Development of Conditionally Specified Statistical Models for Analysis of Environmental Studies

-----------------------------------------------------------------------建议书编号：DMS 9816630 PI：马克·S·凯泽，迈克尔·J·丹尼尔斯，苏门德拉·拉希里研究所：爱荷华州立大学项目：开发用于环境研究分析的有条件指定的统计模型摘要：该项目的目标是制定具有不同形式的相关条件分布的灵活的多变量分布，从这些分布开发统计模型，并根据它们描述和预测环境变量的能力来评估这些模型的性能。这类模型特别适合应用于涉及空间相关过程的问题，例如陆地环境中空气污染物和潜在有毒物质的浓度。这些统计方法将提高我们在监测地点描述这些变量的能力，特别是对未监测地点作出准确和准确的预测。由此产生的污染物的空间和时间平均值将提供更好的环境暴露评估，用于人类健康影响和生态对环境条件的反应的模型。

----------------------------------------------------------------------- Proposal Number: DMS 9816630 PI: Mark S Kaiser, Michael J. Daniels, Soumendra Lahiri Institution: Iowa State University Project: Development of Conditionally Specified Statistical Models for Analysis of Environmental Studies Abstract: The objectives of this project are to formulate flexible multivariate distributions having associated conditional distributions of varying forms, develop statistical models from these distributions, and assess the performance of such models in terms of their ability to describe and predict environmental variables. Such models are particularly attractive for application to problems that involve spatially dependent processes, such as concentrations of air pollutants and potentially toxic substances in terrestrial settings. These statistical methods will improve our ability to describe such variables at monitored locations and, in particular, produce accurate and precise predictions for unmonitored locations. Spatial and temporal averages of pollutants that result will provide better assessment of environmental exposures for use in models of human health impacts and ecological responses to environmental conditions.