Shallow Groundwater Estimation (SAGE) for Groundwater Dependent Ecosystems in NV Battle Mountain District

Using a machine learning algorithm developed by TNC-CA (see Rohde et al. 2021), we applied the Shallow Groundwater Estimation (SAGE) tool to estimate groundwater level trends for GDEs in Nevada. The tool is applied in Google Earth Engine and uses a Random Forest model that is trained using predictor variables from satellite images and climate data, and observed groundwater table depths. A Mann-Kendall trend test is used to determine significance of groundwater level trends. Results are dependent on the quality of the data input to the model, and choice of appropriate predictor variables. For example, locations with very little vegetation like groundwater-dependent playas or lakes may return erroneous results. Python code to apply the tool along with a methods document is available at https://github.com/tnc-ca-geo/SAGE.