PRE_FATE - build Plant Functional Groups (PFG)

See Principle of PFG FATE tutorial

A plant functional group, or PFG, is “a set of representative species [that] is classified based on key biological characteristics, to determine groups of species sharing ecological strategies” (Boulangeat, 2012).

RFate is a R package available on github and designed to provide support functions to the FATE software.

It contains documentation and functions to create and organize all input files required by the model, and building PFG is the first step to run a FATE simulation. The procedure presented below is based on RFate functions.

What are the key steps of this process ?

1. Selection of dominant species

• with the function PRE_FATE.selectDominant

2. Overlap of species environmental niches

• with either Principal Component Analysis (PCA) or Species Distribution Models (SDM) (but see PRE_FATE.speciesDistance)

3. Calculation of species pairwise distance

• by combining overlap and functional distances with the function PRE_FATE.speciesDistance

4. Clustering of species :

• calculate all possible clusters, and the corresponding evaluation metrics
  with the function PRE_FATE.speciesClustering_step1
• choose the best number of clusters from the previous step and find determinant species
  with the function PRE_FATE.speciesClustering_step2
• combine traits data and clustering to calculate mean / median trait values per PFG
  with the function PRE_FATE.speciesClustering_step3

What do you need ?

1. Selection of dominant species

• Gather occurrences for all species within the studied area
• Identify dominant species based on abundances and frequençy of sampling

with the function PRE_FATE.selectDominant

2. Overlap of species environmental niches

with either Principal Component Analysis (PCA) or Species Distribution Models (SDM)

• Option 1: Principal Component analysis
  • Gather environmental data for the studied area
  • Compute PCA over environment to create a climatic/habitat space
  • Calculate the density of each species within this climatic/habitat space from the PCA
  • For each pair of species, compute the overlap of the 2 considered species within the climatic/habitat space

• Option 2: Species Distribution Models
  • Gather environmental data for the studied area
  • For each dominant species, compute a species distribution model (SDM)
    combining environmental data and occurrences to determine the climatic/habitat niche of the species
  • With these SDMs, calculate the niche overlap of each pair of species

3. Calculation of species pairwise distance

by combining overlap and functional distances with the function PRE_FATE.speciesDistance

• Gather traits data for all dominant species within the studied area
  (traits need to be related to fundamental process of growth : light tolerance, dispersal, height…)
• Compute dissimilarity distances between pairs of species based on these traits and taking also into account the overlap of the 2 species within the climatic/habitat space (see previous step)

For further details about the data, please refer to Boulangeat, 2012.

4. Clustering of species

using the dissimilarity distances from previous step :

• calculate all possible clusters, and the corresponding evaluation metrics
  with the function PRE_FATE.speciesClustering_step1
• choose the best number of clusters from the previous step and find determinant species
  with the function PRE_FATE.speciesClustering_step2
• combine traits data and clustering to calculate mean / median trait values per PFG
  with the function PRE_FATE.speciesClustering_step3