The Southwest U.S. comprising of the four states -- Arizona, New Mexico, Colorado, and Utah -- is the hottest and driest region of the United States. Most of the precipitation arrives during the winter season, but the summer precipitation makes a significant contribution to the reliability of water resources and the health of ecology. During summer, this region receives rainfall from convective storms, tropical cyclones, and northern intrusions of the North American Monsoon (NAM) system. The NAM is centered over northern and central Mexico with robust climatological features and interannual variability and has been studied extensively. However, its northern extensions over the southwest U.S in comparison, are weak in their features, thus making the summer rainfall over this region highly variable. Consequently, the skill in forecasting the summer rainfall is low and resources managers -- water and ecological -- ignore this in their planning and operation decisions. However, the summer hydroclimate over southwest U.S. exhibits distinct space-time variability independent of NAM. The importance of summer hydroclimatology and this research gap motivates this dissertation.

Four research questions drive this systematic inquiry of summer hydroclimate variability. (i) What are the moisture sources and pathways that deliver moisture to the summer precipitation in southwestern U.S.? (ii) What is the spatiotemporal variability of the summer season and sub-season precipitation and extremes? And their teleconnections to large scale climate drivers? (iii) What are the patterns of variability of summer streamflow in the river basins of this region? (iv) How do climate models simulate summer precipitation? Four unique contributions emerge from this dissertation upon pursuing these questions. (i) Three major moistures sources - Gulf of California (GoC), Gulf of Mexico (GoM), and land -- were identified for the summer rainfall in the Southwest U.S. The GoC is the dominant source for southern and western Arizona; GoM for eastern New Mexico and these two sources contributed to the region in between. The land source is dominant for rainfall over eastern Colorado and Utah, indicative of moisture recycling. Rainfall from the trajectories originating from GoC showed a decreasing trend while those from land increased during the period 1979-2013. (ii) Dominant modes of spatial and temporal variability of seasonal and sub-seasonal rainfall and extremes over southwest U.S were found to be highly correlated -- indicating that the mean and extremes of rainfall are modulated by similar mechanisms. Leading modes of early season (June) and late season (September) rainfall exhibited correlation with sea surface temperatures in the tropical Pacific sea surface temperatures (SSTs), while the peak season (July-August) had no links to large-scale SSTs. This suggests that the interannual variability of the early and late season has origins to large scale climate features and the peak season rainfall is driven primarily by regional and local scale convection, along with moisture recycling from the land. Furthermore, wet years showed anomalously high moisture transport from the Gulf of California. (iii) The three leading modes of seasonal and sub-seasonal streamflow were found to delineate Upper Colorado River Basin (UCRB), Rio Grande River Basin, and the Salt-Verde River Basins in space and their temporal variability exhibited interannual variability. These modes showed strong correlations with precipitation in their respective basins. The weighted flows of UCRB exhibited high correlations with central Pacific SSTs, and the Salt-Verde flows in the eastern Pacific, albeit, weaker. The Rio-Grande flows showed correlations with SSTs in the Gulf of Mexico and Atlantic Ocean. These findings suggest that the interannual variability of summer streamflow in the southwest U.S. has their origins in the global oceans. (iv) Lastly, a systematic analysis of 27 climate model simulations driven by observed SSTs during the period 1979-2005 was performed. The models simulate the broad climatological features of summer precipitation over this region, but they over simulate the amounts in the eastern parts of the region -- notably, Colorado, and under simulate in the western part. This pattern is persistent through the sub-seasons. Based on a suite of metrics, four models were identified to show good overall performance in simulating the summer rainfall - CMCC-CM, GFDL-HIRAM-C180, HadGEM2-A, and MPI-ESM-MR. The leading modes of variability from these models and the combined mean rainfall showed a stronger relationship with tropical Pacific SSTs as compared to the observed rainfall, indicating their sensitivity to El Nino Southern Oscillation forcings.

The findings from this dissertation offer new insights into the spatial and temporal variability of hydroclimate in the Southwest U.S. They brighten the prospects for improving the forecast skill of summer seasonal and sub-seasonal precipitation and streamflow -- consequently, contributing to efficient management water resources and ecosystems.