Factors influencing storm-generated suspended-sediment concentrations and loads in four basins of contrasting land use, humid-tropical Puerto Rico

Abstract

The significant characteristics controlling the variability in storm-generated suspended-sediment loads and concentrations were analyzed for four basins of differing land use (forest, pasture, cropland, and urbanizing) in humid-tropical Puerto Rico. Statistical analysis involved stepwise regression on factor scores. The explanatory variables were attributes of flow, hydrograph peaks, and rainfall, categorized into 5 flow periods: (1) the current storm hydrograph, (2) the flow and rainfall since the previous storm event, (3) the previous storm event, (4) 2nd previous storm event, and (5) the 3rd previous storm event. The response variables (storm generated sediment loads and concentrations) were analyzed for three portions of the storm hydrograph: (1) the entire storm, (2) the rising limb, and (3) the recessional limb. Hysteresis differences in sediment concentration between the rising and falling limb were also analyzed using these explanatory variables. Sediment availability in the study basins is related to land use and underlying geology. The supply of sediment and its location in the watershed have a strong influence on how current and previous storm events, and flow and rainfall between storm events, affect sediment loads and concentrations. In basins with limited sediment availability (forest and pasture), previous storm events supply sediment to the channel. This in-channel sediment becomes the source of sediment on the rising limb of the next event and clockwise hysteresis occurs. In the cropland basin with high sediment availability, sediment delivered to the channel during events becomes the source of sediment on the rising limb of the next event and clockwise hysteresis occurs. In the urbanizing basin with high sediment availability, counterclockwise hysteresis is prevalent as stormwater runoff dilutes suspended-sediment concentrations on the rising limb and upland sediment arrives on the hydrograph recession. In the urbanized basin, previous storm events flush sediment from the system. The statistical approaches presented here can be used to generate hypotheses on the location and delivery of watershed sediment sources which can improve the design of appropriate field studies.