Age-Related Mortality in a Wintering Population of Dunlin
Total Page:16
File Type:pdf, Size:1020Kb
AGE-RELATED MORTALITY IN A WINTERING POPULATION OF DUNLIN BARBARAE. KUS) PHILIP ASHMAN, 2 GARY W. PAGE,3 AND LYNNE E. STENZEL 3 •Divisionof EnvironmentalStudies, University of California,Davis, California 95616 LISA; 21242East Dayton Street,Madison, Wisconsin,53703 USA; and 3PointReyes Bird Observatory, 4990 ShorelineHighway, Stinson Beach, California 94970 USA ABSTRACT.--Despiteconsiderable evidence that juvenile shorebirdsexperience signifi- cantly higher annual mortality rates than adults, identification and quantificationof the sourcesof mortalityhave received little attention.We found that the proportionof juvenile Dunlins (Calidrisalpina) in the kills of a Merlin (Falcocolumbarius) one winter at Bolinas Lagoon,California was greaterthan the proportionof juvenilesin the lagoon'swinter pop- ulation.This is evidencethat raptorpredation may be one of the factorscontributing to the age differencesin annual mortality ratesof shorebirds.We suggestthat the greatervulner- ability of juvenilesto predationby the Merlin may be causedby age-relateddifferences in Dunlin flocking behavior. Received23 March 1983,accepted 6 September1983. STUDIESof the populationdynamics of shore- Lagoon, reducing the population by as much birds (Charadrii) indicate that, in general, ju- as 21%over a 5-month period (Pageand Whir- veniles experience higher annual mortality acre 1975). Northern Harriers (Circuscyaneus), rates than adults (Goss-Custard1980). This dif- AmericanKestrels (Falco sparverius), Short-eared ferential mortality hasbeen reportedfor a wide Owls (Asiofiammeus), Peregrine Falcons(Falco range of European(Boyd 1962 and references peregrinus),and Merlins (Falcocolumbarius) are therein) and North American species (e.g. among the more frequent predators of small Holmes 1966, Myers 1980,Page et al. 1983).Al- shorebirds at the lagoon, and attacks by Red- though a difference in survivorship between tailed Hawks (Buteo jamaicensis), Cooper's age-classesappears to be widespread among Hawks (Accipitercooperii), and Great Horned shorebirdpopulations, little attention hasbeen Owls (Bubovirginianus) also occur occasionally. directed toward identifying and quantifying We confined our analysisof the effect of pred- specific sourcesof mortality and evaluating ators on mortality rates to an examination of their contribution to this pattern. Heavy mor- Dunlins killed by a female Merlin present at tality among first-year birds is generally the lagoon during the fall and winter of 1979- thought to be the result of their relative inex- 1980. Merlins are the most significant of the perience in dealing with such selectivepres- Dunlins' diurnal predators at Bolinas Lagoon; suresas predation, feeding efficiency,and ex- during one winter (1972-1973) a single Merlin tremes in environmental conditions (see was responsiblefor approximatelyhalf of the referencesabove); these hypotheseshave not observedpredation on Dunlins by all raptors been tested empirically in shorebirds,how- (Page and Whiracre 1975). Unlike many other ever. shorebird predators at the lagoon, the Merlin Here, we report the results of a study in we studied was relatively predictable in her which we attempted to determine the extent to hunting patterns. This enabled us regularly to which predatorsare responsiblefor age-related observeher hunting and eating sandpipersand differencesin the mortality rates of Dunlins to collect the remains from a large portion of (Calidrisalpina). Specifically, we comparethe age her kills. We were therefore able to base our ratio of Dunlins in predator kills with the age examination of age-related mortality rates on ratio of Dunlins in a population wintering at direct observation of the effect of a single, BolinasLagoon, California in order to evaluate known predator. whether or not the mortality rate due to pre- dation is higher for juvenilesthan for adults. STUDY SITE AND METHODS Predationby raptorsis one of the majorcaus- Studysite.--Bolinas Lagoon is a shallow 570-ha es- es of winter mortality for Dunlins at Bolinas tuary on the central California coast,approximately 69 The Auk 101: 69-73. January 1984 70 Ku$ •T ^•. [Auk, Vol. 101 24 km northeastof San Francisco.This study area and 3000- its shorebirdpopulation have been described by Page et al. (1979).It is a majorover-wintering site for large numbersof shorebirdsthat forage on the expansive mudflats.Dunlins are presentat the lagoonfrom late Septemberthrough early May. During 1979-1980, Dunlin numbersincreased to approximately2,000 by November and remained at this level until the onset of migration in early April (Fig. 1). Age-compositionof the population.--We trapped Dun- linsat approximately2-week intervals throughout the winter in order to determine the ratio of juveniles to adults in the BolinasLagoon population. For each period,our estimateof the proportionof juvenilesin the lagoonpopulation is derivedthrough direct ex- trapolationfrom the age ratio of the corresponding sampleof trappedbirds. We capturedbirds by two OCT NOV DEC JAN FEB MAR APR MAY methods and comparedthe results to evaluate the possibilitythat our samplingtechnique, and there- MONTH fore our populationestimates, were biasedtoward a particularage-class. Although it is difficult to deter- Fig. 1. Number of Dunlins at BolinasLagoon, Oc- mine the nature and extent of biases associated with tober 1979-May 1980. differenttrapping techniques, there is someevidence (Pienkowskiand Dick 1976,Goss-Custard et al. 1981) that the compositionof catchesmade by mist-nets, which we used extensively,may be biased in favor primary, the juvenile feather being more pointed, of juveniles.In Novemberand December,we caught worn, and narrower than the more bluntly tipped four sampleswith mist-netsand four with noose- feather of the adult. Because this characteristic is more mats. The mist-netswere positionedat dusk over variable than the first, we testedthe reliability of the channelsthat the Dunlins followed as they flew into primary-shapemethod by determiningthe age of the salt marsh to roost for the night. The noose-mats eachbird we capturedindependently by both criteria were hidden beneaththe sandduring the day in areas and comparingthe results.We found that we were where large flocksof Dunlins were feeding or roost- able to age birds by primary feather characteristics ing. Noose-mattraps have the advantageof elimi- alone with 97% accuracy.Although it was not nec- nating any age-relateddifferences in flying expertise, essaryfor aginglive birds,we relied heavilyon the which hasbeen suggestedas a reasonfor the poten- feather-shapetechnique for aging prey from feather tial bias associatedwith the use of mist-nets (Pien- remains, which often did not include the key coverts kowski and Dick 1976).We regressedthe proportion (see below). of juvenilesin eachtrapped sampleagainst date for Age compositionof Merlin kills.--A female Merlin eachmethod, compared the resultantlines, and found was observedhunting sandpipersat BolinasLagoon no statisticaldifferences in either the slopes(P = 0.29) between 12 October 1979 and 14 March 1980. The or the Y-intercepts(P = 0.18). While this evidence Merlin generallyspent the entire day on the lagoon, suggeststhat our estimatesof the agestructure of the restingor hunting birds on the mudflatsby launch- populationwere not differentiallybiased by the two ingattacks from low perches in theadjacent salt marsh. methods,it doesnot refute the possibilitythat both Followinga successfulhunt, shereturned to one of trapping techniquesmight be similarly biasedto- theseperches where she ate the entire prey except ward catchingjuveniles. However, any biasthat re- for the body feathersand parts of the wings, which suited in our overestimatingthe proportion of juve- were discardedat the baseof the perch.We checked niles in the population would only make the the Merlin's perches regularly and collected the experimenta conservativetest of our hypothesisby feather remains of any kills we found. We aged all making a selective preference by Merlins for juve- Dunlin kills in the manner describedabove, except niles more difficult to detect. for seven for which there were insufficient remains. Eachbird was banded with a unique color combi- Statisticalanalyses.--We used a multiple linear nation for individual recognitionand was aged by regressionmodel to quantify the proportionof ju- plumagecharacteristics. Birds were consideredto be veniles in the population and in the Merlin's kills immatureif they had buffy edgeson their innermost (dependentvariables) as a functionof period during tertials or inner middle-wing coverts and adult if the winter season(independent variable). Because our these feathers were white- or grey-tipped. Age can samplesof trappedbirds were largerin sizethan our also be determined by the shape of the outermost samplesof prey remains,we useda weightedleast January1984] DunlinMortality 71 1,00 0 POPULATION (N=218) ß oMERLIN KILLS (N=66) 0.50 0.00 I I I I I I I I [ I I I I I I I 28 OCT- I7 NOV- 7 DEC- 27 DEC- 16 JAN- 5 FEB- 25 FEB- 17MAR- 6NOV 26NOV 16DEC 5JAN 25JAN 14FEB 6MAR 26MAR PERIOD Fig. 2. Proportionof juvenile Dunlins in the populationand in the Merlin's diet at BolinasLagoon in winter, 1979-1980. squaresanalysis (Dixon 1981)to comparethe slopes Figure 2 illustrates the proportion of imma- and Y-intercepts of the two lines. This technique ture Dunlins in the total population (Y = weightseach point in the regressionaccording to its 75.19- 3.26x) and in the Merlin's diet (Y = associatedsample size and thus correctsfor any in- 86.67