Paper No. 46-7
Presentation Time: 11:05 AM
ASSESSING THE EFFECT OF RAINFALL DATA SCARCITY ON DAILY STREAMFLOW SIMULATION IN SPATIALLY HETEROGENEOUS WATERSHEDS
Rainfall, the main input factor in watershed models, shows high spatial gradient over short distances in Pacific islands. Evaluating the effect of rainfall data on watershed models performance is thus essential for water resources and climate change impact assessment in Pacific island, where limited climatic data are prevalent. We utilized the Soil and Water Assessment Tool (SWAT) model for streamflow modeling in Hawaii and American Samoa. SWAT performance comparison on daily streamflow simulations was carried-out between watersheds that had relatively spatially well distributed rainfall gauging stations within the watershed and those that had rainfall data derived based on records from the surrounding watersheds. After sensitive parameters were identified, the model was calibrated and validated by using measured daily streamflow data at multiple flow gauging stations. Several statistical evaluation criteria, including model prediction uncertainty at 95% confidence interval (95PCI) were used for model accuracy assessment. SWAT showed unsatisfactorily to satisfactorily performance in simulating daily streamflow for those watersheds that had rainfall data from neighboring areas. On the other hand, the model provided good to very good results for those watersheds that had measured rainfall data inside the watershed. In addition, for the latter watersheds, percent of observations bracketed at 95PCI was more than double compared to those watersheds that had rainfall data from outside, signifying a good representation of observations for watershed with relatively well represented rainfall values. Although SWAT performed less for watersheds without rain records, the results were reasonably acceptable. The developed model was used for climate change and variability studies. This study concluded that with methods to resolve data scarcity issues and careful statistical evaluation criteria, data from neighboring areas can be used for watersheds without rain records.