三峡库区小流域修正通用土壤流失方程适用性分析

土壤侵蚀量的定量研究可为国家生态环境建设和水土保持宏观决策的制定提供重要的依据。修正通用土壤流失方程(revised universal soil loss equation,RUSLE)是开展土壤侵蚀定量评价的主要手段。该文在地理信息系统(geographic information system,GIS)的支持下,依据中国土壤流失方程各因子的算法确定RUSLE模型各因子值,估算了三峡库区黄冲子小流域不同时期的土壤侵蚀量,并与基于泥沙平衡原理计算的土壤侵蚀量比较后分析RUSLE模型在库区小流域的适用性。结果表明,基于RUSLE模型估算的小流域1963-2000年(农地小流域)和2001-2014年(林地小流域)的年均土壤侵蚀模数分别为2246.09和868.3 t/(km2·a),其结果与采用137Cs和210Pb技术的塘库沉积物定年结果基本吻合,表明210Pb定年结果可靠。依据泥沙平衡原理计算的小流域1963-2000年和2001-2014年的年均土壤侵蚀模数分别为942.48和811.47t/(km2·a)。RUSLE模型估算小流域1963-2000年和2001-2014年的土壤侵蚀模数相对误差分别为138.32%和7.00%。因此RUSLE模型适用于库区林地小流域,而不适用于库区农地小流域;但是基于地形因子(LS因子)修正的RUSLE模型估算结果相对误差减少至8.14%,其适用于库区农地小流域。

Adaptability analysis on revised universal soil loss equation in small catchment of Three Gorges Reservoir Region

Abstract: Soil erosion by water is a serious problem in southwestern China, particularly in the Three Gorges Reservoir Region (TGRR), which is one of the regions prone to desertification. Soil erosion in the TGRR not only affects soil quality, in terms of agricultural productivity, but also reduces the capacity of flood storage and the projected life span of the Three Gorges Reservoir. The revised universal soil loss equation (RUSLE) has been widely deployed to quantitatively assess soil erosion by water from a small catchment based on a grid cell basis. This study was conducted on the Huangchongzi small catchment in TGRR, to predict different periods' annual soil erosion modulus using the RUSLE model, and to analyze its adaptability with the help of a sediment budget approach. According to the study area data and the calculation methods for each factor of the Chinese soil loss equation applied in the first national water conservation survey, each factor value of RUSLE was determined. The results from the RUSLE model indicated that the annual soil erosion modulus for 1963-2000 (agricultural catchment) and 2001-2014 year (forested catchment) in the Huangchongzi small catchment was 2246.09 and 868.30 t /(km2·a), respectively. The latter was obviously smaller than the former due to the implementation of the Grain-to-Green project, converting slope croplands into forest or grassland. The sediment yields for the 1963-2000 and the 2001-2014 years in the study catchment were 1228.71 and 322.71 t, respectively when 137Cs and 210Pb dating methods had been used as chronometers for sediment deposition in the pond. More importantly, 210Pb-derived dates corresponded well with the results from 137Cs geochronology for pond sediment cores, and this indicated that the dating result by 210Pb technology was correct and reliable. The annual sedimentation amount in the paddy fields was estimated to be 27.11t/a on the basis of the 137Cs tracer method. Therefore, the sediment deposition amount for the 1963-2000 and the 2001-2014 years in the paddy fields were 1003.07 and 352.43t, respectively. Based on the sediment yields and deposition amounts, sediment budgets of the 1963-2000 and the 2001-2014 years for the catchment had been constructed, respectively. The results from the sediment budget approach showed that the annual soil erosion modulus for the 1963-2000 and the 2001-2014 years were 942.48 and 811.47t/(km2·a), respectively. These calculations were directly compared with the RUSLE estimations in different periods since there was no gully and channel erosion in the study catchment. Results showed that the soil erosion modulus by sediment budget approach was reasonably consistent with that provided by RUSLE during the 2001-2014 years; this indicated that the adaptability of RUSLE was good in the forested catchment. However, the soil erosion modulus estimated by RUSLE for the study catchment was much higher than the soil erosion modulus obtained by the sediment budget approach during the 1963-2000 years and its relative error was as high as 138.31%. These findings highlighted that the adaptability of RUSLE was very poor in the agricultural catchment. Nevertheless, the soil erosion modulus for the 1963-2000 years estimated by the revised RUSLE in the study catchment was 1019.18 t/(km2·a), and the error between the revised RUSLE soil erosion rates and that from the sediment budgeting approach decreased to 8.14%. This conclusion demonstrated that the method for correcting the LS factor value proposed by this study in the agricultural catchment was efficient and feasible. In brief, RUSLE with localization parameters can directly evaluate soil erosion in a forested catchment of the TGRR; on the contrary, its estimation error was very large in the agricultural catchment when RUSLE was directly utilized. It was necessarily to revise the LS factor value first and then to assess soil erosion in the agricultural catchment using the revised RUSLE in the TGRR. This study can provide beneficial references for the correct utilization of the soil erosion model and optimum utilization of the soil and water resources in the TGRR.