Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: ASSESSING CLIMATE, SOIL AND LANDSCAPE PROCESSES AFFECTING AGRICULTURAL ECOSYSTEMS Title: Remote Sensing and Modeling Methods for Crop Grain Yield Assessment

Authors
item Doraiswamy, Paul
item Akhmedov, B - SSAI
item Milak, S - SSAI
item Stern, Alan

Submitted to: International Geoscience and Remote Sensing Symposium Proceedings
Publication Type: Abstract Only
Publication Acceptance Date: June 24, 2008
Publication Date: July 6, 2008
Citation: Doraiswamy, P.C., Akhmedov, B., Milak, S., Stern, A.J. 2008 Remote sensing and modeling methods for crop grain yield assessment [abstract]. International Geoscience and Remote Sensing Symposium. 2008 CDROM.

Technical Abstract: Monitoring crop condition and yields at regional scales using satellite imagery from operational satellites remains a challenge for the developed and developing countries. Imagery from the MODIS sensor onboard NASA’s TERRA and ACQUA satellites offer an excellent opportunity for daily coverage at 250 m resolution, that is adequate to monitor field sizes are larger than 25 ha. Studies conducted in the U.S. Corn Belt over the past six years have provided the databases that allow validation of various methods that use satellite imagery to provide rapid assessment of local and regional spatial variability of corn and soybean crop grain yields. The 250 m resolution MODIS 8-day composite surface reflectance data (MOD09) was found to be suitable for developing within-season crop classification in the U.S. Corn Belt and also applied for defining crop condition parameters as well as end of season grain yields. The composite procedure attempts to present the best quality of available data but there are still errors associates with cloud cover, atmospheric influence and BRDF effects. Developing a stable and consistent seasonal NDVI time-series database is critical for potential application in operational crop production assessment for the U.S. Department of Agriculture. The objectives of this research are to (a). Develop a consistent and standardized database for the U.S. Corn Belt to facilitate the application of models and algorithms; develop a decision-tree and algorithm-based crop classification procedure in the U.S. Corn Belt; (b) evaluate methods for producing operational crop yield estimates using primarily the 8-day MODIS composite time-series imagery as well as a climate-based grain yield model that uses biophysical parameters derived from imagery. MODIS classification accuracies were assessed by comparing with Landsat-based classification developed by USDA’s National Agricultural Statistics Service (NASS). The overall accuracy for corn and soybean crop classification using the MODIS imagery was between 75-80 % of the Landsat classification. The two methods developed in this research using MODIS imagery products developed as part of this research. Grain yields results were summarized at the state and county levels to compare with the USDA NASS reports. The grain yield for corn and soybean crops were within 10-15% of the official state yield estimates and within 20 % of the reported county yields for Iowa and Illinois.

Last Modified: 10/31/2014
Footer Content Back to Top of Page