a Crop Condition and Crop Yield Estimation Method Based on Noaa/avhrr Satellite Data
The objective of this research was to develop a crop condition and yield assessment method based on NOAA/AVHRR satellite data for the Sahel and Horn countries of Africa. The method consists of the following steps. (1) Noise in the NOAA/AVHRR satellite data can cause misinterpretation. A proposed procedure reduces the noise by calculating a vegetation index from the satellite data that is a measure of the green biomass. The index is then sampled and averaged in time and space and smoothed by applying an algorithm. (2) NOAA satellite data covering large areas are available every week. A system has been developed to pinpoint areas with abnormal vegetation conditions. (3) The yield estimation method is based on the following idea. In Africa a low percentage of the potential agricultural land is used. During grain-filling of a crop, when the green biomass decreases, the vegetation index of the crop decreases. However, the crop's contribution to the vegetation index of the entire area will be small. As a result, the following rule applies: a high vegetation index during the grain filling stage of a crop indicates a large amount of biomass, favorable growing conditions during that critical period, and a large yield. Regression analysis was performed to establish relationships between crop yields of millet, sorghum, and groundnuts and the vegetation index values during the reproductive phase of these crops. The coefficients of determination (R('2)) between the vegetation indices and yields were found to be 0.71 for millet, 0.52 for sorghum, and 0.76 for groundnuts. Once the vegetation index values during grain filling become available the regression relationships are used to translate the vegetation index values into crop yields. The predicted 1985 yields for millet and sorghum based on satellite data were close to the estimated yields obtained from the Ministries of Agriculture of the Sahel countries of Africa (R('2) of 0.87 for millet and 0.85 for sorghum). The above method produces a crop yield estimate for crops still in the field. In cases where a low yield is expected, an early warning can be provided, which increases the time to take appropriate action.
- Pub Date:
- REMOTE SENSING;
- Physics: Atmospheric Science; Remote Sensing