ada-yl2425/Compounding-climate-isolation-across-Africa-s-road-network
Domain:
geospatialclimatemobilityRecord type:
dataset
Creator:
ada
Host:
# Compounding Climate Isolation Across Africa's Road Network
This repository supports the paper **"Compounding climate isolation across
Africa's road network"**. The study examines how climate-driven degradation of
unpaved roads propagates through African road networks, amplifies travel-time
losses, and changes access to healthcare and cities.
## Study overview
African road systems are highly exposed to climate stress because a large share
of the network is unpaved. The paper defines **compounding climate isolation**
as the process by which widespread weather-driven deterioration on unpaved roads
is amplified by network structure and translated into unequal accessibility
losses for people and places with limited routing alternatives.
The analysis combines:
- A continent-scale road network for 49 African countries, including paved and
unpaved road-surface classification.
- Climate stress indicators from precipitation and soil moisture data.
- Population, health-facility, administrative-boundary, terrain and populated
place datasets.
- Weighted road-network models that compare normal and extreme-weather travel
conditions.
## Analytical workflow
The paper uses a coupled workflow:
1. Estimate climate-conditioned speed and passability for unpaved road segments.
2. Build normal and extreme-weather weighted road-network graphs.
3. Measure travel-time change and network-efficiency loss across international,
national and city-scale networks.
4. Evaluate healthcare accessibility using nearest-facility travel time,
coverage thresholds and within-country inequality metrics.
5. Identify adaptation bottlenecks by combining climate vulnerability with
network importance, then test targeted paving scenarios.
## Key findings
- Extreme weather reduces mean road speed by about 10%, but increases travel
times by 36.3% across connected international city pairs and 41.0% across
within-country city pairs.
- Network structure amplifies direct road-level degradation by roughly
3.7- …