农业旱情遥感指数验证与不确定性分析

以山西省太谷与山东省济宁试验为例,对极轨气象卫星农业旱情遥感监测指数进行验证,并对其不确定性进行分析。山西省太谷试验结果表明,旬温度条件指数(T),植被条件指数(V),植被健康指数(H)与土壤湿度有较高相关性(R2分别达到0.51,0.50,0.56),日部分指数相关性更高(R2分别达到0.58,0.45,0.60);山东省济宁日指数与土壤湿度相关性比太谷的高(T,V,H的R2分别达到0.93,0.66,0.97),但是两地区叶面缺水指数(W)与土壤湿度的相关性都很低,验证结果表明,利用H与T较适合农业旱情监测。另外在对比分析2个地区试验结果的基础上,从作物种植结构、监测时间尺度、植被生理状况,以及指数本身特性等角度分析了农业旱情监测中的不确定性。

Validation of Agricultural Drought Indices and Their Uncertainty Analysis

This study aimed to validate the agricultural drought indices derived from NOAA/AVHRR data and analyzed their uncertainties by taking Taigu County in Shanxi Province and Ji'ning County in Shandong Province for examples.The result from the analysis of relationships between drought indices and soil moisture in Taigu County shows that vegetation health index(H),temperature condition index(T)and vegetation condition index(V)have somewhat low relations with soil moisture(R2 is 0.51,0.50,and 0.56,respectively).H and T have the closer relationship than normalized difference water index(W)and V.Validation using the data in Ji'ning County shows that H,T and V have better relations with soil moisture(R2 is 0.97,0.93,and 0.66,respectively).W has a poor relation with soil moisture in the two cases.These suggest that the uncertainties in drought monitoring should be analyzed in terms of the crop planting pattern,monitoring time scale,vegetation condition and farming activity.