A New Approach of Assessing Soil Erosion using the Remotely Sensed Leaf area Index and its Application in the Hilly Area
Lin Jie, Zhang Jinchi*, Gu Zheyan, Chen Jiadong1, Lyu Heng2
Forestry College of Nanjing Forestry University, Key Laboratory of soil and water conservation and Ecological Restoration, Nanjing-210037, China 1Nanjing Branch of Jiangsu Provincial Hydrology and Water Resources Investigation and Survey Bureau, Nanjing, China 2Nanjing Normal University, Nanjing-210023, China
*Corresponding author E-mail: email@example.com
Vegetation cover is one of the most crucial factors determining soil erosion and is still used for the estimation of the C (the cover management) factor in erosion models such as USLE (Universal Soil Loss Equation) and RUSLE (Revised Universal Soil Loss Equation). However, soil erosion depends not only on the horizontal structure of vegetation, which is usually represented by vegetation cover, but also on its vertical structure. The LAI (Leaf Area Index), as an important structural parameter of the terrestrial ecosystem, can reflect both the horizontal and vertical structure of vegetation effectively. To improve the accuracy of assessing soil erosion, this study develops a new method of remote sensing quantitative inversion of the C factor based on LAI that can truly capture the structural characteristics of vegetation. The modeling results are compared with results based on the conventional input of NDVI (Normalized Difference Vegetation Index) and the horizontal vegetation structure. The results show that the RMSE (Root Mean Square Error) of soil loss based on the conventional input of NDVI and the horizontal vegetation structure is 3.98, while the RMSE of soil loss based on the LAI is 1.62. Using the RMSE criterion, method (b) may be considered to perform better than method (a) for the assessment of soil erosion. The simulation results are the same for the Oak forest and bamboo stands using method (a) and method (b) because these two stands are all multi-stratified forest, which have plant, shrub, litter, roots and so on, but the simulation results are more accurate using method (b) than method (a) in the Chinese fir and Pinus massoniana stands because these two stands are all pure coniferous forest with simple structure. This conclusion indicates that LAI can avoid the limitations of vegetation cover and can provide a better parameterization for evaluating the impact of vegetation on soil erosion when the stands are pure forest. If the stands are multi-stratified forest, evaluation of the impact of vegetation on soil erosion can use NDVI and LAI.