Using a simulation analysis to evaluate the impact of crop mapping error on crop area estimation from stratified sampling

• Published in: International journal of digital earth Vol. 12; no. 9; pp. 1046 - 1066
• Main Authors: Sun, Peijun, Congalton, Russell G., Pan, Yaozhong
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 02.09.2019; Taylor & Francis Ltd; Taylor & Francis Group
• Subjects: Area; area estimation; Crop mapping error; Crops; Decision making; Efficiency; estimation efficiency; Heterogeneity; Land cover; landscape heterogeneity; Mapping; Sample size; Sampling; Simulation analysis; Stratified sampling; Thematic mapping
• ISSN: 1753-8947; 1753-8955
• DOI: 10.1080/17538947.2018.1499827

To analyze the efficiency of area estimations (i.e. estimation accuracy and variation of estimation) impacted by crop mapping error, we simulated error at eight levels for thematic maps using a stratified sampling estimation methodology. The results show that the estimation efficiency is influenced by the combination of the sample size and the error level. Evaluating the trade-offs between sample size and error level showed that reducing the crop mapping error level provides the most benefit (i.e. higher estimation efficiency). Further, sampling performance differed based on the heterogeneity of the crop area. The results demonstrated that the influence of increasing the error level on estimation efficiency is more detrimental in heterogeneous areas than in homogeneous ones. Therefore, to obtain higher estimation efficiency, a larger sample size and lower error level or both are needed, especially in heterogeneous areas. We suggest that existing land-cover maps should first be used to determine the heterogeneity of the area. The appropriate sample size for these areas then can be determined according to all three factors: heterogeneity, expected estimation efficiency, and sampling budget. Overall, extending our understanding of the impacts of crop mapping error is necessary for decision making to improve our ability to effectively estimate crop area.