While the reliability and accuracy of remote sensing data is routinely checked through systematic sampling and ground verification, there has never been a large scale accuracy check that was commodity-based via the supply chain. In this paper the authors have made an attempt to scientifically compare remote sensing-based data with supply chain sources such as millers, exporters and wholesale commodities markets. This first-of-its-kind study was undertaken to assess the accuracy of remote sensing-based production estimations of Basmati rice for the 2005 growing season against the actual arrivals at various grain depots as well as other related sources. The survey was conducted through export, wholesale and marketing boards of Basmati in four leading rice-producing states in India: Uttar Pradesh, Uttaranchal, Punjab and Haryana.
Based on a comparative analysis between Kharif (monsoon growing season) 2005 estimates derived using remote sensing, versus the actual arrives at market, a March 2006 conclusion, based on certain calculated and logical assumptions and limitations found that the remote sensing-based estimated results for Kharif 2005 harvest provided an accuracy of 90% to 94% in the study areas. As anticipated, remote sensing based data as expressed in estimates is higher in a predictable manner than total supply chain sources, thus confirming the reliability of this project's data.
The authors consider this to be a cautious, well calculated and mathematical analyses-based study done for the first time in India. Results are re-defining certain usage of remote sensing techniques in agriculture market research and supply chain study. This study may also have an application in calibrating demand forecasting as well.
For the full paper, write in to us at info@rmsi.com