Heat stress modeling, machine learning, and open tools for climate adaptation in Latin America and beyond.
Systematic Review of Heat Stress in Latin America
A first-of-its-kind region-wide systematic review cataloging heat research across Latin America. We examine how metrics like WBGT and the heat index have been applied across diverse populations: urban residents, outdoor workers, and rural communities. The work identifies geographic and social gaps in coverage to support more inclusive climate adaptation planning. If it hasn't been studied, it can't be protected.
Comparing Heat Stress Metrics for Better Warnings
We compare multiple heat stress measurement approaches against WBGT and other standards, testing their accuracy across different environments: humid tropics, high-altitude cities, and agricultural zones. We use ERA5 reanalysis data to evaluate how each metric performs under real conditions. Choosing the right metric isn't academic; it shapes whether a heat warning is accurate, trusted, and actionable.
Machine Learning for Extreme Heat Forecasts
A stacked ensemble machine learning model that forecasts extreme heat events using ERA5 reanalysis data and bias-corrected CMIP6 projections. The approach emphasizes upper-tail percentiles (95th, 99th) and uncertainty quantification rather than averages, delivering probabilistic risk maps for city planners and health agencies. We blend extreme value theory with quantile modeling to capture the events that actually injure and kill people.
A running record of federal actions affecting climate science, environmental protection, and vulnerable communities.
climate policy rollbacks
A running record of what this administration has done to climate science, environmental protection, and the communities most at risk. An interactive timeline with sources.
We're building in the open. Come build with us.
climacoder@gmail.com