Species Distribution Model: Generalized Linear Models
Wednesday, December 2, 2015
Model Category: Regression.
Model Description: This model can be applied in univariate and multivariate applications, and it is used to estimate an ecological response as a linear combination of independent predictor variables.
Model Assumptions: This model assumes that the response data are drawn from some statistical distribution other than that of the continuous Gaussian (Normal). Model errors require a specific statistical distribution that is paired with a link function that relates the linear function of predictors to some function of the response variable. For instance, the Poisson distribution requires a log link function to model discretely-distributed response variables (e.g., counts).
Model Response Data: Presence-absences and other binary data as well as counts, proportions, and ordinal data.
Model Explanatory Data: Discrete, categorical, and continuous predictor variables are each feasible to apply in these models.
Model Links and Use with R: To fit these models, you can use one of the following functions-
Furthermore, an introductory lecture about generalized linear models is also available at http://statmath.wu.ac.at/courses/heather_turner/glmCourse_001.pdf.
Example Papers: To learn more about how these models can be used to estimate species’ distributions, you can consult the following examples from the literature-
Example with R: