More than 60 per cent of climate envelope studies maybe wrong

September 16th, 2008 - 3:39 pm ICT by ANI  

Washington, September 16 (ANI): A new research has cast doubts on the value of climate envelope models, saying more than 60 per cent of such studies maybe wrong.

According to a report in Nature News, estimates of the impact climate change will have on wildlife may be much less reliable than thought, with a research that is reopening debate over a widely used modelling method endorsed by the Intergovernmental Panel on Climate Change (IPCC).

Climate envelope models use current distributions of species to construct an idea of the climatic conditions that suit them. This envelope can then be used to see where species could live under predictions of future climate.

Use of climate envelope models has been contentious, not least because they omit a number of factors that may be as or more important than climate in controlling species distribution, for example human activity, interactions with other species and pure chance.

Now, a paper from Colin Beale of the UKs Macaulay Institute of Land Use Research, in Aberdeen, and colleagues has cast further doubt on the value of climate envelope models, saying they quite frequently do no better than chance in explaining why species live where they do.

At the moment, people are overconfident in our ability to make these projections, said Beale. We really need to start to think about models that include a lot more biology, he added.

Beale and his colleagues began by generating climate envelopes based on real distribution data for 100 European bird species. Then, they created 99 artificial distribution patterns for each of the real patterns.

These null distributions were designed to mimic the aspatial structure of the birds real distribution, but were created with no reference to climate at all.

This model is no better than a chance association: it is certainly not a model that should inform policy, said Beale.

The researchers then made climate envelope models for both the real and artificial distribution patterns for all 100 species.

Then, they applied a statistical goodness-of-fit test to determine how well the distribution data, both real and artificial, fitted into the climate envelope models based on them, and ranked the distributions for each species according to this measure.

For 68 of the 100 species, the five distributions that fitted their climate envelopes best were null distributions.

So, climate envelopes generated from real distribution data did not describe that data as well as some of the climate envelopes fitted to distribution data made up without any thought of climate.

The team takes this to imply that the association between the climate envelope and the animals range is in many cases no better than chance. (ANI)

Related Stories

Tags: , , , , , , , , , , , , , , , , , , ,

Posted in National |