Comparison of Global Downscaled Versus Bottom-Up Fossil Fuel CO₂ Emissions at the Urban Scale in Four U.S. Urban Areas

Spatiotemporally resolved urban fossil fuel CO₂ (FFCO₂) emissions are critical to urban carbon cycle research and urban climate policy. Two general scientific approaches have been taken to estimate spatiotemporally explicit urban FFCO₂ fluxes, referred to here as “downscaling” and “bottom-up.” Bottom-up approaches can specifically characterize the CO₂-emitting infrastructure in cities but are labor-intensive to build and currently available in few U.S. cities. Downscaling approaches, often available globally, require proxy information to allocate or distribute emissions resulting in additional uncertainty. We present a comparison of a downscaled FFCO₂ emission data product (Open-source Data Inventory for Anthropogenic CO₂ (ODIAC)) to a bottom-up estimate (Hestia) in four U.S. urban areas in an effort to better isolate and understand differences between the approaches. We find whole-city differences ranging from −1.5% (Los Angeles Basin) to +20.8% (Salt Lake City). At the 1 km × 1 km spatial scale, comparisons reveal a low-emission limit in ODIAC driven by saturation of the nighttime light spatial proxy. At this resolution, the median difference between the two approaches ranged from 47 to 84% depending upon city with correlations ranging from 0.34 to 0.68. The largest discrepancies were found for large point sources and the on-road sector, suggesting that downscaled FFCO₂ data products could be improved by incorporating independent large point-source estimates and estimating on-road sources with a relevant spatial surrogate. Progressively coarsening the spatial resolution improves agreement but greater than approximately 25 km², there were diminishing returns to agreement suggesting a practical resolution when using downscaled approaches.