Rice area-yield estimation based on the synergistic use of remote sensing time-series and crop growth modeling in Nigeria

Barbieri, Massimo; Quicho, Emma D.; Ibrahim, Ali; Melchiori, Luca; Cattaneo, Alessandro; Gatti, Luca; Copa, Loris; Holecz, Francesco; Mathieu, Renaud; Saito, Kazuki; Senthilkumar, Kalimuthu

Rice area-yield estimation based on the synergistic use of remote sensing time-series and crop growth modeling in Nigeria

Autores

Editor

Elsevier

Descripción

Context RIICE (Remote sensing-based Information and Insurance for Crops in Emerging Economies) technology, developed since 2010 by sarmap and the International Rice Research Institute, has proven effective in Southeast Asia for monitoring rice production using satellite data. In Nigeria, rice is a key staple crop, yet reliable data on cultivated area and yield are lacking, limiting the ability to design effective interventions for improving productivity. Objective This study aims to apply and improve the RIICE approach to provide accurate and up-to-date estimates of rice cultivated area, yield, and yield gaps in Nigeria. Methods Over 1500 geolocated ground-truth points were collected across Kano, Jigawa, and Benue States during the 2022 and 2023 wet and dry seasons. RIICE technology was used to analyze the temporal signatures of satellite data to detect rice areas and crop season start dates. These outputs, combined with agronomic inputs such as soil characteristics, fertilizer use, and water availability, were fed into a crop model to estimate yields and assess yield gaps. Results and conclusion In the 2023 wet season, estimated rice cultivation areas were 226,702 ha in Kano, 113,871 ha in Jigawa, and 317,282 ha in Benue, with detection accuracy ranging from 85 to 90 % in Kano and Jigawa and 75–85 % in Benue. Yield estimates showed Kano achieving averages of 5.1 t/ha (dry) and 5.0 t/ha (wet) under irrigated system, while Jigawa yielded 3.9 t/ha (dry) and 3.8 t/ha (wet). In Benue, yields averaged 3.5 t/ha, with some areas producing <1.0 t/ha. Yield gaps were significant: 3.0–3.2 t/ha in Kano and 4.0–4.1 t/ha in Jigawa, highlighting the need for targeted agronomic interventions to bridge these gaps and enhance rice productivity. Significance The results demonstrate the effectiveness of integrating remote sensing with ground data and crop models for reliable yield estimation and, consequently, identifying the critical target areas where interventions are the most eventually needed.