Till startsida
To content Read more about how we use cookies on gu.se
Sitemap

Predictive mapping of species & habitats - which methods are useful?

Content:

Introduction.
Mention something general about :

  • how to choose modelling method
  • going from model to map
  • interpretation of map

 

Four appropriate methods
Results from PREHAB indicate that the following methods are useful for finding the statistical relationship between predictor (environmental data) and the biological feature (response variable) you want to map:

* Generalized additive models (GAM),
* Multivariate adaptive regression splines (MARS),
* Random forest (RF) and
* Maximum entropy modelling (MaxEnt)


The ability of a modelling method to produce a reliable map of predicted distribution or abundance depends more on the study sites and number of biological samples, than on differences between the modelling methods themselves.

 

The choice of modelling method should be based on:

* composition of biological data - reliable presences and absences or presences only?
* relationship between response variable and environmental/predictor data - linear or nonlinear?
* interactions between environmental variables/predictor data – absent or present?

 

Vegetation - distribution
Generally, the obtained results from PREHAB indicate that good to excellent classification accuracy of vegetation distribution by these four methods is highly probable, where data consists of more than 400 samples and 30% presences.

There are a number of factors that should be considered when you select an appropriate modelling method:

  • Which type of biological samples (response variable) is available?
  • Which type of environmental (predictor) data is available?
  • What is the relationship between response variable and environmental data - linear or nonlinear?

 

Cover issues, such as:
Advice: (important points, "warnings")
General discussion of the methods: present the methods, basics and their pros and cons
PREHAB results (as examples)

 

Author: Martynas
 

More about choice of method...

  

Cover issues, such as:
• Advice: (important points, "warnings")
• General discussion of the methods: present the methods, basics and their pros and cons
• PREHAB results (as examples)

 

Author: Martynas

 

Highlights from PREHAB

The ability of a modelling method to produce a reliable map of underwater biodiversity depends more on the study sites and number of biological samples, than on differences between the modelling methods themselves.

The following modelling methods are all useful for finding the statistical relationship between predictor (environmental data) and the biological feature (response variable) you want to map:

* Generalized additive models (GAM),
* Multivariate adaptive regression splines (MARS),
* Random forest (RF) and
* Maximum entropy modelling (MaxEnt)

© University of Gothenburg, Sweden Box 100, S-405 30 Gothenburg
Phone +46 31-786 0000, About the website

| Map

The University of Gothenburg uses cookies to provide you with the best possible user experience. By continuing on this website, you approve of our use of cookies.  What are cookies?