[LONDON] A comprehensive 30-year dataset of African rainfall could soon help test climate change predictions and improve climate models, according to a UK researcher.
David Grimes, who studies satellite data at the University of Reading, told SciDev.Net that his group will release the complete, open-access data set within a year.
Researchers can already provide good short-term forecasts of Africa's weather but lack the detailed and consistent long-term data needed for accurate climate predictions.
The new data come from a European Meteosat satellite that has been collecting data over Europe and Africa. The data will supplement the poor ground data on rainfall to help improve climate predictions, which are often contradictory.
"Some models predict an increase in rainfall in some areas, other models predict a decrease of rainfall in the same area, and part of the reason for that is that data coming out of Africa [are] very poor and very sparse," Grimes said.
Many experts think that climate change will make the African climate more variable, with more extreme events, such as this summer's drought in the Horn of Africa.
This increased variability may also raise the risk of floods. Geoff Pegram, emeritus professor at the University of KwaZulu-Natal in South Africa, expects "longer periods of dryness and, when we do have rain, it is likely to be heavier".
The lack of good rainfall data has prevented climate models making robust predictions about how the climate will change at specific locations.
The new data "can tell us whether the rainfall and the climate in particular areas, at particular times of year or seasons, have been changing in the past 30 years, and then we can compare that with what climate models predict," said Grimes. "If the climate models say the same thing as our data sets that would give us much more confidence in their future predictions."
Previous data sets have lacked consistency. The Global Precipitation Climatology Project, for example, has global rainfall data but uses different methods of calculation for different periods, said Grimes, making it harder to understand how the climate has changed.
Tufa Dinku, a researcher at the International Research Institute for Climate and Society at the Columbia University, United States, said: "This data set is unique in that it uses a single algorithm and single satellite sensor, which ensures the consistency of the time series. There is no other satellite rainfall product that goes back to 1983 at ten-daily time scale and spatial resolution of about five kilometres.
"But the dataset is as good as the number of stations used for calibration," he said. Before the data set can be used, it must be calibrated against ground data. Satellites provide rainfall estimates, but they must be compared with ground data to know how they translate to actual amounts of rainfall.
Grimes said that he is planning to run a series of workshops in Africa to calibrate the estimates against their rain-gauge data and train scientists on how to use the data set.
"Africa is the worst continent, outside Antarctica, for the distribution of rain gauges — that's really the reason we do the satellite monitoring, because if you just use the gauges you can't get a complete picture of what's happening," he said.
"If you don't know what the climate is to start with, then you can't really decide whether it's changing or not."