Aquat. Sci. 50: 1289-1298. performance. Wilson Bull. 95:2341. LAGRENADE,M.-C., AND P. MOUSSEAU. 1983. Fe- FORNEY,J. L. 1966. Factorsaffecting first-year growth male-female pairs and polygynousassociations in of walleyesin Oneida Lake, New York. N.Y. Fish a Quebec Ring-billed Gull colony. Auk 100:21 O- Game J. 13:146-167. 212. Fox, G. A., L. J. ALLAN, D. V. WESELOH,AND P. LEACH, J. H. 1993. Impacts of the zebra mussel MINEAU. 1990. The diet of Herring Gulls during (Dreissenu polymorphu) on water quality and fish the nestingperiod in Canadian waters of the Great spawningreefs in western Lake Erie, p. 38 l-397. Lakes. Can. J. Zool. 68:1075-1085. In T. F. Nalepa and D. Schloesser[eds.], Zebra FURNESS,R. W. 1978. Energyrequirements of seabird mussels:biology, -_ impacts, and control. Lewis Pub- communities:a bioenetgeticsmodel. J. Anim. Ecol. lishers, Chelsea, MI: 47139-53. LEWIS.H. F. 1929. The natural historv ofthe Double- GLADSTONE,D. E. 1979. Breeding successof Great crestedCormorant. Ph.D.diss., Co’mell Univ., Ith- Egretsin New Jerseyand Mexico. Am. Midl. Nat. aca, NY. 102:368-373. MADENJIAN,C. P., AND S. R. CARPENTER.199 1. In- BIRD PREDATION ON FISH IN LAKE ERIE 151
dividual-based model for growth of young-of-the- U.S. Fish Wildl. Serv. Rep. on Contract 14-16- year walleye: a piece of the recruitment puzzle. 0009-89-006. EcologicalApplications 1:268-279. SHUGART,G. W. 1980. Frequency and distribution MEATHREL,C. E., ANDJ. P. RYDER. 1987. Sex ratios of polygynyin Great Lakes Herring Gulls in 1978. of Ring-billed Gulls in relation to egg size, egg Condor 82:426-429. sequence, and female body condition. Colonial SOUTHERN,W. A. 1974. The annual range of Ring- Waterbirds l&72-77. billed Gulls in the eastern United States: with NOL, E., AND H. BLOKPOEL.1983. Incubation period comments on potential bird/aircraft collision of Ring-billed Gulls and the egg immersion tech- problems, p. 149-190. In S. A. Gauthreaux, Jr. nique. Wilson Bull. 95:283-286. [ed.], Proceedings:a conference on the biological OWEN,D. F. 1959. Mortality of the Great Blue Heron aspects of the bird/aircraft collision problem. as shown by banding recoveries.Auk 76:464-470. Clemson Univ., Clemson, SC. PALMER,R. S. 1962. Handbook of North American SPAANS,A. L. 1971. On the feeding ecology of the birds, Vol. 1. Yale Univ. Press, New Haven, CT. Herring Gull Larus argentatus Pont. in the north- PAYNTER,R., JR. 1947. The fate of banded Kent Is- em part of The Netherlands. Ardea 59:73-188. land Herring Gulls. Bird-banding 18:156-170. STEIN, R. A., AND M. L. MURPHY. 1976. Changesin PERRINS,C. M., AND T. R. BIRKHEAD. 1983. Avian proximate composition of the crayfish Orconectes ecology. Chapman and Hall, New York. propinquus with size, sex, and life stage. J. Fish. PETERJOHN,B. G., AND D. L. RICE. 1991. The Ohio Res. Board Can. 33:2450-2458. breeding bird atlas. Ohio Dep. Nat. Resour., Co- STEWART, D. J., D. WEININGER,D. V. ROTTIERS,AND lumbus. T. A. EDSALL. 1983. An energeticsmodel for lake P~EROTTI,R. 1981. Male and female parental roles trout, Salvelinus namaycush: application to the in the Western Gull under different environmental Lake Michigan population. Can. J. Fish. Aquat. conditions. Auk 981532-549. Sci. 40:681-698. PIEROT~I,R. 1982. Habitat selectionand its effect on TRAUTMAN,M. B. 1940. The birds of BuckeyeLake, reproductive output in the Herring Gull in New- Ohio. Museum of Zoology, Univ. of Michigan, foundland. Ecology 63:854-868. Misc. Publ. No. 44. Pa~rr, H. M. 1972. Nesting successof Common TREMBLAY,J., AND L. N. ELLISON. 1980. Breeding Egretsand Great Blue Herons in the San Francisco successof the Black-crownedNight Heron in the Bay region. Condor 741447453. St. LawrenceEstuary. Can. J. Zool. 58: 1259-l 263. PRIEGEL,G. R. 1970. Reproduction and early life VERMEER,K. 1970. Breeding biology of California history of the walleye in the Lake Winnebago re- and Ring-billed Gulls: a study of ecological ad- gion. Tech. Bull. 45. Wis. Dep. Nat. Resour. Mad- aptation to the inland habitat. Can. Wildl. Serv. ison, WI. Rep. Ser. No. 12. QU~NNEY,T. E. 1982. Growth, diet, and mortality of WESELOH,D. V., ANDJ. CASSELMAN.1992. Calculated nestling Great Blue Herons. Wilson Bull. 94:571- fish consumption by Double-crested Cormorants 577. in eastern Lake Ontario. Abstract, Colonial Wa- RICKER,W. E. 1975. Computation and interpretation terbird Sot. Annual Meeting, Univ. Mississippi, of biological statistics of fish populations. Fish. Oxford, Oct. 14-18. Res. Board Can. Bull. 191. WESELOH,D.V.C., P. J. EWINS, J. STRUGER,P. MINEAU, SCHAW,W. C., AND J. L. T~APP. 1994. Distribution C. A. BISHOP,S. POSTUPALSKY,AND J. P. LUDWIG. and abundance of heron, egret, Black Tern and In press. Double-crested Cormorants (Phalacro- Forster’s Tern colonies of the U.S. Great Lakes. corax auritus) of the Great Lakes:changes in pop- U.S. Fish Wild]. Serv. Rep. on Contract 14-16- ulation size, breeding distribution, and reproduc- 0009-9 l-933. tive output. Colonial Waterbirds 17. SCHARF,W. C., G. W. SHUGART,AND M. L. CHAM- WIENS, J. A. 1984. Modelling the energy require- BERLIN. 1978. Colonial birds nesting on man- ments of seabird populations, p. 255-284. In G. made and natural sites in the U.S. Great Lakes. C. Whittow and H. Rahn [eds.],Seabird energetics. U.S. Fish Wildl. Serv. Tech. Rep. D-78-10. Plenum Press, New York. SCHARF,W. C., G. W. SHUGART, AND J. C. T~APP. Wrmrs, J. A., AND J. M. Scorr. 1975. Model esti- 1994. A catalogof gull, tern, and cormorant nest- mation of energy flow in Oregon coastal seabird ing colonies of the U.S. Great Lakes, 1989-1990. populations. Condor 77~439-452. 152 CHARLES P. MADENJIAN AND STEVEN W. GABREY
APPENDIX 1. Bioenergeticsmodel used to calculatethe annual consumptionof fish by the nestingpopulation of piscivorous waterbirds in western Lake Erie circa 1990. Resident speciesincluded the Herring Gull, Ring- billed Gull, Double-crested Cormorant, Great Blue Heron, Black-crownedNight-Heron, and Great Egret. Each specieswas modeled separately,and then fish consumptionfor all six populationswas summed to arrive at total fish consumption by the nesting bird populations.
The daily energyexpenditure (DEE) of a post-fledgedbird was calculatedusing the allometric equationpresented by Birt-Friesen et al. (1989): DEE = 1737.8W”.‘*’ where DEE = the daily energy expenditure, in kilojoules, of a post-fledgedbird, and W = mass, in kilograms, of the post-fledgedbird. This allometric equation was based on measurementsof the metabolic rates of free- living seabirds,and therefore this equation was particularly appropriate for estimating fish consumptionby birds in natural ecosystems.Cairns et al. (199 1) used this allometric equation to estimate the DEE of post-fledged Herring Gulls, Ring-billed Gulls, and Double-crested Cormorants in the Gulf of St. Lawrence. DEE was converted to kilocalorie units using the following equation: 1 kilojoule = 0.23892 kilocalories. Daily energy intake (DEI) was calculated by dividing DEE by the assimilation efficiency. We assumed an assimilation efficiency of 0.80 for all birds (Fumess 1978). Daily food intake (DFI) was calculatedby dividing DE1 by the average caloric density of the bird diet. The average caloric density (ACD) of the bird diet was calculatedusing the following equation:
ACD = 2 (CD,).(PRGP,) x-1 where ACD = average caloric density, in kilocalories per kilogram, of the bird diet, CD, = caloric density, in kilocalories per kilogram, of diet category i, PROP, = the proportion of the bird diet, by mass, comprised of diet categoryi, and n = the total number of diet categories.Thus, the daily fish consumption (DFC) was: DFC = DFI’PROP,,, where DFC = daily fish consumption, in kilograms, DFI = daily food intake, in kilograms, and PROP,, = the proportion of the bird diet comprised of fish. The daily energy expenditure of a pre-fledged bird (DEEN) was estimated using the allometric equation presentedin Kendeigh et al. (1977): DEEN = 1.230W”.7749 where DEEN = daily energy expenditure, in kilocalories, of a pre-fledged bird, and W = pre-fledgedbird mass, in grams. Nestlings may show spurts in growth (Kendeigh et al. 1977). Gull chick growth has been modeled using a Gompertz equation (Jehl et al. 1990) or has been described as linear (Spaans 1971). In most cases, nestling growth can be closely approximated by a linear fit (Spaans 197l), and we assumedgrowth to be linear from time of hatching to the completion of fledging. Birds were assumedto reach adult mass at fledging. This was a reasonableassumption becausemass at end of the fledging period is typically within 5 to 10% of adult mass (Kadlec et al. 1969, Jehl et al. 1990). The daily food intake by a pre-fledgedbird was equal to the sum of the food needed to match the DEEN and the food needed to increasepre-fledged bird massby the daily growth increment, DG. Thus, the daily fish consumption by a pre-fledged bird was: DEEN.PROP,,, + DG.PROP,,, DFCN = O.IO.ACD 0.80 where DFCN = daily fish consumption, in kilograms, by a pre-fledgedbird, DEEN = daily energyexpenditure, in kilocalories, by a pre-fledged bird, DG = daily growth increment, in kilograms, by a pre-fledged bird, and PROP,,, and ACD are defined as above. The assimilation efficiency for pre-fledged birds was assigneda value of 0.80 (Cairns et al. 1991). The time step for all of the simulations was one day. The initial day of the simulated time period was the arrival date, and the departuredate marked the final day of the simulation run (seeFig. 1 for arrival and departure dates). At the end of each day, fish consumption for an individual in each of the three life stagecategories (i.e., hatch-year, non-breeder, and breeder) was calculated. Individual fish consumption for a particular life stage categorywas then multiplied by the population size for that life stagecategory to yield the daily fish consumption by that life stagepopulation. Summing daily fish consumptionover all of the simulated days yielded the annual fish consumption for a particular life stage.Within each life stage,simulated bird populationswere assumedto decline exponentially, as described in Perrins and Birkhead (1983). Furthermore, number of eggsand number of pre-fledged birds were assumedto decline exponentially. See Ricker (1975) or Hewett and Johnson (1992) for details on converting a survival rate to an instantaneousmortality rate. BIRD PREDATION ON FISH IN LAKE ERIE 153
APPENDIX 1. Continued.
On the designated egg-laying day (as illustrated in Fig. 1) in the simulation, the energy represented in the eggs was included in the total daily energy expenditure of the breeder population. The energy represented in the eggs was then converted to fish consumption, using the above-detailed algorithm. This computer simulation program for fish consumption by a nesting bird population was written in TURBO PASCAL version 6.0. A copy is available upon request.