Spatial Estimation of Sika Deer Population Density Distribution

Abstract

We estimated the spatial di$tributionf the sika deer (Cerot{s nipPon) population densityinareas around Mt. Hiko inFukuoka Prefecture,Japan.[Ihedatacomprised pelletceunts in86 plots from February toApril1999and frem November 1999 to March 2eOO.Takingthistocorrespond tothepepulation densityintheplots,thesika deer populatien densitywas mapped using a kriginginterpelator. Fivesemivariegrams were examined in a geographic lnformationsy$tem (GIS):spherical, circular, exponential, Gaussian,and linearwith sill models. We used the Akaike InformationCriterion(AIC)and the Bayesian InformatienCriterion(BIC)te select the appropriate semivariogram modeL 1[hespherical rnodel produced the smallest values inbeth the AIC and BIC. The ordinary kriginginterpolatorwith the spherical rnodel previded thedeerpopulation densityat intervalsef 50m. Four distinct high-densityareas, inwhich the deerdensityexceeded 301km2,were found around Mt. Kosho and Mt. Hiko. Using independentdataforthe damage caused by deer and the map of the sika deer density,we estimated the deer den$itiesinal1plets where thedamage by deerhadbeen checked, and determinedtherelation betweenpepulation densityand the degree of browsing dainage,which ismost serious in Fukuoka Prefecture. As the $ika deer populationdensityincreases,the proportien of damaged plots increases.Of plots with the densityexceeding 40 deer!kmE,browsingdamage eccurred in83%.'I/his method provides insightintothe densityof deer over a broadi region and, incembination withthe informationabout deer damage, can be applied togive theprobability of deer damage.