Improving flood hazard and exposure modelling in Rwanda by incorporating community-based flood evidence : A case study of Sebeya River catchment
(2026) In Journal of Hydrology: Regional Studies 66.- Abstract
Study region: Sebeya catchment in northwestern Rwanda. Study focus: The Sebeya catchment is highly vulnerable to flooding due to steep terrain, intense rainfall, and the expansion of human settlements. The aim was to develop a hydraulic model for flood-hazard and flood exposure mapping. The HEC-RAS model was developed for the most exposed stretch of the river, using data on water levels, flows, and topography, along with unique bathymetry data and community-based flood observations from the extreme event in May 2023. Flood hazard maps were developed for different return periods and overlaid with building data to represent exposure. New hydrological insights for the region: The hydraulic model showed satisfactory performance, with RMSE... (More)
Study region: Sebeya catchment in northwestern Rwanda. Study focus: The Sebeya catchment is highly vulnerable to flooding due to steep terrain, intense rainfall, and the expansion of human settlements. The aim was to develop a hydraulic model for flood-hazard and flood exposure mapping. The HEC-RAS model was developed for the most exposed stretch of the river, using data on water levels, flows, and topography, along with unique bathymetry data and community-based flood observations from the extreme event in May 2023. Flood hazard maps were developed for different return periods and overlaid with building data to represent exposure. New hydrological insights for the region: The hydraulic model showed satisfactory performance, with RMSE values of 0.09 m and 0.08 m for calibration and validation, respectively, and NSE values of 0.86 and 0.89. The results revealed that the estimated exposure for the 200-year return period is 534 buildings, while uncertainty analysis indicated that flood exposure may increase considerably under high flood magnitudes. The return period flow that best reproduced the May 2023 flood event was the 200-year event, with a mean bias of −0.36 m and a standard deviation of 1.02 m between community-based flood observations and simulated flood depths. These findings highlight the usefulness of hydraulic modelling combined with community-based information for improving flood-hazard mapping and flood exposure assessment in the Sebeya catchment and similar data-scarce regions.
(Less)
- author
- Hahirwabasenga, Joseph
LU
; Inamdeen, Fainaz
LU
; Knutsson, Linnea
; Nilsson, Erik
LU
; Larson, Magnus
LU
; Bizimana, Hussein
; Wali, Umaru Garba
LU
and Persson, Magnus
LU
- organization
- publishing date
- 2026-08
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- HEC-RAS, Hydraulic simulation, Inundation mapping, Return period estimation, Stakeholder interviews, Water depths
- in
- Journal of Hydrology: Regional Studies
- volume
- 66
- article number
- 103572
- publisher
- Elsevier
- external identifiers
-
- scopus:105039930996
- ISSN
- 2214-5818
- DOI
- 10.1016/j.ejrh.2026.103572
- language
- English
- LU publication?
- yes
- id
- 3e761fbb-8fc6-4526-b960-0fb6092f02ed
- date added to LUP
- 2026-09-02 12:22:49
- date last changed
- 2026-09-02 12:24:00
@article{3e761fbb-8fc6-4526-b960-0fb6092f02ed,
abstract = {{<p>Study region: Sebeya catchment in northwestern Rwanda. Study focus: The Sebeya catchment is highly vulnerable to flooding due to steep terrain, intense rainfall, and the expansion of human settlements. The aim was to develop a hydraulic model for flood-hazard and flood exposure mapping. The HEC-RAS model was developed for the most exposed stretch of the river, using data on water levels, flows, and topography, along with unique bathymetry data and community-based flood observations from the extreme event in May 2023. Flood hazard maps were developed for different return periods and overlaid with building data to represent exposure. New hydrological insights for the region: The hydraulic model showed satisfactory performance, with RMSE values of 0.09 m and 0.08 m for calibration and validation, respectively, and NSE values of 0.86 and 0.89. The results revealed that the estimated exposure for the 200-year return period is 534 buildings, while uncertainty analysis indicated that flood exposure may increase considerably under high flood magnitudes. The return period flow that best reproduced the May 2023 flood event was the 200-year event, with a mean bias of −0.36 m and a standard deviation of 1.02 m between community-based flood observations and simulated flood depths. These findings highlight the usefulness of hydraulic modelling combined with community-based information for improving flood-hazard mapping and flood exposure assessment in the Sebeya catchment and similar data-scarce regions.</p>}},
author = {{Hahirwabasenga, Joseph and Inamdeen, Fainaz and Knutsson, Linnea and Nilsson, Erik and Larson, Magnus and Bizimana, Hussein and Wali, Umaru Garba and Persson, Magnus}},
issn = {{2214-5818}},
keywords = {{HEC-RAS; Hydraulic simulation; Inundation mapping; Return period estimation; Stakeholder interviews; Water depths}},
language = {{eng}},
publisher = {{Elsevier}},
series = {{Journal of Hydrology: Regional Studies}},
title = {{Improving flood hazard and exposure modelling in Rwanda by incorporating community-based flood evidence : A case study of Sebeya River catchment}},
url = {{http://dx.doi.org/10.1016/j.ejrh.2026.103572}},
doi = {{10.1016/j.ejrh.2026.103572}},
volume = {{66}},
year = {{2026}},
}