Status and Habitat Use by American Avocets Wintering at Humboldt Bay, California ’

The Condor 96:178-1X9 0 The Cooper Ornithologml Society 1994

STATUS AND HABITAT USE BY AMERICAN AVOCETS WINTERING AT HUMBOLDT BAY, CALIFORNIA ’

THOMAS J. EVANS~ AND STANLEY W. HARRIS Wildlife Department, Humboldt State University,Arcata, CA 95521

Abstract. The wintering avocet (Recurvirostraamericana) population at Humboldt Bay, California averaged 706 -t 35.6 (n = 30) and 567 1- 67.0 (n = 30) birds from October to May, 1982-1983 and 1983-1984 respectively. Avocets arrived from late-August to mid- November. Birds departedfrom February to late-April and early-May. Avocets are recorded only casuallyat other northern California coastal sites. After constructionof sewageoxidation ponds at the northeastcomer of North Humboldt Bay in 1957, the number of avocets wintering at Humboldt Bay increased. The present wintering population of avocets at Humboldt Bay has increased from 30-35 in 1960 to 500-800 at present. Humboldt Bay probably has the capability to supporta larger wintering population than presently exists. Avocetsused 868 ha in the northeastcomer of North Humboldt Bay, California. Within this area, which covered approximately 25% of available habitat in North Bay, the four major habitats used were: (1) intertidal mud flats, used mainly for feeding and resting; (2) a sewageoxidation pond, usedmainly for feedingand secondarily,as a sourceof fresh water; (3) a section of the North Bay’s highestelevation mud flats, usedas an early high tide roost; and (4) islands in a brackish 6.9 ha lake adjacent to the North Bay, used as the primary high tide roost. Avocets typically roosted on shallow, submergedbars of islands in deep non-tidal ponds or in shallow water or exposedmud near the water’s edgeof tidal mud flats. Islands adjacent to deep, non-tidal ponds were used most frequently as roosting sites. Avocets fed at the oxidation ponds mainly in October when high concentrationsof invertebrate prey, especially Daphne magna, were present. All other foraging took place on intertidal mud flats within 3 km of roosts. Avocets primarily fed using tactile methods in the wettest substrates.They usually fed within 100 m of the tide edge, and most foraging took place when tide levels were between 0.5 and 1.2 m Mean Lower Low Water (MLLW). Key words: Activity budgets;American Avocet;California; habitat use;historical status; Humboldt Bay; intertidal mudflats; Kecurvirostra americana; winteringbiology.

INTRODUCTION ing ecology (Gibson 197 1) of avocets has been The American Avocet (Recurvirostra americana) studied, relatively little is known about their win- breeds in western United States and southern tering behavior and use of wintering habitats. Canada and winters from southwestern United Knowledge of how avocets use resources for States to Guatemala (American Ornithologists’ maintenance, survival, and reproduction at Union 1983). Humboldt Bay, California is the northernmost portion of their wintering range most northerly known wintering site for this spe- could have implications for the management of cies (American Ornithologists’ Union 1983). this species. Since the late 1950s wintering American Avo- cets, formerly rare in northwestern California STUDY AREA AND METHODS (Davis 1939, Grinnell and Miller 1944) have in- Humboldt Bay consistsof two large shallow tidal creasedat Humboldt Bay, California (Yocom and basins, North (Arcata) Bay and South Bay, con- Harris 1975, Evans 1988, Harris 1991). Al- nected by a narrow passage.Humboldt Bay has though the behavior (Hamilton 1975) and breed- two high and two low tides of unequal height eachday, exposingvariable amounts of intertidal mud flats daily and seasonally (Gerstenberg 1979). The highest and lowest tidal levels en- I Received 14 June 1993. Accepted 9 September countered during this study were 2.9 m feet and 1993. * Present address: U.S. Fish and Wildlife Service, -0.5 m Mean Lower Low Water (MLLW), re- Marine Mammals Management, 4230 University spectively. Detailed behavioral studieswere con- Drive, Anchorage, AK 99508. ducted on the intertidal mudflats near the Arcata

[I781 WINTER HABITAT USE BY AMERICAN AVOCETS 179

FIGURE 1. Study area, wintering home range and censuslocations of American Avocets in North Humboldt Bay, California.

Marsh Project and Jacoby Creek delta, the south- tide counts were begun one hour before the pre- ernmost oxidation ponds of the Arcata Public dicted time of highestwater (NOAA 1982-l 984). Works sewagetreatment facility, and the Klopp A compensating polar planimeter and a dot Lake Unit of Arcata Marsh Project (Fig. 1, Evans grid with 64 dots to a square inch was used to 1988). determine the total hectares in the study areas The study was conducted 1 October 1982-l 5 on United StatesGeologic Survey 7.5 topograph- May 1983,3 October 1983-8 May 1984, and 19 ic map (Arcata-South, California). We used sur- December 1984-29 April 1985. A wintering vey data plotted on maps from the weekly (North population estimate was not calculated for 1984- Bay) and bimonthly (South Bay) surveys and a 1985 becausepopulation counts were not made modified version of the modified minimum area prior to 19 December 1984. All data were gath- method (Harvey and Barbour 1965) to plot the ered by direct observations using a 25 x spotting “wintering home range” of the wintering avocet scopeand 9 x 36 binoculars. Counts of avocets population in Humboldt Bay. Since avocetswere were made during 76 weekly surveysat both high never observed in upland sites,upland siteswere and low tide from 11 stations around North Bay excluded and the outermost avocet sightingsad- (Fig. 1) and on 42 bimonthly surveys from eight jacent to the upland areas were connected when stations around South Bay. The location of av- drawing the “modified home range” boundary ocets was marked on maps of the study area. near the shoreline. Low tide counts were begun after falling tides Daytime behavioral observations, activity had exposed 30-50% of the tidal mud flat. High budgets, and habitat use by avocets were deter- 180 THOMAS J. EVANS AND STANLEY W. HARRIS

600

500

400

300

200

100 i

0 + OCT NOV DEC JAN FEB MAR APR MAY FIGURE 2. Maximum number of AmericanAvocets observed weekly in North Humboldt Bay, California, winters 1982-1985.

mined by scan sampling (Altmann 1974) on sev- vided into five 3-hr periods, or three 5-hr peri- en study sites on the intertidal mudflats and ad- ods, starting at 05:OO hr. On any one day, be- jacent brackishand sewageponds. The behavioral havioral observations were conducted on the categories used were described by Hamilton study site with the largest number of birds. This (1975) and included resting, feeding, preening, usually yielded a study population of about 280 comfort movements, alert, swimming, walking, birds. The flock was repeatedly scanned until and flying. A total of 197 hr of behavioral ob- observer fatigue set in, or more than 50% of the servations were made between 1 October 1982 birds initially present had left the area. Obser- and 12 January 1983 and 970 hr between 3 Oc- vation sessionsaveraged 4 hr long (range 10 min- tober 1983 and 3 May 1984. The duration of 5 hr). General observations were made on the scanswas determined by the time needed to rec- intertidal mudflats on 32 dark nights (cloudy ord all birds present on the initial scan. The in- nights or nights with a new moon) and on 25 terval between scanswas equal to the scanlength. bright nights (clear nights with three-fourths to An average of 250-350 birds was scannedduring a full moon) during moderate to low tide heights scans of 5 min or less (> 96% of all scans = to subjectively determine if avocets fed at night. 6,638), taken between 06:OOand 17:00 hr on 292 Four to six nights each month were chosen to days of field work. When scanning large flocks, record the occurrence of nocturnal foraging. up to 750 birds, it was difficult to keep track of Avocets use both visual and tactile feeding the movements of every bird and on occasions methods while feeding in the water column or some individuals or small groups were counted on exposed mudflats (Hamilton 1975). Visual more than once. Longer scans were therefore feeding methods were classified as pecking, consideredless reliable. Consequently, only scans plunging, snatching, and bill pursuit and tactile 5 min or less were used in the analysis. Scans feeding methods included filtering, scraping,and were taken during all daylight hours and at all single and multiple scythe (Hamilton 1975). tide stages.For analysis the time of day was di- Detailed descriptions of study sites, method- WINTER HABITAT USE BY AMERICAN AVOCETS 181 ology, definitions of behavior, weather, and tides RESULTS encountered during the study may be found in POPULATION SIZES Evans (1988). The earliest fall migrant American Avocets ar- STATISTICAL ANALYSIS rived at Humboldt Bay in August in 1982 and We examined differencesin avocet numbers ex- 1983 (Harris 199 1). By 1 October the population hibiting behaviors among habitats and tide had reached about 200 birds in 1982 and 5 11 heights using a Kruskal-Wallis one-way analysis birds in 1983. Between late October and early of variance (ANOVA) (P < 0.05). When analysis February overall population numbers remained demonstrated significant differences, we identi- fairly constant (Fig. 2). The November-March fied the differencesusing Dunn’s distribution free, population averaged 706.4 (SD = 35.6) birds in multiple comparison procedure (Zar 1984). Be- 1982-1983 and 566.8 (SD = 67.0) birds in 1983- cause of potential Type I errors (i.e., behaviors 1984. Peak counts of 827,667,736 occurred on were correlated with each other) all multiple 27 November 1982, 22 January 1984, and 22 comparisons were made at a significancelevel of January 1985, respectively. Weekly fluctuations 0.003 to 0.0008 as suggestedby Bonferroni (Bray of up to 200 birds sometimes occurredespecially and Maxwell 1986). Significance levels were ad- in October 1983 and January 1985. Numbers justed by the Bonferroni method so that the declined steadilybetween February and late April. probability of a Type 1 error was at most 5% for All birds were gone by 29 April 1985 and 8 May each group of comparisons.Alpha values for each 1984, but 57 were still present on 15 May 1983, comparison between a behavior (i.e., resting) and when the last population survey was made. an environmental variable (i.e., tide height) using Dunn’s distribution free, multiple comparison WINTERING HOME RANGE test are indicated in the text as follows (i.e., alpha Nearly all avocetspresent used a “wintering home = 0.0003). range” of 868 ha, comprising only 23.6% ofNorth The proportion of the flock engagedin an ac- Bay (Fig. 1) as determined by changing the mod- tivity was analyzed using a normal approxima- ified minimum area method (Harvey and Bar- tion test for comparing more than two propor- bour 1965). Although the concept of home range tions followed by a Tukey-type multiple usually has been applied to individuals we use comparison test (Zar 1984). When the propor- the term “wintering home range” to describe the tion of the flock engagedin a particular activity area used by the avocet population wintering at was tested against independent variables, signif- Humboldt Bay, California. icant (P < 0.05) differencesoccurred even when Within the northeastern portion of North Bay the differences between the proportions of the avocets used four major habitats: (1) intertidal flock were only one percent. These differences mudflats, used mainly for foraging; (2) the Arcata were related to large sample sizes resulting in sewageoxidation pond, used as a feeding area in extremely high chi-square values. Therefore, in October each year; (3) a high elevation intertidal interpreting the resultswith respectto proportion mudflat near the mouth of Jacoby Creek, used of the flock, biological relevancy, as well as sta- as an early high tide roost; and (4) islands in the tistical significance, was given consideration. Klopp Lake Unit of the Arcata Marsh and Wild- Non-parametric statistical programs from SPSS life Sanctuary, usedas the primary high tide roost (Nie et al. 1975, Hull and Nie 198 1) and BMDP (Fig 1). During the weekly and bimonthly cen- (Dixon and Brown 1979) were usedfor data anal- susesin North Bay and South Bay, respectively, ysis. no avocetswere ever seenon South Bay and none Flight directions were analyzed using para- were recorded from census locations 6 through metric measures(Zar 1984) and circular statistics 11 (Fig. 1) located on the southeast,south, west, (Batschelet198 1). For analysis offlight direction, and north margins of North Bay (Table 1). There heading of avocets leaving the roost were clas- are only a few scattered records of avocets for sified in eight directions: north (23”-338”), north- the north coastaway from North Humboldt Bay east (24”-68”) east (69”-113”), southeast (114” (Evans 1988, Harris 1991). Avocets were en- 158”) south (159”-203”) southwest(204 ”-248”) countered outside the “wintering home range” west (249”-293”) and northwest (294”-338”). only 30 times in 1,142 hours of behavioral ob- 182 THOMAS J. EVANS AND STANLEY W. HARRIS

servation, between October 1982 and April 1985, and always in groups of fewer than 52 birds. None of the 30 encounters occurred on weekly or bimonthly censusesand 26 of the 30 encounters involved birds on the intertidal mudflats along the eastern shoreline of North Bay im- mediately south of the “wintering home range.”

GENERAL HABITAT USE AND BEHAVIOR The general activity pattern was one of foraging on intertidal mudflats at low tide and resting at Klopp Lake at high tide. Temporary or seasonal variations to this general pattern involved early high tide roosting at JacobyCreek, some October foraging at the Arcata sewage oxidation pond, and brief periods of resting on intertidal mudflats during moderate and low tides.

HIGH TIDE ACTIVITIES Roosting. On incoming tides, avocets typically congregatedalong the water’s edge on the high elevation mudflats north of Jacoby Creek (Fig. 1). Even though these were the highest elevation mudflats in the North Bay, they became flooded by most high tides and avocets remained at the JacobyCreek roost throughout the tide cycle only twice in 292 days of observation. When flooded out at Jacoby Creek they flew to the three islands in Klopp Lake 1.35 km away. About two thirds of all birds scanned at the high tide roost were resting (Table 2). Secondary activities at the roost sites were flying, alert behavior, and preening (Klopp Lake) and flying, preening, and feeding (Jacoby Creek) (Table 2). When the tide level was between 1.5 to 2.3+ m MLLW, 87.3% of the scanned birds were re- cordedas resting. At lower tides significantly fewer birds rested and significantly more birds engagedin other activities (alpha = 0.0003) (Evans 1988). Avocets generally stood in shallow (2-l 3 cm) water at roosts. At Jacoby Creek, roosting birds were about evenly distributed between exposed mudflats with visible standing water, flooded mudflats with water just covering their feet, and mudflats with water depths up to their abdomens (Fig. 3). As tides approachedhigh water, avocets consistently roosted on high elevation mudflats until the water levels forced them to move to islands in Klopp Lake. The islands in KJopp Lake ranged from 0.0012 to 0.0017 ha, were unvege- tated, and bordered by a bar submerged under 2-13 cm of water. At Klopp Lake, avocets typ-

184 THOMAS J. EVANS AND STANLEY W. HARRIS

N = 1660 0 Exposed Substrate: No visible water m Exposed Mudflats: Visible standing water @j Flooded Mudflats: Water foot deep N = 96 tj Flooded Mudflats: Water foot to belly deep m Flooded Mudflats: Water > belly deep N = Number of scans

N = 3020

N = 654 L J

Klopp Lake Jacoby Creek Oxidation Pond Intehdal Mudflats

Roost Sites Foraging Areas FIGURE 3. Mean number of avocetsper scanresting relative to water depth.Humboldt Bay, California. Winters 1982-1983, 1983-1984. N = numberof scans. ically rested on exposedmud at the water’s edge prey, probably Daphne magna, where they fed (43%) or on the submergedbars with water nearly intensively using single and multiple scythetech- up to their abdomens (5 1%). The area encom- niques (Hamilton 1975) (784 observations) or passedby the submerged bar was limited by the pecking prey from the water (37 observations). abrupt nature of the island contours and caused Feeding bouts lasted 243 min (X = 8.2 min). some avocets to rest on the higher areas. Often, Even though water levels of the oxidation pond wave action forced roosting avocets to move to were independent of tide level, and theoretically higher ground. Large numbers of Marbled God- avocets could have fed there at any time, the wits (Limosafedoa), Black-bellied Plovers (Plu- pattern of use there remained basically un- vialis squatarola), Least (Caladris minutilla) and changed from the general tidal pattern. They Western (Caladris mauri) Sandpipers, Dunlins roosted elsewhere during the highest tides and (Erolia alpina), and Willets (Catoptrophorus only moved to foraging areas, including the ox- semipalmatus) also used the higher, dryer, center idation ponds, as the tide began to ebb. Although portions of the islands as roosts. the intertidal mudflats first become available for Foraging. The only intense foraging activity foraging at tidal levels between 1.2 and 1.5 m observed during high tides, above 1.5 m MLLW, MLLW, avocets often began to feed at the oxi- occurred at the sewageoxidation ponds. Overall, dation ponds when the tidal levels were still high, 78% of all birds scanned at the oxidation ponds between 1.8 and 2.0 m MLLW. Most avocetsleft were foraging (Table 2). Avocets used the oxi- the oxidation ponds and moved to the intertidal dation ponds only in October and early Novem- mudflats when the tide dropped below 1.2 m ber 1982 and 1983. Dense flocks of up to 360 MLLW. Although this was the general foraging birds swam over concentrations of invertebrate pattern, avocets occasionally fed almost exclu- WINTER HABITAT USE BY AMERICAN AVOCETS 185 sively at the oxidation ponds even when tide avocets occurred in areas with water covering levels were optimum for foraging on the inter- their feet, water up to belly deep, and on mudflats tidal mudflats. with visible surface water, 71%, 64% and 49% of the flock were observed feeding, respectively LOW TIDE ACTIVITIES (Fig. 4). Overall, tactile feeding methods were Avocets typical left the Klopp Lake roost when used much more frequently than visual feeding the tide was between 1.2 and 1.5 m MLLW. methods (x2 = 18.26, df = 1, P < 0.001). Single Avocets departing Klopp Lake flew south (mean scythewas the primary feedingmethod usedwhen angle of departure = 178” + 6.0”) to foraging using tactile methods (R = 89.6% of birds areas more often than in any other direction (x2 scanned).When avocetsfed visually, pecking was = 837.33, df = 7, P < 0.001). On the few oc- the primary visual feeding method used (X = casions when avocets departed in other direc- 93.1% of birds scanned). tions, they avoided flying over upland areas. After dark, no avocets were found feeding in The pattern of avocet distribution in the study October (two dark nights, two bright nights) or area was described only in qualitative terms. April (four dark nights, two bright nights) but an Flocks were describedas either compact or loose- averageof 172 American Avocets was seen feed- ly-knit (Goss-Custard 1969). More than 95% of ing per night on dark nights (28 counts) and 2 17 all avocets foraged in loosely knit groups on the on bright nights (19 counts) between November intertidal mudflats. Occasionally large compact and March. groupsfed along the tidal edgeor in water deeper The proportion of avocets foraging showed no than an avocet’s abdomen. Although most birds significant diurnal differences, when comparing followed the ebb tide edge out into North Bay, different times of day during flood or ebb tides. a few birds always remained on higher exposed mudflats containing visible surface water. DISCUSSION Sixty-one percent of birds on mudflats were foraging (Table 2) with peak periods occurring POPULATION TRENDS when the tide levels were 0.5 and 1.2 m MLLW The first record of an American Avocet in Hum- (mean of peak periods = 69.6%). At tide heights boldt Bay was on 17 August 1935 (Davis 1939). greater than 1.4 m few birds fed and at heights Anderson (1943-1947) recorded one bird in lower than 0.5 m, they remained on mudflats North Humboldt Bay on 27 November 1943. C. and alternated periods of feeding with periods of I. Clay, an active collector at Humboldt Bay be- resting, preening, and comfort movements. Birds tween 1909 and 1982, first recorded an avocet feeding at tides below 0.5 m MLLW usually con- on 26 January 1944. Between 1943 and March centrated along the edgesof tidal channels. 1961 there were only 15 sightings of avocets Birds resting on mudflats at low tide levels (ranging from one to 75 birds) at Humboldt Bay usedexposed sites adjacent to tidal channels. Sig- (Clay 1901-1953, Anderson 1943-1947, Harris nificantly more avocets (X = 84.1%) rested on 1991). After the city of Arcata created a 22-ha mudflats within 100 m of water, in channels or sewageoxidation pond at the northeast comer at the tidal edge, than on mudflats at distances of Humboldt Bay in 1957, wintering avocet greater than 100 m from water (x2 = 5.02, df = numbers increased from 30 to 40 in the early 1, P < 0.05). An average of 84.1% of all birds 1960’s to 250 in 1970 (Gerstenberg 1972), and resting on the mudflats were within 100 m of from 400 to 500 in 1979 (Harris 1991) to 827 water in the channels or at the tidal edge (x2 = on 27 November 1982 (Fig. 2). 5.02, df = 1, P < 0.05). These avocets typically The population at Humboldt Bay begins to restedin loose flocksof two to 75 birds; however, arrive in August (Harris 199 1). Most of the win- the overall number of avocets resting on inter- tering population was presentby November (Fig. tidal mudflats was relatively low compared to 2). This appearsto be somewhat earlier than far- the numbers feeding (Table 2). ther south in coastal California. At Bolinas La- Avocets foraging on the intertidal mudflats goon between 197 1 and 1976, avocets increased mostly waded in water covering their feet (43%), in numbers from November (150 birds) to Jan- in water up to belly deep (34%), or walked on uary (578 birds, Page et al. 1979). At south San mudflats with visible surfacewater (22%). When Francisco Bay, avocets began to arrive in Oc- 186 THOMAS J. EVANS AND STANLEY W. HARRIS

r N=1950 mmT] Rest N=2034 m Feed 1 H Swim N=2694 ( Other I I N = Scans

N=393

N=204 g 25- / I

Exposed Mudflats Flooded Mudflats No Standing Visible Water Between feet Water > water Standing Covering and Body Belly Deep Water Feet FIGURE 4. Mean number of avocets per scan present on intertidal mud flats engagedin various behaviors related to water depth, Humboldt Bay, California. Winters 1982-1983, 1983-1984. N = Number of scans.

tober and November (100 birds) and increased cets to roost near or in shallow water, to avoid until February (1,100 birds, Recher 1966). While flying over land, and to land in water when dis- fluctuating numbers in October suggestedthat turbed or threatened shows that islands are su- some fall birds recorded at Humboldt Bay may perior to shorelinesas roostsand that deep, non- be passagebirds on their way to more southern tidal ponds containing islands for roosts are im- wintering areas, we found no evidence of a sim- portant components of their wintering habitat. ilar spring flight. Rather, beginning in February, SEWAGE OXIDATION PONDS the Humboldt Bay population steadily declined and was essentially gone by early May. Because Avocets used oxidation ponds for foraging, we recorded only one migratory flight in 1,168 drinking, and sanctuary (Evans 1988). Except for hours of diurnal observation, we presume most occasionalphalaropes (Phaluropus spp.), avocets avocet migration occurs at night. were the only shorebirds recorded feeding on concentrations of invertebrates in the oxidation ROOST SITES ponds. The timing and size of the populations Migrating shorebirds prefer staging areas (Pien- of Daphne magna in the sewageoxidation ponds kowski and Evans 1984) and wintering areaswith were highly variable (Frodge 1985). Therefore, safe roost sites (Myers 1984). Typically shore- it would be reasonable to expect that use of this birds in tidal areas roost at high tide and at night food sourceby avocets also would vary. Lee (un- (Goss-Custard 1969, Wolff 1969, Thomas and publ. manuscript) found that concentrations of Dartnell 197 1, Kelly and Cogswell 1979, Puttick Daphne under “swarms” of foraging ducks were 1979, Pienkowski 1982). The tendency of avo- greaterthan in adjacent areaswhere no waterfowl WINTER HABITAT USE BY AMERICAN AVOCETS 187 were present. When avocets fed at the oxidation flats and were never observed feeding (Evans pond, they occurred in dense flocks (less than 1988). During light rains or overcast skies avo- one body length apart) similar to the “swarms” cets fed primarily on mudflats with visible sur- reported by Lee (unpubl. manuscript) for ducks face water and flooded mudflats with lessthan 3 and coots. Lee (unpubl. manuscript) recorded cm of water. Prey availability may increase dur- waterfowl swarms throughout the fall, but avo- ing light rains becausemarine invertebrates be- cetsfed only in swarms during October and early come more active to avoid fresh water (Page et November. This swarming behavior suggeststhat al. 1979) or because the mudflats remain wet, the avocetsat the oxidation ponds were probably enhancing feeding. Under overcast skies prey feeding on denseconcentrations of Daphne mag- availability may increase because negatively na. phototactic benthic invertebrates may move Although avocets occasionally visited the ox- closer to the surface and/or may be more active idation ponds, Jacoby Creek, and Klopp Lake to under low illumination (Hynes 1970). drink or bathe in fresh water it is unlikely that Avocets were never seenfeeding in upland sites fresh water is critical to avocet survival because during or after rains although other shorebird they can exist in hypersaline or alkaline envi- species have been reported to feed on earth- ronments without using fresh water for long pe- worms forced to the surfaceby rain (Gerstenberg riods (Mahoney and Jehl 1985). 1972, Kelly and Cogswell 1979, Townshend 1981). Goss-Custard (1969) and Smith (1975) INTERTIDAL MUDFLATS found that prey capture rates by shorebirds de- Tides exert great influence on the distribution, creasedduring rainfall. Moderate to heavy rains abundance, and behavior of estuarine organisms may decreasesurface activity of prey (Goss-Cus- (Burger et al. 1977, Evans 1979, Connors et al. tard 1969, Smith 1975) and reduce prey avail- 198 1, Burger 1984). For shorebirds, tides affect ability due to increased turbidity and water and the amount of foraging space, length of time mud flat distortion (Dugan et al. 198 1). mudflats are available, and availability of prey Several shorebird specieshave been reported (Reeker 1966, Puttick 198 1, Evans and Dugan to forage at night only because they could not 1984). At Humboldt Bay, large expansesof high find enough food during daylight (Smith 1975, and mid-level mudflats become exposed nearly Evans 1976, Goss-Custard et al. 1977). Dugan simultaneously with a slight drop in the tide. Yet (198 1) and Pienkowski and Evans (1984) found the birds did not spread out over the entire ex- that shorebirds may obtain more than 50% of posed mudflats, but rather remained in the their mid-winter food at night. Many inverte- northeastcomer of the Bay. Prey availability near brates are active at the surface at night (Vader the water’s edge in this relatively small section 1964, Dugan 198 1, Pienkowski 1982). Water- of North Humboldt Bay may have influenced fowl (Swanson and Sargent 1972) and shorebirds which flats were used. (Goss-Custard 1970, Dugan 198 1) have been re- Like other shorebirds (Recher 1966, Smith ported to feed at night in responseto increased 1975, Burger et al. 1977, Evans 1979, Pien- prey activity. Visual feeders may have reduced kowski 198 1, Goss-Custard 1984) foraging av- successat night due to difficulties in detecting ocetsgenerally avoided feeding in dry areas.Sub- prey or locating foragingsites (Smith 1975, Evans strate wetness influences prey behavior and 1976, Pienkowski 1983). Redshanks foraging at availability to shorebirds (Burger et al. 1977, night switched from visual to tactile foraging Evans 1979, Pienkowski 198 1, Goss-Custard methods and changed foraging sites (Goss-Cus- 1984). Carrin (1973) found that 93.5% of the tard 1969). Since avocets primarily used individuals and 60% of the biomass of benthic tactile methods and were observed feeding, from invertebrates occurred in the wettest substrates November to March, in the same locations at in Humboldt Bay. This may explain why the night as during the day, we presume they can proportion of avocets foraging decreasedat low forage effectively at night assuming sufficient tidal levels (-0.5-0.2 m MLLW). Prey also may quantities of prey are available. The ability to be more abundant or active near the tide edge feed at night may be important during extended (Smith 1975) which may explain why avocets periods of severeweather during mid-winter, es- generally foraged near the tide edge. pecially when high winds, heavy rainfall, and During heavy rains avocetsrested on the mud- cool temperatures occur simultaneously. 188 THOMAS J. EVANS AND STANLEY W. HARRIS

The earliest wintering populations of Ameri- BRAY,J. H., ANDS. E. MAXWELL. 1986. Multivariate can Avocets at Humboldt Bay were recorded in analysisof variance. SageUniversity paper series on quantitative applicationsin the socialsciences. the northeast comer of the Bay (Harris 199 1). SagePublications, Beverly Hills, CA. Despite increasing wintering populations since BURGER,J. 1984. Abiotic factors affecting migrant the early 1960s the pattern of habitat use has not shorebirds,p. l-72. In J. Burgerand B. Olla [eds.], changed. The establishment and increase of this Behavior of marine animals: shorebirds-migra- wintering population follows closely the con- tion and foragingbehavior. Vol. 6. Plenum Press, New York. struction of the oxidation pond in 1957. In 1969 BURGER,J., M. A. HOWE,D. C. HAHN, ANDJ. CHASE. the high-level mudflats adjacent to the oxidation 1977. Effects of tide cycles on habitat selection pond were used by avocetsat three times the rate and habitat partitioning by migratory shorebirds. as moderate-elevation mudflats and 10 times the Auk 94~743-758. CARIUN,L. F. 1973. Availability of invertebrates as rate as the low-elevation mudflats farther away shorebirdfood on a Humboldt Bay mudflat. M.Sc. (Gerstenberg 1972). This suggeststhe oxidation thesis, Humboldt State Univ., Arcata, CA. ponds with their seasonalDapne blooms and nu- CLAY, C. I. 1901-1953. Oological Collection of C.I. trient-rich discharge into the northeastern por- Clay. 14 Vols. Humboldt State Univ., Arcata, CA. tion of Humboldt Bay may have contributed to CONNORS,P. G., J. P. MYERS,C.S.W. CONNORS,AND F. A. P~TELKA. 1981. Inter habitat movements both the establishment and spatial pattern of for- by sanderlingsin relation to foraging profitability aging for this population. The later creation of and tidal cycle. Auk 98(1):49-64. the islands in Klopp Lake has reinforced this DAVIS, J. M. 1939. More shorebirdsfrom Humboldt pattern of habitat use by providing secure,suit- Bay region. Condor 4 1:124. DIXON, W. J., AND M. B. BROWN. 1979. Biomedical able islands for roosts near the feeding areas. computer programs. P-Series. Univ. of California It is advantageousfor roost sitesto be adjacent Press, Berkeley, CA. to foraging areas (Zwarts and Wanink 1984). In DUGANP. J. 1981. The importance of nocturnal for- North Humboldt Bay theseavocets foraged with- aging in shorebirds: a consequenceof increased in 3 km of the roosts. If food resources near invertebrate prey activity, p. 251-260. In N. V. Jonesand W. J. Wolff [eds.], Feeding and survival winter roosts are depleted, shorebirds have to strategiesof estuarine organisms. Plenum Press, shift roost locations or travel farther to forage New York. (Zwarts and Wanink 1984). Since the Humboldt DUGAN, P.J., P. R. EVANS,L. R. GOODYER,AND N. C. Bay avocets did not shift foraging patterns or DAVIDSON. 1981. Winter fat reserves in shorebirds: disturbance of regulated levels by severe roost locations and did not forage at high tide, weather conditions. Ibis 123:359-363. we assume the food resourceson the mudflats EVANS,P. R. 1976. Energy balance and optimal for- were adequateto support the 500-800 birds pres- aging strategiesin shorebirds: some implications ent in 1982-1984. Becausethey only used one for their distribution and movement in the non- quarter of North Bay and did not use South Bay breeding season.Ardea 64: 117-l 39. EVANS,P. R. 1979. Adaptations shown by foraging at all, the potential for further increases in the shorebirdsto cyclicalvariations in the activity and wintering population seems high. availability of their invertebrate prey, p. 357-366. In E. Naylor and R. G. Hartnoll [eds.], Cyclic ACKNOWLEDGMENTS phenomenonin marine plants and mammals. Per- gamon Press, Oxford. We are especiallygrateful to Barry Noon for providing EVANS,P. R., ANDP. J. DUGAN. 1984. Coastal birds: advice on the statistical analysis. We thank Anthony numbers in relation to food resources,p. 8-29. In R. Degange, Peter Bergstrom, and Mark Colwell for P. R. Evans, J. D. Goss-Custard,and W. G. Hale comments on earlier drafts of this manuscript. teds.], Coastal waders and wildfowl in winter. Cambridge Univ. Press, London. LITERATURE CITED EVANS,T. J. 1988. Habitat use and behavioral ecol- ALTMANN,J. 1974. Observational study of behavior ogy of American Avocets (Recurvirostra Ameri- sampling methods. Behaviour 49:227-268. cana) Wintering at Humboldt Bay, California. AMERICAN ORNITHOLOGISTSUNION.’ 1983. The M.Sc. thesis, Humboldt State Univ., Arcata, CA. American Ornithologists’ check-list of North FRODGE,J. D. 1985. Studies on Daphnia magna in American Birds. Sixth edition. American Omi- a large oxidation pond. M.Sc. thesis, Humboldt thologists’ Union. State Univ., Arcata, CA. ANDERSON,W. 1943-1947. Bird observations:Hum- GERSTENBERG,R. H. 1972. A study of shorebirds boldt Bay region. On file at Humboldt State Univ. (Charadrii) in Humboldt Bay, California 1968- Library, Arcata, CA. 1969. M.Sc. thesis,Humboldt StateUniv., Arcata, BATSCHELET,E. 1981. Circular statistics in Biology. CA. Academic Press,New York. GERSTENLIERG,R. H. 1979. Habitat utilization by WINTER HABITAT USE BY AMERICAN AVOCETS 189

wintering and migrating shorebirdson Humboldt PIENKOWSKI,M. W. 1981. How foragingplovers cope Bay, California, p. 33-40. In F. A. Pitelka led.],. _ with environmental effectson invertebrate behav- Stud. in Avian Biol. No. 2. ior and availability, p. 179-192. In N. V. Jones GIBSON,F. 1971. Breeding biology of the American and W. J. Wolff [eds.], Feeding and survival strat- Avocet in central Oregon. Condor 73444-454. egiesof estuarine organisms.Plenum Press, Lon- GOSS-CUSTARD,J. D. 1969. The winter feeding ecol- don.

oav of the Redshank./.(Trinau Y totanus). Ibis 111: PIENKOWSKI,M. W. 1982. Diet and energy intake of 338-356. Grey and Ringed Plovers Pluviulis squutarolu and GOSS-CUSTARD,J. D. 1970. The responsesof Red- Charadrius hiaticula in the non-breeding season. shank(Tringu totanus) (L.) to spatial variations in J. Zool. 197:51l-549. the density of their prey. J. Anim. Ecol. 39:91- PIENKOWSKI,M. W. 1983. Changes in the foraging 113. pattern of plovers in relation to environmental GOSS-CUSTARD,J. D. 1984. Intake rates and food factors. Anim. Behav. 3 1:244-264. supply in migrating and wintering shorebirds, p. PIENKOWSKI,M. W., ANDP. R. EVANS. 1984. Migra- 233-270. In J. Burgerand B. Olla [eds.], Behavior tory behavior of shorebirds in the western Pale- of marine animals, shorebirds-migration and arctic. D. 74-124. In J. Burger and B. Olla leds.1, foraaingbehavior. Vol. 6, PlenumPress, New York. Behavior of marine mammals, shorebirds-mi- GOSS-&T&, J. D., A. JENYON,R. E. JONES, P. E. gration and foraging behavior. Vol. 6. Plenum NEWBERG.AND R.L.B. WILLIAMS. 1977. The Press, London. ecology of the Seasonal variation in the feeding PU~~ICK,G. M. 1979. Foragingbehavior and activity conditionsofwading birds (Chardrii).J. Appl. Ecol. budgetsof Curlew Sandpipers.Ardea 67: 11 l-l 22. 14(14):701-719. PUT-TICK, G. M. 1981. Sex related differencesin the GRINNELL,J., AND A. H. MILLER. 1944. The distri- foraging behavior of Curlew Sandpipers. Omis. bution of birds of California. Pac. Coast Avif. 27: Stand. 12:13. 157. RECHER,H. F. 1966. Some aspectsof ecologyof mi- HAMILTON,R. B. 1975. Comparative behavior of the grant shorebirds. Ecology 47:393-407. American Avocet and the Black-neckedStilt (Re- SMITH, P. S. 1975. A study of the winter feeding curvirostridae). Omithol. Monogr. 17:l-98. ecology and behavior of the Bar-tailed Godwit HARRIS,S. W. 1991. Northwestern California birds. (Limosa lupponicu). Ph.D.diss. Univ. of Durham, Humboldt State Univ. Press, Arcata, CA. England. HARVEY,M. J., AND R. W. BARBOUR. 1965. Home SWANSON,G. A., AND A. B. SARGENT.1972. Obser- range of Microtus ochroguster as determined by a vations of night-time feeding behavior of ducks. modified minimum area method. J. Mammal. 46: J. Wildl. Manage. 36:959-96 1. 398-402. THOMAS,D. G., ANDA. J. DARTNELL. 1971. Ecolog- HULL, C. HADW, ANLINORMAN H. NIE. 1981. Sta- ical aspectsof the feeding behavior of two Cali- tistical packagefor the social sciences.Update 7- dridinae sandpipers wintering in southeastern 9. McGraw Hill. San Francisco. Tasmania. Emu 7 1:20-26. HYNES,H.B.N. 1976. The ecologyof running waters. TOWNSHEND,D. J. 1981. The importance of field Liverpool Univ. Press, Liverpool, England. feeding to the survival of wintering male and fe- KELLY. P. R.. AND H. L. COGSWELL.1979. Move- male curlews Numenius arquata on the Tees Es- ments and habitat use by wintering populations tuary, p. 261-273. In N. V. Jonesand W. J. Wolff of Willets and Marbled Godwits, p. 69-82. In F. [eds.], Feeding and survival strategiesof estuarine A. Pitelka [ed.], Stud. Avian Biol. No. 2. organisms.Plenum Press, New York. MAHONEY,S. A., AND J. R. JEHL,JR. 1985. Adap- VADER. W.J.M. 1964. A preliminary investigation tations of migratory shorebirds to highly saline into the reaction of the infauna of tidal flats to and alkaline lakes:Wilson ’s Phalaropeand Amer- tidal fluctuationsin water level. Neth. J. Sea Res. ican Avocet. Condor 871520-527. _ 2:189. MYERS,J. P. 1984. Spacingbehavior of nonbreeding WOLFF, W. J. 1969. Distribution of non-breeding shorebirds, p. 27 1-321. In Behavior of marine waders in an estuarinearea in relation to the dis- animals, Vol 6., shorebirds-migration and for- tribution of their food organisms.Ardea 57: I-25. aging behavior. Plenum Press, New York. YOCOM,C. F., AND S. W. HARRIS. 1975. Birds of NIE, N., C. H. HULL, J. G. JENKINS,K. STEINBRENNER, Northwestern California: status.habitats, and dis- AND D. H. BRENT. 1975. Statistical packagefor tribution of birds of Northwestern California. the social sciences.McGraw Hill, New York. Humboldt State Univ., Arcata, CA. NOAA-NATIONAL OCEANICAND ATMOSPHERICAD- ZAR, J. H. 1984. Biostatisticalanalysis. Prentice Hall, MINISTRATION. 1982-l 984. Tide tables: West NJ. Coast of North and South America. Rockville, ZWARTS,L., AND J. WANINK. 1984. How oystercatch- MD. ers and curlews successivelydeplete clams, p. 69- PAGE,G. W., L. E. STENZEL,AND C. M. WOLFE. 1979. 84. In P. R. Evans, J. D. Goss-Custard, and W. Aspects of the occurrenceof shorebirdson a cen- G. Hale [eds.], Coastal waders and wildfowl in tral California estuary, p. 15-32. In F. A. Pitelka winter. Cambridge Univ. Press, London. [ed.], Stud. Avian Biol. No. 2.