Analysis on Landscape Pattern of Basic Farmland Quality in South Western Plateau Mountain Region at Different Spatial Scales

【Objective】A quantitative research of landscape pattern characteristics and spatial differences was conducted so as to provide a basis for the optimization of scientific layout of basic farmland and land space planning. 【Method】Based on the basic farmland of typical southwestern plateau mountain region in Yunnan Province, taking different spatial scales such as provincial and municipal areas into comprehensive consideration, and in view of comprehensive analysis of basic farmland quality, slope and land types, the landscape ecological theory method was combined with the GIS spatial statistical analysis method.【Result】Dehong Prefecture had the highest grade of farmland quality at 8.49, while Nujiang Prefecture had the lowest grade at 11.70, with the highest number of combinations of cultivated land types being 5. The basic farmlands were distributed in a fragmentary landscape pattern at the provincial scale. Dehong had the highest diversity index at the municipal scale, i.e. 2.0612, while Nujiang had the lowest index, i.e. 0.6215. According to the dominance, uniformity and concentration index, Dehong, Xishuangbanna, Dali and Yuxi, among others, had high quality, highly diversified and uniformly distributed basic farmlands; the global Moran's I index of the basic farmland landscape pattern index and the farmland quality were 0.2052 and 0.0310, respectively, while that of the basic farmland area ratios was 0.6046. Qujing had the highest Moran's I index at the municipal scale, i.e. 0.8145, while Zhaotong had the lowest index, i.e. 0.3541.【Conclusion】The basic farmlands in the southwestern plateau mountain region are distributed in a fragmentary landscape pattern. with great quality difference and few high-quality basic farmlands. The basic farmlands have weak quality correlation at a provincial scale, but strong spatial autocorrelation at a municipal scale, with a cluster distribution of local spatial autocorrelation similar to kernel density distribution. The areas with high kernel density value have positively correlated spatial relations, while those with low kernel density have negatively correlated spatial relation.