Application of Remote Sensing and Crop Modeling for Rice Production Monitoring in Africa: The Mali Experience

Mali is one of the major rice producers given its favorable agro-ecological conditions including access to water and suitable soil types. Rice production plays a crucial role in the country's economy and food security. Under the Accelerating Impacts of CGIAR Climate Research for Africa (AICCRA, with support from CIAT/World Bank) project in Mali, we aim to develop a rice production monitoring system wherein seasonal rice area, start of season (SOS), and yield estimates are generated, as well as climate-driven rice yield losses. Historical flood and drought were being analyzed to generate indicator maps of flood frequency, average duration and timing relative to planting period. The project utilizes the freely available Sentinel-1A Synthetic Aperture Radar (SAR) data (C-band, VV and VH polarization, with 20 m spatial resolution and 12 days temporal resolution) together with Sentinel 2 and LANDSAT 8 and 9, multispectral optical sensors with 30 m spatial resolution and 8 days temporal resolution. During the wet season 2022, rice monitoring was done in Sikasso and Segou Regions in Mali. Seasonal time series imagery were acquired and processed; rice ecosystem maps (rainfed upland and lowland, and irrigated rice) were generated to consider production systems and practices. Ground data collection for start of season map calibration and rice and non-rice area validation were conducted within the season. In addition field experiments were conducted at the Institut d'Economie Rurale, Sikasso to collect phenology data to support the yield mapping. MAPscape-RICE, a fully automated software for generation of rice area, SOS, and leaf area index estimates from multi-temporal SAR and optical data were used in conjunction with RiceYES, an interface linking remote sensing derived information with the ORYZA model. This paper highlights the methods used in generating outputs generated from the 2022 wet season rice cropping. Rice validation results show an accuracy of 97% for Segou and around 85% for Sikasso. This is the first application of the "Remote sensing based Information and Insurance for Crops in emerging Economies" (RIICE, www.riice.org) in West Africa. Experiences and lessons learned in the implementation of the activities during the first monitoring season will be presented.