Babak Naimi is the programmer of sdm-R. Here some questions and answers related to the recent publication of this new software for species distributions modelling.
What is sdm?
sdm stands for “species distribution modelling”, and is the name of an extensible and reproducible R package for modelling, evaluating, and projecting species potential distributions in space and time. It provides a standardised and unified structure for handling species distributions data and supports markedly different modelling techniques. An important capability of the software is that scientists can easily modify existing methods, extend the framework by developing new methods or modelling procedures, and share them to be reproduced by other scientists.
What it aims to respond to? And How?
Species Distributions Models are used in a wide range of fields. The modelling procedure involves several steps and each might involve different methods. A thorough understanding over the whole procedure and different steps is needed to appropriately fit and use the models. In addition, the different methods are needed in the modelling procedure are usually provided by different software/tools following varying standards that make them difficult to integrate. So, an integrated framework enabling multiple methods to be fitted and compared simultaneously is required to move the field of species distribution modelling forward. sdm provides such framework, and makes it easier to to implement protocols that adhere to appropriate standards across the whole modelling procedure.
Why is sdm different from others packages?
Although the features offered by the other relevant packages can also be found in the sdm package, there are several capabilities that make sdm unique and different from the other packages. The most important capability is the generic object-oriented design of the package, which pools together methods and procedures in a way that can be extended by users. The design of the package is such that it facilitates reproducibility of methods and procedures by allowing inclusion of new approaches or modification of existing ones in a way that can be easily shared to the wider community. Another unique capability of the package is its graphical user interface (GUI) that makes it extremely easy to use even by the users who are not familiar with R. In addition, sdm puts all of the steps required in species distribution modelling into a workflow, that can be shared and reproduced whenever it is needed (also can be used to automate the procedure).
Is sdm finished or are you working on new functions to improve it?
Given that the package is an open-source project, I would say that it would never be finished as it can be developed by everyone who is interested in contributing to the project. Nevertheless, it is still actively under development by the authors in two directions. First, we are trying to improve appropriateness of different procedures involved in different steps of species distribution modelling according to the state-of-the-art of knowledge and the standards in the field. Second, additional methods will be included to cover all relevant aspects of the modelling procedure.
Naimi, B. & Araújo, M.B. 2016. sdm: a reproducible and extensible R platform for species distribution modelling. Ecography, 39: 368 – 375. doi: 10.1111/ecog.01881