Spatio-Temporal Statistics and its Application to Remote Sensing

就其本质而言，环境过程涉及强烈的空间和时间变异性。推断因果关系需要在统计模型中纳入空间和时间相关性。这个项目的目标是发展质量平衡的分层时空统计模型，与多变量过程相关的新的损失函数，以及从分层模型的预测分布获得的最优估计。这些方法打算利用来自卫星和其他数据源的大量遥感数据集，用于估算大气二氧化碳的近地表通量。

By their very nature, environmental processes involve strong spatial and temporal variability. Inferring cause-effect relationships requires the incorporation of spatial and temporal dependence in the statistical models. The aims of this project are to develop mass-balanced hierarchical spatio-temporal statistical models, new loss functions that are relevant to multivariate processes, and optimal estimators obtained from the hierarchical model's predictive distribution. These methodologies are intended to be applied to the estimation of near-surface fluxes of atmospheric carbon dioxide, using massive remote sensing datasets from satellites and other data sources.