What's happening at IRRI
Friday, July 31, 2015, 01:15pm - 02:15pm
Remote sensing and crop growth model applications for irrigation management and crop yield monitoring
Mr. Mojtaba Rezaei
Sari University of Agriculture Science and Natural Resource
Social Sciences Division
Water availability is a major threat to sustainable rice production in many parts of the world. A system analysis using crop growth simulation (CGM) model is needed to efficiently evaluate different solutions for different environment in the face of water scarcity. Another useful application of CGM is to provide early estimation of rice production. It is believed that crop growth model (CGM) can be reliable tool to predict crop yield as well as understand and mitigate the effects of new and somehow unknown situation. Despite reasonable accuracy of yield simulations with CGMs in small scales, there are still difficulties in applying CGMs in large scale especially for estimation of rice yield loss due to water shortage. This study was conducted to investigate the possibility of enhancing accuracy of CERES-Rice and ORYZA CGMs in estimating yield at large-scale through the use of MODIS, Landsat 5, and Landsat 7 satellite imageries in Guilan, Iran and Nueva Ecija, Philippines. For study case in Iran, CGM was further applied to find the best irrigation solutions in the face of an on-going water shortage condition in the area. The results in Iran showed high accuracy in yield prediction in small-scale field experiments for calibration and validation of CERES-Rice (RMSEn=8%). Accuracy of yield prediction over large area was improved by assimilation of remote sensing data. Without assimilation of Landsat 5 and Landsat 7, RMSEn for yield estimation was 21% whereas with assimilation of remote sensing data the RMSEn was 12%. The average yield loss in the area are simulated to be 36, 20, 10, and 2% in the case of 250, 300, 350 and 400mm of water application scenarios as compared to fully irrigated condition with 500 mm water application during crop season. In the case of Nueva Ecija, Philippines, MODIS data assimilation resulted in reduction of RMSEn from 38% to 23%, indicating high prospect of such remote sensing data assimilation approach for application of CGM in crop yield monitoring.
SSD Conference Room, Drilon Hall