This dissertation proposes a new numerical model-based methodology to estimate design precipitation sequence and Maximum Precipitation (MP) for long durations on the order of several months over a target region. The estimation of design precipitation for long durations is crucial especially for the regions where extreme floods are caused by long-term hydrological and meteorological processes, such as snow accumulation/melting or a series of extreme precipitation events, as they can all contribute to increasing volumes in reservoirs, leading to large water releases. The proposed approach using the Weather Research and Forecasting (WRF) Model successfully maximized the historical precipitation sequences for the 6-months wet season (i.e., winter season) over large drainage areas in the Columbia River Basin (MP being 961.0 mm for Bonneville Dam drainage area and 1101.7 mm for Libby Dam drainage area). As a result of maximizing precipitation of the identified 115 Atmospheric River (AR) storm events, the change in precipitable water as well as the integrated water vapor transport, was found to be not necessarily consistent with the change in precipitation depths over the basin, which questions the traditional assumption in Probable Maximum Precipitation estimations. Also, the atmospheric boundary condition shifting method was suggested to be an essential method in AR-induced precipitation maximization, rather than by solely increasing atmospheric moisture in model-based approaches, especially for long-duration MP estimates. In this dissertation, we also attempt to better understand the upper tropospheric flow and moisture transport control on fire season precipitation and its deficit, and to provide an insight on how to estimate precipitation deficit conditions during the wildfire season in the US Northern Rockies. Logistic regression analysis and composite analysis showed that the wetting rainy days, which exerted the strong control on wildfire area burned over the region during 1984-2018, may be well explained based on 500mb geopotential height (Z500) anomaly over the Pacific Northwest region and integrated water vapor transport (IVT) anomaly over the US Northern Rockies. In our numerical experiment using the WRF model, fire season precipitation over the US Northern Rockies and Z500 over the Pacific Northwest region were simulated well for the target events. Our numerical experiment suggested a possibility of estimating extreme wetting rain deficit conditions over the US Northern Rockies based on geospatially shifting atmospheric boundary conditions as precipitation depths and wetting rainy days decreased for the target events. Moreover, a numerical experiment increasing the sea surface temperature at the middle and southwest sectors of North Pacific Ocean resulted in changing the Z500 pattern, leading to the decrease in wetting rainy days over the Northern Rockies in a target year. Overall, this dissertation attempts to provide useful insights for estimating physically-based extreme hydrometeorological conditions in the northwestern US region toward regional and seasonal disaster risk assessments.

Water balance and drought analysis are crucial practices to reduce the effects of hydro-meteorological risks on human society in an economically and environmentally sustainable manner. Water withdrawal and water supply are changing dramatically in Vietnam due to the escalating pressure from socioeconomic development, environmental requirements, and climate change impacts. The IPCC has identified Vietnam to be amongst the countries most affected by climate change, predicting an increase in the frequency and magnitude of natural disasters, especially droughts and floods. Water resources in Vietnam are highly dependent on external sources, with over 60 percent of the total average surface water discharge generated outside of the country. The reliance on external sources poses an alarming threat of water scarcity, requiring immediate actions to cope with droughts and increase the resilience of Vietnam’s water resources. Thus, assessment of drought conditions over transboundary river regions is considered one of the most crucial tasks of the Government of Vietnam. The Lo River watershed (LRW), a tributary of the Red River watershed, is a transboundary watershed with a catchment area of 39,000 km2. The watershed is composed of upstream and downstream regions. The upstream region, located in China, accounts for 52% of the total area, while the downstream region, located in Vietnam, accounts for 48% of the total area. The main goal of this study is to assess drought conditions in the transboundary watershed LRW, based on long-term projected water supply and water demand under different 21st century climate change scenarios. The projected water supply is obtained by coupling a hydroclimate model whose inputs are provided by GCM projection outputs. Water demand, including municipal and industrial (M&I), environmental (E), and agricultural (AG), is collected from reliable sources. Future municipal and industrial (M&I) water demand and environmental water demand are collected from the Vietnamese Ministry of Natural Resources and Environment (MONRE), and the Ministry of Agriculture and Rural Development (MARD), respectively. Before projecting water supply, bias correction methods are investigated and applied. Bias correction combined with outputs from the hydroclimate model provide realistic flow data for the downstream region. Once projected water supply and water demand are obtained, a drought analysis is applied over the target watershed. Results of the drought analysis show future drought events are projected to be more intense and severe than past events. Increased understanding of the impact of climate change on future water resources aids policymakers in developing more effective plans to increase the resilience to climate change and sustain the water security of the region and the country.

This study developed a hydrological model for the Upper Trinity Watershed in California. The Upper Trinity Watershed consists of the portion of the Trinity Watershed upstream of Lewiston Dam and contains both Lewiston and Trinity Lakes. The Upper Trinity Watershed is an important source of water supply in California. Lewiston Lake is used for inter-basin water transfer to the Central Valley region of California for agriculture and municipal use. The watershed model was developed using the Watershed Environmental Hydrology Hydroclimate Model (WEHY-HCM). WEHY-HCM was previously developed over many years in the Hydrological Research Laboratory at UC Davis. WEHY-HCM is a physically-based, numerical, integrated, and distributed modeling system and integrates atmospheric, snowmelt, surface and subsurface flow, and hydraulic processes. WEHY-HCM combines physical equations with spatial relationships and observations such as land use/cover, soil type, and digital elevation model datasets. The primary model output in this study is the streamflow into the Trinity Lake reservoir. The model was calibrated using water years (Oct 1st-Sept 30th) 1997-1999 and was validated using water years 2000-2006. In order to calibrate this model, a program was developed using the evolutionary optimization algorithm differential evolution. The results show that a physically-based, integrated, distributed model is an accurate and effective method for hydrological modeling for the Upper Trinity Watershed. While calibration using a sophisticated algorithm can produce acceptable model results, it was comparable to that obtained by manual calibration in other studies. Automated calibration can produce time savings compared with a manual calibration approach.

Groundwater storage in aquifers has become a vital water source due to the water scarcity inthe last years. However, aquifer systems are full of uncertainties, which inevitably propagate throughout the calculations, mainly reducing the reliability of the model output. This study develops a novel two-dimensional stochastic confined groundwater flow model. The proposed model is developed by linking the stochastic governing partial di↵erential equations through their one-to-one correspondence to the nonlocal Lagrangian-Eulerian form of the Fokker- Planck equation (LEFPE). In the form of the LEFPE, the resulting deterministic governing equation describes the spatio-temporal evolution of the probability density function of the state variables in the confined groundwater flow process. This probability evolution is performed by one single numerical realization rather than requiring thousands of simulations in the Monte Carlo simulation. Consequently, the ensemble groundwater flow process’s mean and standard deviation behavior can be modeled under uncertainty in the transmissivity field and recharge and/or pumping conditions. In addition, an appropriate numerical method for its solution is subsequently devised. Then, LEFPE’s solution is presented, discussed, and illustrated through numerical examples. Furthermore, they are compared against the results obtained through the Monte Carlo simulations to evaluate the performance of the LEFPE framework. Results suggest that the proposed model appropriately characterizes the ensemble behavior in confined groundwater systems under uncertainty in the transmissivity field.