Abstract

 

 

Accepted 22nd May, 2014

 

In this study, remotely sensed data and field measurements were used to develop a simple but robust method for estimating the tobacco hectarage and yield in Zimbabwe. The current conventional tobacco yield forecasts rely on seed purchase records, land area record and visual assessment of the crop. This is costly, time consuming and unreliable. Between 2010 and 2013, starting from September, agricultural field boundaries from a pseudo natural colour composite Landsat Thematic Mapper (TM) satellite imagery were visually interpreted and digitized. Cloud free MODIS images covering the period September to end March were downloaded and georeferenced. For each MODIS image, NDVI was estimated. Mean temporal NDVI profiles for these crops using data from sampled tobacco fields were calculated separately and compared. The results of this study indicated that, based on MODIS NDVI data, the third to fourth week of November and the third to fourth week of February are the optimal times for discriminating the irrigated from the non-irrigated tobacco. The crop areas for the three seasons were estimated and yield estimates calculated from the long-term cropped yield- area regression model. The three seasons average yield estimates were 98.8% accurate, as compared to 112% for the traditional method.

 

Keywords: Remotely sensed data, MODIS images georeferenced, NDVI profile yield estimates