Presentation Type
Poster Presentation
Category
STEM (science, technology, engineering, mathematics)
Abstract/Artist Statement
The Great Lakes (Superior, Huron, Ontario, Michigan, and Erie) make up for one-fifth of the freshwater surface area on the Earth (NOAA, 2021). Displacements of the Earth’s surface are caused by quasi-static loads at or near the surface, including surface water, soil moisture and groundwater, and can be predicted by evaluating a convolution integral over the loaded region. In this study, we aim to infer changes in groundwater storage in the Great Lakes region of the central U.S. at sub-monthly time scales. Global Navigation Satellite System (GNSS) data from the Network of the Americas (NOTA; previously known as PBO) are used to estimate 3-D displacements of Earth’s surface (east, north, up) caused by surface loading. GNSS position series represent the superposition of many different signals, from which the groundwater loading signal must be isolated. We therefore model and remove predicted deformation due to soil moisture loading (NLDAS model), snow loading (SNODAS model), atmospheric-pressure loading (ECMWF model), non-tidal oceanic loading (MPIOM model), background global mass load changes outside of the Great Lakes region (GRACE model), and lake loading (NOAA model) from the GNSS position series. Predictions are computed using a Python-based toolkit called LoadDef to model the elastic deformation of the Earth caused by surface loading. Residual displacements between the observed GNSS measurements and the predicted surface deformation from LoadDef are assessed in the context of unmodeled or mismodeled loading sources, and other uncertainties in the GNSS analysis and loading predictions. The root-mean-square error (RMSE) reduction for vertical displacement ranges between -20 to 40 percent, suggesting that the “known” loading models cannot account for most of the scatter in the GNSS time series and that the residual series can be analyzed for signatures of groundwater deformation. Principal component one (PC1) from a principal component analysis (PCA) of 67 GPS stations accounts for 23.24% of the variability in the vertical residual displacements. We hypothesize that PC1 represents the groundwater fluctuations in the Great Lakes region.
Mentor Name
Hilary Martens
Tracking changes in groundwater storage from GNSS geodesy in the Great Lakes region
UC North Ballroom
The Great Lakes (Superior, Huron, Ontario, Michigan, and Erie) make up for one-fifth of the freshwater surface area on the Earth (NOAA, 2021). Displacements of the Earth’s surface are caused by quasi-static loads at or near the surface, including surface water, soil moisture and groundwater, and can be predicted by evaluating a convolution integral over the loaded region. In this study, we aim to infer changes in groundwater storage in the Great Lakes region of the central U.S. at sub-monthly time scales. Global Navigation Satellite System (GNSS) data from the Network of the Americas (NOTA; previously known as PBO) are used to estimate 3-D displacements of Earth’s surface (east, north, up) caused by surface loading. GNSS position series represent the superposition of many different signals, from which the groundwater loading signal must be isolated. We therefore model and remove predicted deformation due to soil moisture loading (NLDAS model), snow loading (SNODAS model), atmospheric-pressure loading (ECMWF model), non-tidal oceanic loading (MPIOM model), background global mass load changes outside of the Great Lakes region (GRACE model), and lake loading (NOAA model) from the GNSS position series. Predictions are computed using a Python-based toolkit called LoadDef to model the elastic deformation of the Earth caused by surface loading. Residual displacements between the observed GNSS measurements and the predicted surface deformation from LoadDef are assessed in the context of unmodeled or mismodeled loading sources, and other uncertainties in the GNSS analysis and loading predictions. The root-mean-square error (RMSE) reduction for vertical displacement ranges between -20 to 40 percent, suggesting that the “known” loading models cannot account for most of the scatter in the GNSS time series and that the residual series can be analyzed for signatures of groundwater deformation. Principal component one (PC1) from a principal component analysis (PCA) of 67 GPS stations accounts for 23.24% of the variability in the vertical residual displacements. We hypothesize that PC1 represents the groundwater fluctuations in the Great Lakes region.