In lieu of an abstract, here is a brief excerpt of the content:

CHAPTER SIX Environmental Data Ecological niche models are built from two sources of input data: (1) the known occurrences of the species of interest discussed in chapter 5, and (2) environmental predictors in the form of raster-format GIS layers. Whereas the quality of and biases in occurrence data have seen considerable documentation and discussion (e.g., Soberón et al. 2000; see chapter 5), the nature, quality, and biases of environmental datasets have seldom been considered in detail in niche modeling analyses, despite the key role that they play in the process of calibrating models (Peterson and Nakazawa 2008). In this chapter, we discuss conceptual and applied aspects of environmental data, in the context of building and interpreting ecological niche models. SPECIES-ENVIRONMENT RELATIONSHIPS In Hutchinson’s conceptualization of the ecological niche, discussed in detail in chapters 2 and 3, persistence and abundance of populations of species are determined by suites of scenopoetic and bionomic variables within an n-dimensional hypervolume of environmental space (Hutchinson 1957). As this concept is critical for ecological niche modeling, we must consider carefully how its details , ambiguities, and difficulties should guide us in implementation. In the first place, it is clear that the type and number of variables comprising the dimensions of the ecological niche vary from one species to another (Leibold 1995, Pulliam 2000). For example, bats are highly sensitive to low temperatures , while felids are more sensitive to vegetation structure (Kitchener 1991, Patten 2004). Moreover, even within one of these broader groups, a particular species may respond to one set of variables, while another responds to other features of the environment. Finally, the relative importance of particular environmental variables for a species may vary according to the geographic and biotic contexts. Environmental variables have been classified in various ways, depending on their relationships with, and influences on, geographic distributions of species. ENVIRONMENTAL DATA 83 While Hutchinson distinguished scenopoetic from bionomic dimensions of environments (Hutchinson 1978), Austin (2002) proposed two different breakdowns , as follows: 1. Idealized variables—This categorization focuses on the degree to which variables have direct physiological effects on organisms, which Austin (2002) subdivided into: • Indirect—Variables with no causal physiological effects on individuals, but that have a correlation with species’ occurrences because of correlations with other factors. Examples would include latitude and elevation. • Direct—Variables that affect organisms physiologically, but that are not consumed by them. Equivalent to Hutchinson’s scenopoetic variables; examples would include temperature. • Resource—Variables that are consumed by or affected by organisms. Equivalent to Hutchinson’s bionomic variables; examples would include food resources, presence of predators or parasites, or light in shade-limited plants. 2. Distal/proximal variables—Here, Austin (2002) sorted variables by the degree of causality of species’ responses to environmental factors (see figure 6.1). He divided variables into: • Proximal—Organisms respond directly to such variables; an example might be freeze durations that affect survival of cacti in northern latitudes directly. • Distal—Responses of organisms to these variables are not direct, but rather via multiple additional causal links; an example would be annual mean temperature, which manifests as a causal factor only via the freeze durations just mentioned earlier. It should be borne in mind that these categories are not mutually exclusive of one another—a variable may be proximal or distal and direct or resource at the same time. Beyond this basic classification, different environmental factors operate at different spatial and temporal scales. As a consequence, their relative importance in defining a species’ distribution and abundance can be highly scaledependent (Mackey and Lindenmayer 2001, Pearson and Dawson 2003, Sober ón 2007). In the early twentieth century, Joseph Grinnell proposed that species’ distributions are hierarchically structured in space, the most inclusive classes consisting of climatic variables (particularly temperature and humidity ) as the main drivers at coarse resolutions, whereas availability of food and 84 CHAPTER 6 refuges are the most important factors at finer resolutions (Grinnell 1917). More recently, it has become generally accepted that different controlling factors typically operate at certain corresponding “scale domains” (Pearson and Dawson 2003): again, macroclimatic variables influence distributions at coarser scales, whereas landscape features (e.g., vegetation cover) act at mesoscales , and specific habitat features and biotic interactions have strongest influences at local scales (Mackey and Lindenmayer 2001, Pearson and Dawson 2003; figure 6.2). Figure 6.1. Example summary of responses of a virtual species to eight environmental characteristics. Adapted from Austin et al. (2006). n 228 n 228 n 228 n 228 n 228 n...


Additional Information

Related ISBN
MARC Record
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.