Mapping the April 2016 Nairobi Flooding from Online Sources

Faith Taylor's picture

Reasearchers Faith Taylor, James Millington and Bruce Malamud investigate the location of floods in Nairobi from social media and online news.

Nairobi flood map april 2016

In late April to early May 2016, Nairobi experienced some very heavy rainfall events, resulting in widespread flooding and a fatal building collapse. There was much talk on Twitter and in the news about this flooding, with the citizens of Nairobi reporting in near real time where floods had occurred.

Throughout the first week of May 2016, we searched through Tweets, digital news and reports to identify and map the areas that were flooded. Our main source of information was Twitter, where we searched through Tweets containing the hashtag #NairobiFloods, in some cases doing further investigation into individual Tweets by online searches to better identify the timing and location of the flood event. We also closely followed Tweets and status reports from the Nairobi branch of the Kenya Red Cross, who are one of the main responders to flood events and searched Google News to find digital articles about the current flooding in Nairobi. This is very much an approximate science, and there are issues such as mis-reporting, differences in place names, language barriers, and biases in reporting (e.g., those who cannot afford access to a computer or smartphone may be less likely to Tweet about flooding in their area). So, on our map, the flood area extent (marked in dark blue) represents uncertainty about the exact location rather than the spatial extent of the water. We have approximately marked the centre point of each flood extent with a yellow star on the map.

Nonetheless, the very approximate map of flood locations that we have assembled tells an interesting story about the topography of Nairobi, and its impact on flooding. This blog post by Panoramicdon neatly explains how Nairobi is essentially split into two areas of differing topography: in the West, there are the higher lands and steeper slopes, which tend to be dominated by the wealthy in leafy areas such as Karen and Westgate. To the East, the land is lower and flatter, which tends to be dominated by industrial and lower-class areas such as Mathare and Eastleigh. This is a result of the Rift Valley system of faults, which in a nutshell is the land of eastern Africa very slowly moving apart over millennia. The division between the steeper highlands and flatter low land is quite distinctive, resulting in a fairly sudden drop that separates the city in two – roughly marked in pink on our map.

The three main rivers of Nairobi run from West to East (flowing from steep, higher land in the West to flatter lower land in the East) marked in light blue on our map. When the rivers cross this sudden change in topography, Panoramicdon describes how the water is able to spread out, rather than being constrained to steep channels, which results in flooding.

When we map the locations of floods reported from Tweets, news and reports (yellow stars and dark blue patches on the map), there is quite a clear spatial pattern of these floods occurring in close proximity to this sudden change in topography. 19 out of the 23 floods we mapped occurred on the eastern side of this topographic division (and 2 others were very close to the western side of division, which could be attributed to lack of certainty in our spatial locations of the flooding). Most of these floods occur near to the major rivers of Nairobi, but we are now looking into other floods possibly intersecting drainage lines identified from the topography.

The next step we are undertaking is to investigate whether such a neat spatial trend is present in previous flood events – such as the heavy rains in January and November 2015. Clearly, there is much potential in near-real time flood location mapping from Twitter and online news, but this will also need to be followed up with more rigorous established techniques to more thoroughly understand the occurrence of flooding in Nairobi.