biodivMapR is an R package aiming at processing remote sensing data provided
as raster files in order to produce spectral diversity metrics.
The first versions of the method were initially dedicated to optical
data processing and the spectral dimension was the key dimension to produce
features to relate to ecological information. However, since biodivMapR v2.0,
the workflow aims at being sensor agnostic:
users can provide raster data obtained from optical sensors (multispectral, hyperspectral sensors), LiDAR or radar data.
Therefore, diversity metrics are no longer spectral. Remotely sensed diversity metrics from a collection of rasters sharing the same projection, spatial resolution and extent but different sensors can also be provided.
biodivMapR was initially developed to produce
Since the earliest versions of the package, our aim is to provide a tool useful for the ecologist community, with the capacity to evolve and provide alternative ways to handle remote sensing data.
Therefore, two main representations of data are handled by biodivMapR:
-
a discrete representation based on clustering techniques, aiming at producing 'spectral species': pixels are grouped into clusters, assuming that user-defined features allow unsupervised clustering to identify ecologically relevant groups corresponding to species, species communities, structurally or functionally similar entities. This representation allows to directly compute diversity metrics usually derived from species inventories (Shannon index, Bray-Curtis dissimilarity...).
-
a continuous representation using directly the feature/trait space, aiming at producing Functional diversity metrics Richness, Evenness, and Divergence as defined in Villeger et al. (2008), as well as Dispersion and Rao entropy. These functional diversity metrics are computed with the package
fundiversity.
biodivMapR aims at providing an open source tool where additional diversity
metrics either based on the discrete or continuous representation of data can be
easily implemented.
If you need specific metrics, want to suggest additional ones, or contribute to their implementation, please feel free to contact the authors of the package
Efforts have been made in the recent versions to allow the integration of
biodivMapR into cloud infrastructures, in order to produce large scale maps.
Please contact the authors to exchange and collaborate on the integration into cloud infrastructures and the production of large scale maps.
As a software developed from researchers and mainly dedicated to researchers,
biodivMapR requires validation to identify its potential and limitations.
biodivMapR uses unsupervised methods. It is not based on modern learning
techniques. The validation is an on-going process, and contributions are welcome
to help us to provide recommendations for the application of biodivMapR in
different contexts.
The package remotes first needs to be installed from the CRAN
install.packages("remotes")
biodivMapR should then be ready for installation with the following command line in your R session:
remotes::install_github('jbferet/biodivMapR')
A tutorial vignette is available here.
If you use biodivMapR, please cite the following references:
Féret, J.-B., de Boissieu, F., 2019. biodivMapR: an R package for α‐ and β‐diversity mapping using remotely‐sensed images. Methods Ecol. Evol. 00:1-7. https://doi.org/10.1111/2041-210X.13310
Féret, J.-B., Asner, G.P., 2014. Mapping tropical forest canopy diversity using high-fidelity imaging spectroscopy. Ecol. Appl. 24, 1289–1296. https://doi.org/10.1890/13-1824.1
This research was supported by