Skip to content

🌍 Alien species risk modelling and mapping

License

Notifications You must be signed in to change notification settings

trias-project/risk-modelling-and-mapping

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

53 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Alien species risk modelling and mapping

This repository contains the framework and R code in development for risk modelling and mapping of alien species throughout Belgium and greater Europe at 1 km2 resolution as part of the TrIAS project. Please go to https://github.com/amyjsdavis/wiSDM for the latest code used in Davis et al., 2024 (https://doi.org/10.3389/fevo.2024.1148895)

Repo structure

β”œβ”€β”€ README.md              : Description of this repository
β”œβ”€β”€ LICENSE                : Repository license
β”œβ”€β”€ risk-modelling-and-mapping.Rproj : RStudio project file
β”œβ”€β”€ .gitignore             : Files and directories to be ignored by git
β”‚
β”œβ”€β”€ data
β”‚   β”œβ”€β”€ external          : external files (e.g. climate rasters, occurrence data, GIS files) required to run the   model. Will be available for download via Zenodo.
β”‚   └── results           : Risk maps for Belgium GENERATED
β”‚
└── src                    : R Code

Although theoretically possible, this workflow is not applied to all species listed in the published unified checklist and whose occurrences are found. We limit our analysis to a list of species labelled as emerging. The emerging status is object of another work package and it is a semi-automated process described in repository indicators: see webpage.

What you need to do run this workflow (see below for more details):

  1. GBIF taxon key and species name in a. csv file (see pra_mammals.csv) for format
  2. Climate and habitat raster data files downloaded from Zenodo (links will be provided when data is available)
  3. R studio installed in your computer.
  4. After cloning this repository, add folders to the existing folder structure shown above as shown below. This will allow you to use the relative path structure in the trias_sdm.R file.
β”œβ”€β”€ data
    β”œβ”€β”€ external
          β”œβ”€β”€ bias_grids (Global taxonomic occurrence grids, downloaded from Zenodo here )
          β”œβ”€β”€ climate (put climate rasters downloaded from Zenodo here)
          β”œβ”€β”€ data cube (European level occurrence cube) 
          β”œβ”€β”€ GIS (GIS data downloaded from Zenodo)
          β”œβ”€β”€ habitat (put habitat rasters downloaded from Zenodo here)
          β”œβ”€β”€ PRA (priority risk assessment species lists)

Workflow

The automated workflow can be divided in three sections:

  1. Develop global scale climate-only species distribution models (SDMs)
  2. Generate European level SDMs
  3. Forecast species distributions under climate change scenarios

What does the Trias modeling workflow do?

  1. Automatically generates IAS risk maps using machine learning. Our workflow requires only a species name and generates an ensemble of machine learning algorithms stacked together as a meta-model to produce the final risk map at 1 km2 resolution. Risk maps are generated automatically for standard IPCC greenhouse gas emission scenarios (RCP).
  2. Automatically generates confidence maps for each IAS risk map. These illustrate confidence of each individual prediction across your study extent.
  3. Addresses geographic sampling bias
  4. Incorporates best practices for the placement of pseudo-absences: pseudo absences are placed in the same ecoregions where presences occur. We use the global model to restrict pseudo absences to areas of low predicted suitability. We use the taxonomic occurrence grid (aka bias grid) to not place pseudoabsences in areas of low sampling effort. The taxonomic occurrence grid summarize the sampling effort of the higher taxon ,the modelled species belongs to.
  5. Detects and removes highly correlated predictors. Highly correlated predictors can have undesirable effects and confuse the interpretation of variable importance
  6. Integrates multiple machine learning algorithms to predict risk. It has been consistently demonstrated that the choice of algorithm has the largest impact on predicted risk and area of predicted risk.
  7. Assesses spatial autocorrelation in the residuals to assess the impacts of clustering. If high, thinning can be employed.

Inputs required run the workflow:

  1. Species name or list of species with name and the GBIF taxon Key (as in pra_mammals.csv)
  2. Can use the list to retrieve global occurrences for each species using the global_download.Rmd
  3. Europe level Occurrence Cube -The same list used in "1." can also be used to run the occurrence cube script. a. Only European level cube is needed for workflow, but can generate Belgium only cube to determine if species occurrence meeting modeling criteria b. Can also determine if there is a minimum number of occurrences; about 100
  4. Also need species name and GBIF key to input directly into the R script to run risk modeling code. I recommend copying and pasting from this list into R.
  5. Predictor data (will be downloadable from Zenodo)

Contributors

List of contributors

License

MIT License

Releases

No releases published

Packages

No packages published

Languages