RESULTS

Result of data processing

In areas above 1500 meters, 99 geographical locations of PPR cases left by 30 km rarefying. The minimum temperature of June was selected as the environmental variant. No multi-collinearity was detected by a VIF values within 0 to 2(<10). The variant of vector distance from the river (river distance) was excluded due to its high standard deviation. Finally, the minimum temperature of June, vegetation, population density and slope were adopted for the final model(fig.2 ).
.
In areas below 1500 meters, 81 geographical locations of PPR cases left by 40 km rarefying. The mean temperature of September was selected as the environmental variant. No multi-collinearity was detected by a VIF values within 0 to 2 (<10). The variant of vector distance from the river (river distance) was excluded due to its high standard deviation. Finally, the mean temperature of September, vegetation, population density and slope were adopted for the final model(fig.3 Table.2).

Result of prediction for the Spatial distribution of PPR

The AUC and SD values of the both models of above and below 1500 m are 0.825, 0.027and 0.890, 0.005 respectively, indicating a better prediction. The Pamirs Plateau and its extended mountains are in high-risk areas, and the Tibet and Xinjiang in western china are surrounded by these risks. From the perspective of administrative divisions, the countries bordering China are all high-risk areas of PPR, while the high-risk areas on the China side are relatively weak(fig.4).

Result of prediction of the maximum available transboundary paths

Were obtained Five groups of PPR distribution points outside China and three groups in China were obtained by ArGis cluster analysis. We got five transboundary paths: 1.Kazakhstan-Confluence of Ili River and Horgos River- Xinjiang (Huocheng county); 2.Tajikistan-WestPamirPlateau-Xinjiang(Kashgarcity);3.Pakistan-WestPamirPlateau-Xinjiang(Kashgarcity);4.Kashmir-Pakistan-WestPamirPlateau-Xinjiang (Kashgarcity); 5.Kashmir-Bangon lake-Tibet (Ritu county)(fig.5).