Keywords
bayesian inference, cellular automata, gulf of finland, logistic regression, phragmites australis
Start Date
1-7-2010 12:00 AM
Abstract
The common reed (Phragmites Australis) has overtaken the Finnish coast of theGulf of Finland in many places influencing the coastal ecology and reducing therecreational value of the coast. Our research aims at developing a spatio-temporalsimulation model of the spread of reed. In order to account for the dynamic nature of thespread, a cellular automata model is employed. We investigate in depth the explanatoryvariables and their relevance to the phenomenon. It was found that water depth, proximityto river mouths, and sea openness are significant explanatory variables of the phenomenon.A predictive logistic regression model is suggested to aggregate the explanatory variablesand provide transition rules for the cellular automata. The results of the regression modelare presented and discussed, raising the necessity of incorporating Bayesian inference inthe logistic regression model. Therefore, more insightful likelihood functions for theexplanatory variables are provided and discussed.
Spatio-temporal modelling of the spread of common reed on the Finnish coast
The common reed (Phragmites Australis) has overtaken the Finnish coast of theGulf of Finland in many places influencing the coastal ecology and reducing therecreational value of the coast. Our research aims at developing a spatio-temporalsimulation model of the spread of reed. In order to account for the dynamic nature of thespread, a cellular automata model is employed. We investigate in depth the explanatoryvariables and their relevance to the phenomenon. It was found that water depth, proximityto river mouths, and sea openness are significant explanatory variables of the phenomenon.A predictive logistic regression model is suggested to aggregate the explanatory variablesand provide transition rules for the cellular automata. The results of the regression modelare presented and discussed, raising the necessity of incorporating Bayesian inference inthe logistic regression model. Therefore, more insightful likelihood functions for theexplanatory variables are provided and discussed.
