Geo DA-based spatial correlation analysis of GDP in Xinjiang

The spatial dependence and heterogeneity is an inherent property and general phenomenon of regional development. And therefore,the spatial relevance has become an important indicator and effective means of revealing the spatial distribution pattern. On the basis of the methods of geographical spatial analysis and Geo DA software,we analyzed the spatial situation of per capita GDP in Xinjiang in 2012,including spatial correlation of each county and the relationship between per capita GDP and urbanization. The results showed that the spatial distribution of per capita GDP in Xinjiang appeared normal distribution. Moran's I was around 0. 4,showing a clear clustering phenomenon in space. The spatial projection indicated a positive gathered correlation distribution pattern,which from inside to outside,from south to north,from west to east. North tropism reflecting the development conditions of the southern and northern Xinjiang were absolutely different. East tropism reflecting the dependence on exchange channel or outgoing location of regional development,natural location and infrastructure was the key to heterogeneity conditions. LISA local autocorrelation analysis discovered that northern Xinjiang hotspots concentrated in Karamay,Wusu,Kuitun and Booker Xaar Mongolia autonomous county,where had better traffic location conditions and abundant resources in center area. Differences between Urumqi- Changji economic circle and the surrounding counties were large and formed a clear fault. Border counties belong to the lag behind and closed space,which showed random development of the state. Southern Xinjiang was more obvious,high value areas were highly concentrated in Korla,where was traffic hub for southern Xinjiang,and the development belt of oil in Tabei. Three regions in the south of Xinjiang and the majority of counties along the border belonged to low value area. Poverty and poor with a spatial coherence,which indicated the regional developments of Xinjiang were homogeneity,isomorphic and convergence. Spatial autocorrelation analysis of per capita GDP and urbanization was also consistent with the conclusions drawn.