RESEARCH at Salford University could have global implications for weather predictions.

It also means that rainfall forecasts in a semi-arid part of Africa could be revised.

Three years ago climate model simulations predicted there could be a 25 per cent fall in rainfall in the Sahel area - a band of land south of the Sahara - by 2100. An increase in greenhouse gases was the cause.

But a new study by Salford University scientist Dr Adrian Chappell has cast doubt on long-held views that rainfall in the Sahel is declining. And this could alter methods of weather prediction across the globe.

Dr Chappell studied data from weather stations in the West African Sahel - a region at the southern edge of the western Sahara which was thought to have suffered from a steep decline in rainfall over the last 40 years.

His new estimates of uncertainty in the rainfall have important implications both in the region and globally.

They show that recent decline in rainfall cannot be detected. It means simple comparisons between a predicted drier climate, dust production and rainfall may be no longer valid and may need to be re-evaluated. In the Sahel the weather stations scientists use for rainfall research are clustered around the more densely inhabited coastal areas.

In contrast, in the drier north and east of the region, there are many fewer stations which rarely see any rain. Also, each year some rainfall stations close down and new ones are set up - resulting in a changing network of rainfall observations.

To compensate for this variation in weather stations, researchers have previously made simple statistical adjustments and results have suggested a long-term decline in the amount of rainfall in the region.

Adrian's studies show previous estimates have been poor at mapping the likely range of rainfall each year.

He said: "It is high-time climate observations contained `uncertainty estimates' so we know how much confidence to place in our models.

"

We are now entering a period where this data will be even more vital."