灌区间盐分变迁与耕地安全特征——以三工河流域农业绿洲为例

Transformation of soil salinity in irrigational regions and cropland security:a case study in agricultural oasis of the Sangong River catchment

针对目前绿洲内部土壤盐渍化问题，以新疆三工河流域农业绿洲为研究区，运用遥感、GIS和地统计学相结合的方法研究了0～20cm土层盐分动态特征与发展趋势。结果表明：流域空间上土壤盐分的理论变异函数拟合符合指数模型，F检验达到极显著水平。Kriging差值及其与同期的土地利用类型图叠加运算表明，整个绿洲区南部阜康灌区耕地土壤盐化面积及盐渍化作用小于北部绿洲阜北灌区，耕地土壤盐分具有显著的局地地带性分布特征，随绿洲耕地南北方向距离的增加土壤盐渍化作用增强。另外南部绿洲耕地农业灌溉具有明显的脱盐作用，随时间尺度的延长，盐渍化作用减弱；而北部绿洲土壤盐渍化作用加强，表现为积盐作用。绿洲受盐害作用的区域主要为北部（阜北灌区）绿洲，通过空间概率模型预测，在长时间范围内，土壤盐渍化依旧是制约北部绿洲耕地生产力的主要因素。通过联合运用变异函数分析、线性回归分析和空间概率分析方法的优势，有效识别了三工河流域农业绿洲空间土壤盐渍化的变化特征与变化趋势，对于认识流域空间单元土壤盐分的变迁具有重要意义。

Aiming at the problems of soil salinization existed in oasis, a case study was carried out on the oasis in the Sangong River catchment in Xinjiang. Spatial characteristics of soil salinity in the topsoil （0 -20 cm） and its relationship with cropland security were studied with the combination of methods in geo-statistics, GIS and RS. The results reveal as follows： on catchment scale, the spatial variation in soil salinity is best fitted by exponential mod- el, with F-test showing high level of confidence. Mapping of soil salinity by Kriging and comparing it with land use maps show that area of soil salinization in Fukang Oasis is smaller than that in Fubei Oasis, and degree of soil salin- ization in Fukang Oasis is also lower than that in Fubei Oasis. There is a local distribution characteristics of soil sa- linity in cropland of the Sangong river watershed, when the distance of cropland from south to north is increased, the degree of soil salinizaiton is intensed in irrigated area. In addition, a clear process of pened washing salinity was hap- soil of cropland by agricultural irrigation water in south oasis, and the degree of soil salinizaiton was weak- en in cropland by long-time scale. Fubei irrigated area in the north oasis is mainly subjected to salinization in the oa- sis of Sangong River catchment, and in a long time, soil salinization is a key factor to limit cropland production using spatial probability model to estimate. This kind of methodology, which is the combination of analysis methods in vari- ogranl, linear regression, and spatial probability, is demonstrated as a useful tool to identity characteristics and trend of soil salinization in agricultural oasis in the Sangong River catchment. Maybe, the combination of methods is important to make a clear estimation for the spatial transformation of soil salinity in irrigational regions of the catchment.