Estimating Irrigated Acres with Missing Data

Abstract

Irrigation for agriculture is a major component for water use planning in the state of Georgia. Despite its imporance in planning, little credible information is available on current water consumption patterns and future demand. The ongoing water rights battle between Alabama, Florida, and Georgia, where the other two states contend that Georgia is wrongfully taping into the waters that belong to them, has necessitated a better understanding of water needs over the next several years. Policy makers are often limited by information available to them which may be lacking due to missing or highly aggregated data. The main problem caused by missing data is that estimators of population characterisitics are assumed to be biased unless evidence to the contrary is provided. Also, the mean square of errors of such estimators are likely to be larger than those obtained from complete data. The challenge, therefore, lies in being able to estimate the missing values in an unbiased and consistent manner. These estimates may then be used to make forecasts. The first objective of the study was to derive a method to disaggregate the available information for the state of Georgia to a county and commodity level. The second objective is to use the disaggregated data to forecast irrigation water demand for the counties by commodity. This analysis focuses on seven commodity groups: corn, cotton, peanuts, rye, sorghum, soybeans, and tobacco for 1970 through 1998.