Description

The source park of the Yellow River (SPYR), as a vital ecological shelter, is suffering different degrees of degradation and desertification, resulting in soil erosion in recent decades. It is urgent to find a soil erosion model suitable for the SPYR to grasp the amount and spatial distribution pattern of soil erosion, and to provide data support for soil erosion zoning and grading management. Base on largescale systematic sampling, 521 measured soil 137Cs data and environmental variables affecting soil erosion were used to verify the applicability of the Revised Universal Soil Loss Equation (RUSLE) and the Revised Wind Erosion Equation (RWEQ) models, meanwhile a custom was validated and compared. Then the optimal model was obtained to estimate the spatial and temporal distribution and dynamic changes of soil erosion in the SPYR. The results show that: 1) The international classical soil erosion models (RUSLE and RWEQ) are not suitable for simulating soil erosion in the SPYR. 2) A custom model by combining random forest algorithm with 11 variables screened by LASSO algorithm achieves the highest accuracy, representing the optimal model. 3) The average soil erosion modulus from 2001 to 2020 was 19.32 t ha-1 y-1, and the total annual average erosion loss was 225.18×106 t y-1 in the SPYR. Spatially, it gradually decreases from southeast to northwest, with the most serious in Qumalai, Maduo and Chengduo Counties. Among the past 20 years the soil erosion has been alleviated. The improved area accounts for 83.43% of the total area of the SPYR. 4) The annual average economic loss caused by soil erosion is 5.196 billion Yuan, accounting for almost 25.6% of the GDP of same year of the SPYR. A new grassland zoning management strategy is proposed for implement graded management of alpine meadow to balance grassland ecological and productive functions.

Share

COinS
 

A New Soil Erosion Model for Alpine Meadow in the Source Park of the Yellow River, Qinghai-Tibet Plateau, China

The source park of the Yellow River (SPYR), as a vital ecological shelter, is suffering different degrees of degradation and desertification, resulting in soil erosion in recent decades. It is urgent to find a soil erosion model suitable for the SPYR to grasp the amount and spatial distribution pattern of soil erosion, and to provide data support for soil erosion zoning and grading management. Base on largescale systematic sampling, 521 measured soil 137Cs data and environmental variables affecting soil erosion were used to verify the applicability of the Revised Universal Soil Loss Equation (RUSLE) and the Revised Wind Erosion Equation (RWEQ) models, meanwhile a custom was validated and compared. Then the optimal model was obtained to estimate the spatial and temporal distribution and dynamic changes of soil erosion in the SPYR. The results show that: 1) The international classical soil erosion models (RUSLE and RWEQ) are not suitable for simulating soil erosion in the SPYR. 2) A custom model by combining random forest algorithm with 11 variables screened by LASSO algorithm achieves the highest accuracy, representing the optimal model. 3) The average soil erosion modulus from 2001 to 2020 was 19.32 t ha-1 y-1, and the total annual average erosion loss was 225.18×106 t y-1 in the SPYR. Spatially, it gradually decreases from southeast to northwest, with the most serious in Qumalai, Maduo and Chengduo Counties. Among the past 20 years the soil erosion has been alleviated. The improved area accounts for 83.43% of the total area of the SPYR. 4) The annual average economic loss caused by soil erosion is 5.196 billion Yuan, accounting for almost 25.6% of the GDP of same year of the SPYR. A new grassland zoning management strategy is proposed for implement graded management of alpine meadow to balance grassland ecological and productive functions.