Uncertainty assessment of multi-parameter, multi-GCM, and multi-RCP simulations for streamflow and non-floodplain wetland (NFW) water storage

Abstract

In this study, we assessed the impacts of uncertainties arising from hydrologic model parameters and climate change data on streamflow and catchment-level non-floodplain wetland (NFW) water storage predictions for the Coastal Plain of the Chesapeake Bay watershed. The hydrologic model used in this study was the Soil and Water Assessment Tool (SWAT) coupled with improved wetland modules. Model uncertainty was represented using 12 parameter sets (PARs) with acceptable model performance. Different projections of future climate conditions (eight global circulation models [GCMs] under three representative concentration pathways [RCPs]) were used to represent climate change uncertainty. The ensemble method and analysis of variance were adopted for uncertainty assessment. The results showed that monthly streamflow projections did not substantially differ with respect to individual PARs, GCMs, and RCPs; by contrast, the projected NFW water storage varied significantly. However, the changes in projected hydrological values relative to historic conditions greatly differed with regard to the PARs, GCMs, and RCPs, leading to a high uncertainty in assessing climate change impacts. The variability of GCM projections was the most significant single contributor, accounting for 46% and 49% of the total streamflow and NFW water storage projection uncertainties, respectively, followed by PARs and RCPs. Our work assessed the impacts of different uncertainty sources on NFW hydrology under climate change, suggesting careful consideration of model and climate change uncertainties for the reliable projections of NFW hydrologic behaviors.