Skip to content
DraconiticResearch
Search Sources API
226,742 studies31 sourcesUpdated 36 min ago
Peer-reviewedCC BY-NC-NDNature Communications

Transfer learning reveals large discrepancies between air and land surface temperatures in cities

9 affiliations
  1. 1Department of Civil and Environmental Engineering, University of Illinois Urbana-Champaign, Urbana, IL, USA
  2. 2Department of Civil and Environmental Engineering, University of Illinois Urbana-Champaign, Urbana, IL, USA. [email protected]
  3. 3Department of Climate, Meteorology & Atmospheric Sciences, University of Illinois Urbana-Champaign, Urbana, IL, USA. [email protected]
  4. 4Institute for Sustainability, Energy, and Environment (iSEE), University of Illinois Urbana-Champaign, Urbana, IL, USA. [email protected]
  5. 5National Center for Supercomputing Applications, University of Illinois Urbana-Champaign, Urbana, IL, USA. [email protected]
  6. 6Pacific Northwest National Laboratory, Richland, WA, USA
  7. 7Department of Electrical and Computer Engineering, University of Illinois Urbana-Champaign, Urbana, IL, USA
  8. 8High Meadows Environmental Institute, Princeton University, Princeton, NJ, USA
  9. 9Department of Earth and Environmental Engineering, Columbia University, New York, NY, USA. [email protected]

Abstract

Understanding of urban weather and extreme events remains severely limited by data poverty resulting from a dearth of true urban weather stations. As a result, land surface temperature (Ts), obtained from remote sensing platforms, has been widely used as a stand-in for near-surface air temperature (Ta) despite their fundamental differences, especially in urban areas. Although Ts provides important scientific insights and practical utility for studying the urban thermal environment, this substitution risks introducing large uncertainties and biased characterization of impact-relevant urban heat stress. Here we develop an urban transfer-learning framework (U-TL) to address this critical gap and to provide urban high-resolution air temperature (U-HAT) data at large scales across the contiguous United States (CONUS). U-TL demonstrates high accuracy and strong robustness in predicting urban Ta, even with limited training data. The resulting U-HAT is a high-resolution urban Ta dataset capable of accurately reproducing observed and well-established urban climatology. U-HAT reveals substantial Ts–Ta discrepancies and therefore cautions the use of Ts to characterize urban heat. We show that satellite-measured Ts substantially overestimates both urban heat stress magnitude and intra-city spatial variability, which have consequential implications for urban heat exposure, vulnerability, and adaptation policy making. This study develops a transfer-learning method to estimate high-resolution urban air temperature across the contiguous United States and shows that satellite land surface temperature substantially overestimates heat stress and temperature variability in those cities.

Related

  1. Discrepancies between Sentinel-2 and LiDAR-derived LAI in temperate deciduous forests: Influence of apparent optical depth, radiative transfer priors, and horizontal heterogeneity
    Nathan Corroyez, Jean-Baptiste Féret, Jérôme Ogée +2
    Remote Sensing of EnvironmentPeer-reviewedCC BY 4.0Full textEnvironmental science & pollutionHAL
  2. Surface temperatures drive strong seasonality in urban reactive carbon emissions
    Michael P. Vermeuel, Dylan B. Millet, Róisín Commane +7
    Science AdvancesPeer-reviewedLicence not statedAtmospheric science & meteorologyCrossref
  3. Systematic Discrepancies Should Be Addressed Before Effective Parameter Learning: Insights From Soil Moisture Simulations Using Differentiable Process‐Based Models
    Peijun Li, Chaopeng Shen, Jiangtao Liu +3
    Water Resources ResearchPeer-reviewedCC BYFull text1 citedEnvironmental science & pollutionDOAJ
  4. An unprecedented Mediterranean tropical cyclogenesis attempt above extreme sea surface temperatures?
    Juan Jesús González-Alemán, Carlos Calvo-Sancho, Mario Marcello Miglietta
    EarthArXivPreprintCC BY-NC-ND 4.0Full textAtmospheric science & meteorologyClimate scienceEarthArXiv
  5. Impact of conduction and convection approaches for urban microclimate assessment: from surface temperatures to outdoor thermal comfort
    Margot Ruiz, Ziang Cui, Diwen Yao +3
    Building and EnvironmentPeer-reviewedCC BY-NC-NDFull textEnvironmental science & pollutionHAL
BibTeX
@article{zhang2026transfer,
  title = {{Transfer learning reveals large discrepancies between air and land surface temperatures in cities}},
  author = {Yiwen Zhang and Lei Zhao and T C Chakraborty and Priyam Mazumdar and Keer Zhang and Pierre Gentine},
  journal = {Nature Communications},
  year = {2026},
  month = may,
  doi = {10.1038/s41467-026-73716-7},
  url = {https://doi.org/10.1038/s41467-026-73716-7},
  note = {CC BY-NC-ND},
}
RIS
TY  - JOUR
TI  - Transfer learning reveals large discrepancies between air and land surface temperatures in cities
AU  - Yiwen Zhang
AU  - Lei Zhao
AU  - T C Chakraborty
AU  - Priyam Mazumdar
AU  - Keer Zhang
AU  - Pierre Gentine
PY  - 2026
DA  - 2026/05/27
JO  - Nature Communications
DO  - 10.1038/s41467-026-73716-7
UR  - https://doi.org/10.1038/s41467-026-73716-7
AB  - Understanding of urban weather and extreme events remains severely limited by data poverty resulting from a dearth of true urban weather stations. As a result, land surface temperature (Ts), obtained from remote sensing platforms, has been widely used as a stand-in for near-surface air temperature (Ta) despite their fundamental differences, especially in urban areas. Although Ts provides important scientific insights and practical utility for studying the urban thermal environment, this substitution risks introducing large uncertainties and biased characterization of impact-relevant urban heat stress. Here we develop an urban transfer-learning framework (U-TL) to address this critical gap and to provide urban high-resolution air temperature (U-HAT) data at large scales across the contiguous United States (CONUS). U-TL demonstrates high accuracy and strong robustness in predicting urban Ta, even with limited training data. The resulting U-HAT is a high-resolution urban Ta dataset capable of accurately reproducing observed and well-established urban climatology. U-HAT reveals substantial Ts–Ta discrepancies and therefore cautions the use of Ts to characterize urban heat. We show that satellite-measured Ts substantially overestimates both urban heat stress magnitude and intra-city spatial variability, which have consequential implications for urban heat exposure, vulnerability, and adaptation policy making. This study develops a transfer-learning method to estimate high-resolution urban air temperature across the contiguous United States and shows that satellite land surface temperature substantially overestimates heat stress and temperature variability in those cities.
SN  - 2041-1723
ER  -