This interdisciplinary research, supported by the National Science Foundation (NSF) and the government of Mexico and led by A. Townsend Peterson of the University of Kansas Natural History Museum and Biodiversity Research Center (UKNHM), relied on extensive collections of species data assembled by Mexico's biodiversity commission, CONABIO, and a powerful software program, the Genetic Algorithm for Rule-Set Prediction (GARP), created by SDSC's David Stockwell, which ran on a supercomputer provided by the NSF-funded National Partnership for Advanced Computational Infrastructure (NPACI).
In a nutshell, what the researchers found is that over the next 50 years the changing climate is predicted to bring about great instability for wildlife, reshuffling ecosystems and throwing new predators and prey together as new diseases and parasites are introduced, with a majority of species having smaller geographical ranges than today. "In some local communities more than 40 percent of species are expected to turn over, which will lead to a cascade of further effects," said Peterson, lead author of the Nature paper. "If you remove enough species from an ecosystem, it's like the old child's game of pick-up-sticks -- there are only so many changes you can make before the ecosystem just rearranges, and maybe strikingly."
Although researchers have long realized that climate change will impact biodiversity, they have previously lacked either comprehensive data on where species live for an area as large as Mexico, or software tools powerful enough to analyze the relationships between species and climate over such a large area.
"This research marks a major step forward in being able to investigate in a quantitative way the initial impacts of climate change on ecosystems and biodiversity," said Peterson. "This is important because the modifications affecting our climate are like a big experiment the whole world is doing without knowing what's going to happen." This innovative modeling research is providing scientific insight into climate impacts on biodiversity that can eventually give crucial guidance to policymakers in establishing effective species conservation and management plans.
"What's unique about what we've done is that for the first time we were able to look at a whole community in detail," said Peterson. Previous research has looked in a broad brush way at how climate change would affect an ecosystem. For example, if a warmer climate will cause a given habitat to move northward or uphill to higher altitude, researchers made the assumption that the various species would simply move along with the ecosystem. Other detailed studies have been limited to looking at only a few species.
In contrast, the Nature article outlines research that not only looks at a great many species across a whole community but also examines them one-by-one, using realistic assumptions about their ability to change where they live in response to climate change. "That species-by-species look has allowed us to appreciate just how big the differences are in the way each species can respond to climate change -- it's quite complex and individual," said Peterson.
The ecological niche models that the researchers developed were overlaid onto current landscapes as well as modeled landscapes transformed by climate change, in order to identify areas of potential distribution for each species as a result of changing climate.
"What we found is that the impacts aren't uniform, and the strongest effects turned out to be in the central Chihuahuan desert, south of New Mexico," said Peterson. In that desert area, the researchers expect the greatest species turnover and disruption of ecosystems, which will unleash a complex chain of consequences that are difficult to foresee in detail but will be highly disruptive to ecosystems and certain species in particular.
The success of this study required collaboration among several disciplines and groups who each brought unique capabilities to the task. The research team included A. Townsend Peterson and Miguel Ortega-Huerta of UKNHM; Jeremy Bartley of the department of Geography, KU; Robert Buddemeier of the Kansas Geological Survey; Victor Sánchez-Cordero, Jorge Soberón, and Adolfo Navarro Sigüenza of UNAM; and David Stockwell of SDSC.
"There are really two stories here -- the results of the research itself, and the interdisciplinary approach we used, which points to a new paradigm of doing science," said Victor Sánchez-Cordero of UNAM. By working in a team with biologists, collections specialists, computer scientists, and others, the researchers created a powerful tool that can be extended to explore the combined effects of climate change and other scenarios. The approach can be extended to other fields to produce, for example, risk maps of emerging diseases, agricultural pests, and invasive species.
Fundamental to this research are the extensive collections of observed species locations for Mexico. "The collections are truly comprehensive," said Peterson. "We used more than 110,000 unique records of observed species locations that include all 1,179 bird species, all 416 mammal species, and 175 butterfly species in Mexico." Peterson and UNAM colleague Navarro developed the data sets for birds, while Sánchez-Cordero helped with the mammal data sets and Jorge Soberón with the data for butterflies.
The collections were assembled through ongoing efforts by CONABIO, a leading model of managing national biodiversity data, which has painstakingly brought together in modern database software information on Mexican species held in more than 40 natural history museums around the world. With advanced information technology to help manage information on species occurrences from far flung museums, the researchers are finding that the "whole is greater than the sum of the parts" and they are able to extract far more insight from large, unified collections than from the separate fragments.
A second piece of the puzzle the researchers needed was predictions of future climate change for Mexico. For this the researchers made use of general circulation climate model predictions made available by the Intergovernmental Panel on Climate Change (IPCC). They considered two scenarios of increasing greenhouse gases over the next 50 years, which predict major modifications of Mexico's climate, with higher average temperatures and lower rainfall.
The third component of this research involved analyzing the data using the GARP software at SDSC, which is accessed through the Biodiversity Species Workshop (BSW) Web interface. The software allows the integration of a range of variable types, including categorical variables such as vegetation and soil and continuous variables such as temperature, rainfall, and topography, into a robust statistical model of the response of a species to its environment, or ecological niche. Predictions of changes in the geographical distribution for each species were the result of developing model sets under a suite of variables including present-day climate, and applying the model to a suite with the climate variables substituted for those produced under simulated greenhouse climate scenarios.
"There are a number of ways in which intensive computations contributed to these results," said Stockwell. "Firstly, the GARP software uses machine-learning methods from artificial intelligence research, called genetic algorithms, to iteratively generate, test, and adopt or discard ecological niche descriptions, thereby evolving robust sets of models. Secondly, analyzing the effects of climate change on each separate species results in thousands of separate models and predictions, necessary to show changes in community composition. Thirdly, we made the spatial scale as fine as possible given the computational resources, in order to achieve accuracy and definition for the species dispersal modeling. The analysis took about 30 days, running on the Sun NPACI multiprocessor parallel machine housed at SDSC."
In the future, the researchers plan to undertake field work to test the validity of the model predictions as well as extending this work beyond biodiversity studies to produce synthetic models that explore the combined effects of climate change and other scenarios such as invasive species. Explains Sánchez-Cordero, "We're excited about the power of this new modeling tool, as well as this interdisciplinary form of research."
Useful URLs:
University of Kansas Natural History Museum and Biodiversity Research Center - http://nhm.ku.edu/
Biodiversity Species Workshop (BSW) - http://biodi.sdsc.edu/bsw_home.html
DesktopGARP - http://beta.lifemapper.org/desktopgarp/
CONABIO, species data for Mexico - http://www.conabio.gob.mx/
UNAM, Universidad Nacional Autónoma de México - www.unam.mx, www.ibiologia.unam.mx, www.ecologia.unam.mx
IPCC, climate model data - http://ipcc-ddc.cru.uea.ac.uk/
About UKNHM: The University of Kansas Natural History Museum and Biodiversity Research Center is the nation's leading university museum and research center in biodiversity research -- discovering, documenting, and disseminating knowledge of life on earth, past and present. Since 1995, the KU museum and research center and its collaborators have innovated information technology to harness biodiversity data from collections worldwide -- the result of 300 years of the biological exploration of the planet -- to advance understanding of natural environments, enhance the power to predict environmental phenomena, and increase knowledge to inform natural resource management.
About UNAM: La Universidad Nacional Autónoma de México (UNAM) is the country's largest university, conducting more than fifty percent of its scientific research. The Instituto de Biología and Instituto de Ecología at UNAM are leading institutions in biodiversity studies that are indispensable in helping Mexico maintain its rich biodiversity, among the highest in the world. UNAM maintains Mexico's largest scientific collections of flora and fauna.
About SDSC: The San Diego Supercomputer Center is an organized research unit of the University of California, San Diego, and the leading-edge site of the National Partnership for Advanced Computational Infrastructure (http://www.npaci.edu/). SDSC is funded by the National Science Foundation through NPACI and other federal agencies, the State and University of California, and private organizations. For additional information about SDSC and NPACI, see http://www.sdsc.edu/ or contact David Hart, dhart@sdsc.edu, 858-534-8314.
Comments:
A. Townsend Peterson, UKNHM, 785-864-3926, e-mail town@ukans.edu
David Stockwell, SDSC, 858-822-0942, e-mail davids@sdsc.edu
Media Contact:
David Hart, SDSC, 858-534-8314, e-mail dhart@sdsc.edu
Journal
Nature