Space-time geostatistical modelling, mapping and simulation of COVID 19: A case study of the Netherlands
Students should be knowledgeable in the R software
The current trend of COVID-19 epidemic threatens to decimate populations even in advance countries. The irony is that advance countries rather suffer the burden perhaps due to their complacency in implementing optimal disease surveillance techniques. Cases in the Netherlands generally keep rising within the different epidemic waves. Vaccination has currently become the only alternative. Currently, there is overwhelming data collected on COVID. Most of these are, however, available in aggregate formats of large administrative units perhaps due to the protection of patients privacy (Figure 1). There are two limitations: (1) The raw rates have biases due to population variability and/or spatial spillovers (spatial autocorrelation), (2) Heterogeneity within the administrative areas of representation is ignored. Maps inform policy and must be devoid of such biases. Also, detail prediction maps can help the identification of hotspots and detailed areas necessary for immediate vaccination.
The main objective of this study is to develop a geostatistical model for filtering the observed risk, granular mapping, and generation of realizations through simulations. The structure of the method depends on the data structure of the data available. The maximum likelihood estimate of the risk, also the crude risk, is to be decomposed into a mean component (stationary or nonstationary) and residual component. This crude risk may have to be smoothed first because of the prevalence of variance instability. The residual component is to be modelled as a space-time gaussian process with covariance functions deduced from empirical semi-variograms. With the appropriate space-time covariance parameters determined, prediction weights can be obtained for ordinary space-time kriging. Simulation based uncertainty characterization and error propagation should be conducted to ascertain the quality of the predictions.