Hydrology, Cryosphere & Earth Surface Research Spotlight

A Database of African Precipitation from 1983 Onward

Satellite infrared observations are used to reconstruct African precipitation records for the past 30 years in an attempt to infer rainfall variability.


Many observation-based climate monitoring systems do a poorer job representing precipitation variability over Africa than other continents. When it comes to important measurements like precipitation, records for Africa just aren’t as long, dense, or consistent as for some other regions. Even in the modern era, political and social turmoil mean that, in some parts of Africa, ground-based measurements often suffer gaps or delays.

Many Africans rely on accurate rainfall estimates. Unsurprisingly, rainfall estimation algorithms designed specifically for Africa tend to improve performance better compared to other algorithms, although there are few that are tuned for the continent.

Using infrared satellite observational records, Maidment et al. have built a climatological database that reconstructs precipitation trends since 1983. The database could serve as a baseline for interpreting and anticipating future conditions. The authors’ project builds from earlier work in which researchers designed an algorithm to turn infrared observations from the Meteosat satellites into precipitation estimations for all of Africa.

Unlike other previous attempts to produce an African precipitation climatology, the authors’ system relies exclusively on calibrated satellite-based observations to determine year-to-year changes in rainfall. Not using real-time ground-based gauge data strips the records of some details but avoids issues around inconsistent or inaccurate gauge measurements. The authors suggest that their climatology should be seen as complementary to existing records rather than as a replacement.

Testing their precipitation climatology against existing records, the scientists find that it accurately reflects temporal and spatial patterns, and captures interannual variability. The data set does, however, have a consistent dry bias, underestimating precipitation by roughly 20%. (Journal of Geophysical Research: Atmospheres, doi:10.1002/2014JD021927, 2014)

—Colin Schultz, Freelance Writer

Citation: Schultz, C. (2015), A database of African precipitation from 1983 onward, Eos, 96, doi:10.1029/2015EO023415. Published on 6 February 2015.

© 2015. The authors. CC BY-NC 3.0