Research On Factors Affecting Red-crowned Crane’s Breeding Sites And Its Distribution Forecasting

With the developement of statistics technique and Geographic Information System, Species Distribution Model is widely used in the research of distribution of red-crowned crane’s breeding sites. The implement of ensemble models could reduce the uncertainty of SDMs, which could advance the precision of the imitation. Red-crowned crane is first class national protected wild animals of China. Analysing the factors impacting the distribution of breeding sites and predicting the change in the climate environment in the future could provide a detailed basis for the conservation of red-crowned crane, which is meaningful for the diversity of species and sustainable development of environment.Firstly, clearing up the data of breeding sites of red-crowned crane, which is choosing and merging the effective breeding sites. Choosing the typical explanatory variables by rank correlation effecient cluster analysis, which referencing the varibles of researches in th past. Constructing variable species distribution models to compare and choose the best model. Ensembled models is constructed by ROC and different sets of variables are compared to analysis the effect of simulation. Then the contribution of variables to the breeding sites of red-crowned crane is analysised and the ensembled models are adopted to predict the spatial distribution of high probability in the future.Random Forest performed the best among the models, while SRE was the worst, indicated that complicated models performed better. Selecting the explanatory variables before modeling will achieve better fitting. The predictions coincided with the actual distribution of red-crowned crane’s breeding sites and the influence of climate change is measured. The application of ensemble model could lower the uncertainty and improve the performance of SDMs. The prediction showed the breeding sites moving toward higher altitude and the range will be smaller in the given future climate environment.