Citation:

Vanderhoof, M., L. Alexander, AND J. Christensen. Tracking patterns of inundation across the Prairie Pothole Region using Sentinel-1 and Sentinel-2 image collections in cloud-based computing platforms. SWS Annual Meeting 2021 - Virtual, Virtual, June 01 - 04, 2021.

Impact/Purpose:

Mapping of wetland and river dynamics is relevant to understanding important hydrological issues of flooding, drought, and water quality. This study uses high-resolution imagery and cloud computing capacity to derive multi-source inundation datasets to characterize dynamics.

Description:

Wetland and river hydrology across the Prairie Pothole Region depend on not just patterns of precipitation but also surface storage capacity and the agricultural re-directing of water through tile drainage and ditching, among other drivers. A critical precursor to analyzing the relative role of climate and land use on wetland hydrology is a robust dataset representing inundation dynamics at a spatial and temporal scale relevant to a landscape dominated by small, dynamic wetlands. We present preliminary results of inundation frequency and duration across the Prairie Pothole Region derived from time series analysis of Sentinel-1 and Sentinel-2 image collections. Our approach takes advantage of cloud-based computing platforms including Google Earth Engine and Jupyter Notebook to produce inundation extent at 10 m resolution. While a reliance on only multi-spectral data can underestimate the response of landscapes to precipitation and flood events, due to cloud interference, time series analysis of SAR imagery can help strengthen estimates of inundation timing and duration. Deriving multi-source inundation datasets are therefore critical to characterize the episodic, seasonal and interannual response of inundation to both climate and land use activities.

Record Details: