Flood Risk Map of Kenya using GIS

For the doc version: Kenya Flood Risk Map

Abstract

The Republic of Kenya has recently witnessed a series of devastating hydrometeorological events, transitioning from a severe multi-year drought to catastrophic, El Niño-enhanced flooding between 2024 and 2025. These events have underscored a critical need for high-resolution spatial data to inform disaster risk reduction and humanitarian response. This research, produced as a Legacy Project for the Humanitarian OpenStreetMap Team (HOT) Community Working Group (CWG) Mentorship 2025, presents a comprehensive national-scale flood risk assessment for Kenya. The study employs a Geographic Information System (GIS) and Multi-Criteria Decision Analysis (MCDA) framework to synthesize six influential factors: rainfall intensity, elevation, slope, Land Use/Land Cover (LULC), distance to water bodies, and distance to road networks. Utilizing a weighted overlay methodology, the study reclassifies these parameters based on their hydrological and anthropogenic influence to produce a final flood risk map categorized into five classes: Very High, High, Moderate, Low, and Very Low. The analysis reveals that high-risk zones are predominantly concentrated in low-lying river basins and informal urban settlements, where high rainfall accumulation coincides with poor drainage and high exposure. The findings provide a strategic foundation for the OpenStreetMap community and disaster management agencies to prioritize anticipatory actions, refine field data collection, and enhance the resilience of vulnerable populations.

Keywords

flood, Kenya, flood risk, mapping, GIS, Multi-Criteria Decision Analysis, OpenStreetMap

Introduction

The Kenyan Paradox: Historical Context and Emerging Flood Dynamics

The geographical and climatic landscape of Kenya is defined by extreme variability, a characteristic that has become increasingly pronounced in the context of global climate change. In the years leading up to 2025, Kenya experienced what has been described as a “climatic seesaw,” swinging from the worst drought in forty years to unprecedented deluges that submerged vast sections of the country. This volatility is not merely a localized weather phenomenon but a manifestation of broader regional shifts in the East African climate, influenced by the Indian Ocean Dipole and the El Niño Southern Oscillation (ENSO). Historically, Kenya has navigated recurrent cycles of droughts and floods, but the events of 2024–2025 reached a threshold that challenged both national infrastructure and community resilience. The March-April-May (MAM) long rains of 2024, intensified by El Niño patterns, resulted in flooding that affected 40 out of 47 counties. By June 2024, official reports from the National Disaster Operations Centre (NDOC) indicated that 293,200 individuals had been displaced, and approximately 250,000 learners were out of school due to the destruction of educational facilities and the use of schools as temporary shelters. The fatalities recorded during this period exceeded 290, with many individuals still missing as of late 2024. The economic impact of these floods is equally profound. The agricultural sector, which provides the livelihood for a significant portion of the Kenyan population, suffered immense losses. Over 65,000 acres of cropland were damaged, and 11,000 heads of livestock were lost, exacerbating food insecurity in regions that were already struggling to recover from the preceding drought. Critical infrastructure, including 68 roads and 45 health facilities, sustained heavy damage, creating logistical barriers to humanitarian aid delivery. This report addresses the need for a predictive and diagnostic tool that identifies the spatial distribution of these risks, allowing for more efficient resource allocation and targeted disaster preparedness.

The Role of Open Mapping and the HOT CWG Mentorship 2025

This research project is situated within the institutional framework of the Humanitarian OpenStreetMap Team (HOT) and its Community Working Group (CWG). The HOT CWG Mentorship Program was established to foster peer-to-peer learning and knowledge exchange within the humanitarian open mapping space. By pairing experienced geospatial professionals with emerging mappers, the program aims to build local capacity in priority countries, ensuring that those most affected by disasters are equipped with the tools to map their own vulnerabilities. As part of the “Audacious Project,” HOT has committed to mobilizing one million volunteers to map areas home to one billion people by 2025. This ambitious goal is predicated on the belief that maps and data, while not directly saving lives, provide the essential infrastructure for those who do. The transition from project-based work to a community-centered approach is vital for the sustainability of these efforts. This Legacy Project, authored by Yakubu Enoch and Alex Muruthi, represents a bridge between academic research and community-driven action. It utilizes open-source software and humanitarian datasets to create a reproducible model for flood risk assessment that can be adopted by OSM communities across Sub-Saharan Africa. The mentorship program emphasizes professional development, open geospatial skills, and data in humanitarian work. This project specifically addresses the “Open Geospatial Skills” and “Data in Humanitarian” focus areas by demonstrating how advanced GIS techniques, such as Weighted Overlay Analysis, can be applied to real-world crises. By publishing this work on the OpenStreetMap diary, the authors contribute to a global repository of knowledge, encouraging transparency, public reflection, and the continuous improvement of data quality within the OSM ecosystem.

Aim

To produce a Flood risk assessment map of Kenya.

Objective

a. Criteria Dataset gathering. b. Map creation and analysis using weighted overlay/raster calculation. c. Flood risk data interpretation.

Study Area Map of Kenya.

The study area encompasses the entire landmass of the Republic of Kenya, located in East Africa, spanning approximately 580,367 square kilometers. Kenya’s geography is marked by its diversity, ranging from the low-lying coastal plains along the Indian Ocean to the high-altitude Central Highlands, divided by the Great Rift Valley.

Literature Review

GIS and Flood Risk Assessment in Kenya

The application of Geographic Information Systems (GIS) and remote sensing in flood management has undergone a revolution in the past two decades. In the Global South, where ground-based meteorological stations are often sparse, these technologies provide a vital alternative for risk delineation and vulnerability analysis.

The Evolution of Flood Risk Modeling

Flood risk is traditionally conceptualized as the product of hazard, exposure, and vulnerability. Early models relied heavily on hydraulic and hydrological simulations that required extensive localized data. However, recent research has favored a multi-parametric approach using Multi-Criteria Decision Analysis (MCDA) and the Analytical Hierarchy Process (AHP). Studies in areas like the Eldoret Municipality have demonstrated that combining factors such as rainfall distribution, elevation, slope, and soil type can produce reliable risk maps with high validation accuracy. In the Western Region of Kenya, particularly the Budalangi sub-county, research has shown a strong correlation between altitude and flood risk. Analysis of Landsat satellite images reveals that 90% of flood risk zones are located below 1,144 meters above sea level. These zones are often covered by natural vegetation or farmlands, while “safe zones” are predominantly occupied by human settlements and administrative centers. This settlement pattern suggests a degree of historical adaptation, but the rapid expansion of populations into marginal lands is eroding these traditional safety margins.

Challenges of Data Quality and Uncertainty

A recurring theme in the literature is the challenge of data quality and uncertainty in flood prediction. The accuracy of flood maps is entirely dependent on the quality of input datasets, such as Digital Elevation Models (DEMs) and satellite rainfall estimates. Overlooking data uncertainty can lead to significant errors in estimating the intensity and timing of floods, resulting in misleading policy decisions. In Kenya, studies in the Lake Victoria Basin have utilized Satellite Rainfall Estimates (RFE) to overcome the lack of ground data, finding that while daily accumulations may vary, the products are effective for detecting rainfall occurrence and seasonal surges.

Methodology

Weighted overlay sum and raster calculation

The methodology for this national flood risk assessment follows a structured GIS workflow designed for reproducibility and technical rigor. The core of the analysis is the Multi-Criteria Decision Analysis (MCDA) framework, which allows for the integration of diverse environmental and social variables into a single risk index.

Methods

A. Criteria Selection and Data Acquisition

Six criteria were selected based on their established hydrological influence on flood generation and the availability of national-scale datasets:

i. Rainfall (30%): The primary hydrological trigger. Data was sourced from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), providing 35+ years of quasi-global rainfall time series.

ii. Distance from Water (25%): Proximity to the drainage network is a primary indicator of riverine flood exposure. Derived from OpenStreetMap waterbody and waterway layers.

iii. Elevation (20%): Determines the direction of flow and accumulation points. Data was obtained from the Shuttle Radar Topography Mission (SRTM) DEM.

iv. Slope (10%): Affects the velocity of runoff and the rate of infiltration. Calculated directly from the DEM.

v. Land Use/Land Cover (10%): Influences surface permeability. Data was sourced from the Copernicus Global Land Service 100m land cover maps.

vi. Distance from Road (5%): Acts as a proxy for infrastructure exposure and can influence local drainage patterns.

B. Dataset Pre-processing

To ensure compatibility, all datasets were projected to a consistent coordinate reference system and resampled to a uniform spatial resolution. Vector layers, such as roads and water bodies, were converted to raster format using GDAL vector-to-raster tools. Proximity analysis was conducted for roads and water bodies using the “Multi ring buffer” tool to create continuous distance surfaces.

C. Reclassification and Standardization

Because the input rasters have dissimilar units (e.g., millimeters for rainfall, meters for elevation, degrees for slope), they must be standardized into a common evaluation scale. A scale of 1 to 5 was adopted, where 5 represents the highest flood risk and 1 represents the lowest risk.

![Texte alternati] (https://commons.wikimedia.org/wiki/File:Reclassification_of_kenya_flood_risk_table.png)

For Land Use/Land Cover (LULC), specific weights were assigned to each class based on runoff potential:

1. Waterbodies: 5 (Source of hazard)
2. Bare land: 4 (Poor absorption)
3. Urban/Built-up: 3 (Low infiltration, high runoff)
4. Cropland: 2 (Moderate runoff)
5. Vegetation: 1 (Natural sponge, high absorption).

Weighted Overlay Calculation

The final flood risk map was generated using the Raster Calculator to apply a weighted sum overlay. Each reclassified raster was multiplied by its assigned influence weight and summed according to the following mathematical expression:

Use of Raster calculator in QGIS to get Flood Risk Index. Flood Risk Index = (Rainfall {reclass} x 0.30) + (Distance from water {reclass} x 0.25) + (Elevation {reclass} x 0.20) + (Slope {reclass} x 0.10) + (LULC {reclass} X 0.10) + (Distance from road {reclass} x 0.05)

The result is a continuous raster with values between 1 and 5, which was then categorized into five discrete risk classes: Very High, High, Moderate, Low, and Very Low.

Result

LULC

The LULC map, provided by the Copernicus Global Land Service, reveals the impact of human modification on flood dynamics. The expansion of built-up areas, particularly in Nairobi and Mombasa, has replaced natural pervious surfaces with asphalt and concrete, leading to rapid runoff during flash floods. Conversely, the preservation of forest cover in the Central Highlands remains a critical factor in mitigating the downstream impact of heavy rainfall.

Topography: Elevation & Slope

The DEM analysis shows that a large proportion of Kenya’s landmass consists of high-elevation plateaus. However, the low-lying coastal plains and the internal drainage basins (like the Tana River and Lake Victoria basins) are significant. The reclassified elevation map highlights these lowlands (below 500m) as “Very High” risk zones. Similarly, the slope map reveals that much of the Tana River basin and the coastal strip is remarkably flat (slope < 2°), which facilitates the ponding of water and slow-moving riverine floods.

Distance from Water and Road

The “Distance to Water” map creates a high-risk buffer along Kenya’s perennial rivers. This factor is critical for identifying areas susceptible to riverine flooding. The “Distance to Road” map highlights the intersection of human infrastructure and flood risk. Because major transport routes often follow valley bottoms or low-lying corridors, they represent significant exposure points for economic disruption.

Rainfall

The rainfall map, derived from CHIRPS data, highlights the stark contrast between the humid highlands and the arid north. During the 2024 El Niño cycle, coastal regions and the Lake Victoria basin saw anomalies significantly above the long-term average. The reclassification of this layer identifies these regions as the primary engines of flood risk, where the volume of water entering the system frequently exceeds the capacity of natural and artificial drainage.

The National Flood Risk Map of Kenya

The synthesis of these factors produces the final National Flood Risk Map of Kenya. This map provides a prioritized visualization of susceptibility, allowing for a nuanced understanding of where hazards are most likely to intersect with human activity.

Spatial Distribution of Risk Classes

1. Very High Risk: Concentrated in the Tana River Delta, the Lower Kano Plains near Kisumu, and the coastal strip including Mombasa. Urban informal settlements in Nairobi (Mathare, Kibera) also fall into this category due to high runoff and proximity to riparian zones.
2. High Risk: Includes the northern Rift Valley lakes (Turkana, Baringo) and the floodplains surrounding the Ewaso Ng’iro and Sabaki rivers. These areas face frequent inundation during strong “short rains” and El Niño seasons.
3. Moderate Risk: Covers transitional zones between the highlands and the lowlands, including the Arid and Semi-Arid Lands (ASALs) where flash flooding is a periodic threat.
4. Low to Very Low Risk: Located in the Central Highlands, the Aberdare Range, and the high-altitude regions of the Rift Valley. These areas have steep slopes and good drainage, making them safe from large-scale inundation, though they remain vulnerable to localized landslides.

Statistical Profile Of National Risk

The weighted overlay results indicate that while only a small percentage of Kenya’s total land area is at “Very High” risk, these zones contain a disproportionately high amount of the country’s population and critical infrastructure.

Recommendation

1. Ecosystem Restoration and Nature-Based Solutions The significant influence of Land Use/Land Cover on flood risk underscores the need for ecosystem restoration. The government should prioritize the restoration of degraded forests and wetlands in “water tower” regions to enhance natural water absorption. Implementing nature-based solutions, such as wetland restoration and the creation of riparian buffers, will improve the resilience of riverine communities.

2. Strategic Urban Planning and Infrastructure There is an urgent need for the implementation of spatial and development plans in urban areas. This includes improving garbage collection to prevent the blockage of drainage systems and relocating residents from the highest-risk riparian zones to safer, planned housing. Infrastructure like roads and bridges must be designed to withstand projected increases in rainfall intensity, particularly in the “Very High” risk coastal and lake regions.

3. Empowering the Open Mapping Community As a HOT CWG Legacy Project, this study recommends the continued empowerment of local OSM chapters. The mapping of “Very High” risk areas should be prioritized in the HOT Tasking Manager, with a focus on capturing high-resolution building footprints and drainage networks. Community-led data integration and participatory mapping will ensure that risk maps are grounded in the lived reality of vulnerable populations.

4. Early Warning Systems and Anticipatory Action Kenya needs to adopt more effective early warning systems that provide timely and accurate information to residents in high-risk zones. By integrating real-time satellite rainfall estimates (like CHIRPS) with the risk map produced in this study, disaster managers can trigger “anticipatory action” protocols—providing cash transfers or evacuation assistance before the floodwaters arrive.

Conclusion

The 2025 National Flood Risk Assessment of Kenya provides a rigorous spatial framework for understanding the nation’s vulnerability to hydrometeorological extremes. By synthesizing environmental and anthropogenic factors using GIS and Multi-Criteria Decision Analysis, the research has identified the specific geographical clusters where risk is most acute. The transition from the 2022–2023 drought to the 2024–2025 floods serves as a stark reminder that the traditional patterns of climate in Kenya are shifting, requiring more sophisticated and data-driven approaches to disaster management. The Legacy Project for the HOT CWG Mentorship 2025 fulfills its aim of producing a functional flood risk map and providing a technical pathway for future researchers and humanitarian mappers. The findings reveal that while the highlands remain relatively safe, the low-lying basins and urban informal settlements face a “Very High” risk of catastrophic flooding. Addressing these risks requires a multi-faceted approach: restoring ecosystems, climate-proofing infrastructure, and, most importantly, empowering local communities with open geospatial data. This map is not a final product but a living tool for the OpenStreetMap community and its partners as they work toward a more resilient and prepared Kenya. Through the continued exchange of geospatial skills and the mobilization of global volunteers, the vision of the “Audacious Project” can be translated into tangible safety for those living on the front lines of the climate crisis.

Work Cited

1. Kenya Crisis Response Plan 2024, https://crisisresponse.iom.int/response/kenya-crisis-response-plan-2024
2. Life after Kenya’s floods of 2024 - PreventionWeb, https://www.preventionweb.net/news/life-after-kenyas-floods-2024
3. Flood Havoc in Kenya Underscores Climate Adaptation Need, https://climateadaptationplatform.com/flood-havoc-in-kenya-underscores-climate-adaptation-need/
4. Kenya Floods Recovery Needs Assessment - United Nations Development Programme, https://www.undp.org/sites/g/files/zskgke326/files/2025-05/kenya_floods_recovery_needs_assessment_2024.pdf
5. Humanitarian OSM Team/Working groups/Community/Mentorship - OpenStreetMap Wiki, osm.wiki/Humanitarian_OSM_Team/Working_groups/Community/Mentorship
6. Assessment of flood risk using space technology in Matuga state …, https://www.space4water.org/news/assessment-flood-risk-using-space-technology-matuga-state-kenyas-coastal-area

7. Creating a Flood Map Simulation with QGIS

   by K Yashima

   Helios-techblog - Medium, https://medium.com/helios-techblog/creating-a-flood-map-simulation-with-qgis-8a9beed0a2b8. Flood conditioning factors: (a) Distance to roads, (b) Distance to streams, (c) Rainfall, (d) Lithology. - ResearchGate, https://www.researchgate.net/figure/Flood-conditioning-factors-a-Distance-to-roads-b-Distance-to-streams-c-Rainfall_fig4_378610335

8. Multi Criteria Overlay Analysis (QGIS3) — QGIS Tutorials and Tips, https://www.qgistutorials.com/en/docs/3/multi_criteria_overlay.html
9. The Risk of Flooding to Architecture and Infrastructure amidst a Changing Climate in Lake Baringo, Kenya - Scirp.org., https://www.scirp.org/journal/paperinformation?paperid=123556
10. The Risk of Flooding to Architecture and Infrastructure amidst a Changing Climate in Lake Baringo, Kenya - Scirp.org., https://www.scirp.org/journal/paperinformation?paperid=123556
11. Mapping Land Use Land Cover within Flood Risks and Safe Zones …, https://www.journals.eanso.org/index.php/eajenr/article/view/2943