Time-space Kriging to address the spatiotemporal misalignment in the large datasets

Area of Science:

• Environmental Science
• Geostatistics
• Bayesian Statistics

Background:

• Classical Kriging methods are computationally intensive for large spatiotemporal datasets.
• Spatiotemporal data often exhibit misalignment, scale mismatches, and missing values.

Purpose of the Study:

• Introduce Markov Cube Kriging (MCK), a novel Bayesian hierarchical spatiotemporal interpolation method.
• Address computational limitations and data complexities in large spatiotemporal datasets.

Main Methods:

• Developed MCK, a computationally efficient Bayesian hierarchical spatiotemporal interpolation technique.
• MCK handles non-separable structures and nonstationary covariance across hierarchical scales.
• Applied MCK to estimate daily fine particulate matter (PM2.5) concentrations.

Main Results:

• MCK provides robust predictions for spatiotemporal random effects.
• Successfully captured hierarchical and nonstationary spatiotemporal structures in air pollution data.
• Generated daily PM2.5 estimates at a 2.5 km grid for Cleveland (2000-2009).

Conclusions:

• MCK offers a computationally efficient and flexible solution for spatiotemporal interpolation.
• MCK has significant implications for exposure quantification in environmental epidemiology.
• Facilitates data collocation from diverse sources across different spatiotemporal scales.