The correlation analysis on the landscape pattern index and hydrological processes in the Yanhe watershed, China

Abstract

Yanhe watershed, as a typical and experimental district of Soil and Water Conservation District, has long been plagued by soil erosion due to severe human disturbances. Exploring the relationship between watershed landscape pattern and hydrological processes can find effective ways to solve soil erosion problems. At first, with remote sensing and GIS (Geographic Information System) technology and based on SWAT model, this paper analyzed and simulated ecological hydrological processes in Yanhe watershed. It is on subbasin scale that the runoff and sediment yields were simulated monthly in Yanhe watershed using SWAT model. Secondly, it quantified landscape pattern with landscape indices. The seven landscape indices at the landscape level were selected with principal component factor analysis, including Disjunct Core Area Density (DCAD), Radius of gyration (GYRATE_SD), Patch Cohesion Index (COHESION), Shannon’s diversity index (SHEI), Total Core Area (TCA), Perimeter-Area Fractal dimension index (PAFRAC), Interspersion and Juxtaposition Index (IJI), etc. Thirdly, a new composite landscape index was constructed on the basis of eco-hydrological processes, which was closely related to soil erosion. The results are as follows: (1) Coupled analysis on the relationship of landscape indices and annual runoff as well as annual sediment yields in each subbasin, the correlation coefficient of seven selected landscape indices and runoff is very small, no passing all significant tests. But the correlation between sediment yields and the indices except for TCA and IJI is significant, and the absolute value of the correlation coefficient is between 0.3 and 0.5. (2) According to the “source-sink” theory of soil erosion, Slope-HRUs landscape index (SHLI) was built and can reflect the relationship between landscape pattern and soil erosion processes to a certain extent. The coupling relationship between Slope-HRUs landscape index (SHLI) and annual sediment yields in each subbasin is very clear, and correlation coefficient is −0.6, which is significantly negatively correlated.