Spatial Hierarchical Bayes Small Area Model for Disaggregated Level Crop Acreage Estimation

Abstract

Agriculture is the key livelihood for the vast majority of population in India. The sector is such a crucial that prosperity of agrarian community is essential for Government/Institutional stability. Therefore, the accurate estimation of production in terms of harvested area and yield are equally important in ensuring the accurate determination of their product. Although the yield estimation gets most of the attention, there are many complexities to the estimation of area that might not be readily apparent. Crop area statistics in most of the states are furnished based on complete enumeration or census method. But, shortage of man power, failure of the primary and revenue staffs to devote adequate time and attention in collection and compilation of data has deteriorated the quality of area statistics as well as increased the time lag in availability of data in hand. In the view of above problem, a well-designed sample survey has the ability to cater the need of accurate crop area information and is especially important in developing countries which have very limited resources to apply to the collection of agricultural data. A pilot experiment conducted by ICAR- Indian Agricultural Statistics Research Institute, New Delhi attempts to estimate district level crop yield based on reduced number of Crop Cutting Experiments (CCEs) while crop acreage estimation has been done through well designed sample survey approach. But, traditional sampling theory has also some limitations in providing reliable and valid estimates particularly for districts with few or negligible sample sizes. To tackle the need of representative crop acreage estimation at disaggregated level, Small Area Estimation (SAE) approach has been considered in this paper. In particular, using Hierarchical Bayes spatial small area model disaggregated level crop area has been estimated for two major crops, rice and wheat respectively in the state of Uttar Pradesh for Agriculture year 2015-16. Estimates produced using SAE technique has acceptable precision level and is a positive attempt of crop acreage estimation at micro or local level through SAE approach in India.