Mark P. Kumler, Ph.D.
James Ciarrocca, M.S.
Crop yield prediction can play an important role in developing the agriculture sector in Colombia. Remote sensing and GIS have proven to be an effective mechanism for this purpose in developed economies. This project created a proof-of-concept application for the Colombian Ministry of Agriculture and other related governmental institutions. The project used existing methodologies including the classification of satellite imagery, interpolation of climate data into continuous surfaces, the extraction of Normalized Difference Vegetation Index, and the computation of multiple linear regressions. ESRI ArcGIS provided the interface, software, tools and functions to build the application, and to integrate and automate the application‟s functionalities.
Cloud coverage in the imagery and the lack of specialized data affected the accuracy of the crop yields estimates. Nevertheless, the application predicts corn yields with an estimated accuracy of 71% when cloud coverage is minimal. The application can use both Landsat and Spot preprocessed images, and in less than six minutes yield predictions for areas inside Cordoba, a major corn producing state in Colombia.
Lemos, M. F. (2008). Using GIS to Predict Corn Yields in Colombia (Master's thesis, University of Redlands). Retrieved from https://inspire.redlands.edu/gis_gradproj/177