Altitude referenced biological databases - a tool for understanding mountain biodiversity.
A GMBA-SCOPE Rapid Assessment Project

http://gmba.unibas.ch/index/index.htm

Contact: Eva Spehn, Sylvia Martinez (GMBA@unibas.ch)

The Global Mountain Biodiversity Assessment Network was initiated by the Swiss Academy of Sciences in 2000 in fulfillment of Agenda 21 with DIVERSITAS. The network’s activities were officially taken up by convening the First International Conference on Mountain Biodiversity in Rigi-Kaltbad LU, Switzerland in September 2000. The research network focuses on biological richness, its function and change at the high elevation end of the biosphere.

Modern approaches to biodiversity assessments are increasingly being built around database technology. The Global Mountain Biodiversity Assessment (GMBA) aims for an internationally concerted action among electronic database entries by major collections worldwide to produce altitude referenced subfiles as far as available from their database. Worldwide databases on organisms until now have not been mined for such an assessment, designed to elaborate global conditions and trends in organismic taxonomic diversity with altitude.

This provides the opportunity for SCOPE to collaborate more closely with the Swiss SCOPE Committee and with DIVERSITAS, through a Rapid Assessment Process workshop. GMBA focusses on full elevational transect data (altitudinal comparison). Across the globe, the elevation of the alpine treeline will be used as the biogeographic reference line, which should allow comparison among climatologically coherent data (latitudinal comparison). All data need geographical coordinates to be represented by region, continent or mountain system. Also the latitude-altitude link is of interest. Given that treeline elevation is relatively robust for a given region (200 x 200 km) a single entry could help eco-referencing thousands of regional collections.

GMBA specific objectives

Based on a standardized procedure, the GMBA aims for a global assessment of geo-referenced mountain biodiversity information. Project leaders envisage frequency profiles by elevation which in order to distill altitudinal trends in coherent organismic groups. A major challenge will be the handling of collection bias, because high elevation biota are often sampled less intensively. Another challenge is that altitude is often confounded with variable trends in water availability and land use intensity affecting biodiversity, which are overlapping altitudinal trends in biodiversity.

Data availability

Global biodiversity informatics initiatives (such as GBIF) have already established biodiversity information networks, data exchange standards and an information architecture that enables interoperability and facilitates data-mining, one of the pre-requisites for this project. Increasing numbers of museum collection databases are geo-referencd and even provided on the web to users in GIS-linked spatio-temporal coverages (e.g. MaPSTeDI of biodiversity data in the southern and central Rocky Mountains by the University of Colorado Natural History Museum, Denver Museum of Nature and Science, and Denver Botanic Gardens, or plant, bryophyte and fungal diversity data in China's Hengduan Mountains region, with the The Field Museum in Chicago, Arnold Arboretum at Harvard and the Kunming Institute of Botany in China, http://hengduan.huh.harvard.edu/fieldnotes). In addition, a worldwide project coordinated by the University of California at Berkeley is developing a universal system for geo-referencing the diverse specimen records in natural history collections, in order to increasing the availability of geo-referenced species distribution data (BioGeomancer project, www.biogeomancer.org).

Geo-mapping biodiversity data

Databases allow for great flexibility in a wide range of questions that can be asked and biological and ecological applications. With geo-referencd biodiversity data sets, it is possible to examine geospatial patterns linked to features such as habitat, elevation and climate where these data are recorded in the species treatments, thus permitting selecting mountain-relevant data. Modern statistical techniques on such large datasets, with the possibility to merge with other biological and non-biological databases, is an entirely new possibility of data mining and more recently awareness has increased that such annotations such as georeferencing are key to ecological interpretation of records. A recent workshop of the National Evolutionary Synthesis Center of Duke University (Wrightsville Beach, NC, May 31 – June 3, 2005) gave strong support for the concept that link biodiversity databases with geographic data, to provide a basis for geo-mapping of diversity information.

Foci

Data requirements:
Best data entries (A) would include -- in hundred meters (or better) of altitude above sea level -- (1) center of distribution, (2) lower range limit, (3) upper range limit. However, because such information is normally not contained in taxon oriented files, (B) the minimum information required would be the elevation of the collecting site. In cases where only a range is given, the mean will be used for narrow ranges (<500 m) and data with larger ranges only will be dismissed. For taxa which had been collected repeatedly, B-type data may permit estimation of A-type data. A very important key for geo-referencing in montane areas is that we can use digital elevation models in a GIS context to increase the accuracy and reduce the uncertainty of a geo-referenc by indexing against elevation values or ranges provided by collectors.

Modelling species distribution and ecosystem boundaries:
Species distribution (niche) modeling may also prove to be an effective method for comparing climatic and elevation data for certain candidate organisms. Using both climate and elevation surfaces, models allow statistical extraction of climatic, elevation, and addititonal remote sensing data for each taxon that is modeled. It should also be possible to model not just species, but features such as alpine treelines within a given region. Models for individual taxa can also be combined into community surfaces that could potentially demonstrate association or mutual exclusivity between taxa. Recent advances in algorithm development allow for generation of species distribution models using as few as six geo-referencd collection records, as well as incorporation of bias surfaces that can correct for oversampling near populated areas and/or along roads.

Final Products:

• A freely accessible mountain data portal which will include a working list of known mountain species according to the implementation of the CBD programme of work (PoW) for the Global Taxonomy Initiative and mountain biological diversity PoW.
• A synthesis volume including chapters by mountain regions or by major organismic groups provided by workshop participants, and other results of the workshops (methodological approaches, comparisons of mountain regions on a continental scale). The synthesis volume, provisionally entitled "E-mining for global trends in mountain biodiversity“, should form the third publication of the GMBA series with CRC Press (Vol I: Mountain Biodiversity. A global assessment, Eds Ch. Körner & EM Spehn, 2002, Parthenon Publishers; Vol II: Land Use Change and Mountain Biodiversity, Eds: EM Spehn, M Liberman, Ch. Körner, 2005, CRC Press).

Scientific Advisory Committee (SAC) nominees

A small advisory group with database experience has been identifed for the GMBA RAP project:

Christian Körner, University of Basel, Switzerland (chair)
Michael Donoghue, University of Yale, USA
Jose Sarukhan Kermez, National Institute of Ecology, UNAM, Mexico
Mary Kalin Arroyo, University of Chile, Chile
Hang Sun, Kunming Institute of Botany, China

Last up-dated May 2008