An Integrated Approach for Estimating Groundwater Storage Variations in Regional Aquifers, 20-R8051
Inclusive Dates: 04/01/09 – 12/01/10
Background — This project tackled a strategically important research area related to using Earth observation products for societal benefit. The availability of adequate and safe drinking water supplies has been and will continue to be critical. However, water resources are not evenly distributed. Some countries are water rich, while others are water poor. In recent years, remotely sensed data have been used extensively to monitor surface and near-surface components of the water cycle. This project evaluated the potential use of precision gravity data to measure temporal and spatial changes in the storage of water in aquifers.
Approach — The technical objectives of this project were to develop and test an integrated approach for combining GRACE (Gravity Recovery and Climate Experiment) satellite data with in-situ measurements and a land surface model to better constrain estimates of the water budget in regional aquifers; and demonstrate the utility of the developed approach for managing the interconnected Edwards-Trinity Plateau and Pecos Valley aquifers system in west Texas. The aquifers cover a combined area of 115,000 km2. At the global level, several worldwide initiatives focusing on Earth observations and the hydrologic cycle have been launched. Characterizing the storage and fluxes of sub-surface water has proven less tractable to typical satellite methods. A noticeable and important exception is GRACE, a joint satellite mission launched in March 2002 by NASA and its German counterpart, Deutsches Zentrum für Luft-und Raumfahrt. GRACE comprises two identical satellites at about 500 km altitude, separated by about 220 km, in identical near-polar orbits. GRACE measures Earth's gravity field and its changes over time using range-rate perturbations between the two satellites sensed with a microwave interferometer. Over land, and accounting for atmospheric circulation, changes in the gravity field are mainly attributed to temporal variations in terrestrial water storage (TWS), which is a vertically integrated measure of water storage that includes groundwater, soil moisture, surface water, snow water equivalent and biomass water. Hence, GRACE gravity data can be used to monitor temporal (near-monthly) TWS variations.
Accomplishments — This project resulted in three main technical achievements. First, SwRI researchers, in collaboration with NASA and the National Center for Atmospheric Research, developed an algorithm for estimating aquifer storage parameters using GRACE data. Previous studies focused mainly on validating GRACE observations. This study not only showed the responses of GRACE signals to TWS in the study area, but also demonstrated for the first time the utility of GRACE data for solving groundwater inverse problems. Second, a multiobjective optimization strategy was developed to calibrate regional groundwater models using GRACE data by adjusting hydraulic conductivity, specific yield and recharge multiplier simultaneously. Finally, a webGlS prototype application was developed to visualize Groundwater Availability Model (GAM) results online.
Aquifer storage parameters play an important role in transient groundwater flow simulations and in water resources planning. However, determining representative storage values using traditional techniques (e.g., pumping test) is challenging, especially at the regional scale. To address this problem, a robust optimization method for estimating aquifer storage parameters (specific yield or storativity) using the GRACE data, in-situ well level observations and other ancillary information was developed. Uncertainty inherent in the remotely sensed and in-situ time series can adversely affect the parameter estimation process and, in the worse case, make the solution completely meaningless. An estimation problem was formulated to directly minimize the negative impact of data uncertainty by incorporating bounds on data variations. This method is demonstrated for the interconnected Edwards-Trinity Plateau and Pecos Valley aquifers in central Texas. The study area was divided into multiple zones based on the geology and monitor well coverage. The estimated aquifer storage parameters are consistent with previous results obtained from pumping tests and model calibration, demonstrating the potential of using GRACE data for validating regional groundwater model parameters.
Previous GRACE studies highlight the importance of identifying distributed storage values instead of assuming a uniform global value for all wells in the study area. This work complements such a goal. Finally, the importance of maintaining high-quality in-situ monitoring networks is emphasized, which will significantly help extend the benefits of GRACE data for groundwater resources analyses.