List of Tables

• 1.1 Recent studies attempting to estimate a balanced water budget via remote sensing observations.
• 2.1 Datasets compiled for the four major components of the water cycle.
• 2.2 GRACE datasets available for total water storage.
• 2.3 GRACE observations begin and end dates, showing the errors in the begin date and duration compared to calendar months.
• 2.4 CAMELS runoff data sources that may be used in future studies.
• 2.5 Ancillary environmental data compiled at the river basin and pixel scale for use as inputs to an NN model.
• 2.6 Global remote sensing-based vegetation datasets.
• 3.1 Statistics for basin masks.
• 4.1 Comparison of the Nash-Sutcliffe Efficiency, NSE, and the Coefficient of Determination, R².
• 4.2 Assumptions necessary for the purposes to which ordinary least squares (OLS) regression is applied.
• 4.3 Ladder of powers.
• 4.4 Percentage of observations among pixel-scale EO variables that are outside the range of basin-averaged training dataset.
• 5.1 Summary of the cross-validation experiment for surface fitting methods.
• 5.2 Experimental neural network configuration trials.
• 5.3 Potential utility of the ancillary environmental variables.
• 5.4 Statistical significance for the effect of the ancillary variables on the NN calibration model.
• 5.5 Summary of the fit of calibrated EO data to the OI solution.
• 5.6 Evaluation of the NN model predictions for P, E, and R.
• 5.7 Advantages and disadvantages of the two main methods for water cycle closure explored in this thesis.
• 6.1 Goodness of fit to evapotranspiration estimated by various methods.
• 6.2 Goodness of fit to GRACE observations for total water storage change chapter estimated indirectly by the water-budget method at the pixel scale.
• 6.3 Goodness of fit between runoff estimated indirectly by the water-budget method and observed river discharge at 1,781 river gages.
• 6.4 Fit statistics for monthly runoff for the Mississippi River at Vicksburg calculated from EO datasets, pre- and post-calibration by the NN model.
• A.1 Global precipitation datasets.
• A.2 Global evapotranspiration datasets.