南方红壤区林下水土流失的遥感判别——以福建省长汀县为例  

Detection of Soil Erosion Area Under Forest Canopy in the Red Soil Region of Southern China Using Remote Sensing Techniques:Changting County, Fujian Province

作　　者：徐涵秋[1,2] 张博博[1,2] 关华德[3] 胡秀娟[1,2] 陈明华 付伟[1,2] 

机构地区：[1]福州大学环境与资源学院/遥感信息工程研究所,福建福州350116 [2]空间数据挖掘与信息共享教育部重点实验室/福建省水土流失遥感监测评估重点实验室,福建福州350116 [3]澳大利亚弗林德斯大学环境学院 [4]福建省水土保持监测站,福建福州350001

出　　处：《地理科学》2017年第8期1270-1276,共7页Scientia Geographica Sinica

基　　金：国家科技支撑项目(2013BAC08B01-05);福建省教育厅重点项目(JA13030)资助~~

摘　　要：针对林下水土流失缺乏有效判别方法的问题,提出了一种遥感判别方法。该方法以植被覆盖度、植被健康度、土壤裸露度和坡度为判别因子,采用规则法来建立林下水土流失遥感判别模型,并将其应用于福建省长汀县。结果发现,长汀县有311.66 km^2的林地发生不同程度的林下水土流失,其中有13.35%的土壤侵蚀强度达到中度。通过遥感方法识别出的林下水土流失区的空间分布位置可为该县今后深入治理水土流失提供目标靶区。This study proposes a new rule-based method to locate soil erosion under moderate-to high-density forest canopy using remote sensing techniques. Five factors that are closely related to the soil erosion in forest are specially selected and used as discriminators to develop the discrimination rules. The selected five factors include fractional vegetation coverage, nitrogen reflectance index, yellow leaf index, bare soil index, and slope degree. The selection of these five factors aims to detect vegetation density, vegetation health status, soil exposure degree, and terrain steepness. These five factors can all be derived from remote sensing imagery based on related thematic indices or algorithms. The proposed method was applied to the forest areas in Changting County of Fujian Province, southern China. The Hetian town of the county is one of the most typical soil loss areas in the red soil regions in southern China. A Landsat 8 OLI image acquired on October 15, 2014, in-situ measured spectral data, and nutrient data from soil samples have been used as data sources for the study. The result reveals a total area of 311.66 km^2 of soil erosion in forest of the county. Of them, 13.35% is in moderate erosion intensity and the rest is in light erosion intensity. The accuracy assessment against ground truth indicates that the method can achieve an overall accuracy of 88.45%, with a Kappa coefficient of 0.731. The revealed locations of soil erosion in forest provide very useful information for Changting County to develop land management plans to further reduce soil loss in forest.

关 键 词：林下水土流失 规则判别 遥感 植被覆盖度 植被健康状况 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置] S127[自动化与计算机技术—控制科学与工程]