MaxEnt is a Java application developed by Robert Schapire and Miro Dudik of Princeton University and Steven Phillips of AT&T's Research Lab. It can be downloaded from Schapire's website and is free for educational and non-profit research use. It is intended to model the geographic distribution of plant and animal species. In the developer's words, “this software takes as input a set of layers or environmental variables (such as elevation, precipitation, etc.), as well as a set of georeferenced occurrence locations, and produces a model of the range of the given species”.
Uses a maximum-entropy approach to predicatively model species distribution
MaxEnt will run on any computer running Java 1.5 or later. It requires at least 512 mb of available memory. Windows users can manually increase its memory usage up to approximately 1.3 gb.
Presence localities (samples) are selected in the “Samples” tab. Samples files must be in .csv format in the following syntax:
Coordinate systems other than longitude/latitude may be used, but the x coordinate must precede the y coordinate. Duplicate samples are deleted by default. This feature can be disable if the user wishes to retain duplicate values.
Environmental variables are selected from the “Environmental layers” tab. They must be in ESRI ASCII raster (.asc) grids. If multiple variables (e.g. elevation, mean annual temperature, and ground cover) are inputed, each must be in the same resolution and within the same geographic bounds.
The gain is closely related to deviance, a measure of goodness of fit used in generalized additive and generalized linear models. It starts at 0 and increases towards an asymptote during the run. During this process, Maxent is generating a probability distribution over pixels in the grid, starting from the uniform distribution and repeatedly improving the fit to the data. The gain is defined as the average log probability of the presence samples, minus a constant that makes the uniform distribution have zero gain. At the end of the run, the gain indicates how closely the model is concentrated around the presence samples; for example, if the gain is 2, it means that the average likelihood of the presence samples is exp(2) ≈ 7.4 times higher than that of a random background pixel (Phillips, 2010, p. 112).
Once the independent variables are entered, MaxEnt will calculate the probability of suitable conditions occurring for the species, rather than the likelihood of the species' presence. MaxEnt produces an html report containing a map as a .png image as well as links to additional reports in tabular format.
Output html file with links to MaxEnt reports (Phillips, 2010, p. 112)
A map output of expected probability of conditions suitable for the Brown-throated sloth Bradypus variegatus occurring. Warmer colors indicate more suitable conditions (Phillips, 2010, p. 113).
One of the most powerful features of MaxEnt is the ability to project a species' potential range in the future as a result of environmental changes, e.g. an increase in mean annual temperature and precipitation due to climate change. In order to project a species' range, MaxEnt must run training sessions based on past and/or present conditions.
Schapire, R.E. 2011. Maxent software for species habitat modeling. Department of Computer Science, Princeton University. Link to download application: http://www.cs.princeton.edu/~schapire/maxent/. Also provides links to download full-text articles on MaxEnt (see Additional Resources) written by its developers.
Phillips, S.J. 2010. Species’ distribution modeling for conservation educators and practitioners. American Museum of Natural History, Lessons in Conservation. http://ncep.amnh.org/linc.
Young, N., Carter, L. and Evangelista, P. 2011. A MaxEnt model v3.3.3e tutorial (ArcGIS v10). International Biological Information System (IBIS), Colorado State University. http://ibis.colostate.edu/WebContent/WS/ColoradoView/TutorialsDownloads/A_Maxent_Model_v7.pdf.
Elith, J., Phillips, S. J., Hastie, T., Dudík, M., Chee, Y. E. and Yates, C. J. 2011. A statistical explanation of MaxEnt for ecologists. Diversity and Distributions, 17: 43–57. doi: 10.1111/j.1472-4642.2010.00725.x. http://onlinelibrary.wiley.com/doi/10.1111/j.1472-4642.2010.00725.x/pdf
Phillips, S.J., Anderson, R.P., Schapire, R.E. Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190:231-259, 2006. http://www.cs.princeton.edu/~schapire/papers/ecolmod.pdf
Phillips, S.J., Dudík, M., Schapire, R.E. 2004. A maximum entropy approach to species distribution modeling. In Proceedings of the Twenty-First International Conference on Machine Learning, pp. 655-662. http://www.cs.princeton.edu/~schapire/papers/maxent_icml.pdf
Bedia, J., Busqué, J. and Gutiérrez, J.M. (2011), Predicting plant species distribution across an alpine rangeland in northern Spain. A comparison of probabilistic methods. Applied Vegetation Science, 14: 415–432. doi: 10.1111/j.1654-109X.2011.01128.x.
You must have an account and be logged in to post or reply to the discussion topics below. Click here to login or register for the site.