Yield forecast modeling For wheat crop Using agro meteorological data
Loading...
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Management and Technology Lahore
Abstract
This thesis explores the possibility of using remote sensing images and its derived products, Normalized Difference Vegetation Index (NDVI), to assess and measure wheat yield potential. Panel regression models were developed for the Gujranwala division at different stage of spatial aggregation. Through the panel regression analysis, the study identified the relationship between the crop vegetation over the growing period and the final crops yield. The vital goal of the research is to inspect whether the NDVI model can generate precise and timely yield forecasts.
The present study primarily focuses on the variation of crop yield through NDVI based panel regression analysis is used as a tool to achieve this objective. The NDVI model is used to represent crop vegetation density or greenness (reflectance of all agro metrological parameter) of the vegetation cover. It proves that, crops yield can be predicted more effectively and accurately by using NDVI base model as an alternative tool against the conventional models. The Redundant and Housman test which guides about the fixed effect model is suitable in this study. The NDVI base fixed effect model fits the crop yield data extremely well, with high R-square value at about 90%. The results suggest that NDVI for the month of January, February and March were ideal for wheat yield assessment.
Market participants rely mainly on apt and precise crop yield forecasting to make informed buying and selling decisions which results in well organized resource distribution. It is an important point to mention that weather forecast network is not being used effectively and official crop production and yield estimations are not based on the accurate and timely information in Pakistan where agricultural growth is crucial for eradicating poverty. A practical and cost effective solution is provided for them by using Remote sensing and NDVI. The relationship between the accumulation of crop vegetation over the growing season and crop yields is confirmed in the present research. The NDVI panel regression model suits the crop yields data quite well in spite of a relative difference between forecast and actual official wheat yield. The three independent NDVI variables had positive impact on crop yield though to a varying degree.
Description
Supervised by:Prof. Dr. Ahmad Faisal Siddiqi
Keywords
Redundant and Housman test, Vegetation cover, MS Thesis