Do landscape features predict the presence of barn owls in a changing agricultural landscape?

Download Do landscape features predict the presence of barn owls in a changing agricultural landscape?

Post on 29-Nov-2016




0 download

Embed Size (px)


  • Landscape and Urban Planning 107 (2012) 255 262

    Contents lists available at SciVerse ScienceDirect

    Landscape and Urban Planning

    jou rn al h om epa ge: www.elsev ier .com/ l

    Do landscape features predict the presence of baragricul

    So Hind id Ja Centre for Wi tish Cob Pacic Wildli , Canac Pacic Wildlif

    h i g h l i g h t s

    We investigated how changes to the agricultural landscape inuences nest site use by barn owls. Thirty percent of nest sites used in the early 1990s were lost by 2007. Loss of grassland habitat did not predict the continued use of remaining nest sites. Barn owls The length

    a r t i c l

    Article history:Received 11 OReceived in reAccepted 17 JuAvailable onlin

    Keywords:Farmland birdTyto albaUrbanizationAgricultural laRoad ecology

    1. Introdu

    Agricultplant and anrotations anduce a strucof biodivers& McCracke

    CorresponTel.: +1 604 94

    E-mail addElsie.Krebs@ec(D.J. Green).

    0169-2046/$ http://dx.doi.owere less likely to persist at sites with increased highway trafc exposure. of highway within a 1-km radius of available nest sites inuences whether they are occupied by barn owls.

    e i n f o

    ctober 2011vised form 7 June 2012ne 2012e 10 July 2012



    a b s t r a c t

    Population declines of farmland birds have been linked to the loss and fragmentation of grassland habi-tats resulting from changes in agricultural practices and urbanization. We investigated how changes tolandscape attributes in the Fraser Valley, British Columbia, Canada, inuenced the persistence and cur-rent occupancy by barn owls at roosting and nesting sites. There has been considerable development inthe agricultural landscape of the Fraser Valley between the early 1990s and 2007/2008: grassland coverdeclined by 53%, the area of urban cover increased by 133%, length of secondary roads increased by 18%,and the volume of highway trafc increased by 33%. We also found that 30% of the sites used by barnowls in the early 1990s have been lost. Although the availability of grasslands are thought to inuencethe distribution of barn owls, in our study, barn owls were not more likely to persist at sites with littleloss of grass cover, or to currently occupy sites surrounded by more grassland. The only variables thatpredicted the continued use and current occupancy of sites were trafc exposure and the length of high-ways. Barn owls were most likely to persist at sites with lower increases in trafc exposure and occupiedsites containing fewer kilometers of highway within a 1-km radius. We conclude that the distribution ofbarn owls in the Fraser Valley is inuenced by the loss of suitable roosting and nesting sites and locationof highways.

    2012 Elsevier B.V. All rights reserved.


    ural landscapes provide important habitat for manyimal species. Traditional farming practices such as cropd the maintenance of hedgerow and grassy verges pro-turally varied landscape capable of maintaining levelsity similar to that of many natural ecosystems (Bignaln, 1996; Bohlen & House, 2009). However, changes in

    ding author at: 5421 Robertson Road, Delta, BC, Canada V4K 3N2.0 4690; fax: +1 604 946 7022.resses: (S. Hindmarch), (E.A. Krebs), (J.E. Elliott),

    agricultural practices over the last 50 years (e.g. use of agrochem-icals, modernized machinery) have reduced the heterogeneity ofagricultural landscapes, resulting in large, heavily utilized, mono-culture elds and an overall reduction in the quality of land ashabitat for wildlife.

    In addition to changes in agricultural practices, in many areasagricultural land has been lost and fragmented due to urbanizationand its associated infrastructure (Forman et al., 2003; Underhill& Angold, 2000). Although urbanization results in habitat loss,infrastructure such as roads and railways can have a dispropor-tionate impact on wildlife due to mortality from collisions withvehicles, restricted access to resources through the barrier effect, orbecause populations become subdivided and isolated into smallerand more vulnerable fractions (Forman et al., 2003; Jaeger et al.,2005). These changes have been implicated in the declines both in

    see front matter 2012 Elsevier B.V. All rights reserved.rg/10.1016/j.landurbplan.2012.06.010tural landscape?

    marcha, Elizabeth A. Krebsb,, John E. Elliott c, Davldlife Ecology, Department of Biological Sciences, Simon Fraser University, Burnaby, Brife Research Centre, Environment Canada, 5421 Robertson Road, Delta, British Columbiae Research Centre, Environment Canada, Delta, British Columbia, Canada V4K 3N2ocate / landurbplan

    n owls in a changing

    . Greena

    lumbia, Canada V5A 1S6da V4K 3N2

  • 256 S. Hindmarch et al. / Landscape and Urban Planning 107 (2012) 255 262

    Fig. 1. Study a featulocations of av Speci

    range and alandscapes Devictor, Cl

    There isare particulIn Britain, 8over the lasreported froKuvlesky, 2Peterjohn, 2sible for the(Newton, 2(Miliaria calinked to a elds, a key2002; Wilstrast, poputo be linkeincreased g& Crick, 20tion declinthat more sappropriate2003).

    The barnbeen associWarburtonencing rang1985; Tomhave been to have drrea within the Fraser Valley, British Columbia, Canada used for assessing landscapeailable nest sites are not indicated because barn owls are a species listed under thebundance of many species associated with agricultural(Benton, Vickery, & Wilson, 2003; Filippi-Codaccioni,obert, & Julliard, 2008; Robinson & Sutherland, 2002).

    accumulating evidence that farmland birds as a grouparly sensitive to changes to the agricultural landscape.6% of farmland birds have shown range contractionst 40 years (Fuller et al., 1995). Similar trends have beenm elsewhere in Europe and North America (Brennan &005; Donald, Sanderson, Bureld, & van Bommel, 2006;003). However, no single factor appears to be respon-

    observed range contractions and population declines004). For example, population declines in corn buntinglandra) and tree sparrow (Passer montanus) have beenreduction in winter survival due to loss of fallow grain

    winter food supply (Siriwardena, Robinson, & Crick,on, Boyle, Jackson, Lowe, & Wilkinson, 2007). In con-lation declines in lapwing (Vanellus vanellus) appeard to higher rates of nest predation resulting fromrazing intensity on marginal grasslands (Chamberlain03). The variation in underlying causes for popula-es of farmland birds has led some authors to arguepecies-specic research is required in order to design

    conservation measures (Fuller et al., 1995; Peterjohn,

    owl (Tyto alba) is an iconic farmland bird that hasated with humans and agriculture for centuries (Bunn,, & Wilson, 1982). However, barn owls are now experi-e contractions and declines across their range (Colvin,s, Crick, & Shawyer, 2001). In Britain, where theystudied most intensively, their numbers are reportedopped by 69% over the last 50 years (Toms et al.,

    2001). Threbarn owl pover the laand potentSecond, suiversion of the remova(Ramsden, increased dvolume in 2003).

    PopulatiNorth Ameis thought America (Ewestern Caon the Statu2010). To eimpacted texamine hohave impacine how cunest/roost scally we usof four feathypothesiz(i) the amothe fragmedevelopmeBurnside, M2003).res that inuence nest-site use by barn owls in 2007 and 2008. Exactes at Risk Act in Canada and most sites are located on private land.e major factors have been associated with declines inopulation. First, the loss of moderate-length grasslandst 50 years has decreased small mammal populationsially prey availability (Colvin, 1985; Taylor, 1994).table nest/roosting sites have been lost due to the con-old wooden barns into inaccessible steel barns andl of old trees as part of eld enlargement programs1998; Taylor, 1994). Third, adult mortality may haveue to increases in the number of roads and trafc

    agricultural areas (Preston & Powers, 2006; Ramsden,

    on trends for barn owls are not well documented inrica but, based on breeding bird surveys the barn owlto be declining by up to 3% per year in western Northnvironment Canada, 2010). The barn owl population innada was designated as Threatened by the Committees of Endangered Wildlife in Canada in 2010 (COSEWIC,

    valuate how changes to the agricultural landscape havehe distribution of barn owls in Western Canada, wew changes in landscape features over the past 15 yearsted occupancy at historical nest/roost sites, and exam-rrent landscape features inuence the use of suitableites across the Fraser Valley, British Columbia. Speci-e an information theoretic approach to assess the roleures of the landscape surrounding available nest sitesed to impact the distribution and breeding of barn owls:unt of grassland cover (Colvin, 1985; Taylor, 1994), (ii)ntation of grassland habitat (Taylor, 1994), (iii) urbannt and (iv) trafc exposure, or length of highways (Bond,etcalfe, Scott, & Blamire, 2005; Lod, 2000; Ramsden,

  • S. Hindmarch et al. / Landscape and Urban Planning 107 (2012) 255 262 257

    2. Methods

    2.1. Study site

    We monrey, an areaCanada (49key importencompassWildlife Aretlement theand low shrcovered priSince Europfor pasture tural practields and gulation of D1960s popu(7.5%) of lan(Agricultura

    2.2. Survey

    To evaluowl nestingof the 129 (Andrusiak,mission to aremained arent distribnesting andrey (n = 143Andrusiak additional 6sites includable openinproperties (

    Sites webreeding peand March that occupasites at leasthe inside ations of thecarcasses). beams/platcould not bducted at dthe structuwas sightedthree survewhere no bwere foundquently visithey were ution of wheapproximating the incuthe breedinability to acfor 106 of th

    2.3. Land u

    We quaradius of ea

    layers using Geographic Information System software (ArcGIS 9.2).We used a 1-km radius around each site because this results in anarea that approximates the home range of a barn owl (3 km2; Bondet al., 2005; Shawyer & Shawyer, 1995; Taylor, 1994).

    ssland layp tyas n

    y insing thmmeancorea (, andyer m

    caton tandall, Aasslaed f

    htly gour td 4ualityrass

    regulpes s vaatoryecondataeverafrom


    to dihesehe Vver,

    the yer mition08, ays

    owexpomaingth ite bfc n avbtairy of averaighw

    ta a

    charrveyehe fo(ii) ctanc

    in tafc itored barn owls in the municipalities of Delta and Sur- of 681 km2 within the Fraser Valley, British Colombia,80 North, 122180 West; Fig. 1). This area is ofance for migratory and residential farmland birds andes important wildlife areas such as the Alaksen Nationala, Burns Bog and Boundary Bay. Prior to European set-

    oodplain would have been dominated by grasslandub vegetation while higher elevations would have beenmarily by coniferous forest (North & Teversham, 1983).ean settlement the area has traditionally been usedand hay production, but over the last 35 years agricul-ces have shifted to vegetable and berry production inreenhouses (Statistics Canada, 2006). The current pop-elta and Surrey combined is seven times larger than thelation. This has coincided with the removal of 5000 had from agricultural production in this area since 1974l Land Reserve, 2009).


    ate changes in the availability and occupancy of barn/roosting sites in Delta and Surrey we revisited 116sites occupied by barn owls in this area in the 1990s

    1994). We were unable to obtain the landowners per-ccess the remaining 13 sites. In total, 80 of these sites

    vailable for barn owls (see results). To evaluate the cur-ution of barn owls we surveyed all potentially suitable

    roosting sites that we could access in Delta and Sur- sites). This sample included the 80 sites surveyed by(1994) that were still available for barn owls and an3 sites not surveyed in the 1990s. Potentially suitableed old wooden barns or other tall structures with suit-gs near the roof and old single-standing trees on farmBunn et al., 1982; Taylor, 1994).re monitored for seven months over the peak barn owlriod in BC for two years (March to September 2007to September 2008; Campbell et al., 1990). To ensurency was not underestimated, we surveyed all potentialt three times per year. During each survey we searchednd perimeter of the structure for barn owls or indica-ir presence (such as fresh pellets, feathers and/or preyWe also climbed up the side of the structure to checkforms for any nesting activity. If the beams/platformse accessed at a site, a 30 min. observation was con-usk, to determine if any barn owls exited or enteredre. A structure was considered occupied if a barn owl

    or if we found fresh pellets or feathers during any of theys. Interviews with landowners conrmed that all barnsarn owls were sighted and no fresh pellets or feathers

    were not currently being used by barn owls. We subse-ted all occupied sites every 12 months to determine ifsed for breeding. Visits of this frequency enable evalua-ther breeding occurred, as incubation and rearing takesely 2.5 months, and females and chicks beg loudly dur-bation and nestling periods. We could not determineg status at all occupied sites due to limitations on ourcess and search sites; breeding status was determinede 115 occupied sites.

    se characteristics and spatial analysis

    ntied current and historical land use within a 1-kmch potential nesting/roosting site from digitized data

    Graual eland crolayer wvisualldigitizing, co2006 Vstudy arampsroad la

    Webased (AschwTattersnent grdisturbare ligup to f(mowehigh qlands, gthat is land tyin eldexplan

    To rcreate from s1990s pixel).these punabletat in tfrom tVancouwithinroad laNo addand 20highwof barnin the that rethe leneach sage traData owere oMinistas the way/h

    2.4. Da

    Wesite suusing tcover, age dischangeway trd and crop cover in Delta was obtained from an individ-er from 2007 that contained information on grasslandpe (Ducks Unlimited, unpublished data). A similar dataot available for Surrey, so we created a data layer bypecting individual elds for grassland or crop type andese data (n = 1747 elds). Data on urban cover (hous-rcial and industrial land use) were obtained from auver Regional District land-use layer map for the entireMetro Vancouver, 2008). Data on highways, connecting

    roads within the study area were obtained from a BCap (GeoBC, 2010).

    egorized grass cover into ve different categorieshe intensity of use, a factor linked to vole densitiesen, Holzgang, & Jenni, 2007; Butet & Leroux, 2001;vundo, Manley, Hart, & MacDonald, 2000): (i) perma-nds, (ii) grassland set asides (areas of grass that are notor up to four years), (iii) pastures (areas of grass thatrazed), (iv) hayelds, areas of hay crop that is mowed

    imes/year, (v) grassy verges along roads and eld edges6 times/year). We further categorized grass cover into

    foraging habitat that is not mowed (permanent grass-land set asides and pasture) versus lower q...


View more >