Parental behaviour of a precocial species: implications for juvenile survival

Transcription

1 Journal of Appled Ecology 29, 46, Parental behavour of a precocal speces: mplcatons for juvenle survval Vctora J Dretz* Colorado Dvson of Wldlfe, 317 West Prospect Road, Fort Collns, CO 8526, USA do: 11111/j x Summary 1 Parents determne habtat selecton for precocal young by leadng ther young to foragng areas untl the chcks attan full ndependence There are potental benefts and costs to reproductve success assocated wth changng habtats whle carng for young Ths study nvestgated the relatonshp between dfferent types of habtats and ther qualty on chck survval and brood movements of a declnng upland shorebrd, the mountan plover Charadrus montanus 2 From 24 to 26, a total of 153 mountan plover broods were montored on the prmary breedng habtats n eastern Colorado, USA; two shortgrass prare habtats that were ether occuped or unoccuped by black-taled prare dogs Cynomys ludovcanus and agrcultural lands Habtat qualty hypotheses were tested usng newly developed statstcal applcatons to estmate survval of chcks and brood movement patterns 3 Chck survval and brood movements were nfluenced by habtat Chck survval over the 3-day brood-rearng perod was substantally hgher on nestng habtat of shortgrass occuped by prare dogs compared wth agrcultural land and shortgrass unoccuped by prare dogs The rate of brood movement away from shortgrass wth prare dogs was lower than shortgrass wthout prare dogs, but hgher than agrcultural lands for each year of the study 4 Ths study suggests that complex processes nfluence how dfferent habtats affect brood-rearng actvty of mountan plovers Even though broods moved off nestng habtat of shortgrass occuped by prare dogs, ths habtat had the hghest survval rate and s hghly mportant to mountan plover reproductve success 5 Synthess and applcatons In order to develop effectve conservaton strateges, the provson of adequate breedng habtat should nclude nformaton on patterns of habtat selecton for all stages of the breedng cycle, ncludng the nestng and dependent young perods From a conservaton perspectve, understandng the habtat use of young brds s crtcal when populaton dynamcs show great senstvty to survval of young Prevous studes on mountan plovers have suggested that nest success s smlar among shortgrass prare habtats and agrcultural lands Thus, conservaton measures that ncrease nest success may be neffectve for mountan plovers unless they are accompaned by measures promotng chck survval Key-words: brood movements, Charadrus montanus, chck survval, detecton probablty, habtat selecton, mountan plover, mult-strata, precocal speces, shorebrds Introducton Interacton between habtat-specfc demography and habtat selecton may be a crucal component n populaton dynamcs and regulaton The value of a habtat depends not only on the *Correspondence author E-mal: vctoradretz@statecous Re-use of ths artcle s permtted n accordance wth the Terms and Condtons set out at authorresources/onlneopenhtml current resources avalable and the ablty to provde resources for some tme n the future (Orans & Wttenberger 1991) but also the lfe stage of the ndvdual For example, n brds, the stage of the breedng cycle (e nest and fledgng) determnes the value of a habtat A habtat wth hgh nest survval s not always the same type of habtat supportng the hghest rate of fledgng success To complete the breedng effort successfully, habtats must support more brds as the breedng cycle advances to enhance the persstence of the populaton, decrease energetc costs of foragng, as well as reduce exposure to predators Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety

2 Juvenle survval of precocal speces 871 Habtat use of fledglng brds, and ts relatonshp to survval, s poorly understood for most avan speces From a conservaton perspectve, understandng the habtat use of young brds may be crtcal because populaton dynamcs often show great senstvty to survval of young (Anders et al 1997; Colwell et al 27), although ths depends on the lfe hstory of the speces Habtat selectvty of young precocal speces s usually determned by the attendng parent(s) that lead ther young to dfferent habtats untl they attan full ndependence For example, n many precocal shorebrds, the parent(s) lead ther chcks from the nest to dfferent habtats where chcks feed, grow and fledge (eg Lengyel 26; Colwell et al 27; Kosztolány et al 27) There are several potental benefts that young may gan n movng to dfferent habtats Frst, they can be better protected from predators Second, they may avod competton wth conspecfcs (Lengyel 26) Thrd, they may explot better foragng resources There are also costs such as ncreased chck mortalty durng extensve movements because of hgher rates of starvaton Furthermore, habtats wth more food resources may also have hgher predaton rsks Ths study used newly developed quanttatve methods to compare how habtat nfluenced survval of chcks and brood movement actvty of a declnng North Amercan upland shorebrd, the mountan plover Charadrus montanus Habtats ncluded the prmary mountan plover breedng habtat n eastern Colorado, USA; agrcultural land, shortgrass prare occuped by prare dogs and shortgrass prare unoccuped by prare dogs Chck survval was estmated wth an analytcal method that accounts for mperfect detecton of chcks Brood movements were estmated usng a mult-state model allowng for mssng values (or data) n the encounter hstores Habtat qualty hypotheses were tested based on the ecology of ths speces The dstrbuton of adult plovers across eastern Colorado durng the breedng season suggests that densty on agrcultural land s smlar to densty on shortgrass unoccuped by black-taled prare dogs Cynomys ludovcanus (hereafter referred to as grassland), but substantally lower than shortgrass occuped by black-taled prare dogs (hereafter referred to as prare dog) (Tpton et al 29) Mountan plover chck survval was hypotheszed to be the hghest on prare dog nest habtat Knopf & Rupert (1999) reported that two of three broods that nested on grassland moved to an agrcultural feld wthn 2 km of the nest ste By contrast, Dretz et al (25) suggested broods that nested on agrcultural land moved to dfferent habtats, and broods that nested on habtats other than agrcultural land stayed on the nest habtat Because studes on other shorebrd speces suggest that broods move to the habtat wth the hghest chck survval and stay at that habtat (Lengyel 26; Schekkerman & Bentema 27), I hypotheszed that broods that nested on prare dog habtat would tend to stay on ther nest habtat andbroodsthatnestedonotherhabtatswouldtendtomove to prare dog habtat Mountan plover chcks are ndfugous and feed themselves on a wde range of arthropods (Baldwn 1971; Knopf & Wunder 26) Foragng condtons for plover chcks may be affected by changes n food avalablty between habtats (Knopf 1998) No prevous study has compared plover prey resources (e arthropod populatons) on dfferent habtats Prey avalablty of each habtat durng the brood-rearng perod s hypotheszed to explan the observed patterns n chck survval and brood movement actvty between habtats Specfcally, plover prey densty and bomass on prare dog nest habtat s predcted to be hgher resultng n hgher chck survval and movement to prare dog habtat Explanatory varables relatng to ndvdual characterstcs and temporal varaton were ncluded to explore the relevance of habtat qualty n explanng patterns of chck survval and brood movements of mountan plovers The mountan plover breedng system s descrbed as double clutch wth two clutches lad per par per year; the frst clutch s tended by the male and the second by the female (Graul 1973) Unparental behavour contnues through the brood-rearng stage (Knopf & Wunder 26) Dnsmore & Knopf (25) found that chcks tended by females had hgher survval than those tended by males Therefore, the sex of the tendng adult was ncluded as a varable In addton, yearly effect was ncluded to explan temporal varablty such as weather condtons and dfferent groups of observers assstng each year of the study Ths work provdes new nformaton on the relatve roles played by dfferent habtats n a hghly fragmented landscape on the brood-rearng ecology of mountan plovers and wll help land managers develop conservaton agendas for ths declnng shorebrd Materals and methods FIELD SITE The study area covered 21 5 km 2 n eastern Colorado, USA (see Supportng Informaton Fg S1) on prvately owned lands Clmate s characterzed by low relatve humdty, abundant sunshne, large daly temperature range and low average annual precptaton (Æ35 m) that manly (7 8%) occurs durng the growng season, Aprl to September, largely from thunderstorm actvty The landscape s relatvely flat and domnated by pastures of shortgrass prare and dryland agrcultural felds The two shortgrass prare habtats were vegetated by low-growng buffalograss Buchloe dactylodes and blue grama Bouteloua gracls Shortgrass that was grazed to varyng degrees by domestc ungulates, prmarly cattle, was defned as grassland and that grazed by a natve herbvore, the black-taled prare dog, was defned as prare dog Agrcultural land was comprsed of felds of dryland crops, prmarly wheat Trtcum aestvum L and sorghum Sorghum bcolor L and fallow strps wth varyng structure of crop stubble Specfc feld stes were selected such that a habtat was juxtaposed <2 km from one of the other two habtats to permt brood movements between habtats Boundares dstngushng agrcultural felds from the other habtats were easy to determne vsually Defnng the boundares between prare dog and grassland was problematc When adjacent to grassland, prare dog habtat was defned as the prare dog colony plus a 4 km buffer based on smlartes n vegetatve structure (eg heght, amount of bare ground) The total area of each habtat ncluded n the study vared between years Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

3 872 V J Dretz but followed a general pattern of agrcultural land prare dog grassland FIELD DATA COLLECTION From Aprl to August 24 26, data were collected on mountan plover brood-rearng ecology Mountan plover chcks leave the nest wthn 3 h of the last egg hatchng, and by the end of ther frst day, chcks appear capable of catchng small arthropods (Graul 1973) Mountan plover chcks fledge at days posthatch (Graul 1975; Mller & Knopf 1993) However, the broodrearng perod was defned from hatch to 3 days post-hatch to avod potental premature fledgng of chcks Montored broods were from known nests located <2 km from one of the other two habtats of nterest, allowng potental movement between habtats Plover prey resources Ptfall traps were used to measure prey densty and bomass of surface-actve arthropods (Work et al 22) Prey samplng was n known areas of plover nestng actvty at the same nne stes (three replcates per the three habtats) each year Samplng stes were selected at random but restrcted to one ste per habtat patch A trappng lne transect grd (Lukacs et al 24b) of 2 1 m 2 wth 6 ptfall traps at Æ5-m ntervals (see Supportng Informaton Fg S2) was nstalled at each ste Indvdual ptfall traps were placed <Æ1 m of the target locaton Ptfall traps were 16-oz Solo Ò plastc cups (95 mm dameter, 115 mm depth) that were bured n the sol wth the top of the cup flush wth the sol s surface The nne grds were nstalled 1 week pror to feld samplng Durng samplng, the traps were flled wth 2 ml of water and <1 ml of lqud dshwashng detergent to reduce the surface tenson of the water and prevent arthropods from escapng Samplng was conducted between the hours of sunrse (5: hours MST) and 11: hours MST because plovers are most effectve at foragng early n the mornng (Knopf & Wunder 26) The contents of each ptfall trap were dentfed to order and famly; sorted by lfe stage; and dred at 6 C for 48 h to obtan dry bomass of each trap to the nearest Æ1 mg The stomach contents from three mountan plover chcks yelded prey tems <15 mm n length (Baldwn 1971) Flyng arthropods were rare prey tems because chcks are not capable of flght untl days (Graul 1975; Mller & Knopf 1993) Hence, mountan plover chck prey tems (hereafter, smply prey tems) were defned as arthropods that were <15 mm n length and not capable of flght Montorng brood-rearng actvty Lghtweght 1Æ8 g rado transmtters (A245, Advanced Telemetry Systems, Inc) were ftted to nest-tendng adults 1 5 days pror to hatchng Adults wth nests <2 km from one of the other two habtats were captured wth a walk-n trap placed over the nest Transmtters were affxed by applyng a lght coatng of waterproof epoxy (e cyanoacrylate glue) and sldng t under the upper layer of mantle feathers so that the transmtters was postoned between feathers (Dretz et al 25) Transmtters were lkely to be shed durng moultng pror to autumn mgraton The battery lfe of transmtters was expected to be 56 days Addtonally, a feather sample was collected from each adult for sex determnaton usng DNA analyss After hatchng, adults were located every h to record ther locaton, habtat and the number of chcks present untl chcks were fledged at 3 days post-hatch Observatons were <3 mn, dependent on the age of chcks, wth less tme spent observng younger ndvduals due to ther vulnerablty STATISTICAL ANALYSIS Plover prey resources An approach based on dstance samplng wth trappng lne transects was used to calculate the densty of plover prey avalablty (Lukacs et al 24b) Trappng lne transects are passve n that ndvduals (or groups) are detected by enterng a trap whose dstance from the centre lne s known The detecton functon s estmated from nformaton obtaned on the perpendcular dstance from the centre lne to a trap off the centre lne Trap densty s the hghest on the centre lne and decreases wth ncreasng dstance away from the centre lne The densty estmator for trappng lne transects was: ^D ¼ n 2wL ^P a where n s the number of dfferent ndvduals trapped, w s the wdth of the trappng grd, L s the length of the grd and ^P a s the estmated probablty that an ndvdual s trapped gven area a The varance of the estmated densty was: cvar½ ^DŠ ¼ ^D 2 ½fcvðnÞg 2 þfcv½ ^P a Šg 2 Š: Program dstance 5 (Thomas et al 26) was used to estmate the prey densty Dstance samplng uses a set of sx models each composed of a key functon or general shape functon to ft the detecton functon, and a nonparametrc flexble form called a seres expanson that adjusts the key functon (Thomas et al 26) The number and arrangement of ptfall traps for the trappng lne transect grd (see Supportng Informaton Fg S2) was based on smulatons usng the hazard rate key functon wth a smple polynomal expanson seres model Ths same model was used to produce densty estmates of plover prey per 1 ha for each habtat per year of the study The dry weght of each ptfall trap was computed to obtan a measure of bomass Data were grouped across samplng days such that there were replcates (three grds sx or nne samplng days) per habtat for each year The mean and standard error was calculated for each habtat and year n sas proc means (SAS Insttute 23) Chck survval Chck survval was calculated usng a lkelhood-based extenson of the Cormack Jolly Seber (CJS) model (Cormack 1964; Jolly 1965; Seber 1965) n whch nformaton collected on the attendng adult and ts chcks s used to estmate survval of dependent young (Lukacs et al 24a) The approach requres that the attendng adult be unquely marked, and that the mark be read wthout error each tme the adult s re-sghted The young are not requred to have any form of mark The parameters of nterest nclude the probablty of survval of an ndvdual chck from tme to tme + 1gventhechcksalve at tme and remans on the study area, /, and the probablty that a chck wll be re-sghted gven t s alve and the adult was re-sghted at tme, p Because of all possble outcomes of / for a chck wthn a brood, / becomes a matrx, / : Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

4 Juvenle survval of precocal speces / m m / m 1 m m 1 ð1 / Þ / m 2 m 2 ð1 / Þ 2 m ð1 / Þ m / m 1 m 1 / m 2 m 1 m 2 ð1 / Þ ð1 / Þ m 1 / m 2 m 2 ð1 / Þ m where m s the maxmum number of chcks n a brood The rows represent the number of chcks alve at encounter occason and the columns represent the number of chcks alve at + 1 The possble outcomes of p are handled such that p s: 2 p m ; f h ¼ m ; m p m 1 m 1 ð1 p Þ p m 1 ; f h ¼ m 1;; The model set explctly examned f nest habtat or characterstcs of nest habtat nfluenced survval of chcks The sex of the adult was ncluded because t has been reported to nfluence chck survval of mountan plovers n other parts of the speces range (Dnsmore & Knopf 25) The effect of year was ncluded because chck survval for most avan speces exhbts annual varaton The model set also ncluded addtve or multplcatve combnatons of the dfferent factors Model selecton and nference was based on nformaton-theoretc methods and scored usng Akake s Informaton Crteron adjusted for small sample szes (AIC c ) (Burnham & Anderson 22) The models were ranked and compared n terms of ther ablty to explan varaton n the emprcal data usng DAIC c and AIC c weghts (Buckland et al 1997; Burnham & Anderson 22) The DAIC c for a gven model s the dfference n AIC c between the best approxmatng model and all other models (Burnham & Anderson 22) Further, to better nterpret the relatve lkelhood of a gven model over a set of models, models are normalzed (by summng to 1) to yeld a set of AIC c weghts Evdence ratos were computed based on the weght of the best approxmatng model over the next best model (Burnham & Anderson 22) Chck survval was computed over the 3-day brood-rearng perod of nterest as: / 3 day ¼ / 3 daly where / daly s the daly estmate based on the best model The top model carred most of the AIC c weght; so, ths procedure provdes a reasonable estmate of / 3-day Based on a delta approxmaton, the varance of the estmate / 3-day follows: ; f h s unobserved ð:þ 1 where h s the number of chcks seen at each samplng occason for a gven brood, ether, 1, 2,, m, or A represents a brood that was not sampled on a gven occason dfferng from a n whch a brood was sampled by observng the adult but no chcks were detected Because data are obtaned on ndvdual chcks wthn a brood, ths approach allows for an estmaton of chck survval, not brood survval The lkelhood functon s proportonal to a mult-nomal probablty mass functon Lð/; pjn; hþ / Yn P½h Š: ¼1 The lkelhood functon was optmzed to obtan the maxmum lkelhood parameter estmates n program r (R Development Core Team 27) Usng the above approach, a sute of models were developed to test the specfc habtat qualty hypotheses examnng brood-specfc effects on ndvdual chck survval (see Supportng Informaton Table S1) For all models, p was ether constant () orvaredbyyear (year) Whle the probablty of detectng a chck s condtonal on detectng the adult, the data used n the analyss are dependent on locatng the attendng adult usng rado telemetry Thus, t s unlkely that p would vary by habtat or other envronmental covarates The ncluson of year accounts for potental dfferences between the dfferent observers collectng data each year For /, the effects of nest habtat, prey densty and bomass wthn nest habtat, sex of the adult and year were nvestgated cvar½/ 3 day Š¼ð3/ 29 daly Þ2 cvar½/ daly Š: Movement probabltes A mult-state (or mult-strata) modellng approach was used to nvestgate daly brood movements Ths approach s an extenson of the CJS (Cormack 1964; Jolly 1965; Seber 1965) lve captures model extended to multple states (Hestbeck et al 1991; Browne et al 1993; Schwarz et al 1993) States can be geographcal areas, physologcal states or behavoural status In ths study, states refer to the dfferent habtats A daly brood movement event (or transton) was defned as the movement from one habtat to a dfferent habtat wthn a 24-h perod The parameters of nterest n ths mult-state analyss nclude p s, the probablty of re-sghtng a brood n state s at tme, andw rs,the probablty that a brood n state r moves to state s at the end of the nterval startng at tme, condtonal on the brood remanng alve and avalable for re-sghtng An addtonal parameter ncluded n the mult-state modellng approach s the probablty of survval durng nterval on state r, / r For ths study, /r s the probablty of brood survval durng nterval on state r Because the study focused on chck survval and not brood survval, / r was constraned to be constant across all strata, treatng / as a nusance parameter rather than a parameter of nterest States were defned as the three habtats; agrcultural felds (A), grassland (B) and prare dog (C), represents a sampled but undetected brood, and represents a brood that was not sampled on a gven occason (e mssng value) For example, an encounter hstory such as AC B ndcates a brood ntally observed on agrcultural land that moved to and was re-sghted on prare dog durng the second Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

5 874 V J Dretz occason, was not sampled durng the thrd occason, was sampled but not re-sghted durng the fourth occason, then was re-sghted on grassland durng the last occason A set of canddate models was developed to test the habtat qualty hypotheses on daly brood movement probabltes (see Supportng Informaton Table S2) Smlar to the chck survval analyss, p was modelled as constant or varyng by year and w was modelled as a functon of the habtat of locaton (not necessarly the nest habtat), prey densty and bomass wthn the nest habtat, sex of the adult, and year, or addtve combnatons AIC c was used for model selecton and nference as descrbed above (see Statstcal Analyss: Chck Survval) The mult-state analyss was conducted n program mark (Whte & Burnham 1999) The probablty of at least one movement from one habtat to a dfferent habtat was computed over the 3-day brood-rearng perod as: w 3 day ¼ 1 ð1 w daly Þ 3 where w daly s the daly movement probablty Because the best approxmatng model contaned all of the AIC c weght (1Æ), the daly movement estmates from ths model were used to compute w daly The varance s based on a delta approxmaton as: cvar½w 3 day Š¼ð 3ð1 w daly Þ 29 Þ 2 cvar½w daly Š: Results PLOVER PREY RESOURCES Ptfall trappng Data were collected on prey tems n ptfall traps In 24, samplng occurred only twce (May and July) because substantal ran hndered access to the grds n md-june Of the ptfall traps, 4Æ8% (621 ptfall traps) contaned no plover prey tems A total of plover prey tems were collected: (22Æ8%) on agrcultural land, (44Æ7%) on grassland and (32Æ6%) on prare dog Speces n the Order Hymenoptera, Famly Formcdae (e ants) comprsed 5Æ5% ( prey tems) of the total prey tems The Order Coleoptera comprsed the second hghest number of total prey tems wth prey tems (26Æ8%) Sap beetles (Famly Ntduldae) comprsed 7Æ9% (22 55 prey tems) of ths Order The largest number of the prey tems collected on agrcultural land were speces n the Order Coleoptera (41Æ5%, prey tems of ) and on grassland and prare dog were speces n the Order Hymenoptera (63Æ8%, prey tems of and 53Æ3%, 2 4 prey tems of respectvely) Densty and bomass To estmate densty, prey tems were clustered by Famly for each ptfall trap for a total of 4 67 prey observatons: observatons on agrcultural land, on grassland and on prare dog The estmates of prey densty vared among habtats and years (Fg 1a) Agrcultural land had less varaton n densty across years than grassland or prare dog (Fg 1a) Bomass of plover prey vared among habtats and ncreased each year of the study (Fg 1b) The pont estmates for bomass and densty of prey tems per habtat and year were used as covarates n the chck survval and the brood movement analyses PLOVER BROOD-REARING ACTIVITY A total of 153 adults wth broods were montored (Table 1) Fewer broods from grassland nest habtat (n = 2 broods) were montored than broods from agrcultural (n =66 broods) and prare dog nest habtat (n =67broods)The effort to locate nests was the hghest on grassland; therefore, low numbers of nests on grassland may have been due to low reproductve effort on ths habtat durng the study More broods were tended by males (n = 79 broods) than females (n = 48 broods); ths dd not dffer by year and nest habtat except for on prare dog n 26 n whch an equal number of broods were tended by males and females (Table 1) The sex of 26 adults tendng to a brood was unknown due to the lack of an approprate sample for DNA analyss At least 43 chcks from 28 broods (17 broods fledged at least one chck, seven broods fledged at least two chcks and four broods fledged at least three chcks) survved from hatch to 3-day post-hatch (Table 1) Of the 153 broods montored, 38 broods (25%) moved off the nest habtat at least once (Table 1) The number of broods that moved were smlar for each of the three habtats but was hgher n 26 than n 24 and 25 (Table 1) Of the 38 broods that moved from ther nest habtat, 76% (n =29 broods) were observed most often on ther nest habtat and averaged 279 (SD 2Æ12) movement events suggestng these (a) Densty (*1/ha) (b) Bomass (g) Agrcultural Felds Grassland Prare Dog Fg 1 Estmates of densty (a) and bomass (b) and 95% confdence ntervals for mountan plover prey tems on agrcultural felds, grassland and prare dog colones n Colorado from 24 to 26 Estmates are dark shaded; 95% confdence ntervals are lght shaded Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

6 Juvenle survval of precocal speces 875 Table 1 Summary of mountan plover Charadrus montanus broods montored n Colorado, USA from 24 to 26 broods moved back and forth from ther nest habtat The average age of chcks at the frst movement event occurred when chcks were slghtly older n 25 (9Æ25 days post-hatch, SD 7Æ21)thann24(4Æ67 days post-hatch, SD 5Æ4) and n 26 (4Æ88 days post-hatch, SD 465) Chck survval Year Total Broods montored Nested on agrcultural lands Males Females Unknown Nested on grassland Males Females Unknown Nested on prare dog colones Males Females Unknown Broods that fledged chcks Broods moved from nest habtat On agrcultural felds On grassland On prare dog colones Based on AIC c, the most parsmonous model that best explaned the data was a model wth chck survval dfferng among nest habtats and re-sghtng probablty dfferng by year (Table 2, also see Supportng Informaton Table S1) The results also ndcated that dfferences n chck survval among nest habtats were greater than can be explaned by our measures of prey densty and bomass wthn the nest habtat, year or sex of the tendng adult In addton, the best approxmatng model contaned most of the AIC weght (87%) compared wth the next best model (8%) (Table 2, also see Supportng Informaton Table S1) and the evdence rato suggests that the 3-d survval probablty Agrcultural felds Grassland Prare dog Habtat Fg 2 Estmates of 3-day chck survval and 95% confdence ntervals for mountan plovers nestng on agrcultural felds, grassland and prare dog colones n Colorado from 24 to 26 Estmates are based on the top model n whch chck survval dffered among nest habtats and re-sghtng probablty dffered by year best approxmatng model was 1 tmes more lkely than the second best model Daly chck survval estmates ranged from Æ95 (CI Æ93 Æ96) for chcks hatched on agrcultural land and grassland to Æ99 (CI Æ98 1Æ) for chcks hatched on prare dog Takng these estmates over the 3-day brood-rearng perod, substantal dfferences n chck survval exst between chcks from prare dog nest habtat and chcks from nests on other habtats (Fg 2) Thrty-day chck survval for prare dog nest habtat (Æ75, CI Æ54 Æ87) s approxmately three tmes hgher than agrcultural nest habtat (Æ23, CI Æ14 Æ33) and grassland nest habtat (Æ24, CI Æ8 Æ45) wth no overlap n 95% confdence ntervals Re-sghtng probabltes of ndvdual chcks decreased through the years from 51 (CI Æ46 Æ56) n 24 to Æ25 (CI Æ26 Æ36) n 26 Movement probabltes The model selecton crteron suggested that daly movement probabltes were nfluenced by the addtve effects of habtat and year, and re-sghtng probabltes were affected by year Table 2 Summary of model selecton results for mountan plover chck survval and brood movement actvty n Colorado, USA, Model K AIC c DAIC c w Chck survval /(habtat)p(year) Æ62 Æ Æ87 /(year + prey densty)p(year) 7 254Æ14 6Æ52 Æ8 /(year + habtat)p(year) 8 257Æ 9Æ38 Æ5 /(sex of tendng adult)p(year) Æ22 19Æ61 Æ Brood movement w(habtat + year)p(year) Æ13 Æ 1Æ Models wth a DAIC c <2 unts are presented n ascendng by AIC c, wth DAIC c ndcatng the dfference between each model and the model wth the lowest AIC c value For chck survval, apparent survval (/), ncluded the effects of habtat, year, sex of the tendng adult, and plover prey densty Re-sghtng probablty (p) ncluded the effect of year For brood movement, brood movement (w) ncluded the effects of habtat and year and re-sghtng probablty (p) ncluded the effect of year The number of parameters Akake s nformaton crtera Akake weght Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

7 876 V J Dretz (Table 2, also see Supportng Informaton Table S2) The top model contaned 1% of the AIC weght (Table 2, also see Supportng Informaton Table S2) Models contanng the effects of prey densty or bomass wthn the nest habtat or sex of the tendng adult ranked substantally lower, >32 DAIC c unts, from the best approxmatng model (see Supportng Informaton Table S2) Daly movement probabltes for each habtat per year were relatvely small rangng from Æ to Æ12 The 3-day brood-rearng perod estmates suggest that plover broods on grassland were more lkely to leave grassland and move, at a smlar rate, to ether agrcultural land or prare dog (Fg 3, also see Supportng Informaton Table S3) Movement probabltes from agrcultural land to other habtats had the lowest estmates each year, but these estmates were not sgnfcantly dfferent from the probabltes of movng from prare dog to other habtats Further, the results suggest that broods moved among habtats more frequently n 26 than n ether 24 or 25 (Fg 3, also see Supportng Informaton Table S3) Re-sghtng probabltes for broods, not ndvdual chcks as n the chck survval analyss, were relatvely hgh rangng from 88 to 98 Dscusson The habtats used by mountan plover broods strongly nfluenced chck survval and movement patterns Smlar fndngs have been observed n other shorebrds wth precocal young (eg Lengyel 26; Colwell et al 27; Kosztola ny et al 27; Schekkerman & Bentema 27) Daly chck survval estmates for each nest habtat are wthn the range of those reported n past studes for mountan plovers (Mller & Knopf 3-d movement probablty Habtat transton Fg 3 Estmates of mountan plover 3-day brood movement probabltes and 95% confdence nterval n Colorado from 24 to 26 A movement event s defned as movng from one habtat to a dfferent type of habtat wthn a 24-h perod The habtats are defned as agrcultural felds (AG), grassland (GR) and prare dog colony (PD) Estmates are based on the top model n whch daly movement probabltes were nfluenced by the addtve effect of habtat and year, and re-sghtng probabltes dffered by year 1993; Knopf & Rupert 1996; Lukacs et al 24a; Dnsmore & Knopf 25) Studes on other shorebrd speces suggest that broods move to the habtat wth the hghest chck survval and stay at that habtat (Lengyel 26; Schekkerman & Bentema 27) By contrast, some mountan plover broods moved away from the nest habtat wth the hghest chck survval Further, broods moved off all the habtats ncluded n the study, smlar to fndngs by Knopf & Rupert (1999), and brood movement was not dependent on the type of habtat as suggested by Dretz et al (25) Prey avalablty of a habtat has been suggested to nfluence both chck survval and brood movements of shorebrds (Lengyel 26; Colwell et al 27; Kosztola ny et al 27; Schekkerman & Bentema 27), but ths was not supported n ths study PREY AVAILABILITY Contradctory to the hypothess, densty and bomass of plover prey resources wthn the nest habtat dd not explan patterns of mountan plover chck survval or brood movements among habtats Although prey resources may reflect habtat qualty, drect measurement of prey resources s challengng Samplng of arthropod populatons wthn and between habtats s complcated by spatal and temporal varaton n prey abundance (Smth & Rotenberry 199) Prey denstes on grassland and prare dog nest habtats vared more between years than between denstes on agrcultural nest habtat Ths result supports the hypothess that the dstrbuton of prey resources can be aggregated, and these aggregatons shft durng the breedng season and between years as resources fluctuate (Smth & Rotenberry 199; Schekkerman & Bentema 27) Prey bomass also vared between nest habtats and ncreased on each nest habtat durng the study Varaton n bomass may be a result of temporal changes n arthropod lfe cycles ncludng those dctated by ether long- or short-term weather condtons or habtat-specfc management (Vckery et al 21) The study area experenced a severe drought n 22 The observed ncreasng yearly trend n prey bomass may be n response to ths weather event Fnally, the measures of densty or bomass of prey used n ths study may be poor predctors of habtat qualty for mountan plover chcks Habtat qualty may be more nfluenced by a combnaton of factors, ncludng predaton pressures and natural dsturbances assocated wth weather CHICK SURVIVAL AND BROOD MOVEMENTS Ths study supports the relatve mportance of habtats occuped by prare dogs to mountan plover reproductve success Breedng habtat for mountan plovers s assocated wth areas of hgh dsturbance such as those ntensvely grazed by natve or domestc herbvores, recently burned areas and agrcultural lands (Knopf & Wunder 26) As predcted, adults nestng on prare dog habtat were three tmes more lkely to fledge chcks (Æ75, CI Æ54 Æ87) than adults nestng on agrcultural land (Æ23, CI Æ14 Æ33) and grassland (24, CI Æ8 Æ45) Ths result suggests that habtats occuped by prare dogs may be provdng Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

8 Juvenle survval of precocal speces 877 the dsturbance needed for mountan plovers to successfully rase chcks In terms of breedng success, Dretz & Knopf (27) dd not fnd that habtat nfluenced nest success when categorzed as ether agrcultural land or grassland wth or wthout prare dogs Habtat-related reproductve success between nestng and chck survval suggests that mountan plovers may ncorrectly assess nest stes on agrcultural lands and grassland, whch may functon as ecologcal traps gven the low rate of chck survval on these habtats Smlar fndngs have been suggested for other shorebrd speces (Sze kely 1992; Lengyel 26) It s unknown why mountan plover broods moved way from prare dog habtat, even though ths habtat had the hghest success n fledgng chcks when used as a nestng habtat Brood movement can have extensve costs to chcks and the parent A total of 3 broods n ths study were censored due to mortalty of the tendng adult (n = 11), or where the rado transmtter was found but the fate of the adult could not be confrmed (n = 19) Of the adult mortaltes, seven were from broods that stayed on ther nest habtat, and four from broods that moved to dfferent habtat Whle the sample szes are small, our results suggest that mortalty of the tendng adult, and most probably the chcks, does not ncrease as a result of movement to dfferent habtats In fact, movement to a dfferent habtat may ncrease survval of chcks and the parent The brood movement fndngs from ths study suggest that complex processes nfluence how dfferent habtats affect brood-rearng actvty of mountan plovers and further study on factors nfluencng mountan plover brood movement s warranted Beyond the habtat qualty hypotheses, the sex of the attendng adult and yearly varablty were also ncluded to potentally explan the observed patterns n both chck survval and brood movement On breedng grounds n Montana, USA, mountan plover broods tended by females had hgher chck survval rates than males (Dnsmore & Knopf 25) By contrast, ths study dd not fnd that the sex of the attendng adult nfluenced chck survval Dfferent factors, such as habtats (Colorado shortgrass vs Montana mxedgrass), breedng denstes or landscape confguraton may be contrbutng to these spatal dfferences between breedng populatons Addtonally, sex of the tendng adult dd not nfluence mountan plover brood movement whch has been observed n the Kentsh plover Charadrus alexandrnus (Kosztolány et al 27) The reason that year of the study nfluenced brood movements but not chck survval s unknown There were many dfferences between the study years, such as weather condtons For example, frequent lght ran events occurred throughout the broodrearng perod durng 26, when relatvely hgh levels of brood movement between habtats were observed Durng 24 and 25, weather events were less frequent but more severe ncludng halstorms and floodng, causng mortalty of chcks and, consequently, lower rates of brood movement occurred Ths study s the frst to quantfy mportant aspects of the brood-rearng perod of mountan plovers among the most commonly used habtats Whle prey densty and bomass dd not nfluence chck survval or brood movements, other physcal or ecologcal characterstcs of the dfferent habtats may nfluence these reproductve parameters Dfferences n predaton pressure between habtats have been demonstrated to nfluence chck survval and brood movements n many speces Speces exposed to predators can change ther habtat use by movng to habtats wth lower encounters wth predators or they can aggregate n groups usng varous defence mechansms to fend off predators Competton also plays a role n chck survval and brood movements n shorebrds (Lengyel 26; Kosztolány et al 27) Intraspecfc competton among ped avocets Recurvrostra avosetta for brood-rearng terrtores s thoughttocausebroodmovementtolesssutablehabtats (Lengyel 26) Smlar behavour may be occurrng wth mountan plover broods, especally when the densty of broods n a habtat patch exceeds the avalablty of resources CONSERVATION IMPLICATIONS An mportant fndng of ths study s that conservaton strateges amed at ncreasng nest success may be neffectve for mountan plover unless they are accompaned by measures promotng chck survval For some shorebrd speces, fledgng of chcks appears to be more closely related to populaton dynamcs than nestng success (Lengyel 26; Colwell et al 27) The three habtats evaluated n ths study are thought to be hghly sutable for nestng success (Dretz & Knopf 27) However, the results presented here support the ecologcal trap hypothess n that seasonal changes can decrease the sutablty for fledgng of two of the three nestng habtats, thus reducng the chances of successful reproducton However, further assessment of brood-rearng ecology on a larger spatal scale s necessary to confrm ths result The conservaton of mountan plovers should concentrate on habtats where survval of chcks s the hghest, e shortgrass prare occuped by prare dogs, to determne whch partcular attrbutes of these habtats (eg vegetaton structure, predaton pressure and prare dog behavour) nfluence reproductve success Acknowledgements I sncerely thank the prvate landowners who provded access to ther lands Numerous ndvduals asssted wth data collecton B Kondrateff provded advce on nvertebrate dentfcaton P M Lukacs provded statstcal assstance I thank G C Whte for modfyng program mark to allow for mssng values n the mult-state model Fnancal and logstcal support was provded by the Colorado Dvson of Wldlfe I thank C J Bshop, E Cam, P F Doherty, Jr, F L Knopf, JA Royle, J P Runge, B K Sandercock and an anonymous revewer for useful comments on early drafts of ths manuscrpt References Anders, AD, Dearborn, DC, Faaborg, J & Thompson, FR, III (1997) Juvenle survval n a populaton of neotropcal mgrant brds Conservaton Bology, 11, Baldwn, PH (1971) Det of the Mountan Plover at the Pawnee Natonal Grassland, Report No 134 US Internatonal Bologcal Program, Grassland Bome Program, Fort Collns, CO Browne, C, Hnes, JE, Nchols, JD, Pollock, KH & Hestbeck, JB (1993) Capture recapture studes for multple strata ncludng non-markovan transtons Bometrcs, 49, Buckland, ST, Anderson, DR, Laake, JL, Burnham, KP, Borchers, DL & Thomas, T (1997) Advanced Dstance Samplng: Estmatng Abundance of Bologcal Populatons Oxford Unversty Press, Oxford Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,

9 878 V J Dretz Burnham, KP & Anderson, DR (22) Model Selecton and Multmodel Inference: A Practcal Informaton-Theoretc Approach, 3rd edn Sprnger, New York Colwell, MA, Hurley, SJ, Hall, JN & Dnsmore, SJ (27) Age-related survval and behavor of snowy plover chcks Condor, 19, Cormack, RM (1964) Estmates of survval from the sghtng of marked anmals Bometrcs, 51, Dnsmore, SJ & Knopf, FL (25) Dfferental parental care by adult mountan plovers, Charadrus montanus Canadan Feld-Naturalst, 119, Dretz, VJ & Knopf, FL (27) Mountan plovers and the poltcs of research on prvate lands BoScence, 57, Dretz, VJ, Wunder, MB & Knopf, FL (25) Comparatve movements and home ranges of mountan plover broods n three Colorado landscapes Wlson Bulletn, 117, Graul, WD (1973) Breedng adaptatons of the mountan plover (Charadrus montanus) PhD thess, Unversty of Mnnesota Graul, WD (1975) Breedng bology of the mountan plover Wlson Bulletn, 87, 6 31 Hestbeck, JB, Nchols, JD & Maleck, R (1991) Estmates of movement and ste fdelty usng mark-resght data of wnterng Canada geese Ecology, 72, Jolly, GM (1965) Explct estmates from capture recapture data wth both death and mmgraton stochastc models Bometrka, 64, Knopf, FL (1998) Foods of mountan plover wnterng n Calforna Condor, 1, Knopf, FL & Rupert, JR (1996) Reproducton and movements of mountan plovers breedng n Colorado Wlson Bulletn, 18, Knopf, FL & Rupert, JR (1999) Use of cultvated felds by breedng mountan plovers n Colorado Studes n Avan Bology, 19, Knopf, FL & Wunder, MB (26) Mountan plover (Charadrus montanus) The Brds of North Amerca (eds A Poole & F Gll), no 211 The Academy of Natural Scences, Pennsylvana, and the Amercan Ornthologsts Unon, Washngton, DC Plover Kosztola ny, A, Sze kely, T & Cuthll, IC (27) The functon of habtat change durng brood-rearng n the precocal Kentsh plover Charadrus alexandrnus Acta Ethologca, 1, Lengyel, S (26) Spatal dfferences n breedng success n the ped avocet Recurvrostra avosetta: effects of habtat on hatchng success and chck survval Journal of Avan Bology, 37, Lukacs, PM, Dretz, VJ, Knopf, FL & Burnham, KP (24a) Estmatng survval probabltes of unmarked dependent young when detecton s mperfect Condor, 16, Lukacs, PM, Frankln, AB & Anderson, DR (24b) Passve approaches to detecton n dstance samplng Advanced Dstance Samplng (eds ST Buckland, DR Anderson, KP Burnham, JL Laake, DR Borchers & L Thomas), pp Oxford Unversty Press, New York Mller, BJ & Knopf, FL (1993) Growth and survval of mountan plovers Journal of Feld Ornthology, 64, 5 56 Orans, GH & Wttenberger, JF (1991) Spatal and temporal scales n habtat selecton Amercan Naturalst, 137, S29 S49 R Development Core Team (27) R: A Language and Envronment for Statstcal Computng R Foundaton for Statstcal Computng, Venna, Austra accessed 1 June 27 SAS Insttute (23) sas Verson 91 SAS Insttute Inc, Cary, NC Schekkerman, H & Bentema, AJ (27) Abundance of nvertebrates and foragng success of black-taled godwt Lmosa lmosa chcks n relaton to agrcultural grassland management Ardea, 95, Schwarz, CJ, Schwegert, JF & Arnason, AN (1993) Estmatng mgraton rates usng tag-recovery data Bometrcs, 49, Seber, GAF (1965) A note on the multple recapture census Bometrka, 52, Smth, KG & Rotenberry, JT (199) Quantfyng food resources n avan studes: present problem and future needs Studes n Avan Bology, 13, 3 5 Sze kely, T (1992) Reproducton of Kentsh plover Charadrus alexandrnus n grasslands and fsh-ponds: the habtat mal-assessment hypothess Aqula, 99, Sze kely, T (1996) Brood deserton n Kentsh Plover Charadrus alexandrnes: an expermental test of parental qualty and rematng opportuntes Ibs, 138, Thomas, L, Laake, JL, Strndberg, S, Marques, FFC, Buckland, ST, Borchers, DL, Anderson, DR, Burnham, KP, Hedley, SL, Pollard, JH, Bshop, JRB & Marques, TA (26) dstance 5 Release 2 Research Unt for Wldlfe Populaton Assessment, Unversty of St Andrews, UK accessed 2 March 27 Tpton, HC, Doherty, PF, Jr & Dretz, VJ (29) Abundance and densty of mountan plover and burrowng owl n eastern Colorado Auk, n press Vckery, JA, Tallown, JR, Feber, RE, Asterak, EJ, Atknson, PW, Fuller, RJ & Brown, VK (21) The management of lowland neutral grasslands n Brtan: effects of agrcultural practces on brds and ther food resources Journal of Appled Ecology, 38, Whte, GC & Burnham, KP (1999) Program mark: survval estmaton from populatons of marked anmals Brd Study, 46(Suppl), Work, TT, Buddle, CM, Kornus, LM & Spence, JR (22) Ptfall trap sze and capture of three taxa of ltter-dwellng arthropods: mplcatons for bodversty studes Envronmental Entomology, 31, Receved 13 September 28; accepted 22 Aprl 29 Handlng Edtor: Andy Royle Supportng Informaton Addtonal supportng nformaton may be found n the onlne verson of ths artcle Table S1 Summary of model selecton results for Mountan Plover chck survval Table S2 Summary of model selecton results for Mountan Plover brood movements Table S3 Parameter estmates of Mountan Plover 3-day brood movement probabltes between habtats Fg S1 Map llustratng the general study area n eastern Colorado, USA Fg S2 Schematc dagram and photograph of feld applcaton of trappng lne transect grd Please note: Wley-Blackwell are not responsble for the content or functonalty of any supportng nformaton suppled by the authors Any queres (other than mssng materal) should be drected to the correspondng author for the artcle Ó 29 The Author Journal complaton Ó 29 Brtsh Ecologcal Socety, Journal of Appled Ecology, 46,