Modeling species richness controlled by community‐intrinsic and community‐extrinsic processes: coastal fish communities as an example

Abstract

This article focuses on the analysis of coastal fish communities along the Norwegian Skagerrak coast. Species numbers are estimated based on annual samples of the fish communities within 12 fjords from 1953 to 1994. On this basis, a community dynamics model (incorporating both community-intrinsic and community-extrinsic processes) was developed and analyzed. This model is then discussed on the basis of other community models available through the literature, both phenomenologically oriented and process-oriented models.

Pacific sardine (Sardinops sagax, Jenyns 1842) landings prediction. A neural network ecosystemic approach

“…The GAM models performed better than the MLR models, which along with the results of the neural approaches, support the hypothesis pointed by Stenseth et al (2002) about the stochasticity and high non-linear dependence between the abundance and ecosystem characteristics. Similar results can be found in other ecological applications that compare the prediction capacity of MLR with CNN (Lek-Ang et al, 1999;Gutiérrez-Estrada et al, 2008).…”

“…To reach this objective requires the selection of adequate statistical and mathematical tools because, as suggested Steele (1985), Murdoch (1994) and Stenseth et al (2002), the marine ecosystems are non-equilibrium systems characterized by the presence of a high level of stochasticity and non-linear relationships between environmental, biological and ecological variables. James and McCulloch (1990) reported that conventional models (like a wide range of classic multivariate techniques) assume continuity, steady-states and linear or simple polynomial relationships between variables which is a serious shortcoming when highly non-linear relationships must be modelled (Lek et al, 1996).…”

Forecasting of jack mackerel landings (Trachurus murphyi) in central‐southern Chile through neural networks

“…Marine ecosystems remain in a constant imbalance, and they are characterized by the presence of high stochastic levels and non-linear relations (Murdoch, 1994;Stenseth et al, 2002). Sometimes, the non-linear relations are not evident and present difficulties for modeling.…”

Short-term forecasting of halibut CPUE: Linear and non-linear univariate approaches

“…These properties have favoured the application of this kind of model to predict different variables in numerous fields of engineering and the sciences, including fishery science (Phillips, 1983;Stocker and Noakes, 1988;Stergiou, 1989Stergiou, , 1990Stergiou, , 1991Stergiou et al, 1997;Lloret et al, 2000;Becerra-Muñoz et al, 2003;Hänninen et al, 2003;Punzón et al, 2004;Gutiérrez-Estrada et al, 2004). Stenseth et al (2002) reported that ecological systems (like fisheries) may be heavily influenced by substantial stochastic processes (as climate conditions), as well as by non-linearities, leading to non-equilibrium systems as the results. Thus, Lek et al (1996a,b), Guégan et al (1998), Gozlan et al (1999) and Gutiérrez-Estrada et al (2000) propose a non-linear approach to the forecasting of ecological-biological variables.…”