Improving assessments of water resources characterization and relationships to climate variabilities

With the increasing vulnerability of water resources due to climate change and the associated more frequent and intensive extreme events, there is a growing need to monitor the spatial and temporal variability in water resources. Coastal aquifers are a major source of freshwater for almost 40% of the world’s population living in coastal regions. Changes in groundwater level impact groundwater discharge to surface waters, which in turn affects the amount of streamflow resulting in variability in freshwater inflow to estuaries. However, lack of continuous spatial and temporal coverage of groundwater elevation data hinders direct measurements of groundwater storage (GWS). The first objective of this dissertation is to quantify the spatial and temporal variation in coastal GWS using uninterrupted monthly Gravity Recovery and Climate Experiment (GRACE) derived terrestrial water storage (TWSGRACE) utilizing the Texas Gulf Coast Aquifer as a benchmark. To accomplish this, the data gap in TWSGRACE was filled with reconstructed values from multi-linear regression (MLR) and artificial neural network (ANN) models. The reconstructed TWSGRACE was integrated with land surface model products and ground-based measurements to examine the long- and short-term trends in GWS in response to changing climate conditions. This study found a significant decline (0.35 ± 0.078 km3·yr?1, p-value: < 0.01) in GWS during the study period 2002-2019 and was able to capture extreme climate events (i.e., drought and flood). The second objective provides an innovative approach to deriving a continuous TWSGRACE record for improved evaluation of GWS at a global scale. The approach integrates three machine learning techniques (deep-learning neural networks [DNN], generalized linear model [GLM], and gradient boosting machine [GBM]) and eight climatic and hydrological input variables to fill GRACE record gaps and reconstruct the TWSGRACE data record at both global grid and basin scales. This study’s models’ performances were found to be superior to those from 61% of previous similar studies and comparable to 21%. The reconstructed TWSGRACE data captured the occurrences of extreme hydro-climatic events over the investigated basins and grid cells. In addition, the third objective evaluates the impact of individual and coupled ocean-atmospheric phenomena (El Niño Southern Oscillation [ENSO], Pacific Decadal Oscillation [PDO], and Atlantic multidecadal oscillation [AMO]) on hydrological variables (i.e., precipitation, streamflow, and freshwater inflow) across several major estuaries located along the northwestern Gulf of Mexico (nGOM), which span a significant climatic gradient. Results show that the individual and coupled climate variability phenomena have a significant impact on all three hydrological variables and the magnitude of impact varies seasonally and spatially, with the strongest modulation on cold seasons and in the wet region (of Texas). The severity of droughts increases significantly when the La Niña phase of ENSO is coupled with the PDO/AMO cold/warm. This dissertation 1) offers a reliable approach to examine the long- and short-term trends in GWS, 2) provides a robust and effective approach to fill the data gaps in TWSGRACE that can serve as a reference tool to fill the data gaps in any hydrological system across the globe, and 3) improves the understanding of the relationship between estuarine hydrology and climate variability. The results from these studies serve as valuable tools for water management strategies and benefit water resources planning and disaster management in coastal areas.