Impact of Artificial Reservoir Size and Land Use/Land Cover Patterns on Probable Maximum Precipitation and Flood: Case of Folsom Dam on the American River

Abstract

The design of the dams usually considers available historical data for analysis of the flood frequency. The limitation of this approach is the potential shift in flood frequency due to physically plausible factors that cannot be foreseen during design. For example, future flood extremes may change, among other factors, due to strong local atmospheric feedbacks from the reservoir and surrounding land use and land cover (LULC). Probable maximum flood (PMF), which is the key design parameter for hydraulic features of a dam, is estimated from probable maximum precipitation (PMP) and the hydrology of the watershed. Given the nonlinearity of the rainfall-runoff process, a key question that needs to be answered is How do reservoir size and/or LULC modify extreme flood patterns, specifically probable maximum flood via climatic modification of PMP? Using the American River Watershed (ARW) as a representative example of an impounded watershed with a large artificial reservoir (i.e., Folsom Dam), this study applied the distributed variable infiltration capacity (VIC) model to simulate the PMF from the atmospheric feedbacks simulated for various LULC scenarios (predam, current scenario, nonirrigation, and reservoir-double). The atmospheric feedbacks were simulated numerically as PMP using the regional atmospheric modeling system (RAMS). The RAMS-generated PMP scenarios were propagated through the VIC model to simulate the PMFs. Comparison of PMF results for predam and current scenario conditions showed that PMF peak flow can decrease by about 105 m3/s, while comparison of current scenario with nonirrigation PMF results showed that irrigation development has increased the PMF by 125 m3/s. On the other hand, the reservoir size had virtually no detectable impact on PMP and consequently on PMF results. Where downstream levee capacity is already underdesigned to handle a dam’s spillway capacity, such as for the case study, such increases indicate a likely impact on downstream flood risk to which any flood management protocol must adapt. The premise that modern dam design and operations should consider an integrated atmospheric-hydrologic modeling approach for estimating proactively potential extreme precipitation variation due to dam-driven LULC change is well-supported by this case study.