The authors of this book have been giving statistics courses to ecologists for 15 years. We have taught more than 8,000 scientists. During our courses we cover topics such as R, data exploration, data visualizsation, multiple linear regression, generalised linear models, linear mixed-effects models, generalised linear mixed-effects modelling
(GLMM), generalised additive models (GAM), generalised additive mixed-effects models (GAMM), Bayesian analysis and MCMC, and multivariate analysis, among many other topics. Over the years a large number of participants have asked us to teach a module that covers the analysis of spatial, temporal, and spatial-temporal data. Although random effects in GLMM and GAMM can be used to deal with dependency, such
an approach is not optimal for spatial, temporal or spatial-temporal data. Although there were various tools available in R, they either required expertise knowledge or required extensive computing time (e.g. MCMC in WinBUGS or OpenBUGS). We therefore elected to stay away from teaching and writing about spatial, temporal, and spatial-temporal data analysis. It was only after we became aware of material described in Lindgren et al. (2011) that we realised that GLMs and GLMMs, and all their zero- inflated cousins and smoothing cousins, can be extended to spatial, temporal, and spatial-temporal data. Unfortunately, the literature describing the approach (Integrated Nested Laplace Approximation, abbreviated as INLA) is rather technical. A book published in 2015 by Blangiardo and Cameletti helped us understand the INLA world better. Although we find it an excellent book, it still requires a fair amount of statistical knowledge in order to fully comprehend the material.
