Presentation

To follow biomod2 evolution of organization and features, the presentation below details the different steps and modules of the package.

Download pdf of WORKFLOW GRAPHIC

Download pdf of PRESENTATION

Data formatting : BIOMOD_FormatingData

• Formatting data : combine observations, coordinates and explanatory variables,
  and sample pseudo-absences for presence-only data (through bm_PseudoAbsences)

• Cross-validation : create calibration / validation datasets (through bm_CrossValidation)

• Modeling options (with the help of ModelsTable and OptionsBigboss datasets)

  • Define specific options for single models (through bm_ModelingOptions)
  • And tune with your own data (through bm_Tuning)

Single models : BIOMOD_Modeling

• Evaluation : transform predicted probabilities in binary values (through bm_BinaryTransformation),
  and compute evaluation metrics (through bm_FindOptimStat)

• Variables’ importance : estimate the impact of each explanatory variable on predictions (through bm_VariablesImportance)

Ensemble models : BIOMOD_EnsembleModeling

• Filter single models based on one evaluation metric
• Combine them according to different combinations
• And build ensemble models

Exploring outputs

• Evaluation : bm_PlotEvalMean, bm_PlotEvalBoxplot
• Variables’ importance : bm_PlotVarImpBoxplot
• Response curves : bm_PlotResponseCurves

Projecting models : BIOMOD_Projection, BIOMOD_EnsembleForecasting

• Project models onto new conditions
• Transform predicted probabilities in binary values (through bm_BinaryTransformation)

Species range change : BIOMOD_RangeSize