27/11/2012 10:19 GMT | Updated 24/01/2013 05:12 GMT

The Hard Facts: Weather Kills People in Africa and Asia

As the world's climate experts gather in Doha, Qatar, for the COP-18 Climate Talks , major new findings on the effects of weather on deaths in Africa and Asia has just been published by the INDEPTH Network. Detailed findings can be downloaded.

Sometimes the global discussions on climate have been criticised for being too speculative and possibly based on questionable data. That's not the case here, though, because these analyses are based on more than 100,000 deaths carefully documented across seven INDEPTH sites in Burkina Faso, Ghana, Kenya, Tanzania, India and Bangladesh. These deaths occurred during more than 10 million observed person-years (that means, for example, a million people observed for 10 years).

What the analysts across the African and Asia sites have done with this huge dataset (to put it simply) is to take each death, knowing the date it occurred and the basic characteristics of the person who died (age, sex, etc.) and look at that alongside detailed weather records (temperature, rainfall, etc.) observed at the closest weather station to where the death occurred. With a lot of complex mathematics, it's then possible to work out whether deaths occur more at hotter or cooler, wetter or dryer times.

Firstly in West Africa, the Burkina Faso site found that mortality was much higher around the hottest days, particularly for young children. Nearby in Ghana, the picture was similar, with heat also increasing mortality among the elderly. In Nairobi, Kenya, high temperatures were associated with deaths among children, and non-communicable diseases among adults, while low temperatures affected babies, and rainfall increased deaths from pneumonia and non-communicable diseases. In a rural area of Tanzania, mortality was strongly seasonal, with child mortality being higher after rainfall. Moving to Asia, in India child deaths were associated with high and low temperatures and heavy rainfall, while adult deaths increased at high temperatures. In Bangladesh, two sites reported increased mortality with lower temperatures, and to some extent with hotter periods.

What can't be emphasised too strongly here is that these are analyses of real deaths and actual weather. They are not simulations or models - and it reflects the great strength of the INDEPTH Network that it is possible to analyse factual information in this way from parts of the world where reliable data are usually in short supply.

What does all this tell us? Apart from documenting what has actually happened, these results are really important when "What if...?" questions about possible future changes in climate are asked (by changes in climate, we mean long-term variations in weather). There is a lot of speculation about various parts of the world becoming hotter, or wetter, as a result of possible long-term changes. It's hard to interpret these possibilities with certainty. But these results from INDEPTH enable us to judge what the effects in terms of increasing numbers of deaths might be if certain levels of climate change were to come about in the future.

These results will be made available to delegates at the Doha climate meeting - hopefully contributing another piece to the complex jigsaw of the world's understanding of climate change.