flex-damage

Documentation | Developer docs | Reports | Parameters on Zenodo

Internal tool for the Climate Impact Lab. Runs on the University of Chicago RCC cluster (Midway) with access to CIL project data.

This implementation is based on the original CIL FlexDamage code at ClimateImpactLab/flexdamage.

flex-damage fits statistical emulators of CIL climate impact projections. Each region gets its own quadratic in temperature; a globally fitted income elasticity captures adaptation. The damage function is

$$D_{it} = (\alpha_i T_t + \beta_i T_t^2) \cdot Y_{it}^{\gamma}$$

with $T$ in °C anomaly, $Y$ in 2005 USD PPP per capita, and outcome units that depend on the sector. See the documentation for methodology, per-sector units, and parameter format.

Sectors

Released at impact-region (~24k regions) and country resolution:

• Agriculture: corn, rice, soy, sorghum, cassava, wheat (combined, spring, winter)
• Mortality: all-cause all-age
• Labor: combined, high-risk, low-risk
• Energy: total, electricity, non-electricity

Quick start

git clone https://github.com/ClimateImpactLab/flex-damage.git
cd flex-damage
pip install -e .

# Run one sector end-to-end (estimation + parameter export)
python scripts/run.py configs/agriculture/corn.yaml

To regenerate the documentation site after editing docs:

bash scripts/deploy_docs.sh

License

CC-BY-4.0. Climate Impact Lab.