Main content area

Estimating soil erosion intensity in susceptible areas based on a comprehensive stratified sampling system associated with the geographical information system

Wang, Liang-Jie, Zhang, Hongda, Qi, Fei, Liu, Xia, Zhang, Qinghong, Zhang, Chunqiang, Yuan, Li, Yao, Xiaoyou
Canadian journal of soil science 2019 v.99 no.3 pp. 321-333
geographic information systems, land use, local government, models, prediction, soil erosion, vegetation cover
An error-free and fast approach for the identification of the area dynamics of soil erosion intensity (SEI) is essential for local governments. In this study, a simple method combination of the vegetation cover, gradient slope, and land use coupling index (vegetation and slope coupling index) model with field sampling unit schemes was explored to estimate the area variations of SEI between the years 1998 and 2009. The obtained results show that the minimum average prediction accuracy (PA) for sampling units was 94% among different land use types, whereas the maximum PA was 98.5%. Although PAs for different land use types showed discrepancies, the trend clearly depicted higher rates of sampling units and lower relative errors. The average fitting accuracies (FAs) of the sampling units were 98.84%, 97.92%, and 97.26%, respectively, based on different proposed strategies, whereas the FAs of the extrapolation results were 94.77%, 87.57%, and 92.41%, respectively. In addition, the extrapolation results were found to be less efficient than the sampling units. However, this is acceptable, considering the field observation workloads and time consumption. Therefore, this study provides a promising scheme for the rapid estimation of the area dynamics of SEI, which will be useful for estimating the SEI in other areas.