摘要：In Germany, since several decades the RAUMIS modelling system is applied for policy impact assessments to measure the impact of agriculture on the environment. A disaggregation at the municipality level with more than 9.600 administrative units, instead of currently used 316 counties, would tremendously improve the environmental impact analysis. Two sets of data are used for this purpose. The first are geo-referenced data, that are, however, incomplete with respect its coverage of production activities in agriculture. The second set is the micro census statistic itself, that has a full coverage, but data protection rules (DPR) prohibit its straightforward use. The paper show how this bottleneck can be passed to obtain a reliable modelling data set at municipality level with a complete coverage of the agricultural sector in Germany. We successfully applied a Bayesian estimator, that uses prior information derived a cluster analysis based on the micro census and GIS information. Our test statistics of the estimation, calculated by the statistical office, comparing our estimates and the real protected data, reveals that the proposed approach adequately estimates most activities and can be used to fed the municipality layer in the RAUMIS modelling system for an extended policy analysis.
关键词：Highest Posterior Density estimator (HPD);RAUMIS;Down scaling
