Utilization of Generalized Linear and Generalized Additive Models to Predict Plant Species Distributions in Peatlands

Authors:
Rémy Pouliot

Book:
Proceedings of the 14th International Peat Congress

Venue:
Stockholm

Keywords:
bogs-gam-glm-niches-predictive-curves

Documentfile:
Pouliot 2012: Utilization of Generalized Linear and Generalized Additive Models to Predict Plant Species Distributions in Peatla

Summary:

Theme I.  Inventory, stratigraphy and conservation of mires and peatlands

SUMMARY

Predictive models as generalized linear or generalized additives models (GLMs or GAMs) are tools that provide statistical relationships between species abundances and a series of climatic or chemical variables. The prediction of the probability to find a species in non-surveyed locations and the quantification of species presence are thus possible. GLMs and GAMs were used for the prediction of the distribution of ericaceous shrubs and Sphagnum mosses in natural peatlands along many environmental gradients. In addition, we will discuss of the possibility to use GLMs or GAMs results as a tool evaluating the success of peatland ecological restoration or in a goal of species conservation.