Winter Habitat Selection By Ungulates In Huangnihe Nature Reserve

Winter habitat selection by ungulates named wild boar(Sus scrofa), roe deer (Capreolus capreolus) and red deer(Cervus elaphus) was investigated using transects line method in Huangnihe Nature Reserve in two winters from2009to2011. Forty-seven transect lines with179.2km total length were investigated from November in2009to February in2010as well as12transect lines with182.3km toatl length from November in2010to February in2011. Fifteen kinds of ecological factors, namely elevation, slope direction, gradient, slope location, canopy, thickets coverage, hiding level, food abundance, distance far from human disturbance, diameter at breast height, forest type, conifer, snow depth, fallen tree and bare rock were taken into consideration to evaluate the winter habitat. Eighty-one habitat quadrats selected by wild boar,72habitat quadrats selected by red deer,109habitat quadrats selected by roe deer,72bed-site quadrats selected by roe deer and109control quadrats were investigated in the first winter as well as57habitat quadrats selected by wild boar,56habitat quadrats selected by red deer,124habitat quadrats selected by roe deer,115bed-site quadrats selected by roe deer and157control quadrats in the second winter. The availability of the ecological factors measured in the research was estimated with comparing the differences of control quadrats in two winters and, in Huangnihe Nature Reserve, the regulations of winter habitat selection by ungulates were assessed by means of analyzing the differences of quadrats selected by ungulates as well. The differences of the15kinds of ecological factors determined in the research between the quadrats selected by ungulates and control quadrats were analyzed to appraise the effects upon the winter habitat selection by ungulates. The results were as follows:1) The results of Chi-square test and two-samples Mann-Whitney U rank test indicated that there were no significant differences of the availability of the15kinds of ecological factors in the research between the two winters and, the availability maintained a relatively stable status and the selection regulations on the measured ecological factors by ungulates presented no significant changes in short term in Huangnihe Nature Reserve.2) The results of winter habitat selection by wild boar obtained through the Chi-square test combining with two-samples Mann-Whitney U rank test and Bonferroni Confidence Interval Discriminant showed that wild boar preferred the habitats located in gentle and sunny slope with582m of average elevation, moderate hiding level, higher food abundance and diameter at breast height, lower snow depth, distance far from human disturbance, fallen tree available and broad-leaved forest while avoided half sunny slope, shady slope, upper slope, mixed forest and coniferous forest in Huangnihe Nature Reserve. The Binary Logistic Regression demonstrated that food abundance, diameter at breast height, forest type, distance far from human disturbance, snow depth and slope direction were the key factors to winter habitat selection by wild boar and that, hiding level and slope location were the secondary factors. A Binary Logistic Regression model which could predict correctly with96.5%of overall could be constructed by these8kinds of ecological factors as follow:Z=-17.09+0.661×slope direction (1)+2.455×slope direction (2)-2.184×slope location (1)-0.899×slope location (2)-0.213×hiding level+0.164×food abundance+0.370×diameter at breast height+0.001×distance to human disturbance+6.805×forest type (1)+4.506×forest type (2)-0.075×snow depth. The selection probability was that:P (Z)=ez/1+ez.3) The results of winter habitat selection by red deer got by the Chi-square test combining with two-samples Mann-Whitney U rank test and Bonferroni Confidence Interval Discriminant indicated that red deer preferred the habitats located in gentle slope and mixed forest with652m of average elevation, lower snow depth, moderate diameter at breast height and hiding level, higher food abundance and thickets coverage, distance far from human disturbance, while avoided the habitats characterized by shady slope and coniferous forest. The Binary Logistic Regression presented that food abundance, distance far from human disturbance, forest type, slope direction and snow depth were the key factors to winter habitat selection by red deer and that, gradient and hiding level were the secondary factors in Huangnihe Nature Reserve. A Binary Logistic Regression model which could predict correctly with98.7%of overall could be constructed by these7kinds of ecological factors as follow:Z=-46.834+0.060×gradient+1.815×slope direction (1)+8.865×slope direction (2)+0.175×hiding level+0.512×food abundance+0.003×disturbance+11.808×forest type (1)+12.097×forest type (2)-0.118×snow depth. The selection probability was that:P(Z)=ez/1+ez.4) The results of winter habitat selection by roe deer achieved through the Chi-square test combining with two-samples Mann-Whitney U rank test and Bonferroni Confidence Interval Discriminant demonstrated that roe deer preferred the habitats located in gentle, middle and sunny slope with596m of average elevation, lower snow depth, moderate hiding level and canopy, higher food abundance and thickets cover, fallen tree available and mixed forest, while avoided the habitats located in upper, lower and shady slope with broad-leaved forest, coniferous forest and bare rock available in Huangnihe Nature Reserve. The Binary Logistic Regression showed that food abundance, snow depth, canopy, forest type, elevation, gradient and hiding level were the key factors to winter habitat selection by roe deer and that, slope direction was the secondary factor. A Binary Logistic Regression model which could predict correctly with89.8%of overall could be constructed by these8kinds of ecological factors as follow:Z=-0.008×elevation-0.043×gradient+1.027×slope direction (1)+1.652×slope direction (2)+0.093×hiding level+0.080×canopy+0.116×food abundance-1.018×forest type (1)+1.337×forest type (2)-0.078×snow depth. The selection probability was that:P(Z)=ez/1+ez. The results of winter bed-site selection by roe deer indicated that roe deer preferred the habitats located in middle, sunny and flat slope with586m of average elevation, moderate diameter at breast height and hiding level as well as canopy, lower snow depth, higher food abundance and thickets coverage, fallen tree available and mixed forest, especially preferred the conifers in mixed forest for rest while avoided the habitats located in upper, lower and shady slope with broad-leaved forest, coniferous forest and fallen tree available. The Binary Logistic Regression demonstrated that diameter at breast tree, food abundance, conifer, elevation and snow depth were the key factors to winter bed-site selection by roe deer in Huangnihe Nature Reserve and that, gradient, bare rock, hiding level and slope direction were the secondary factors. A Binary Logistic Regression model which could predict correctly with98.9%of overall could be constructed by these9kinds of ecological factors as follow:Z=27.765-0.079×elevation-0.231×gradient+25.342×slope direction (1)+19.331×slope direction (2)+0.623×hiding level-0.546×diameter at breast height-14.378×bare rock-26.207×conifer (1)-1.063×snow depth. The selection probability was that:P(Z)=ez/1+ez.