Modeling washoff of total suspended solids in the tropics

Washoff behavior in the tropics is expected to behave differently from temperate areas due to differences in rainfall characteristics. In this study, rainfall, runoff and total suspended solids (TSS) were monitored from 9 catchments distinguished by different types of land use, in Singapore. The catchments ranged in size from 5.7ha to 85.2ha. Over 120 rain events were studied and more than 1000 storm samples were collected and analyzed. Monte Carlo analysis was applied to obtain the best fit values of the washoff model parameters consisting the washoff coefficient c₃, washoff exponent c₄ and initial mass on surface B_ini. The exponent c₄ was found to be approximately unity for all the events monitored, in agreement with other studies. The values of c₃ and B_ini were found to vary between events. Among all the rainfall and runoff characteristics studied, rainfall depth of the current event (d) was found to be the single parameter that significantly influenced the values of c₃ and B_ini. Contrary to expectations, B_ini did not correlate well with antecedent dry period or with rainfall depth of the prior storm event. The results show that the common modeling practice where B_ini is assumed to vary with antecedent dry period and previous rainfall depth should be reassessed when applied to catchments in the tropics. ANCOVA analysis showed that land use was not significant, but rather the variation of c₃ and B_ini with d was found to correlate well with the catchment area.