ABSTRACT
This chapter focuses on modelling and the prediction of various features of marine environments. In order to properly calculate such forecasts, it is necessary to apply data-based or physically-based prediction techniques, often combined with geospatial tools provided, for instance, by a geographic information system (GIS). The combination of the two is usually attained by programming. In particular, there exist programming languages which allow the user to integrate advanced mathematical modelling packages needed for computing predictions with the GIS software required to capture the spatial character of marine and coastal analysis. Among modelling and prediction techniques we focus on a review of several empirical time series methods (e.g. data transformations, polynomial-harmonic models, autoregressive processes, prediction equations) and some physical approaches (limited herein to general circulation models and coupled air-sea models). This review will be complemented by a few examples covering different aspects of marine studies. In addition, geospatial methods will be shown as tools for visualising model and prediction performance as well as for carrying out spatial modelling which may strengthen the entire forecasting exercise. The most common statistical measures used to evaluate models and the resulting predictions will be discussed.
