The measure of data aggregation bias for crop production risk assessment

ABSTRACT: Crop risk assessment has been a hot issue for many years. Due to the lack of sufficient low level data, in practice, low level risk is usually measured based on large level aggregated data which will lead to data aggregation bias and risk underestimation. During the past decades, various studies have been made to explore and settle the problem of risk underestimation. In this paper, through quality and quantity analysis, two questions have been answered: (i) what factors will have an impact on the data aggregation bias and (ii) to what extent these factors can influence the risk underestimation level.