The Brooks-Corey functions are commonly used in hydrologic models, with parameters obtained by fitting the functions directly to measured soil water retention data or by conversion methods from the van Genuchten functions which are continuous across the domain of matric suctions. Problems in fitting the BC functions directly to the retention data motivated use of the conversion methods. However, differences in converted parameters could significantly influence model predictions. We compared the direct fitting method and the conversion methods of Lenhard et al., Morel-Seytoux et al., and van Genuchten using measured water retention data during drainage and determined the influence of these methods on hydrological predictions when the converted parameters were used in the root zone water quality model. The conversion methods had significant influence on predictions of water retention, hydraulic conductivity, runoff, and evapotranspiration, with the observed level of significance (p≤0.006) much lower than the test level of significance (&agr;=0.05). The method of Morel-Seytoux et al. inadequately described measured water retention data (p=0.027), whereas the other two methods adequately described the data at relatively high suctions (p≥0.687), deviations occurred around the air-entry suction. The method of Lenhard et al. best reproduced the characteristics of the Brooks-Corey functions (p≥0.31) and could be used to obtain the Brooks-Corey parameters simply and reproducibly. ¿ 1999 American Geophysical Union