Enhancing Ecosystem Types Using Museum Data and ED Models
Enhancing Environmental Types Using Museum Data and ED Models Improves Assessments of Climate/Land-use Change Impacts
Daniel P. Faith1, Kristen J. Williams2, Simon Ferrier3 and Ben E. Lawson4
1Austalian Museum, Sydney, N.S.W., Australia; 2 CSIRO Sustainable Ecosystems, Atherton, Queensland,
Australia; 3CSIRO Entomology, Canberra, A.C.T., Australia; 4Griffith University, Brisbane, Queensland,
One strategy for assessing biodiversity losses due to climate and land-use change is based on estimated area losses for different biomes, vegetation types, or other environmental classes. Fractional area loss can be linked to fractional species loss within each class through species-area curves. Studies such as the Millennium Ecosystem Assessment simply take the sum, over all classes, of the estimated fractional species losses to determine total species losses. A weakness is that species overlap among classes is ignored. An alternative method estimates overlap (dissimilarities among classes) using museum collections data, places the classes in environmental space, and applies the ED approach. ED methods use p-median and related criteria to optimally sample environmental space, under an assumption of unimodal response of species to gradients. ED can use probabilities or fractions assigned to each site. When the points in environmental space are classes, the fractions may be inferred from species-area curves.
We explore simple scenarios to compare ED and the separate-classes approaches. For 3 classes, i, j, and k, suppose that estimated dissimilarities imply that j and k are 1 unit apart, while i differs from these two by 100 units. Each class is equally species rich. Species-area curves produce valid estimates for the fractional species losses in each class. For some scenarios of species fractional losses, the classes-only approach can indicate
that biodiversity has increased, when in fact it has decreased, as indicated by taking overlap into account through the ED method. ED provides a general strategy for calibrating given classes or types, based on species collections data from GBIF and other sources. The approach may assist global biodiversity monitoring within GEO BON (http://www.earthobservations.org/cop_bi_geobon.shtml), where remote sensing could provide temporal information on changes in land condition/extent, for input into ED calculations for different biomes.
reference: Faith D. P., Williams K, Ferrier S., and Lawson, B. (2009) Enhancing environmental types using museum data and ED models improves assessments of climate/land-use change impacts. INTECOL 10. Ecology in a Changing Climate, Two Hemispheres - One Globe. Abstracts. http://www.intecol10.org/abstracts/pdf/0908015Abstract01219.pdf
ED calculated among types, combined with probabilities, will have other applications.
The emerging IUCN Red List of Ecosystems (http://www.iucn.org/about/union/commissions/cem/cem_work/tg_red_list/ ) will provide ecosystems red list ratings for threatened ecosystem types. These ratings typically would indicate fractional losses of area-extent. A species-area relationship can convert these area-fractions to fractions of species lost (or probabilities of loss). Using probabilistic ED, threatened ecosystem types that are more distinctive (little compositional overlap with other types) might be given higher conservation priority. This could form an “EDGE” program (“ecologically distinctive, globally endangered”) analogous to the program for threatened species. In both cases, priorities reflect the need to conserve some lower-level units: for the “ecologically distinctive, globally endangered” EDGE, those units are species; for the “evolutionarily distinctive, globally endangered” EDGE, those units are species traits or features. For an “ecologically distinctive, globally endangered” program, the Faith et al INTECOL study reveals the importance of taking compositional overlap among ecosystems into account.
Probabilistic ED analyses may be useful within ecosystem types, to take heterogeneity within the ecosystem types into account. ED provides hypothetical species for “biodiversity viability analysis” (BVA). These “species” can be mapped, based on environmental predictors. The species then could be assessed by the standard range-related IUCN criteria, assigning threat status and an inferred probability of extinction. An ecosystem type therefore can be assigned an overall rating that reflects the extinction probabilities of both its actual (known) and hypothetical (unknown) species - for example, estimating the expected number of species that will be lost.
For more discussion on ED applied to types or classes, see:
Faith, D.P & Walker, P.A (1996) Environmental diversity: on the best-possible use of surrogate data for assessing the relative biodiversity of sets of areas. Biodiversity and Conservation, 5:399–415.
Faith, D.P & Walker, P.A (1996) Integrating conservation and development: incorporating vulnerability into biodiversity-assessment of areas. Biodiversity and Conservation, 5:417-429
Faith, D.P, Ferrier, S & Walker, P.A (2004) The ED strategy: how species-level surrogates indicate general biodiversity patterns through an ‘environmental diversity’ perspective. Journal of Biogeography, 31:1207–1217.
Faith, D.P & Walker, P.A (1994) Diversity: a software package for sampling phylogenetic and environmental diversity. Reference and user’s guide. v 2.1 CSIRO Division of Wildlife and Ecology, Canberra. Available at www.amonline.net.au/systematics
Dr Dan Faith , Principal Research Scientist