Review of modelling techniques

Ecological communities are the expression of complex biological processes (reproduction, nutrition, rest, interspecific relationships, et cetera) and abiotic processes (nutrient cycling, discharge regimes, erosion, et cetera), both being expressed on various scales of time and space. To analyse all these processes (i.e., to include and understand the relationships which exist in the community) and to characterise their relationships using environmental parameters, their degree of importance, and their structuring, require the observation of variables related to the operation of the system. The complexity of the ecological systems often results in complex relations between the biological and abiotic variables, justifying the use of multiple modelling techniques. These models are based on different statistical and simulation techniques, designed to predict community structure from environmental variables. This chapter aims to review current ecological models that predict community structure in aquatic ecosystems for the selection of the appropriate models, depending on the type of target community. Ecological water management is designed to enhance the value of aquatic ecosystems. Such management requires the understanding of how these ecosystems function, and thus how communities are related to the environment. To learn communityenvironment relationships, data-analytical approaches are explored: Conventional statistical models, Artificial neural networks, Bayesian and Mixture models, Support vector machines, Genetic algorithms, Mutual information and regression maximisation techniques, and Structural dynamic models. In the following sections, we have summarized these modelling techniques and presented their applications in ecological studies, together with their strengths and weaknesses.