We use daily, county-level gas consumption data and interpretable machine learning models to assess spatial patterns in natural gas consumption of households as a function of outdoor air temperature, as well as factors such as household income, employment rates, the size, age and type of homes, employment rates, and a range of other characteristics.

Edible molecules can be directly synthesized without agricultural feedstocks, saving land, water and GHG emissions. Dietary fats made from natural gas feedstock and average U.S. electricity would emit ~0.8 gCO₂-eq/kcal (and 0 if using atmospheric carbon and renewable electricity) compared to 1-3 gCO₂-eq/kcal from agricultural fat production.

Using a newly developed database of iron and steel plants worldwide, we identify the most cost-effective regional priorities for decarbonization. Our database will be regularly updated and improved to support future plant-by-plant strategic mitigation planning for global net-zero steelmaking.

We assess nine possible pathways to achieve net-zero emissions from aviation, including changes and trade-offs in demand, energy efficiency, propulsion systems, and alternative fuels for both passenger and freight transport, as well as atmospheric carbon removal to offset non-CO₂ radiative forcing.

We estimate how many tons of steel, copper, silver, rare earth metals, and other materials will be needed to build power generation facilities across a wide range of scenarios. Although wind and solar energy require lots of such materials and current production will need to increase, availability of metals and materials will not constrain the projected expansion.

Global modeling shows that potential climate benefits of farming seaweed are large but sensitive to uncertain yields and competition with phytoplankton. We also find that carbon removal by sinking seaweed is much costlier than avoiding emissions by substituting seaweed for land-based crops.