21st Century Surface Water Trends for the Conterminous United States Using Monthly Image Composites

Abstract:

Information about surface water trends on both national- and regional-scales can provide land managers with the ability to better manage water resources. There are many challenges to monitoring surface water across broad geographic extents, including regional variations in availability of imagery, cloud cover, and topography. This project describes the creation and analysis of a consistent national surface water product for the purposes of, 1) mapping the geographic distribution of surface water, 2) quantifying trends in areal coverage across various spatial and temporal scales, and 3) quantifying correlations between precipitation and surface water. We created monthly composites of surface water extent (e.g. water bodies, lakes, reservoirs, rivers) for the conterminous United States (U.S.) using the Dynamic Surface Water Extent (DSWE) algorithm, which we adapted for use with Moderate Resolution Imaging Spectroradiometer (MODIS) data. The DSWE algorithm produces a raster product that classifies surface water into four confidence-based classes for each image. We processed daily 250-m MODIS imagery from 2003 to 2019 in Google Earth Engine to produce 204 monthly composites. To explore spatiotemporal variability over the past two decades, we analyzed changes within all Level I and Level III Environmental Protection Agency ecoregions. Results show regional differences in surface water trends including decreasing surface water in mountain ranges across the western U.S., across the east-central U.S., and Appalachian Mountains; we observed increasing surface water trends across the Great Plains. We found generally positive correlations with precipitation for most locations except for the western mountain ecoregions which had significant negative correlations. Overall, by summarizing the status and trends of surface water at a national scale, these findings provide important context to help support informed water-management decision-making at state and federal levels.