source: google research: expanding our heat resilience data to 50+ global cities
level: technical
google research has expanded its heat resilience dataset to cover more than 50 cities across nine countries, including london, athens, rio de janeiro, and los angeles. the data maps rooftop reflectivity, or albedo, at a 30-centimeter resolution, allowing city planners to identify individual buildings that would benefit most from cool roof retrofits. this granularity moves beyond neighborhood averages, enabling precise targeting of heat-trapping surfaces.
the method fuses freely available sentinel-2 satellite data with high-resolution commercial imagery from airbus pléiades neo. machine learning models and radiometric calibration blend the radiometric accuracy of sentinel-2 with the spatial detail of the commercial images. validation against airborne hyperspectral measurements in boulder, colorado, showed a root mean square error of just 0.04, confirming the maps' precision. the approach reconstructs a full spectral reflectance profile for each urban pixel, making building-level analysis possible.
a new heat resilience earth engine app provides interactive access to the data. users can visualize albedo as building centroids, analyze baseline reflectivity, monitor changes over time, and download data for local use. the app includes dynamic zoom from census-tract aggregates to individual building insights. google research developed the tools in collaboration with the world resources institute, aiming to support cool roof ordinances and adaptation plans that could lower urban surface temperatures by up to 0.5°c globally.
why it matters: high-resolution albedo data helps cities prioritize cool roof installations to reduce urban heat island effects, directly informing climate adaptation strategies and public health interventions.
source: google research: expanding our heat resilience data to 50+ global cities