58-11 Assessing Yield Potential At the Large Scale with Remote Sensing.

See more from this Division: ASA Section: Climatology & Modeling
See more from this Session: Symposium--Satellites Serving Agriculture and the Environment: Honoring the Achievements of Paul Doraiswamy
Monday, October 22, 2012: 4:25 PM
Duke Energy Convention Center, Room 260-261, Level 2
Share |

Urs Schulthess, CIMMYT, Texcoco, Mexico, Andrew McDonald, CIMMYT, Kathmandu, Nepal, Farida Perveen, Global Conservation Agriculture Program, CIMMYT-Bangladesh, Dhaka, Bangladesh and Jagadish Timsina, Melbourne School of Land and Environment, The University of Melbourne, Melbourne, Australia
Knowledge of the yield potential or attainable yield of a given field has many practical applications. It is relevant for optimal nutrient management, choice of maturity type and cropping sequence. One of the biophysical parameters that can be best estimated with remote sensing is ground cover, which is a direct measure of the amount of the intercepted photosynthetically active radiation (IPAR). In maize, that parameter, measured at around tasseling, in turn has been linearly related to number of kernels per m2. We used the SALUS crop simulation model to estimate yield as a function of different ground cover levels for irrigated Rabi (winter) maize production in Northwestern Bangladesh. Based on a RapidEye satellite image acquired around tasseling in late March we identified the maize fields and calculated ground cover for the entire image, which measured more than 500 km2. For the identification of the maize fields, we first created segments with the aid of eCognition.  Based on a visual analysis of the image, we then created a training set consisting of two classes: maize and others. The classification algorithm called Random Forest followed by a visual quality control was used to identify the maize segments. The average yield of each segment was then predicted using the prior established function between ground cover and yield. In the final step, we generated zones to map the productivity potential based on the maximum yielding segments in each zone. That value will then serve to set the yield goal for nutrient recommendations.
See more from this Division: ASA Section: Climatology & Modeling
See more from this Session: Symposium--Satellites Serving Agriculture and the Environment: Honoring the Achievements of Paul Doraiswamy