Integrating SAR and optical imagery for regional mapping of paddy rice attributes in the Poyang Lake Watershed, China

Abstract

The development of agricultural monitoring tools is a focus of the Group on Earth Observations for the Global Earth Observation System of Systems (GEOSS). This requires combining synthetic aperture radar (SAR) and optical satellite sensors to provide more meaningful information in an operational context. The goal of this research was to characterize paddy rice agro-ecological attributes in an operational context. A decision tree framework combined multitemporal and multiscale PALSAR, MODIS, and Landsat observations to map rice extent, hydroperiod, crop calendar, and cropping intensity. The study was carried out in the Poyang Lake Watershed, Jiangxi Province, China. A multiseason field campaign was carried out to calibrate algorithms and validate the rice maps. The field data corresponded relatively well with the remote sensing metrics and validation found that the derived rice paddy maps possessed a high overall accuracy of 89%. The rice maps indicate that the watershed has 25% rice agriculture with 85% of all paddies undergoing a double-crop management pattern. Remotely sensed metrics showed that inundation periods for early rice were typically twice as long as inundation periods between crops, which corresponded to site level measurements at double crop locations. Using hydroperiod and crop intensity information a crop calendar indicated that the day of year (DOY) planting and harvesting activities was typically around DOY 77 and 329, respectively. The automated approach combining SAR and optical platforms with continental-scale acquisition strategies might allow for large-area, operational agricultural rice mapping. This can contribute to a GEOSS framework for improved agricultural monitoring.