Laboratory experiments have been conducted to study the effects of various simulated rain properties on soil detachment due to raindrop impact. Two soil materials were tested: a fine sand and a silt loam. The mass of detached sediment was measured with the splash cup technique. The splash cups filled with soil material were exposed to simulated rain intensities ranging between 10 and 140 mm h-1. The detached sediment was collected and weighed, whereas rain intensity, drop diameter, and fall velocity of each raindrop were measured with an optical spectropluviometer. A statistical analysis has been made in order to evaluate which rain parameter best predicts the splash detachment rate. Linear fits between the rate of detached sediment and the product of drop size (D) and drop velocity (V), that is, D&agr;V&bgr; with values of &agr; varying between 1 and 6 and values of &bgr; varying between 0 and 2, have been computed. The results indicate that values for &agr; ranging between 3 and 6 and for &bgr; smaller than or equal to 2 best describe the rate of detached sediment. Although all indices containing the mass of water (i.e., D3) predicted splash detachment rates relatively well, the product of momentum and drop diameter (D4V) was slightly superior in describing splash detachment. The existence of a rain erosivity threshold (critical kinetic energy) to initiate soil detachment is confirmed. This threshold energy equals 5 &mgr;J for the fine sand and 12 &mgr;J for the silt Ioam and compares relatively well with the threshold energy data for other soils reported in the literature. ¿ 2000 American Geophysical Union