Improving flood and drought prediction in Africa
In many vulnerable areas of the world knowing when, where and how much rain has fallen often proves a trickier problem than expected.
Now, research at the University of Hull will potentially allow for more accurate flood and drought estimations to be made.
For many vulnerable areas of the Earth, local agencies rely heavily on satellite observations to measure rain to fill a gap in ground based instrumentation.
However, the observations by satellites often do not directly observe rainfall, and so they have uncertainties about where it has rained and how much rain has fallen.
To make accurate computer models of the environment – and so to predict droughts or floods – this detail is essential.
‘Accounting for the uncertainty’
Now PhD student Chris Skinner has developed a method to better utilise these uncertain rainfall estimates when predicting river flows.
Chris said: “In sub-Saharan Africa there is a lack of ground-instrumentation that can record rainfall, yet many areas are particularly vulnerable to changes in rainfall. It is important to have timely and accurate estimates of rainfall and satellite data can fill this gap.”
“However, when used at small scales the satellite data can often miss areas of rainfall, or mis-detect rainfall where there is none. To account for this we produce several hundred unique, but equally possible predictions of the rainfall each day.”
“The research developed a method to setup hydrological models, used to predict river flows from rainfall estimates, for use with these ensemble representations of rainfall. This method was found to improve the accuracy and reliability of the river flow predictions when compared to more commonly used methods.”
‘Another small step forward’
Understanding how rivers flow will have huge implications on predicting where and when there may be floods or droughts.
“The recent devastating flooding in Malawi really highlights just how important it is to improve the predictions of river ws in Africa,” Chris continued.
“Ensemble rainfall estimates have been used by scientists for some time now, but a lot of research into them is still needed, particularly how they are used with other environmental models. Accurate forecasting of rainfall and river flows in Africa is a massive issue, and this is another small step in improving the situation.”
Chris undertook the project under the supervision of Tim Bellerby, working with colleagues from the TAMSAT (Tropical Applications of Meteorology using SATellite data and ground-based observations) team, University of Reading.
The research, carried out as part of a University of Hull 80th Anniversary Scholarship, has recently been published in the Journal of Hydrology. The paper can be viewed for free until the 27th February 2015.