Environmental and Habitat Correlates of Pasture Use by Nonbreeding Shorebirds

The Condor 9X337-344 0 The Cooper Ornithological Society 1997

ENVIRONMENTAL AND HABITAT CORRELATES OF PASTURE USE BY NONBREEDING SHOREBIRDS

MARK A. COLWELL AND SARAH L. DODD~ Department of Wildlife, Humboldt State University, Arcata, CA 95521, e-mail: [email protected]

Abstract. Throughoutcoastal regions of the world, pasturelandsoften augmentimportant and declining intertidal foraging habitats for nonbreedingshorebirds (Suborder Charadrii). Little is known, however, about factors influencingbird use of pastures.Hence, we examined the spatialdistribution of shorebirdsin coastalpastures near Humboldt Bay, California from October 1991-May 1992 and correlatedspatial distributionpatterns with environmentaland habitat variables. Shorebirdsused pasturesin a nonrandom (clumped) fashion. Pastureuse variedseasonally for Killdeer (Charadrius vociferus), Marbled Godwit (Limosa fedoa), and Common Snipe (Gallinago gallinago); Dunlin (Calidris alpina) and Black-bellied Plover (Pluvialis squatarola) use increased when it rained, and Dunlin, Black-bellied Plover, and Killdeer use increasednearer the time of a new moon. At the level of the individual pasture, the likelihood of encounteringmost species(5 of 6) increasedas vegetationheight decreased. The likelihood of pastureuse by Dunlin and Killdeer increased when shorebirdsused pastures the previous week. Extensive, nonrandom use of coastal pastures by nonbreeding shorebirdsindicates that conservationplanning for shorebirdsshould consider habitat characteristics. In the vicinity of Humboldt Bay and other coastal bays, use of pastures by shorebirdscan be increasedby practicesthat provide short vegetation. Although shorebird use did not correlate with use of pasturesby cattle, grazing by livestock is probably the means by which to achieve habitat characteristicsattractive to shorebirdswhile maintaining compatible human uses on private lands. Regular use of some pasturesby shorebirdsindicates that site faithfulnessmay be important and that “traditional” sitesneed to be identified and protected. Kev words: shorebirds. Charadrii. oastures, spatial distribution, logistic regression, bird-habitat relationships, Vegetation height.

INTRODUCTION lance is facilitated by short vegetation. It also The value of coastal pastures to nonbreeding has been suggestedthat some waterbirds asso- waterbirds is well known (Townshend 1981, ciate with livestock because of foraging benefits Goss-Custard and Durell 1983, Colwell and (Thompson et al. 1982, Colwell and Dodd Dodd 1995), especially in areas where intertidal 1995). For some waterbirds, pasture use has habitats have been mostly “reclaimed” by hu- been shown to increase when standing water is mans. In coastal areas, loss of important inter- present (Colwell and Dodd 1995). However, less tidal foraging habitats has been partially offset is known about environmental conditions and by agricultural land use practices, which provide habitat characteristics associated with pasture habitats for waterbirds, especially shorebirds use by shorebirds. (suborder Charadrii). Although the importance Elsewhere, we showed that a waterbird com- of pasturelands to nonbreeding shorebirds is munity numerically dominated by shorebirds widely recognized, little is known about envi- had greater species richness and densities of ronmental and habitat features associated with most species in grazed pastures with short veg- shorebird distributions in these habitats. Several etation (Colwell and Dodd 1995). However, wa- studies have demonstrated that nonbreeding terbird species did not respond uniformly to shorebirds frequent short-vegetation pastures short vegetation. Here, we expand our analysis (Fuller and Youngman 1979, Milson et al. 1985, of variables influencing shorebird use of pas- Colwell and Dodd 1995), presumably because tures to encompass environmental conditions prey are easier to capture and/or predator vigi- prevailing at the time of use, as well as habitat characteristicsof pastures themselves. Our ob- jectives are to: (1) examine the spatial distribu- I Received29 July 1996.Accepted 17 January1997. 2 Current address:South Carolina Departmentof tion of shorebirds within coastal pastures of Natural Resources,Samworth WMA, 420 Dirleton northern California adjacent to important non- Rd., Georgetown,SC 29440. breeding habitats (Colwell 1994), and (2) eval-

[3371 338 MARK A. COLWELL AND SARAH L. DODD

- I \

ACIFIC‘ : ICEAN

(

0 N

FIGURE 1. Locationof 20 pastureplots within the largerstudy area in northernCalifornia. uate the relative contributions of variables ex- these study plots because cooperative landown- plaining nomandom shorebird distributions. We ers permitted access for research; hence, the include in our analysis variables representing study plots are not a random sample and may site faithfulness of birds to pastures in order to not be representative of local pasture habitats. understand whether or not birds use pastures Colwell and Dodd (1995) provide a complete consistently, at least over short periods of time. description of the area. Finally, we discussconservation implications of our findings, focusing on bird-habitat relation- BIRD SURVEYS ships that can be manipulated. Each week from 1 October 1991 to 1 May 1992, we collected data on shorebird distribution with- METHODS in pastures in two ways. First, during weekly STUDY AREA surveys (see below), we recorded opportunisti- We studied shorebirdsin coastal pastureslocated cally the location (t 100 m) and number of between the Mad River Slough of Humboldt shorebirds observed on the ground. Second, on Bay and the Mad River estuary in northern Cal- each of the 20 pasture plots we surveyed shore- ifornia (40”5’ N, 124”O’ W, Fig. 1). Much of this birds weekly by scan sampling (Altmann 1974) area was coastal salt marsh, but humans con- during daylight hours, beginning surveys within verted intertidal habitats to pastures nearly a 2 hours of mean high tide for Humboldt Bay. century ago (Haynes 1986); these pastures are Compared with intertidal habitats (Colwell currently grazed year-round by cattle (Bos tuu- 1993, 1994), pasturesappear to be supplemental TUS).Within the larger study area defined by Fig- feeding sites used by shorebirds during high ure 1, we examined shorebird use of 20 pastures tides (Hoff 1979), especially during winter. We varying in size from 2.3-13.6 ha. We selected used 8.5 X 44 mm binoculars and a 22X spot- SHOREBIRD USE OF PASTURES 339 ting scope to survey each pasture from the same location 1 m outside a perimeter fence, although on some occasions we surveyed at greater distances if we thought birds would be frightened by our presence. Surveys lasted the time re- quired to scan pastures and record data on species, their abundances, and behaviors (not pre- sented in this article). After each survey, we walked a zig-zag path through each pasture to flush unseen birds (0.7% of all bird detections), which may have been hidden by vegetation.

HABITAT AND ENVIRONMENTAL MEASUREMENTS We analyzed variables potentially influencing shorebird use of pastures at two levels-environmental and habitat (pasture) characteristics (see below). At both levels of analysis, variables varied temporally and spatially in a similar man- FIGURE 2. Perspective plot showing spatial varia- ner. tion in shorebird distribution across the larger study area. Grid cells are approximately 150 X 150 m. Data Environmental variables. Prior to each sur- representsummed observations for the 1 October 1991 vey, we recorded the following environmental to 1 May 1992 study period. variables: (1) maximum height of high tide co- incident with the survey based on National Oce- anic and Atmospheric Administration (NOAA) birds could penetrate the soil with their bills us- data (Tidelogs 1991, 1992), (2) time of day (sun- ing a device similar to that described by Myers rise-10:00; lO:Ol-14:OO; 14:01-sunset), (3) et al. (1980; see Colwell and Dodd 1995 for de- week of the study (1 October = week 1,27 April tails). We summarized habitat data by calculat- = week 29), (4) moon phase (new, quarter, half, ing means of 10 random samplesfrom each pas- three quarters, full; ignoring waxing and waning ture visit, yielding a single data point for each conditions) based on NOAA data (Tidelogs pasture per week. 1991, 1992), (5) wind speed estimated in km hri, (6) wind direction measured in compass DATA SUMMARY AND ANALYSIS bearing, (7) temperature (“C), and (8) whether or We examined the spatial distribution of shore- not it rained. Finally, we recorded the amount of birds in two ways. First, we collated shorebird precipitation in the 24-hr period preceding the locations (x, y coordinates) for the entire study survey date based on data collected approxi- period and graphically portrayed the distribution mately 2 km south of the study area (B. Alden, (Fig. 2) using geostatisticalmethods (Rossi et al. pers. comm.). 1992). Second, we summarized the average (2 Pasture characteristics. At the start of a sur- SD) density, as well as the prevalence (propor- vey, we tallied the number of cattle using a pas- tion of observations; n = 29 weeks) of shore- ture. To assessthe importance of shorebird site birds in each pasture (Fig. 3). faithfulness to pasture use, we also noted wheth- We used logistic regression(Glantz and Slink- er or not shorebirdshad used a pasture the pre- er 1990, Trexler and Travis 1993) to examine vious week. After each bird survey, we quanti- correlatesof shorebirduse (presenceor absence) fied vegetation height, water depth, and soil pen- of pasturesbecause most (83%) surveys yielded etrability at 10 random locations within each no shorebirds, and multiple individuals tended pasture. We measuredvegetation height and wa- to occur in pastures.We based analyses on 580 ter depth to the nearest cm by dropping a sam- observations (20 pastures surveyed each of 29 pling rod (3 mm diam.) vertically into the veg- weeks), although sample sizes for analyses of etation and recording maximum height at which environmental (n = 571) and pasture (n = 560) vegetation and water contacted the rod (Colwell correlates are smaller because of missing data and Dodd 1995). We gauged the degree to which points. Shorebirds move to pastures from inter- 340 MARK A. COLWELL AND SARAH L. DODD

2.5 Density (per ha) W ';: 2 2.0 Prevalence q .I

z 6 1.5

3 13 6 20 2 16 1 19 5 4 7 6 9 11 10 12 14 15 16 17 Pasture FIGURE 3. Average (2 SD) shorebird density and prevalence (percent of 29 weekly observationsin which shorebirdsused pastures)for pastureplots. Prevalence scale (0.0-1.0) is embedded in the densityscale; n = 29 for all datapoints. Pasture numbers correspond to Figure 1. tidal habitats at least twice daily in response to the 20 pasture plots. Of the latter sample, 95% rising (semi-diurnal) tides (Tidelogs 199 1, of observations (ranked in order of decreasing 1992), hence we considered each survey of a density; Table 1) consisted of Dunlin (Culidris pasture to represent an independent “choice” of alpina; 54%), Black-bellied Plover (Pluvialis habitat by birds. Additionally, to assuretemporal squaturolu; 15%), Killdeer (Charadrius vocifer- independence of survey data collected at one us; 13%), Marbled Godwit (Limosafedoa; 7%), week intervals, we screened the 29 weekly ob- Long-billed and Short-billed Dowitchers (Em- servations for each pasture using time series nodromus scolopaceus and L. griseus, respec- analysis (Hinze 1995), and found little evidence tively; 6%), and Common Snipe (Gallinago gal- of autocorrelation (data available from senior linago; 2%). Speciesdensities varied from 0.04- author at mac3Qaxe.humboldt.edu). 1.7 birds per ha. Most (78%) shorebirds ob- We used stepwise forward logistic regression served in pastures foraged (Colwell and Dodd (Harrell 1986) to examine the relative impor- 1995). tance of variables to pasture use by shorebirds, Shorebirds occurred on 17% of surveys, but including only those variables explaining a sig- we encountered individual species on l-8% of nificant (P < 0.10) amount of variation. We min- occasions(Table 1). Species’ occurrencesin pas- imized spurious correlations with dependent ture plots (prevalence) varied widely: Killdeer variables by stopping model building when occurred in 55% of pastures, whereas dowitch- x*(Wald)-statistic for testing all variables jointly ers used 15% of pastures. (residual x2) had a P-value exceeding 0.10 (Har- rell 1986). We used the fraction of concordant SPATIAL DISTRIBUTION OF SHOREBIRDS pairs of observations, including one-half the number of tied pairs, to evaluate each logistic Shorebirds varied greatly in their use of pas- regression model, which is considered more re- tures. Over the entire study area, we recorded liable than a classification table for assessingthe considerable variation in shorebird distribution predictive ability of the model (Harrell 1986). (Fig. 2). On the four most heavily used areas, total counts exceeded 3,000 (n = 3) and 2,000 RESULTS shorebirds (n = 1); many areas had little or no Within the larger study area, we recorded 28,099 shorebird use. A nonrandom (clumped) spatial shorebird sightings, 8,314 of which occurred on distribution was obtained for the 20 pastures as SHOREBIRD USE OF PASTURES 341

TABLE 1. Results of stepwise forward logistic regressionexamining the relationshipbetween pastureuse by shorebird speciesand environmentaland habitat variables.

Scale of Vanables Beta x’ (Wald) Concord- Speaes (% use) analyris selected coeffiaent SE statistic P ant pairs Killdeer (8) Environmental Week -0.09 0.02 20.44 0.001 0.71 Moon phase -0.38 0.15 5.90 0.015 Pasture Previous use 1.98 0.38 27.45 0.001 0.90 Vegetation height -0.33 0.08 22.38 0.001 Dunlin (4) Environmental Rain during survey 1.37 0.49 7.78 0.005 0.58 Moon phase -0.53 0.25 4.61 0.032 Pasture Previous use 1.38 0.49 8.02 0.005 0.82 Vegetation height -0.25 0.08 8.56 0.003 Common Snipe (3) Environmental Week -0.11 0.03 10.21 0.001 0.71 Pasture Vegetation height -0.24 0.08 8.75 0.003 0.73 Black-bellied Plover (3) Environmental Rain during survey 1.77 0.56 9.90 0.002 0.66 Moon phase -0.88 0.32 7.47 0.006 Pasture Vegetation height -0.56 0.14 16.43 0.001 0.84 Marbled Godwit (2) Environmental Week 0.15 0.06 6.79 0.009 0.77 Dowitchers (1) Pasture Vegetation height -0.33 0.16 4.11 0.04 0.80

dThe percentage of surveys (n = 580) in which we observed the specm using pastures. well (Fig. 3). Shorebird densities ranged from where vegetation was less than 10 cm. Finally, O-20.9 birds per ha and shorebirds occurred on the likelihood of pastureuse by Dunlin and Kill- O-62% of surveys. deer increased when shorebirds used a pasture We detected similar distributions of bird ob- the previous week. servations for the larger study area and pasture plots (Kolmogorov-Smimov test, D = 0.16, P > DISCUSSION 0.05); most observations (67% of study area and The nonrandom distribution of shorebirds in 83% of pasture plots) yielded no shorebirds. In coastal pasturelands of northern California (see those cases in which shorebirds occurred, 75% also Colwell and Dodd 1995) appears to be in- of observations within the larger study area had fluenced by a small number of variables oper- > 10 birds, whereas 47% of pasture plot obser- ating at several spatial and temporal scales. vations had > 10 birds. These variables parallel the series of choices proposedin hierarchical models of avian habitat CORRELATES OF PASTURE USE selection (Svardson 1949, HildCn 1965, Wiens Environmental variables. Pasture use by shore- 1985) in which an individual’s choice of habitats birds correlated with three environmental vari- is influenced by geographical position and hab- ables (Table 1): (1) rain during a survey, (2) itat configurations, modified by interspecific in- phase of moon, and (3) week of study. Dunlin teractions and site fidelity. and Black-bellied Plovers were more likely to Elsewhere, we reported that landscape effects use pastureswhen it rained during a survey. The were relatively unimportant with none of the five likelihood of pasture use by Dunlin, Black-bel- shorebird species showing a preference (P > lied Plover, and Killdeer increasednearer in time 0.15; stepwise regression of average species’ to the new moon. From October to May, pasture densities) for pasturesnear the bay, estuary, and use decreasedby Killdeer and Common Snipe, beaches(Colwell and Dodd 1995), which are the but Marbled Godwit use increased. main foraging areas for most shorebirds (Col- Pasture characteristics.Two characteristicsof well 1994). Movements of shorebirds from in- pastures contributed significantly to pasture use tertidal areas into pastures prompted by rising by shorebirds (Table 1). For five of six species tides can be observed almost daily (Colwell, un- (not Marbled Godwit), use varied inversely with publ. data). However, the conclusion that land- vegetation height. All shorebirdobservations oc- scapeposition does not influence pasture use by curred in pastures with vegetation less than 20 shorebirds is tempered by the restricted spatial cm; 93% occurred in pastures with vegetation distribution of pastures we sampled-all pas- less than 15 cm, and 82% occurred in pastures tures were within approximately 3 km of inter- 342 MARK A. COLWELL AND SARAH L. DODD tidal habitats. In the vicinity of Humboldt Bay, duced access of birds to polychaete prey inter- however, some pastures lie at distances greater tidally and increased the availability of earth- than 5 km from intertidal areas.We strongly sus- worms as prey in pastures (Townshend 1981). pect that inclusion of pasturesmore distant from We found similar patterns in coastal northern intertidal areas than the 20 plots we studied California. Both seasonal and daily use of pas- would have yielded pronounced landscape ef- tures by shorebirdscorrelated with precipitation. fects. With winter rains, shorebird use of pasturesin- Three environmental variables (week of the creased in association with decreasedavailabil- study period, moon phase, and rain at the time ity of prey in intertidal habitats (Carrin 1973, of a survey) correlated with pasture use by Colwell 1993) and increased prey availability in shorebirds. Three species varied seasonally in pastures(Colwell, unpubl. data). Moreover, pas- their use of pasturesbut patterns differed among ture use by Dunlins and Black-bellied Plovers species. From October to May, pasture use by increased when it was raining, but not in asso- Killdeer and Common Snipe decreased,whereas ciation with the amount of rain in the day prior Marbled Godwit use increased. Seasonal to a survey. Precipitation presumably influences changes in pasture use (White and Harris 1966, pasture use by either increasing the availability Hoff 1979, Colwell and Dodd 1995) are proba- of prey in pasturesand/or decreasingprey avail- bly associated with migration phenology (Hoff ability in intertidal habitats (Evans 1976, Town- 1979), as well as variation in profitability of for- shend 1981). aging in intertidal and pasture habitats (Town- Regression models consistently identified shend 1981, Colwell, unpubl. data). vegetation height as being strongly associated Pasture use correlated inversely with moon with pasture use by shorebirds-short vegetation phase for three species: Dunlin, Black-bellied increased the likelihood of pasture use. For all Plover, and Killdeer occurrence increased at shorebirds (except Marbled Godwits), pasture times nearer the new moon. Our results, based use increased significantly as vegetation height on diurnal observations, are similar to those re- decreased. These results corroborate findings ported by Milsom (1984) who showedLapwings based on coarse-scale analyses of pasture char- (Vanellus vanellus) foraged more during day- acteristics averaged over the 29-week study pe- light hours when more time elapsed to the next riod (Colwell and Dodd 1995). We suspectthat full moon. The diurnal relationship between pastureswith short vegetation are used by shore- moon phase and pasture use differs from studies birds because prey availability is greater in less of nocturnal foraging in intertidal habitats where structured vegetation (Eiserer 1980), and/or de- shorebird occurrence correlates positively with tection of predatorsis enhanced in open habitats moon phase (Robert et al. 1989, Swennen 1990, (Colwell and Dodd 1995). The second of these Zwarts et al. 1990). Shorebirdsmay forage more two nonmutually exclusive explanations is prob- on nights nearer the full moon because prey are ably most applicable to small shorebirds,which more active and/or more visible compared with are more affected by vegetation height than larg- nights of low illumination (Pienkowski et al. er species.At a given vegetation height (i.e., 10 1984). The inverse relationship between diurnal cm), smaller birds would have their vision ob- pasture use and moon phase may arise owing to structed by vegetation to a greater extent than related changesin the availability of prey in pas- larger birds. Interestingly, the one exception to tures and/or intertidal habitats. the vegetation height-pasture use relationship is Precipitation has been shown to influence pas- the largest species, Marbled Godwit. ture use by a variety of waterbirds (Reed et al. Our analyses indicate that pasture use by two 1977, Hirst and Easthope 1981), but mostly of the most abundant species-Dunlin and Kill- through its effects on habitat characteristics,es- deer-correlated strongly with previous pasture pecially standing water. In intertidal habitats, use. These results, together with the clumped precipitation reduces prey availability (Goss- spatial distribution of shorebirds(Fig. 2) and ob- Custard 1984, Pienkowski et al. 1984), but it servations of similar numbers of a speciesusing may increase prey availability in pastures. For the same pasture repeatedly (Colwell, unpubl. example, Curlews (Numenius arquata) used pas- data), suggest that pasture use by shorebirds is tures adjacent to the Tees estuary in England “traditional,” at least over short periods of time. when cold temperatures and winter rains re- Furthermore, we often observed shorebirdsleav- SHOREBIRD USE OF PASTURES 343 ing intertidal habitats and repeatedly moving dicates the need to identify and protect these into the same pastures. Similar patterns of lo- habitats. calized and predictable shorebirddistributions in Finally, the strong relationship between veg- grasslands of southern England led Fuller and etation height and shorebird use of pasturespro- Youngman (1979) to suggestthat Golden Plov- vides useful information to integrate shorebird ers (&vi&is apricaria) exhibit strong site fi- conservation with management of lands to ben- delity over periods of several years. efit livestock and dairy interests. However, given Elsewhere, we examined bird use of these the narrow taxonomic focus of this paper, we do same 20 pasturesbased on an analysis of a more not necessarily wholeheartedly endorse grazing diverse waterbird community including water- as a land use practice fully compatible with oth- fowl, wading birds, gulls and shorebirds (Col- er conservation interests (e.g., raptors, wading well and Dodd 1995). Results presentedhere of- birds). On the other hand, we wish to emphasize fer broad support for conclusions of our earlier the conservation value of land use practices as- paper, especially the strong inverse relationship sociated with dairy and livestock industries in between vegetation height and shorebird use of coastal habitats. We argue that these practices pastures. are infinitely more desirable than other land use options (e.g., housing development) that cause CONSERVATION IMPLICATIONS greater, irrevocable damage to natural and semi- natural landscapes. Like other large bays and estuariesalong the Pa- cific coast of North America, Humboldt Bay is ACKNOWLEDGMENTS considered critical habitat for migrating and wintering shorebirds (Senner and Howe 1984, We thank R. Collenburg, E Laurence, L. Laursen, and C. Pifferini for allowing accessto their lands, B. Alden Colwell 1994). Pastures adjacent to many of for weather data, and the Nielson Foundation, Hum- these intertidal areasprovide foraging habitat for boldt State University Department of Wildlife, and waterbirds (Colwell and Dodd 1995), especially College of Natural Resourcesand Sciences for finan- during the nonbreeding season when intertidal cial and logistical support.B. Harrington, B. Kus, W. Koenig, and several anonymous reviewers provided prey abundance declines (Canin 1973, Colwell helpful comments on the manuscript. 1993) and energetic demands of over-wintering increase at temperate latitudes (Evans 1976). LITERATURE CITED Nevertheless, the value and importance of these ALTMANN,J. 1974. Observationalstudy of behaviour: habitats to nonbreeding waterbirds remains sampling methods. Behaviour 49~227-267. largely unknown. CARRIN,L. E 1973. Availability of invertebrates as Nonrandom use of coastal pastures by non- shorebirdfood on a Humboldt Bay mudflat. M.Sc. thesis, Humboldt State Univ., Arcata, CA. breeding shorebirds and bird-habitat relation- COLWELL,M. A. 1993. Shorebird community patterns ships provide the basis for conservation action in a seasonallydynamic estuary. Condor 95: 104- involving local, regional and national agencies 113. working with private landowners adjacent to in- COLWELL,M. A. 1994. Shorebirdsof Humboldt Bay, California: abundanceestimates and conservation tertidal habitats. At Humboldt Bay, the conser- implications. West. Birds 25: 137-145. vation value of pastures for shorebirds varies COLWELL,M. A., AND S. L. DODD. 1995. Waterbird spatially, and the effectiveness of habitat manip- communities and habitat relationships in coastal ulations to benefit shorebirds and other water- pasturesof northern California. Conserv. Biol. 9: 827-834. birds may vary acrossthe landscape. But, land- EISERER,L. A. 1980. Effects of grasslength and mow- scape effects require further study. We suspect ing on foraging behavior of the American Robin that in other coastal areas of the world with im- (Turdus migratorius). Auk 97:576-580. portant shorebirdwintering habitats and adjacent EVANS,I? R. 1976. Energy balance and optimal foraging strategiesin shorebirds:some implications pasturelands a cursory knowledge of spatial for their distributionsand movements in the non- movements of shorebirds may enhance local breeding season.Ardea 64: 117-139 conservation efforts. Our analyses indicate that FULLER,R. J., AND R. E. YOUNGMAN.1979. The uti- some shorebird speciesuse pasturesconsistently lization of farmland by Golden Plovers wintering in southernEngland. Bird Study 26:37-46. (at least over short periods of time), suggesting GLANTZ,S. A., AND B. K. SLINKER.1990. Primer of site-faithfulness is an important conservation applied regression and analysis of variance. Mc- consideration. “Traditional” use of pastures in- Graw-Hill, New York. 344 MARK A. COLWELL AND SARAH L. DODD

GOSS-CUSTARD,J. D. 1984. Intake rate and food sup- in the St. Lawrence Valley, Quebec. J. Appl. Ecol. ply in migrating and wintering shorebirds,p. 233- 14:667-680. 270. In J. Burger and B. L. Olla [eds.], Shorebirds: ROBERT,M., R. MCNEIL, AND A. LEDUC. 1989. Con- migration and foraging behavior. Plenum Press, ditions and significanceof night feeding in shore- New York. birds and other waterbirds in a tropical lagoon. GOSS-CUSTARD,J. D., AND S. E. A. LE V. DIT DURELL. Auk 106:94-101. 1983. Individual and age differences in the feed- ROSSI,R. E., D. J. MULLA, A. G. JOURNEL,AND E. H. ing ecology of OystercatchersHaematopus os- FRANZ. 1992. Geostatistical tools for modeling tralegus wintering on the Exe Estuary, Devon. and interpreting ecological spatial dependence. Ibis 125:155-171. Ecol. Monogr. 621277-314. HARRELL,E E., JR. 1986. The LOGIST procedure,p. SENNER,S. E., A&D M. A. HOWE. 1984. Conservation D. 269-294. In R. I? Hastings Ted.1._I SUGI sutmle- of nearctic shorebirds. 379-421. In J. Burger y - II mental library user’s guide, version 5. SAS lnsti- and B. L. Olla [eds.], Shorebirds:breeding behav- tute, Inc., Cary, NC. ior and populations.Plenum Press, New York. HAYNES,C. S. 1986. The Arcata Bottoms: flooding on SVARDSON,G. 1949. Competition and habitat selection a changing landscape. M.A. thesis, Humboldt in birds. Oikos 1:157-174. State Univ., Arcata, CA. SWENNEN,C. 1990. Oystercatchersfeeding on giant HILDBN,0. 1965. Habitat selection in birds. Ann. bloody cockles on the Bane d’ Arguin, Maurita- Zool. Fennica 2:53-75. nia. Ardea 78:53-62. HINZE,J. L. 1995. Number cruncherstatistical system. THOMPSON,C. I?, S. M. LANYON, AND K. M. THOMPSON. Kaysville, UT [publishedby author] 1982. The influence of foraging benefits on as- HIRST,S. M., AND C. A. EASTHOPE.1981. Use of ag- sociationof Cattle Egrets (Bubulcus ibis) with cat- ricultural landsby waterfowl in southwesternBrit- tle. Oecologia 52: 167-l 70. ish Columbia. J. Wildl. Manage. 45:454-462. TIDELOG. 1991, 1992. Northern California. Pacific HOFF,C. J. 1979. Bird use of agriculturallands around Publishers,Bolinas, CA. North Humboldt Bay, California. M.Sc. thesis, TOWNSHEND,J. D. 1981. The importanceof field feed- Humboldt State Univ., Arcata, CA. ing to the survival of wintering male and female MILSOM,T. I? 1984. Diurnal behaviourof lapwings in Curlews Numenius arquatu on the Tees estuary, relation to moon phase during winter. Bird Study p. 261-274. In N. V. Jonesand W. J. Wolff [eds.], 31:117-120. Feeding and survival strategies of estuarine or- MYERS,J. P, S. L. WILLIAMS,AND E A. PITELKA.1980. ganisms. Plenum Press, New York. An experimental analysis of prey availability for TREXLER,J. C., AND J. TRAVIS. 1993. Nontraditional sanderlings (Aves: Scolopacidae) feeding on regressionanalyses. Ecology 74: 1629-1637. sandy beach crustaceans.Can. J. Zool. 58:1564- WHITE, M., AND S. W. HARRIS. 1966. Winter occur- 1574. rence, foods, and habitat use of snipe in north- PIENKOWSKI,M. W., P. N. FERNS,N. C. DAVIDSON,AND western California. J. Wild]. Manage. 30:23-34. D. H. WORRALL. 1984. Balancing the budget: WIENS,J. A. 1985. Habitat selection in variable en- measuring the energy intake and requirementsof vironments:shrub-steppe birds, p. 227-25 1. In M. shorebirdsin the field, p. 29-56. In F! R. Evans, L. Cody [ed.], Habitat selectionin birds. Academ- J. D. Goss-Custard,and W. G. Hale [eds.], Coastal ic Press, Orlando, FL. waders and wildfowl in winter. Cambridge Univ. ZWARTS.L.. A. BLOMERT.AND R. HUPKES.1990. ln- Press, Cambridge. crease of feeding time in waders preparing for REED,A., G. CHAPDELAINE,AND P. DUBLJIS.1977. Use spring migration from the Bane d’ Arguin, Mau- of farmland in spring by migrating Canada Geese ritania. Ardea 78~237-256.