Investigating surface water trends across the conterminous United States using daily MODIS imagery

Keywords: surface water, inundation, remote sensing, MODIS, Google Earth Engine
Session Type: Virtual Poster Abstract
Day: Saturday
Session Start / End Time: 2/26/2022 09:40 AM (Eastern Time (US & Canada)) - 2/26/2022 11:00 AM (Eastern Time (US & Canada))
Room: Virtual 69

Abstract

Spatial information about surface water trends at national- and regional- scales can provide land managers with the ability to better manage water resources and understand the spatial extent of flooding events. Satellite imagery is commonly used to map and monitor surface water over time, but Landsat-based products (16-day return interval) may result in continuity issues during periods of excessive cloud cover. USGS researchers with the Patterns in the Landscape – Analyses of Cause and Effect (PLACE) project have created a series of high-frequency surface water map composites derived from daily Moderate Resolution Imaging Spectroradiometer (MODIS) imagery. Using Google Earth Engine, the team developed customized image processing steps and adapted the Dynamic Surface Water Extent (DSWE) model to generate 5-, 10-, 15-day, and monthly surface water map products for 2003-2019 at a 250-m pixel resolution. The 5-day map composites have proven valuable for identifying short term changes in surface water extent associated with flooding, while the monthly composites create a continuous record better suited for investigating long term trends. To explore spatiotemporal variability over the past two decades, we analyzed changes within all Level I and Level III Environmental Protection Agency Ecoregions. This poster describes the creation of the DSWEmod maps, an assessment of the regional trend analyses generated from 204 months of national surface water maps data, and the role of precipitation on observed trends.