Spatial Analysis Using Big Data

4.8. Spatiotemporal model

4.8.1. Outline

Geostatistics has traditionally developed targeting space with two (x, y) or three dimensions (x, y, z), but studies on spatiotemporal geostatistics considering the time axis have been actively conducted (Cressie and Wikle, 2011; Wikle et al., 2019). Assuming continuity in space, these studies have been roughly classified into approaches that view the time axis as continuous and approaches that view the time axis as discrete.

4.8.2. Approaches that view time axis as continuous

Now, we consider the spatiotemporal process {y(s,t): s   ∈   D, t   ∈   T} (see Chapter 1 about the definition). Here, let us assume that the spatiotemporal process is composed of the trend component m(s,t) and the error ...