Date of Award


Document Type

Open Access Dissertation


Civil and Environmental Engineering

First Advisor

Jonathan L. Goodall


Hydrologic models are widely used to determine the occurrence and severity of droughts, and to aid in the decision making process for reducing drought impacts. These models apply approaches including water and energy balance equations to determine water availability at field to continental scales. One challenge in implementing hydrologic models at larger scales (regional to continental) is the lack of data with sufficient spatial and temporal resolution to represent actual conditions within the modeling domain. If such data is available, the challenge remains how best to find, access, and process the data so that it can be used within hydrologic models. The overarching goal of this work is related to these challenges of modeling regional-scale hydrologic systems and, more specifically, to quantify water storage within regional hydrologic systems by applying both models and observational datasets. Three research studies are presented that address specific challenges as they relate specifically to estimating high resolution, large scale water storage and drought severity at seasonal and interannual scales in South Carolina.

The first study adopts a water balance approach whereby the rate of change of water stored within sub-watersheds, defined as the drainage area between streamflow gaging stations, is estimated from observed streamflow and precipitation along with evapotranspiration estimates from North American Regional Reanalysis (NARR) program. Water storage estimates from the water balance approach were compared to groundwater level observations across the state as a means for evaluating the accuracy of the estimates. The results of this study provide insights into how water storage decreases due to drought conditions and, more importantly, the complexity of the spatial patterns in drought impacts across the state.

The second study builds on the first study by improving on estimates of evapotranspiration across the state, the most uncertain of the hydrologic fluxes used in the water balance framework. Two methods for quantifying evapotranspiration at regional scales were investigated as part of the work: simulation of evapotranspiration using the Variable Infiltration Capacity (VIC) macro-scale hydrologic model and remote sensing of evapotranspiration using Moderate Resolution Imaging Spectroradiometer (MODIS) derived estimates. The results of this analysis show that, when VIC generated evapotranspiration are applied within a water balance framework, estimates of changes in water storage match better with estimates of water storage from the Gravity Recovery and Climate Experiment (GRACE) Terrestrial Water Storage (TWS) anomalies product.

The third study advances on prior work by addressing methodological challenges associated with applying large-scale hydrologic simulation models. Specifically, the work presents an application of the VIC model within a data management cyberinfrastructure system to automate the necessary data gathering and processing steps. This cyberinfrastructure allows for the creation of a larger, more detailed, and better tested VIC model to analyze drought conditions in North and South Carolina river basins. The integration of VIC model data processing using the integrated Rule Oriented Data Systems (iRODS) within DFC data grid facilitates automation of the data collection, processing, curation, preservation, and replication of VIC input and output files. Finally, this improved methodology allows for easier sharing, reproducibility, and documentation of large-scale hydrologic simulation models.

In summary, the research presented in this dissertation advances knowledge of South Carolina water resources, specifically how drought impacts water availability within the State. It also advances hydrologic analysis of regional-scale systems through the application of data-driven water budget estimations. These types of analysis techniques are well-suited for state-wide water resources management, and are data intensive, but do not require sophisticated models. The final contribution of this work is an approach for building cyberinfrastructure needed to address data management challenges when applying state-of-the-art hydrologic models to regional-scale hydrologic systems. Given that data availability is rapidly advancing through the deployment of sensor networks and remote sensing instruments, approaches like this are needed to better techniques for handling the data life-cycle associated with performing hydrologic model simulations.