UC Santa Barbara

Understanding the rainfall-runoff patterns has been a central focus in watershed hydrology. This dissertation investigates the effects of climate variability, wildfire, and subsurface storage on rainfall-runoff responses of small watersheds in coastal California under a Mediterranean climate. I examine orographic rainfall distributions calculated with multiple spatial interpolation methods, which rely on topographic information, and assess the performance and bias of these methods. I present evidence that geographic weighted regression (GWR) performs well in generating spatial distributions of event-based rainfall.I use time series analysis to examine event-based rainfall and runoff data under the impacts of synoptic climate variability (atmospheric rivers; AR, Madden Julian Oscillation; MJO, El Niño- Southern Oscillation; ENSO). The inter-relationships of synoptic climate variability indicate that ENSO phases are the dominant factor in altering AR-associated large-rainfall events in the wet seasons (Nov.-Apr.), especially in wet years. Wavelet power spectra in the runoff data indicate that large floods, generated under the AR events and MJO phases, coincide in time with strong El Niño events.I assess post-fire hydrologic responses using two metrics, differences in event-based storm runoff and flow duration curves (FDC) and event-based rainfall-runoff regression models. I present several findings during post-fire periods: large inter-storm and inter-annual variability of rainfall and runoff, declining differences in event-based runoff, and a downward shift of annual FDCs. These observations can be attributed to precipitation variability, and large coverage of bedrock within the burned areas. Large variation in high flows and similar values of low flows evident in annual post-fire FDCs reflect the runoff response to fire is smaller than the effects of inter-annual variation in rainfall during the post-fire period. I identify water storage dynamics in soil and fractured bedrock using simple water balance models. Recession constants generated from the linear reservoir model indicate that water released from soils and bedrock behaves similarly in four selected watersheds. Water stored in fractured bedrock zones remains slightly changed in the dry seasons. However, during the progression of wet seasons, it may react quickly to rain storms, and lead to large changes to groundwater storage.