Heat Risk Assessment in Kinshasa, Democratic Republic of Congo

A geospatial and climate-risk assessment project analyzing urban heat vulnerability in Kinshasa, DRC, using remote sensing, socioeconomic indicators, and GIS-based spatial modeling.

The study integrates Land Surface Temperature (LST), vegetation cover, population exposure, and socioeconomic vulnerability to map heat-related health risks across Kinshasa using Crichton’s Risk Triangle framework.

---

Overview

Rapid urbanization, population growth, and climate change have intensified heat-related risks in African cities, particularly in densely populated informal settlements.

This project assesses spatial heat risk patterns in Kinshasa by combining:

• Land Surface Temperature (LST)
• Population density
• Vulnerable age groups
• Relative Wealth Index (RWI)
• Vegetation cover (NDVI)

The resulting heat risk maps identify high-risk urban zones and provide insights for climate adaptation and urban resilience planning.

---

Objectives

• Assess urban heat hazard distribution in Kinshasa
• Analyze exposure using population density
• Evaluate vulnerability using socioeconomic and environmental indicators
• Develop a spatial heat risk model
• Identify high-risk urban communities
• Support heat adaptation and mitigation planning

---

Methodology

Risk Assessment Framework

The study applies Crichton’s Risk Triangle, where heat risk is determined by the interaction of:

• Hazard
• Exposure
• Vulnerability

$$Heat\ Risk = f(Hazard,\ Exposure,\ Vulnerability)$$

Hazard Component

• Land Surface Temperature (LST) derived from Landsat-8 thermal imagery

Exposure Component

• Population density from WorldPop datasets

Vulnerability Indicators

• Vulnerable age groups (<10 and >65 years)
• Relative Wealth Index (RWI)
• NDVI (vegetation cover)

Analytical Workflow

1. Landsat image preprocessing in Google Earth Engine
2. LST extraction from thermal imagery
3. NDVI generation
4. Population and socioeconomic data integration
5. Data normalization and resampling
6. Vulnerability modeling
7. Heat risk mapping and classification

---

Data Sources

Dataset	Source	Purpose
Landsat-8 Imagery	Google Earth Engine	Land Surface Temperature & NDVI
WorldPop Population Data	WorldPop	Exposure analysis
Age Structure Data	WorldPop	Vulnerability analysis
Relative Wealth Index (RWI)	Humanitarian Data Exchange	Socioeconomic vulnerability

---

Technologies Used

• Google Earth Engine (GEE)
• QGIS

---

Key Equations

NDVI Formula

$$NDVI = \frac{NIR - R}{NIR + R}$$

Vulnerability Layer

$$V = \frac{(va - NDVI - RWI)}{3}$$

Final Heat Risk Layer

$$FHRL = Hazard \times Exposure \times Vulnerability$$

---

Key Findings

Heat Hazard Distribution

• Eastern and southeastern Kinshasa showed the highest heat hazard levels
• Urban centers exhibited strong Urban Heat Island (UHI) effects
• Vegetated areas showed lower temperatures and reduced risk

Vulnerability Patterns

High vulnerability was concentrated in:

• Bumbu
• Selembao
• Ngaba
• Ngiri-Ngiri

These areas are characterized by:

• High population density
• Low vegetation cover
• Lower socioeconomic status
• Limited access to cooling infrastructure

Heat Risk Distribution

• Urban centers experienced medium to very-high heat risk
• Rural and peripheral regions showed low to very-low heat risk
• Dense informal settlements were the most vulnerable to heat-related health risks

---

Applications

This project supports:

• Urban climate resilience planning
• Heat-health risk assessment
• Climate adaptation strategies
• Urban sustainability planning
• Public health decision-making
• Environmental justice analysis

---

Future Improvements

• Integrate air quality and humidity data
• Include healthcare accessibility indicators
• Use higher-resolution thermal imagery
• Develop machine learning-based heat vulnerability models
• Conduct temporal heatwave trend analysis