An overview of spatial sampling procedures for crop area estimation

Abstract

Information on crop sown acreage is an important basis for the formulation of national food policies and economic planning. Timely and accurate knowledge of crop acreage plays a very important role in enhancing agriculture management and ensuring national food security. The complete statistical survey method has been used to obtain the crop sown area information at the national scale in China for a long time. However, there are many shortcomings using the method, for example, the enormous inventory workload, the high investigation cost and the very long survey time and so on. With the rapid development of the economy and the increasing of the government decision-making departments, the social public demand for agricultural statistical data, the complete statistical survey has been not able to meet the need that the new rural development situation and the crop planting structure changes. In order to improve the survey efficiency, the traditional list sampling method (the operation procedure is as follows: sampled counties were drawn from the Province, sampled towns were drawn from the sampled counties, sampled villages were drawn from the sampled towns, and finally, sampled famers were drawn from the sampled villages) has been employed to investigate crop acreage by the Chinese statistical department since 1984. Although the traditional list sampling has solved part of the problems appearing in the complete statistical survey, however, limited by itself operational mechanism, the new problems are that the update of the sampling frame is very slow and that spatial information is not adequately employed in the investigation process, when the traditional list sampling is used to estimate the crop area. With the development of “3S” technology (Remote Sensing, Geographic Information System, Global Positioning System), spatial sampling methods constructed by combining traditional sampling and “3S” technology have been gradually used for estimating the crop area at large scales. Many countries have currently used the spatial sampling to monitor and estimate the crop area in the world, and the estimation accuracy and monitoring timeliness of crop area information have been improved significantly. In this paper, the previous studies on the samples selection in space, the optimization of the sampling unit size and the reasonable design of the sample layout involved in the spatial sampling scheme are comprehensively summarized in the past thirty years. According to the summary of previous studies, the problems existing in the spatial sampling survey can be analyzed as follows: the first is that the quantitative evaluation on the spatial sampling efficiency is still lack at the different agricultural area; the second is that the sampling unit size is formulated mainly by experience, however the studies on its optimization is not enough yet; the last is that the rationality of sample layout is deficient, due to that the spatial correlation and variability of the sampling units have not been quantitatively estimated. Furthermore, the development trends of spatial sampling for crop area estimation are also presented in this study in the future.