在简要阐述3S技术和神经网络方法(3S-ANN)在农地石漠化预警方面原理和计算的基础上,通过遥感影像进行解译获取广西都安县各地类的植被覆盖率、基岩裸露率、地表起伏指数、≥25°坡地面积比等资源环境信息,并辅以农业人口密度、人均耕地面积、土地垦殖指数和坡耕地指数等经济、社会等指标数据,采用神经网络中的BP算法对都安县不同时期农地石漠化警情进行分析预测.结果表明,3S-ANN分析与实际相符:1977-2005年农地石漠化警情不断加剧,无警情乡镇由11个减少为0个,重警乡镇由0个增加到9个,但2007年的石漠化警情较2005年所减轻.应用2005、2007年数据对未来的2015年进行预测,结果显示,该县东、西部地区的东庙、加贵、保安、大兴等乡镇石漠化程度将有所加强,可作为今后防治的重点.%On the basis of 3S and ANN method, by means of RS interpretation, the resources and environment information such as vegetation coverage, exposing rate of rock, relief degree of land surface, the ≥ 25° slope land area of different land use types is obtained. Then early warning for karst rocky desertification in agricultural land during different periods is done in light of economic and social data such as agricultural population density and farmland areas per person as well as reclaimed indexes of land with the BP Algorithm of ANN method. The results show that the analyses of 3S-ANN are consistent with the actual situation. The warning condition of karst rocky desertification in agricultural land increased constantly from 1977 to 2005. The no-warning condition townships decreased from eleven to zero and the severe-warning condition townships increased from zero to nine, but the warning condition in 2005 relieved slightly compared with that in 2007. In light of the prediction with the data in 2005 and 2007, it is concluded that the rocky desertification degrees in Dongmiao, Jiagui, Baoan and Daxing townships will aggravate in 2015. so should pay more attention in the prevention and control in future.
展开▼
