The Condor95~817-830 0 The CooperOrnithological Society 1993
IMPORTANCE OF LANDFILLS TO NESTING HERRING GULLS
JERROLDL. BELANT~,THOMASW. SEAMANS,STEVEN W. GABREY,AND SHERI K. ICKES U.S. Department of Agriculture,Denver Wildlife ResearchCenter, 6100 ColumbusAvenue, Sandusky, OH 44870
Abstract. We determined the importance of three landfills to a population of nesting Herring Gulls (Larus urgent&us, 3,250 pairs) on Lake Erie, Ohio, from May-July 1992. Fish was the dominant food ofadults and chicksthroughout the study.Occurrence ofgarbage in the diet of adults and chicks remained low through chick-rearing then increased after fledging.Presence of telemeteredadults at their nest sitesdecreased from incubation through post-fledging,in contrast to their increased presenceat landfills during the same periods. Overall, females visited landfills more frequently and stayed longer than males; however, use of landfills by both sexeswas minimal (~4% of total time) during all periods. Overall, gulls spent 43% and 4% of their time daily at the nest site and landfills, respectively. We estimate ?80% of the time remaining was spent on Lake Erie, presumably to forage. The estimated daily mean number of adult Herring Gulls at the landfills increased from incu- bation (143) to chick-rearing (723) to post-fledging (1,912). We estimate that 5-7%, 12- 19%, and 35-55% of the adult nestingpopulation was presentat landfills at least once during incubation, chick-rearing, and post-fledging,respectively. The population turnover rate of adult Herring Gulls at one landfill decreased50% from incubation and chick-rearingto post- fledging. Significantlymore gulls at the landfills were observed on areasother than exposed refuseand were not actively foraging,suggesting that landfills are important to Herring Gulls for other reasons such as loafing or social interaction. We conclude that landfills are un- important to nestingHerring Gulls when alternate,higher quality food (e.g., fish) is available. The increaseduse of landfills by Herring Gulls during post-fledging,however, suggeststhat gull activity at landfills located near airports could dramatically affect aircraft safety during this time of year. Key words: Herring Gull:food habits;landfill; Larus argentatus;Ohio
INTRODUCTION a concurrent increase in conflicts with humans. Populations of several species of gulls have in- These conflicts include transmission of patho- creased throughout North America and Europe gens and parasites through contamination of wa- in recent years (Harris 1970, Spaans 197 1, Drury ter sources (Mudge and Ferns 1982) damage to and Kadlec 1974, Conover 1983, Blokpoel and buildings (Bradley 1980, Vermeer et al. 1988, Tessier 1986). Suspected causes for these in- Belant 1993), and a hazard to aircraft at airports creased populations include the protection of (Blokpoel 1976, 1983; Dahl 1984; Sherigalin breeding colonies (Kadlec and Dntry 1968, 1990). Therefore, there is critical need for data Spaans 197 l), an increase in nesting habitat from on the influence of various types of landfills have the creation of dredge disposal islands (Patton on gull activity and the importance of landfills and Hanners 1984), and exploitation of landfills to gulls. by gulls as dependable sources of food (Verbeek Although there is general agreement that pu- 1977,Burger 1981,Patton 1988,BelantandDol- trescible waste at landfills contributes to the beer 1993). overall diet of Herring Gulls (Larus urgent&us; Sanitary landfills frequently attract large num- Kihlman and Larsson 1974, Horton et al. 1983, bers of gulls and other birds, at least seasonally Pierotti and Annett 1987, Patton 1988, Pons (Horton et al. 1983; Patton 1988; Belant, unpubl. 1992), there are conflicting conclusions from pre- data). Because landfills are often located near ur- vious studies regarding the importance of land- ban areas, their increased use by gulls has caused fills to gulls during the breeding season. Several authors have suggested that the availability of I Received 12 February 1993. Accepted 28 April garbage increases (Kadlec and Drury 1968, Hunt 1993. 1972, Pons 1992), or is essential for (Sibley and Z Presentaddress: Great LakesIndian Fish and Wild- McCleery 1983), reproductive success. In con- life Commission, 1908YzWest SuperiorStreet, Duluth, MN 55806. trast, Pierotti and Annett (1987) have suggested 18171 818 JERROLD L. BELANT ET AL.
LAKE ERIE
HERRING GULL NESTING COLONIES LANDFILL
ERIE CO. LANDFILL -
FIGURE 1. Locationof threemixed solid wastelandfills in relation to nestingcolonies of Herring Gulls in SanduskyBay, Lake Erie, Ohio. that garbage is a low-quality food compared to Three mixed solid waste landfills are located other “natural” foods, and could reduce repro- within 36 km of the nesting concentration (Fig. ductive success. 1). The Erie County Landfill, 7 km south of Lake The objectives of this study were to determine Erie and 19 km southeastof the nesting concen- food habits, flight patterns, and population dy- tration, averaged 275 metric tons of refuse per namics at landfills of adult Herring Gulls and day during the study. The Ottawa County Land- their chicks during incubation, chick-rearing, and fill, 2 km from Lake Erie and 28 km northwest post-fledging periods. Our goal was to determine of the nesting concentration, averaged 563 met- the importance of landfills to a population of ric tons of refuse per day, May-September 199 1. nesting Herring Gulls. The Huron County Landfill, 30 km from Lake Erie and 35 km south the concentration, aver- STUDY AREA aged 84 metric tons per day during the study. At The study was conducted in north-central Ohio each landfill, refuse was spread and compacted from 1 May-3 1 July 1992. The Herring Gull throughout the day, then covered with soil at the nesting concentration (one of the largest on the end of the workday using bulldozers. Each land- Great Lakes, 4,250 nesting pairs in 1989) is lo- fill generally had ~0.5 ha of exposed refuse on cated on Sandusky Bay, Lake Erie (Dolbeer et any given day. These were the only landfills with- al. 1990, Fig. 1). The gulls nest on Turning Point in 40 km of the nesting concentration. Island (TPI), coal piles, and breakwalls and have recently expandedto rooftopsin Sandusky,Ohio. METHODS TPI is a 2.7-ha dredge disposal island created in Population censusand reproduction, We deter- 1900 and is bordered by riprap (Scharf et al. mined the adult breeding population of Herring 1978). About 50% of the island has herbaceous Gulls by conducting two complete ground counts vegetation. Dominant shrub and tree speciesin- of nests containing 2 1 egg on TPI, coal piles, clude red mulberry (Torus rubru), red-osier dog- and breakwalls between 1 and 21 May. Nests wood (Corms stolonijkra), and eastern cotton- with 2 1 egg on rooftops were counted weekly wood (Populus deltoides,Scharf et al. 1978). form early May to early July. The largestnumber HERRING GULL USE OF LANDFILLS 819 of nests counted in each area was combined to abdominal feathersby applying 4-5 ml of a mix- obtain a total count of nests for the population. ture of rhodamine-B dye and silica gel to one or We individually marked 110 three-eggclutch- three eggsof each clutch or by using a dummy es on TPI using 0.6-m wire surveying flags and egg (Belant and Seamans, in press). We applied checkedthem one to two times weekly until eggs the dye mixture to 535 clutches between 1 and had hatched, were destroyed, or considered 13 May, thus marking 16% of the nesting pop- abandoned or inviable (eggs not hatched 26 ulation. Becausegulls were marked over a two- weeks after our initial visit). Nests on rooftops week period, for data collected during observa- were also individually marked using wood blocks tions conducted at the landfill (see below), we and monitored weekly using the same criteria. adjustedthe proportion ofthe population of nest- We defined hatch successas the number of eggs ing gulls color-marked after each dye application. hatched divided by the total number of eggslaid We captured nesting Herring Gulls on TPI us- for TPI and rooftops. Mean hatch date was es- ing walk-in traps (Weaver and Kadlec 1970). We timated by interpolation based on the date of the measured head and bill length, and bill depth of previous check, the number of eggs that had captured individuals to determine sex (Fox et al. hatched or were pipping, and the relative age of 198 1). Each of 20 gulls received a U.S. Fish and chicks(Kadlec et al. 1969). We defined the length Wildlife Service leg band and a radio transmitter ofincubation and chick-rearing periods as 28 and (Advanced Telemetry Systems,Inc., Isanti, MN) 42 days before and after mean hatch dates, re- with a backpack harness made of 6-mm-wide spectively (Kadlec et al. 1969, Drent 1970, Hay- teflon ribbon. The completed radio package cock and Threlfall 1975, Pierotti 1982, Paynter weighed 30 g, or about 3% of the mass of an 1949). Post-fledging data were collected through adult Herring Gull. We did not apply the dye 31 July. mixture to eggsof telemetered gulls. Food habits. We collected food remains and Telemetry. We estimated locations of 17 tele- pellets of undigested material found 5 1 m from metered gulls (nine females, eight males) that nests. Food remains and pellets were collected successfully hatched e 1 egg using standard on TPI, rooftops, and breakwalls. We recorded ground and aerial telemetry techniques (Mech date and location for each sample collected.Food 1983). Ground telemetry was conducted2 3 times items were collected once or twice each week each week and aerial telemetry was conducted from 4 May-14 July. once per week between 07:30-17:00 hr, weather We also collected boli from chicks on TPI one permitting. For ground telemetry, all receiver lo- to two times weekly. Chicks were captured op- cations were plotted using Universal Transverse portunistically by hand or with a net. If a chick Mercator Grid System (UTM) coordinates from did not regurgitate upon capture, we inserted a 7.5 min U.S. Geological Survey topographic finger into its proventriculus and removed the maps. For each gull location estimate, we ob- contents (Hunt 1972). We recorded date, body tained two azimuths within 15 min. The program mass,and ageclass (pre-fledging [age classes 1-3B] Locate II (Nams 199 1) was then used to estimate or fledging [age class 41, Kadlec et al. 1969) for gull locations. For aerial telemetry data, gull lo- each chick from which a bolus was obtained. Boli cations were plotted directly onto topographic were stored in 80% ethyl alcohol until analyzed. maps and UTM coordinates were determined. The contents of each sample were identified For each location we recorded date, time, and and initially classifiedinto broad categories(fish, when possible, habitat in which found. For each garbage, etc.). Except for pellets containing re- sex, we used the location estimates away from mains of fish, we identified boli and food remains the colony to calculate mean daily distance (km) to the lowest taxon possible using reference col- from the colony during each reproductive period. lections. Frequency of occurrence of each food A data collection computer (DCC II Model type was recorded. Because mean hatch dates D5401, Advanced Telemetry Systems, Inc., Is- were not determined for gulls nesting on break- anti, MN) was used at the Erie County landfill walls, these data were excluded from analyses and the nesting colony on TPI to record fre- related to reproductive periods. quency and duration of presence of 15 teleme- Capture and marking. During the censuses,a tered gulls (eight females, seven females). We sample of adult Herring Gulls nesting on TPI excluded data from one female and one male and breakwalls were marked on their breast and becausetheir nest sites were outside the range of 820 JERROLD L. BELANT ET AL.
DCC reception. Reception range of the DCC at for 10 min to obtain estimates of ingress and TPI was about 15 m, and did not extend to open egress(i.e., population turnover rates). To esti- water. Because Herring Gulls are territorial at mate the number of flying gulls which were adult their nest sites (Burger 1986) we assumed that Herring Gulls, we assumed that the proportion signalsreceived were of telemetered gulls at their of flying gulls observed during a lo-min period nest sites. Reception range of the DCC at Erie that were adult Herring Gulls was equal to the County Landfill was about 300 m, which includ- proportion of adult Herring Gulls that were pres- ed all available areas at the landfill where gulls ent at the landfill during the corresponding total had been previously observed (Belant, unpubl. population count. data). Other areas within the range of reception After completing observations of flying gulls, of the DCC were unsuitable for gulls (woodlots, observersselected 10 gulls from throughout one corn fields); therefore, we were confident that we of the three predetermined areas. Each gull was only received signals of telemetered gulls that observed for 5 set and the most prevalent be- were at the landfill. The DCC was moved be- havior was recorded. Behavior categories used tween sites every 3-7 days and was calibrated to were: (1) foraging, (2) maintenance, (3) loafing or receive the signal of each telemetered bird at 1O- alert, (4) aggressiveand (5) other. Upon com- min intervals throughout the day. pletion of observations in one area, the sequence Observations at landfills. Observations were was repeated for the two remaining areas. If the conducted at the Erie County Landfill by one or behavior of a selectedgull could not be observed two individuals twice per week (from 07:00-12:00 for 5 set, behavior of the gull nearestthe selected or 12:00-17:00 hr) on randomly selected days. bird that was completely visible to the observer We divided the landfill into three areas: (1) ex- was recorded. Each series of observations (i.e., posedrefuse, (2) partially covered refuse, and (3) 5 30 gulls, 5 10 for each area) was separatedby non-refuse areas.Areas containing partially cov- 3 min. This procedurewas conductedthree times. ered refuse correspond to “secondary feeding” Excluding behavior data, additional observa- areas described by Greig et al. (1985). At the tions were conducted at the Erie County Landfill beginning of each hour, observers used binocu- twice each day, five days per week. The twice lars to identify a group of known size (~200). daily observations were conducted at randomly The total population of gulls and the number of selectedtimes, one each during the morning and gulls in each of the three areas was estimated by afternoon. Two of these observations were part counting the number of groupsof the known size of the two 5-hr observations periods conducted and multiplying by the group size. We usually each week. Thus, we obtained 18 estimates of conducted total counts when gull populations population turnover data and 18 total counts of numbered ~500 individuals. Only gulls on or adult Herring Gulls at the Erie County Landfill within 30 m of the ground were included in each week. We similarly conducted two to three counts. The age and speciescomposition of gulls observations per week at the Huron and Ottawa (Grant 1986) at the landfill was determined by County landfills. counting throughouteach of the three areas 2 10% Populations turnover at IandJills. We estimated ofthe estimated population. The number ofadult the mean daily number of individual adult Her- Herring Gulls counted was extrapolated to the ring Gulls (G) that visited a landfill during in- entire population to estimate the total number cubation, chick-rearing, and post-fledging, using of adult Herring Gulls present. We also estimated the formula: the number of marked adult Herring Gulls pres- G = M x h/D, ent by counting the number of marked gulls ob- served from 2 10% of the adult Herring Gull where A4 is the mean number of adult Herring population, counting only thosegulls whoseven- Gulls present per observation at the landfill from tral surface was visible. Because of the limited 07:00-l 7:00 hr during a reproductive period, h longevity (4-5 weeks) of color-marks and mor- is the number of hours (10, a constant) within tality of eggs(Belant and Seamans,in press),data the range of time gulls were observed at the land- from color-marked gulls at the landfills were used fill (07:00-17:00), and D is the mean duration of only during incubation. visits recorded for telemetered Herring Gulls for Immediately after each count, we determined each reproductive period. For example, during the number of gulls flying to and from the landfill the incubation period, if A4 = 20 gulls (Table 6) HERRING GULL USE OF LANDFILLS 821 and D = 1.4 hr (Table 5), then G = 143 gulls and garbage (Table 1). Occurrence of fish was (Table 6). We also determined the number of higher (x2 = 7.02, 1 df, P < 0.01) for pre-fledging hours required for the mean number of gulls ob- chicks (84%) than for fledgedchicks (64%). Con- served at the landfill at any one time to be re- versely, occurrenceof garbagewas higher (Fish- placed by the same number of different gulls er’s exact, P = 0.05) for fledgedchicks (21%) than within each reproductive period by dividing the for pre-fledging chicks (8%). Plant material con- mean daily number of individual adult Herring sisted primarily of red mulberry fruit, which co- Gulls presentat the landfill (G) by the mean num- incided with the maturation of this fruit on TPI. ber of gulls recorded per observation during that We collected 317 food remains from all lo- same period (M) and then dividing this quotient cations. Fish was the most abundant item, fol- into h (10 hr). For example, during incubation, lowed by garbage (Table 1). The diet of gulls at if G = 143 and A4 = 20, then the population TPI (n = 117) and rooftops (n = 106) was similar turnover rate (hr) is: lo/( 143/20), or 1.4 hr (Ta- with one exception: occurrence of birds was ble 6). greater (x2 = 8.23, 1 df, P < 0.05) at nests on We estimated the total number of individual rooftops (14%) than at nests on TPI (3%). Be- adult Herring Gulls that used the landfills during causeoccurrence of fish (72% vs. 72%) and gar- a reproductive period (r) using the formula: bage (18% vs. 17%) was similar (x2 < 0.01, 1 df, P > 0.90) between TPI and rooftops, respec- T=Mx hl(Vx D), tively, data for these food items (n = 223) were where V is the mean number of visits per gull pooled to compare among reproductive periods. each day (from radio telemetry data, Table 5). Frequency of fish was lower (Fisher’s exact, P -C For example, during incubation, if M = 20, h = 0.0 1) during post-fledging(O%, n = 9) than during 10, V = 0.3, and D = 1.4, then T = 20 x lO/ incubation (65%, n = 3 1) or chick-rearing (77%, (0.3 x 1.4) or 476 gulls. n = 183). Frequency of fish was similar between Statistical analyses. Chi-squared tests of in- incubation and chick rearing (x2 = 2.02, 1 df, P dependence(or Fisher’s exacttwo-tailed testwhen = 0.16). Percent occurrenceof garbagewas sim- necessary)were used to detect dietary differences ilar between incubation (3%, n = 3 1) and chick- of adults among reproductive periods and be- rearing (17%, n = 183, Fisher’s exact, P = 0.06); tween unfledged and fledgedchicks. We used the however, occurrence of garbage was greater General Linear Models Procedure (SAS Insti- (Fisher’s exact, P -C 0.01) during post-fledging tute, Inc., 1988) and Tukey multiple comparison (67%, n = 9) than during incubation or chick- tests for all other analyses. All means are re- rearing. ported with -t one standard deviation. Occurrence of garbage in boli obtained from chicks increased temporally, although occur- RESULTS rence of garbagedid not exceed that of fish (Fig. Nesting population censusand reproduction.We 2). The occurrence of garbage in food remains counted 3,250 gull nests in all areas: 1,918 on also increased temporally, however, its occur- TPI, 1,026 on breakwalls, 176 on rooftops, 122 rence surpassedfish about 10 days before mean on coal piles, and eight in other areas. Hatch fledging.The increasedoccurrence of garbageand successof three-egg clutches on TPI (7 l%, n = food remains corresponded with increasing oc- 105 nests) was similar to hatch successof three- currence of adult Herring Gulls at Erie County eggclutches on rooftops (62%, n = 138 nests, x2 Landfill. = 1.14, 1 df, P > 0.10). Hatch successof eggsof Of 542 pellets collected, 98% contained fish. telemetered gulls was also similar (62%, n = 15 With the exception of vegetation (8%) and fish, nests, x2 = 0.24, 1 df, P > 0.50). Mean (*SD) no other food item was represented in > 1% of hatch date for nests on TPI and rooftops was 19 pellets (Table 1). May + 6 days and 30 May f 8 days, respectively; Nest-siteattentiveness. Daily presenceof males therefore, the respectivechick-rearing period for (X = 11.3 f 2.5 hr per day) and females (X = 9.3 TPI and rooftops was 20 May-30 June and 31 -t 3.3 hr per day) at the nest site from May-July May-l 1 July. was similar (F = 2.98; 1,39 df; P = 0.09). Equita- Food habits.We collected 160 boli from chicks bility of nest-site attendance remained constant on TPI. Fish was the most frequently recorded among reproductive periods (F = 0.14; 2, 39 df; item, followed by earthworms, plant material, P = 0.87). Daily individual presenceof teleme- 822 JERROLD L. BELANT ET AL.
TABLE 1. Occurrence(O/o) of food items in boli from Herring Gull chicks (n = 160), and in food remains (n = 3 17) and pellets of undigestedmaterial (n = 542) found on or 5 1 m from Herring Gull nests in Sandusky Bay, Lake Erie, Ohio, May-July 1992.
Food item Bali Food remains Pellets
All fish species 79 74 98 Alewife (Alosa pseudoharengus) 11 17 - Gizzard shad (Dorosoma cepedianum) 9 7 - Unidentified Clupeidae 11 3 - White bass (Morone chrysops) 4 11 - White perch (Morone americana) 13 26 - Unidentified Morone spp. 1 1 - Emerald shiner (Notropis atherinoides) 24 8 - Spottail shiner (Notropis hudsonius) 0 loo-a I 80 80 40 20 -\ /- A, / ’ / ’ _/ / ’ I \ \ _a _d 4‘ I Ow--r-rs_’ - A#- I '0 4MAY 11 18 25 1 JUN 8 15 22 29 40 - 20 - 4MAY 11 18 25 1 JUN 8 15 22 29 v FISH -+- GARBAGE--c HERRING GULLS FIGURE 2. Percent occurrenceof fish and garbagein the diet of Herring Gulls, SanduskyBay, Lake Erie, and mean number of adult Herring Gulls observedat the Erie County Landfill (ECL), Ohio, 1992: (a) regurgitations collected from chicks on Turning Point Island (TPI), (b) food remains collected from nests on TPI, breakwalls, and rooftops. < 0.01). Movements by adult female Herring cubation, chick-rearing, and post-fledging, re- Gulls increasedfrom incubation to chick-rearing spectively. and then remained similar through post-fledging. In contrast, movements by males increased USE OF LANDF1LLS through all 3 periods. Using aerial telemetry data No color-marked Herring Gulls were observed excluding locations at the nesting colony, gulls at the Huron or Ottawa County Landfills. Only were present on Lake Erie or Sandusky Bay dur- one telemetered gull was located at the Huron ing 83%, 82%, and 90% of locations during in- County landfill twice during the post-fledgingpe- 824 JERROLD L. BELANT ET AL. TABLE 2. Mean hours (k SD) spentby radio-telemeterednesting Herring Gulls (Larus urgent&us; eight females, seven males) at their nest site by reproductive period and time of day, Turning Point Island, Lake Erie, Ohio, May-July 1992. Reproductive 02:00- 06:00- lO:OO- 14:00- 18:OS 22:00- Total hr period OS:59 09:59 13:59 17:59 21:59 01:59 per day Incubation 3.5 3.0 2.8 2.6 3.1 3.1 18.9 * 2.0 (1-19 May) Chick-rearing 1.5 1.4 1.9 1.6 1.6 1.8 9.7 +- 5.1 (20 May-30 June) Post-fledging 1.0 1.0 0.9 0.8 0.9 1.1 5.7 + 3.0 (1-31 July) All periods 1.7 1.6 1.7 1.5 1.7 2.0 10.3 * 3.1 (1 May-3 1 July) riod. Therefore, we concluded that theselandfills visits per day; F = 3.83; 1, 39 df; P = 0.06); this were not important to nesting Herring Gulls from visitation rate was constant among reproductive the Sandusky Bay concentration. Further anal- periods (F = 0.32; 2, 39 df; P = 0.73). Mean ysesincludes data from Erie County Landfill only. duration of visits at the Erie County landfill by Gull behavior. Overall, 62% of gull activity at telemetered Herring Gulls during incubation (1.4 the landfill at a given time was in activities other & 1.7 hr) and brood-rearing (1.1 IfI 1.3 hr) was than foraging (Table 4). Loafing or alert postur- less(F = 4.15; 2,39 df; P = 0.02) than time spent ing, followed by maintenance activities were the during post-fledging (2.3 f 2.6 hr). Ninety-eight most prevalent behavior exhibited by gulls in percent of telemetered gull use of the Erie County areas other than exposed refuse. Foraging was landfill occurred between 06:OO and 17:00 hr. the most prevalent (70% overall) behavior ex- Approximately 50% of use was between 10:00 hibited by gulls on exposed refuse and was in- and 14:00 hr. No telemetered gulls were at the frequently observed in other areas. Frequency of landfill from 22:00-02:OO hr. aggressivebehavior declined from open refuse The mean number of adult Herring Gulls ob- to recently covered refuse to non-refuse areas.In served during each observation at the landfill areas of exposedrefuse, occurrenceof aggressive increased from incubation to chick-rearing to behavior appeared to be related to the number post-fledging (F = 162.3; 2, 842 df; P < 0.01; of birds present. Table 4). The number of gulls among exposed Gull abundance. On average, females spent refuse, covered refuse, and non-refuse areas at more (F= 8.83; 1, 39 de P < 0.01) time (1.3 f the landfill also differed (F = 65.52; 2, 842 dc P 0.9 hr) each day at the landfill than did males < 0.01). The mean number of gulls observed on (0.5 f 0.4 hr) which was constant among repro- the recently covered refuse (n = 27) was lessthan ductive periods (F = 0.78; 2, 39 df; P = 0.47; the number observed on exposedrefuse (n = 62) Table 5). Females also frequented the landfill (0.7 which was lower than the mean number ob- visits per day) almost twice as often as males (0.4 served in non-refuse areas (n = 130, Tukey test, TABLE 3. Mean distance (*SD [n]) radio-telemetered Herring Gulls (Lams argentutus; nine females, eight males) were located away from nesting colony by reproductive period, Lake Erie, Ohio, May-July 1992. Distance (km) from nesting colony Reproductive period Female Male Combined Incubation 4.6 + 2.7 (11) 6.5 + 5.8 (15) 5.7 ? 4.8 (26) (1-19 May) Chick-rearing 10.3 * 5.7 (50) 8.7 f 4.1 (55) 9.5 f 4.9 (105) (20 May-30 June) Post-fledging 10.8 ? 5.5 (31) 14.7 + 8.5 (39) 13.0 + 7.5 (70) (1-31 July) All periods 9.8 f 5.7 (92) 10.5 + 6.9 (109) 10.2 & 6.4 (201) (1 May-3 1 July) HERRING GULL USE OF LANDFILLS 825 TABLE 4. Mean number (&SD) of adult Herring Gulls (Larus argentatus) present and occurrence (%) of behavior at Erie County Landfill, northern Ohio, by reproductiveperiod and location at landfill, May-July 1992. Behavior (%) Reproductive Location Number of Mainte- Lo;a;@ AggXS- period at landfill gulls present Foraging nance six Other Incubation Exposed refuse 9 & 16 61 1 17 2 13 (1-19 May) Covered refuse 8 & 12 2 16 66 1 15 Non-refuse 3& 11 1 34 40 2 23 Combined 20 i 29 31 12 41 2 15 Chick-rearing Exposed refuse 33 * 51 66 2 25 4 3 (20 May-30 June) Covered refuse 14 f 28 11 11 75 1 2 Non-refuse 46 f 84 4: 14 76 P < 0.05). There was also an interaction of the We also divided observations into three time nesting period and location effects (F = 40.21; periods: 07:00-09:59, lO:OO-13:59, and 14:00- 2, 842 de P < 0.01) with similar numbers of 17:OOhr, to approximate that done for analyses gulls observed on exposed and covered refuse of telemetered gulls. The mean number of birds during the incubation period in contrast to great- observed among time periods was similar (F = er numbers of gulls observed on exposed refuse 1.10; 2, 842 de P = 0.33), from 258 (07:00- during chick-rearing and post-fledging periods. 09:59) to 231 (lO:OO-13:59)to 192 (14:00-17:OO). The number ofgulls present in all areasincreased There was no interaction of the time period and through post-fledging; however, the number of reproductive period effects(F = 0.40; 4, 842 de gulls present in non-refuse areas increased at a P = 0.81). rate 7 times greater than did the number of gulls Population turnover.The estimated daily mean observed on exposed refuse (Table 4). number of individual gulls using the landfill in- TABLE 5. Mean hours (*SD) spent, and duration and frequencyof visits by radio-telemeterednesting Herring Gulls (Lanls argentatus;eight females, seven males) at Erie County Landfill, northern Ohio, by reproductive period and time of day, May-July 1992. Number of Reproductwe 02:00- 06:00- lO:Oo- 14:00- l8:00- 22:00- Duratmn of “lslts Total hr period Sex 05:59 09:59 l3:59 l7:59 21:59 Ol:59 visits (hr) per day per day* Incubation Female 0.0 0.2 0.4 0.1 10.1 0.0 1.7 f 1.8 0.4 * 0.4 0.6 & 0.7 (1-19 May) Male co.1 co.1 10.1 co.1 co.1 0.0 0.5 f 0.6 0.2 i 0.3 0.1 ? 0.2 Combined a Calculated from mean duration of visits and mean number of visits per day 826 JERROLD L. BELANT ET AL. TABLE 6. Mean populationturnover of adult Herring Gulls (Lnrus argentatus)at Erie County Landfill, northern Ohio, by reproductiveperiod, May-July 1992. Seemethods for descriptionsof calculations. Estimated Estimated Mean total number of movements” Number Popula- daily number total number (arrivals and departures) present tion of indiwdual of indivigual Reproductive per obser- turnover P”:a;;;;;” g;;;g;;pg 07:00- lO:OO- 14:00- 07:00- period vation rate (hr) 09:59 13:59 17:oo 17:oo Incubation 20 1.4 143 416 140 86 12 298 (l-19 May) Chick-rearing 94 1.3 123 1,221 426 429 395 1,250 (20 May-30 June) Post-fledging 491 2.6 1,912 3,601 1,296 868 1,023 3,188 (l-3 1 July) All periods 221 1.7 1,300 23,601 729 441 515 1,691 (1 May-31 July) Calculated from direct observations of gulls arriving and departing from the landfill. see methods creased substantially from incubation to chick- ing, and radio telemetry data, minimum (from rearing to post-fledging, representing a mean color-marking) and maximum (from radio te- maximum of 2%, 1 lo/& and 29% of the nesting lemetry) values for the proportion of nesting population, respectively (Table 6). The hourly Herring Gulls using the landfill during incuba- population turnover rate of gulls at the landfill tion, chick-rearing, and post-fledging are 5-7%, was about twice as fast during incubation and 12-19%, and 35-55%, respectively. chick-rearing than during post-fledging. A max- During incubation, we observed a daily mean imum of 55% of the adult nesting population of 298 movements (arrivals and departures) of used the landfill at least once during post-fledg- individual gulls at the landfill, which indicates a ing, as compared to 19% during chick-rearing mean of 149 gulls used the landfill daily. This is and 7% during incubation. similar to the estimated total number of gulls From observations of color-marked gulls dur- (143) that used the landfill daily based on turn- ing the incubation period, on average, we esti- over rate estimates. During chick-rearing and mated that 10% of the adult Herring Gulls sam- post-fledging, the number of gulls arriving and pled at the landfill were marked. Therefore, departing the landfill from direct observations approximately 63% ([ 1OO%/ 16%] x 10%) of the was lower than the total number of gulls (number adult Herring Gulls observed at the landfill dur- of movements/two movements per gull) that used ing the incubation period were from the nesting the landfill daily basedon turnover rate estimates colony. Multiplying this value (63%) by the mean during these same periods (Table 6). Except dur- number of gulls observed at the landfill at any ing incubation and early chick-rearing, gulls were one time during incubation (20) yields a mean generally present at the landfill before 07:OOand of 13 adult Herring Gulls from the nesting colony after 17:O0. The estimated number of move- present per observation at the landfill during in- ments observed that were attributed to adult cubation. We therefore estimate that during the Herring Gulls generally correspondedwith direct incubation period, on average, 93 nesting Her- counts of gulls conducted at the landfill. There ring Gulls from the Sandusky Bay concentration were more (F = 46.83: 2, 261 df; P < 0.01) used the landfill each day and 3 10 gulls used the movements recorded during post-fledging than landfill during the entire period. These values during incubation or chick-rearing periods. There represent 1% and 5% of the nesting population. was neither a difference (F = 0.87; 2, 261 df; P If we assume that the 63% of the population of = 0.27) among time periods or an interaction of gulls at the landfill during the incubation period time period and reproductive period effects(F = were from the nesting colony applies to the pro- 1.35; 4, 261 df; P = 0.25). portion of gulls observed at the landfill during other periods, about 12% and 35% of the nesting DISCUSSION population used the landfill during chick-rearing Fish was the most prevalent food in the diet of and post-fledging periods, respectively. There- Herring Gulls during incubation and chick-rear- fore, based on population estimate, color-mark- ing. In contrast, garbage apparently contributed HERRING GULL USE OF LANDFILLS 827 little (< 20% overall) to their diet during this same increasein garbagein boli and food remains after period. Quality and availability of food can in- fledging, possibly a consequence of decreased fluence clutch size and eggsize and mass(Pierotti availability of fish. For example, white perch and Annett 1987, Hiom et al. 1991), which af- (Morone chrysops),a specieswe frequently found fects hatch successand offspring survival (Par- at Herring Gull nests and in chick boli, spawn sons 1970, 1972; Pierotti 1982). Garbage has near shore in Lake Erie during May, after which been considereda low quality (low-protein) food the adults move to deeper water (Scott and compared to fish; adult Herring Gulls that spe- Crossman 1973). cialized on garbagefledged fewer chicks than did Time spent at the nesting colony by adult Her- adults that specialized on other foods (Pierotti ring Gulls decreased from incubation through and Annett 1987). Also, Murphy et al. (1984) post-fledging and was inversely related to the concluded that Glaucous-winged Gull (L. gluu- mean distance adults were observed away from cescenschicks fed fish had higher survival rates the colony. Maximizing time at the nest site im- than chicks fed other foods. Thus, the high pro- proves reproductive success,allowing increased portion of fish in the diet of chicks during our attentiveness to eggsand care of chicks (Hunt study likely increased their physical condition 1972, Morris and Hunter 1976). That the mean and subsequent survival. distance adults traveled from the nesting colony The number of adult Herring Gulls observed increased from incubation to chick-rearing is at the Erie County Landfill increased just prior probably related to the additional food required to mean fledging (Fig. 2). This change may be to feed their chicks. Increasing foraging distance related to a reduction in energetic demands, as from the colony may reduce intraspecific com- adults no longer had to forage extensively for petition by dispersing adults over a larger area, chicks. Pierotti and Annett (1987) found that the which could improve foraging efficiency (Gorke diet of adult Herring Gulls changed to a lower and Brand1 1986). The greater flight radii from quality food after chicks had fledged.Pierotti and the colony observedduring post-fledgingare likely Annett (1987) have also suggestedthat non- a consequence of a reduction in fidelity to the breeding Herring Gulls can probably survive well nest site. Adults frequently leave the colony after on the protein and caloric values present in gar- chicks have fledged (see Coulson and Butterfield bage, and that the availability of garbage may 1986). increase longevity. Garbage is a dependable Frequency and duration of visits to the landfill sourceof food for gulls, requiring no special han- by female Herring Gulls was greater than that of dling techniques(Davis 1975). Thus, garbagemay males. However, overall use of landfills was min- be a readily available and nutritionally adequate imal (< 4% of total time). Coulson and Butter- source of food for adults after the breeding sea- field (1986) and Coulson et al. (1987) found no son. In addition, Annett and Pierotti (1989) sug- significant difference in the average number of gested that garbage may not be detrimental to visits made by males and females to landfills. A older gull chicks. We observed an increasein the greaterpresence at landfills by females during our occurrenceof garbagein boli from chicks during study may be attributed in part to aggressivebe- mid-June, 4-6 weeks after the mean hatching havior during foraging. In competitive feeding date. Spaans (1971) also found garbage in the situations at undisturbed sites, the smaller fe- diet of older Herring Gull chicks. males are subordinate to the more aggressive A change in diet may also be due to the avail- males (Monaghan 1980, Greig et al. 1985). Fe- ability of food rather than to a preference for males, however, are more maneuverable and are specific types of food. Spaans (197 1) concluded able to compensate somewhat by foraging when that a change in the diet of breeding Herring landfill equipment is being operated (Greig et al. Gulls in the Netherlands was not related to the 1985). Therefore, the greater frequency and du- food requirements of chicks, but to the avail- ration of visits observed for females may be nec- ability of food. In Witless Bay, Newfoundland, essaryto provide adequate opportunities to for- occurrenceof fish increasedsix-fold after the peak age. in chick hatching (Haycock and Threlfall 1975). Most studies of gull behavior at landfills have However, this change in diet coincided with a emphasized foraging behavior on exposedrefuse migration of spawning fish into the Bay. We ob- (Verbeek 1977; Burger 198 1; Burger and Goch- served a decline in occurrence of fish and an feld 198 1, 1983). We determined that most (62% 828 JERROLD L. BELANT ET AL. overall) adult Herring Gulls observedat the land- gulls are not fully understood, but appear related fill at any one time were not actively foraging. in part to the availability of alternate food and Rather, the majority were engagedin loafing or possibly, the distance landfills are from the col- maintenance activities, suggestingthat landfills ony. provide opportunities suitable for other activi- During our study, landfills were unimportant ties, such as social interaction. During the non- to nesting Herring Gulls, probably because al- breeding seasonin England, Coulson et al. (1987) ternate, higher quality food (fish) was available. similarly observed60% of the adult Herring Gulls Reutter and Hartman (1988) have reported that at any one time in activities other than foraging. Lake Erie is the world’s largestfreshwater fishery, Landfills in our study contained large, open areas with an annual harvest frequently greater than with sparse vegetation that were relatively un- the combined harvests of the other four Great disturbed by people and thus ideal for loafing or Lakes. That differencesexist in previous studies social interaction. Additional research is re- on the importance of landfills to gulls suggests quired to determine the proportion of gulls that that results of a study apply in part only to the use landfills for foraging in relation to the total area studied. Therefore, the importance of land- population present. Also, as the majority of adult fills to breeding Herring Gulls currently should Herring Gulls observed at the landfill during this be considered on a site-specific basis. study were on non-refuse areas, habitat manage- ment techniques such as elimination of standing ACKNOWLEDGMENTS water and vegetation manipulation (Blokpoel We thank personnelat the Erie County, Huron County, 1976) should be investigated to assesstheir ef- and Ottawa County landfills for accessto their facili- fectiveness in reducing overall gull use of land- ties. Personnel from American Quality Stripping, Ce- dar Point, ChesapeakeDisplay and PackagingCom- fills. pany, and Sandusky Dock Corporation granted Few studies have presented data on the pop- permission to conduct nest counts. R. A. Dolbeer pro- ulation turnover of Herring Gull populations at vided valuable suggestionsfor study design;R. A. Dol- landfills. The turnover rate of gulls at the Erie beer, P. Ewins, and D. V. Weseloh critically reviewed this manuscript; M-K. W. Belant, G. E. Bemhardt, E. County Landfill was substantial. We estimated J. Bly, B. N. Buckingham, E. C. Cleary, L. C. Gibbs, that the daily number of individuals that used L. C. Gibbs, Jr., L. A. Hegyes, K. D. Madaras, M. R. the landfill during the reproductive period was Rutger, J. A. Shieldcastle,and P. P. Woronecki pro- 4-7 times greater than the mean number of in- vided field assistance.Funds and support for this re- dividuals presentat a given time during that same searchwere provided by the Federal Aviation Admin- istration(FAA). Office of Airnorts Safetvand Standards. period. This is much faster than the daily tum- Washington, D. C., and Anports Division, Airport over of 1.5 times for adult Herring Gulls during Technology Branch, FAA Technical Center, Atlantic fall and winter in England (Coulson et al. 1987). City International Airport, New Jersey. Although gulls in our study arrived and departed LITERATURE CITED slightly earlier in the day, the mean duration of visits (102 min) was similar to the 90 min de- ANNETT,C., ANDR. PIERO~X. 1989. Chick hatching scribed by Coulson et al. (1987). as a trigger for dietary switching in the Western Gull. Colonial Waterbirds 12:4-l 1. We estimated that 5 19% of the nesting Her- BELANT,J. L. In press. Nest-site selectionand repro- ring Gull population used landfills during incu- ductive biology of roof- and island-nestingHerring bation and chick-rearing and 3555% used the Gulls. Trans. N. Am. Wildl. and Nat. Resour.Conf. landfill during post-fledging. In contrast, Pons 58. BELANT,J. L., ANDR. A. DOLBEER.1993. Population (1992) observed that 7 1% of a marked popula- status of nesting Laughing Gulls in the United tion used a landfill during incubation and chick- States 1977-l 991. A&. Biids 471220-224. rearing. Coulson and Butterfield (1986) and BELANT,J. L., AND T. W. SEAMANS.In press. Eval- Coulson et al. (1987) recorded relatively few ob- uation of dyes and techniques to color-mark in- servations of banded adult Herring Gulls at land- cubating Herring Gulls. J. Field Omithol. 64. BLOKPOEL,H. 1976. Bird hazards to aircraft. Can. fills outside the breeding season and concluded Wildl. Serv., Environ. Can., and Pub. Centre., that landfills were not preferred feeding sites. Supply and Serv. Can. Clarke, Irwin & Co. Coulson and Butterfield (1986) also stated that BLOKPOEL,H. 1983. Gull problems in Ontario. Can. landfills in their study area were infrequently used Wildl. Serv., Ottawa, Ontario. BLOKPOEL,H., AND G. D. TESSIER.1986. The Ring- by gulls from April through July. Factors affect- billed Gull in Ontario: a review of a new problem ing the use of landfills by various populations of species.Can. Wildl. Serv., Occas. Pap. 57. HERRING GULL USE OF LANDFILLS 829 BRADLEY,T. W. 1980. Gull-scaring trials for landfill HORTON,N., T. BROUGH,AND J.B.A. ROCHARD.1983. studies. Surveyor 1556-7. The importance of refusetips to gullswintering in BURGER.J. 1981. Feeding competition between an inland area of southeastEngland. J. Appl. Ecol. Laughing Gulls and Her&g Gulls at a sanitary 20:751-765. landfill. Condor 83:328-335. HUNT, G. L., JR. 1972. Influence of food distribution BURGER,J. 1986. Selection for equitability in some and human disturbanceon the reproductive suc- aspectsofreproductive investment in Herring Gulls cessof Herring Gulls. Ecology 53: 105l-106 1. L&us argentatus.Omis Stand. 18:1 l-23. _ KADLEC.J. A.. AND W. H. DRURY. 1968. Structure BURGER,J., AND M. GOCHFELD. 1981. Age-related of the New England Herring Gull population. differences in piracy behavior of four species of Ecology 491644-676. gulls, Lams. Behaviour 77~242-267. KADLEC,J. A., W. H. DRURY, JR., AND D. K. ONION. BURGER.J.. AND M. GOCHFELD.1983. Behavior of 1969. Growth and mortality of Herring Gull nine avian speciesat a Florida garbagedump. Co- chicks. Bird-banding 901222-233. lonial Waterbirds 654-63. K~HLMAN.J.. AND L. LARSSON.1974. On the impor- CONOVER.M. R. 1983. Recent changesin Ring-billed tance of refuse dumps as a food source for win- and California Gull populations in the western tering Herring Gulls Lams argentatusPont. Ornis United States.Wilson Bull. 95:362-383. Stand. 5:63-70. COULSON, J. C., AND J. BUTTERFIELD.1986. Studies MECH, L. D. 1983. Handbook of animal radiotrack- on a colony of colour-ringed Herring Gulls Larus ing. Univ. of Minnesota Press, Minneapolis. argentatus:II. Colony occupationand feedingout- MONAGHAN,P. 1980. Dominance and dispersalbe- side the breeding season.Bird Study 33:55-59. tween feeding sites in the Herring Gulls (Larus COULSON,J. C., J. BUTTERFIELD,N. DUNCAN,AND C. argentatus).Anim. Behav. 28:521-527. THOMAS. 1987. Use ofrefuse tips by adult British MORRIS,R. D., AND R. A. HUNTER. 1976. Monitoring Herring Gulls Larus arnentatusduring the week. incubation attentiveness of ground-nesting co101 J. App: Ecol. 24:789-800. nial seabirds.J. Wildl. Manaae. 40:354-357. DAHL. H. 1984. The bird strike situation and its eco- MUDGE, G. P., AND P. N. FERNS.-1982. The feeding logical background in the Copenhagen airport. ecology of five speciesof gulls (Aves: Larini) in Kastrup, p. 287-290. In M. J. Harrison, S. A. the inner Bristol Channel. J. Zool. (Lond.).~ 197: Gauthreaux.Jr.. and L. A. Abron-Robinson leds.1. 497-5 10. Proceedings’wildlifehazards to aircraft conference MURPHY,E. C., R. H. DAY, K. L. OAKLEY, AND A. A. and workshop. U.S. Dep. Transport., Fed. Avia- HOOVER. 1984. Dietarv changesand noor reuro- tion Admin., Washington, DC. ductive performance in Glaucous-winged Gulls. DAVIS, J.W.F. 1975. Specialization in feeding loca- Auk 101:532-541. tion by Herring Gulls. J. Anim. Ecol. 44:79%804. NAMS. V. 0. 1991. Locate II users manual. Truro. DOLBEER,R. A., P. P. WORONECKI,T. W. SEAMANS,B. Nova Scotia. N. BUCKINGHAM,AND E. C. CLEARY. 1990. Her- PARSONS,J. 1970. Relationship between eggsize and ring Gulls, Lams argentatus,nesting on Sandusky post-hatching mortalitv in the Herrina Gull. Na- Bay, Lake Erie, 1989. Ohio J. Sci. 90:87-89. ture 228:122-l-1222. . DRENT,~R.H. 1970. Functional aspectsof incubation PARSONS,J. 1972. Egg size, laying date and incuba- in the Herring Gull. Behaviour Suppl. 17:l-l 32. tion period in the Herring Gulls. Ibis 114:536- DRURY,W. H., AND J. A. KADLEC. 1974. The current 541. statusof the Herring Gull populationin the United PATTON,S. R. 1988. Abundance of gulls at Tampa States. Bird-Banding 45:297-306. Bay landfills. Wilson Bull. 100:43l-442. Fox, G. A., C. R. COOPER,AND J. P. RYDER. 1981. PATTON,S. R., ANDL. A. HANNERS.1984. The history Predicting the sex of Herring Gulls by using ex- of the Laughing Gull population in Tampa Bay, ternal measurements.J. Field Omithol. 52: l-9. Florida. Fla. Field Nat. 12:49-57. GORKE,M., AND R. BRANDL. 1986. How to live in PAYNTER,R. A., JR. 1949. Clutch-size and the egg colonies: spatial foraging strategiesof the Black- and chick mortality of Kent Island Herring Gulls. headed Gull. Oecologia 70:288-290. Ecology 30:146-166. GRANT,P. J. 1986. Gulls: a guide to identification. PIEROTTI,R. 1982. Habitat selectionand its effect on Buteo Books, Vermillion, SD. reproductive output in the Herring Gull in New- GREIG,S. A., J. C. COU~SON,AND P. MONAGHAN.1985. foundland. Ecology 63:854-868. Feeding strategiesof male and female adult Her- PIEROITI, R., AND C. ANNETT. 1987. Reproductive ring Gulls (Larus argentatus).Behaviour 94:4 l- consequencesof dietary specialization in an eco- 59. HARRIS,M. P. 1970. Rates and causesof increases logical generalist, p. 417-442. In A. C. Kamil, J. Krebs, and R. Pullium teds.], Foraging behavior. of some British gull populations. Bird Study 17: 325-335. Plenum, New York. HAYCOCK,K. A., AND W. THRELFALL. 1975. The PONS,J. 1992. Effects of changesin the availability breedingbiology of the Herring Gull in Newfound- of human refuseon breedingparameters in a Her- land. Auk 92678-697. ring Gull Larus argentatuspopulation on Brittany, HIOM, L., M. BOLTON, P. MONAGHAN, AND D. France. Ardea 80: 143-150. WORRALL. 1991. Exuerimental evidencefor food REUTTER,J. M., AND W. L. HARTMAN. 1988. A his- limitation of eggproduction in gulls. Omis Stand. tory of human impacts on the Lake Erie fish com- 22194-97. munity, p. 163-176. In J. F. Downhower [ed.], 830 JERROLD L. BELANT ET AL. The biogeographyof the island region of western SIBLEY,R. M., ANDR. H. MCCLEERY. 1983. Increase Lake Erie. Ohio State Univ. Press,Columbus, OH. in weight of Herring Gulls while feeding. J. Anim. SAS INSTITUTE,INC. 1988. SAS/STAT user’s guide, Ecol. 52:35-50. release 6.03 edition. SAS Institute. Carv. NC. SPAANS,A. L. 1971. On the feeding ecology of the SCHARF,W. C., G. W. SHUGART,AND M. ‘i. CHAM- Herring Gull Lams urgent&u Pont. in the north- BERLIN. 1978. Colonial birds nesting on man- em part of the Netherlands. Ardea 59:75-188. made and natural sites in the U.S. Great Lakes. VERBEEK,N.A.M. 1977. Comparative feeding ecol- U.S. Dep. Int., Fish &Wild. Serv., FWYOBS-78/ ogy of Herring Gulls Laws argentatus and Lesser 15, U.S. Army Engineer Waterways Exp. Stn., Black-backedGulls Larusfuscus. Ardea 65~25-42. Vicksburg, MS. VERMEER,K., D. POWER,AND G.E.J. SMITH. 1988. SCOTT,W. B., ANDE. J. CROSSMAN.1973. Freshwater Habitat selectionand nestingbiology of roof-nest- fishesof Canada.Fisheries Research Board of Can- ing Glaucous-wingedGulls. Colonial Waterbirds ada Bull. 184. 11:189-201. SHERIGALIN,J. E. 1990. Resultsofornithofauna study WEAVER,D. K., AND J. A. KADLEC. 1970. A method at some Soviet airfields 1972-1988. Bird Strike for trapping breeding gulls. Bird-Banding 4 1:28- Committee Europe 20:201-206. 31.