Skip to content
No. 023Ruled outDisaster early warningUpdated Nov 16, 2026

Building a satellite-based record of past rain-triggered landslides in the East African highlands, so that the first rainfall warning thresholds for the region can be set and tested

Good it would do
2 / 5
Can AI help
2 / 5
How neglected
3 / 5
Can we check it
3 / 5

Why we stopped

Rejected on prior art: rainfall thresholds for the western East African Rift were published in 2019 (with Rwanda and Malawi thresholds since), and a BELSPO-funded RMCA/VUB group has already built the satellite-dated landslide inventory the scout proposed (101 events, 2017-2021). What is left (an honest held-out false-alarm test, an IMERG refit, the Ethiopian and Kenyan highlands) is held back by missing no-landslide data, coarse satellite dating and the lack of any agency that issues landslide warnings, not by compute; the outside skeptic dissents that narrower versions are not yet ruled out.

Why it's stuck

Landslide warning research has passed Africa by, and the bottleneck is data, not interest. A 2022 study of Rwanda (Uwihirwe and others, Natural Hazards and Earth System Sciences 22, 3641) states that, according to Guzzetti and others (2020), there are no landslide early warning systems in African countries, despite the high number of fatal landslides recorded there.

What we'd try

Run landslide detection over the full Sentinel-1 radar and Sentinel-2 optical archives (2015 onward, free) for the Kivu-Rwanda-Burundi highlands, Mount Elgon and the Malawi highlands. Adapt the open SAR-LRA model to rain-triggered landslides, then date each detected landslide to a window of days.

Source: https://nhess.copernicus.org/articles/22/3641/2022/

Think we got it wrong?

Ruled-out problems can come back. Tell us what we missed.

Members can read the full record for this problem, check its sources and comment. Open the member view → Not a member? Here is how to join.