Abstract: The runoff and sediment yield of typical grassland slope is unique, while the erosion process is complex. It is of great significance to quantify the contribution of the factors affecting the runoff and sediment yield on typical grassland slope. In this study, we carried out field experiments to examine runoff and sediment yield on a typical grassland under artificial simulated rainfall conditions, with three rainfall intensities (30, 60, 90 mm·h^-1), three slopes (4°, 9°, 14°), and three vegetation coverages (0, 35%, 66%). The effects of vegetation coverage and slope on initial runoff time and total sediment yield were quantitatively analyzed under different rainfall intensities. The results showed that the total runoff and sediment yield were positively correlated with rainfall intensity and slope, but negatively correlated with vegetation coverage. The pattern of the initial runoff time was opposite to that of the total runoff and sediment yield. Rainfall intensity, vegetation coverage and slope had significant effects on initial runoff time, total runoff, and total sediment yield (P<0.05). The initial runoff yield time was mainly controlled by rainfall intensity, with a contribution rate of about 60%. Vegetation coverage was the main factor affecting the total runoff, with a contribution rate of more than 75%. Slope was closely related to total sediment yield, with a contribution of more than 55%. The runoff sediment concentration model was established based on the effects of rainfall intensity, slope and vegetation coverage on the runoff process of typical grassland. The runoff sediment concentration model could be described by a three element exponential function, with a determination coefficient of 0.824 and a significance level of less than 0.001, implying good simulation accuracy. This study provides a reference for the simulation and prediction of total runoff and sediment yield on typical grasslands.

Key words: typical grassland, soil erosion, slope runoff, artificial rainfall, runoff sediment concentration model.