中文摘要：

以黄土高原北部的横山县为例,利用ETM＋遥感数据和RUSLE水土流失模型实现了基于“3S”技术的黄土丘陵区水土流失定量反演,并在神经网络技术支持下对定量反演结果进行评价.结果表明,横山县平均水蚀模数为103.23 t·hm^-2·yr^-1,年水蚀量为4.38×10^7 t,总体为中强度水蚀.水蚀程度在空间上呈自西、北向东、南逐渐增强的态势.水蚀随地形部位分异也较为显著,即川坝地、墚峁顶、风沙区地形平坦,以溅蚀为主,有微弱水蚀甚至无水蚀现象;墚峁坡上,坡度平缓,以面蚀为主,为轻度侵蚀;墚峁坡中下部,坡度增大,以线蚀为主,为中度侵蚀;沟缘线以下,坡面陡立,以冲蚀为主,为重度侵蚀.本文研究技术和方法切实可行,具有推广价值,研究结果对流域治理和水土保持具有指导意义.

英文摘要：

With Hengshan County of Shanxi Province in the North Loess Plateau as an example, and by using ETM ＋ and remote sensing data and RUSLE module, this paper quantitatively derived the soil and water loss in loess hilly region based on ＂3S＂ technology, and assessed the derivation results under the support of artificial neural network. The results showed that the annual average erosion modulus of Hengshan County was 103. 23 t·hm^-2, and the gross erosion loss per year was 4. 38× 10^7 t. The erosion was increased from northwest to southeast, and varied significantly with topographic position. A slight erosion or no erosion happened in walled basin, fiat-headed mountain ridges and sandy area, which always suffered from dropping erosion, while strip erosion often happened on the upslope of mountain ridge and mountaintop fiat. Moderate rill erosion always occurred on the middle and down slope of mountain ridge and mountaintop fiat, and weighty rushing erosion occurred on the steep ravine and brink. The RUSLE model and artificial neural network technique were feasible and could be propagandized for drainage areas＂ control and preserved practice.