2020 Irrigated Lands for the Eastern Snake Plain Aquifer: Machine Learning Generated

This raster file represents land within the ESPA study boundary classified as either “irrigated” with a cell value of 1 or “non-irrigated” with a cell value of 0 at a 10-meter spatial resolution. These classifications were determined at the pixel level by using random forest, a supervised machine learning algorithm. To build a random forest model and supervise the learning process, IDWR staff create pre-labeled data, or training points, which are used by the algorithm to construct decision trees that will be later used on unseen data. Model accuracy is determined using a subset of the training points, otherwise known as a validation dataset. Several satellite-based input datasets are made available to the random forest model, which aid in distinguishing characteristics of irrigated lands. ESPA Irrigated Lands 2020 employed the following input datasets: US Geological Survey (USGS) products, including Landsat 8/9 and 10-meter 3DEP DEM, European Space Agency (ESA) Copernicus products, including Harmonized Sentinel-2, and Global 30m Height Above Nearest Drainage (HAND) (Donchyts et al., 2016). Evapotranspiration data from the METRIC model (Mapping Evapotranspiration at high Resolution with Internalized Calibration) was provided by IDWR and used as an input. IDWR staff used the following datasets to aid in the labeling of training data: NDVI derived from Landsat 8/9, Sentinel-2 CIR imagery, US Department of Agriculture National Agricultural Statistics Service (USDA NASS) Cropland Data Layer, Active Water Rights Place of Use and METRIC data from IDWR, and USDA’s National Agriculture Imagery Program (NAIP) imagery. NAIP imagery from 2019 was used as a reference; all other datasets were available for 2020. ESPA Irrigated Lands 2020 model runs were processed on six separate subregions covering the study boundary. This was done to reduce processing time and better train the model on specific climatic regions. Each subregion may undergo 1-4 model iterations, where at each iteration IDWR staff added or removed training points to help improve results. The northeast section of the study boundary, spanning from Kilgore, Island Park, and Ashton, was largely hand-delineated due to a lack of quality training data or inaccuracy of modeled results. Post-processing of model output included a wetland mask derived from the Fish and Wildlife Service’s National Wetlands Inventory wetlands dataset, as well as a manually created mask specific to issues found in the ESPA 2020 model results. The masking datasets and pre-labeled training points are available on request. References:Donchyts, G., Winsemius, H., Schellekens, J., Erickson, T., Gao, H., Savenije, H., & van de Giesen, N. (2016). Global 30m height above the nearest drainage (HAND). Geophysical Research Abstracts, 18, EGU2016-17445-3. EGU General Assembly 2016.