Artificial Intelligence for Computational Sustainability: A Lab Companion/Biodiversity

Overview

Biodiversity is the "totality of genes, species, and ecosystems of a region". The scale of biodiversity ranges from the variation of genes within a given species population, to the variation of life forms across an ecosystem, biome, or an entire planet. Biodiversity is considered to be a measure of the health of ecosystems, since it depends on climate and interactions among different species.

Species Distribution

Partial map of North America, showing the Native range of the painted turtle (C. picta), separated into its four sub-species.      Eastern (C. p. picta)      Midland (C. p. marginata)      Southern (C. p. dorsalis)      Western (C. p. bellii)      Mixing of Western and Midland Painted Turtles      Mixing of Eastern and Midland Painted Turtles      Mixing of Southern and Eastern Painted Turtles

The distribution of a species describes the spatial arrangement of a biological taxon. The pattern of distribution is determined by a variety of factors, including climate, resources, and dependence on other species. Distribution patterns can change seasonally in response to the availability of resources, and over time in response to environmental changes. The distribution of a species, also called a species range, can be represented using a species range map.

Note that species distribution should not be confused with dispersal, which is the movement of individuals away from their area of origin or from centers of high population density. Dispersion usually takes place at the time of reproduction, or through translocation by people, animals, wind and water. Globalization and the transportation industry are significant factors in species translocation, causing a wider distribution for many species.

Species distribution modeling is the process of using computational methods to predict the geographic distribution of one or more species based on a combination of environmental factors (including climate, land cover, water depth, etc.), resource availability, and the distribution of other species. Species distribution modeling can also incorporate dispersal factors, migration patterns, and environmental changes to predict changes over time. The particular distribution of a species also depends on the scale at which it is viewed. Species distribution models can be used to assess climate change impacts, evaluate conservation policies, measure the effects of deforestation, and prioritize areas for policy intervention.

Lab: Species Distribution Modeling Using Maximum Entropy

See Machine Learning For Prediction / Lab: Species Distribution Modeling Using Maximum Entropy