Mapping and quantifying agricultural irrigation in heterogeneous landscapes using Google Earth Engine

Abstract

Geospatial analysis using remote sensing data has emerged as an effective tool to monitor irrigated lands over a variety of climatic conditions and locations. Humid and heterogeneous landscapes such as those in the South Atlantic Coastal Plain are challenging for satellite image classification over large spatial extents. The objectives of this study are to: 1) compile relevant vegetation indices for mapping and quantifying irrigated and dryland agricultural areas across the coastal plain region of South Carolina, USA; 2) identify the spatial and temporal change of the agricultural irrigation areas in the region over time; 3) compare the results to county-level statistics reported by the United States Department of Agriculture (USDA) in the study area. In this study, a framework has been developed to regularly monitor irrigated agriculture's extent and distribution using high-spatial-resolution Sentinel-2 imagery for the coastal plain of South Carolina region using the Google Earth Engine (GEE) platform. The maps were produced using random forest supervised classification for three years with overall accuracy assessments of 83.73% (2016), 86.18% (2017), and 84.55% (2018). The major irrigated crops in the study area are corn, cotton, soybeans, and peanuts. Overall, the irrigated crops' distribution areas were mainly concentrated in counties near the center of the studied region. The produced maps will provide valuable information for the development of irrigated agriculture and the optimization of irrigation water use in the region. The methodology followed in this study can be applied to any other regions around the world with similar landscapes and climatic conditions.