A semi-empirical method for estimating complete surface temperature from radiometric surface temperature, a study in Hong Kong city

Jinxin Yang, Man Sing Wong (Corresponding Author), Hung Chak Ho, E. Scott Krayenhoff, P. W. Chan, Sawaid Abbas, Massimo Menenti

Research output: Journal article publicationJournal articleAcademic researchpeer-review

7 Citations (Scopus)


The complete surface temperature (Tc) in urban areas, defined as the mean temperature of the total active surface area, is an important variable in urban micro-climate research, specifically for assessment of the urban surface energy balance. Since most vertically-oriented building facets are not observed by a nadir-viewing remote imaging radiometer, the radiometric surface temperature (Tr) measured at a specific view angle cannot be used with existing heat transfer equations to estimate radiative and convective fluxes in the urban environment. Thus, it is necessary to derive Tc for city neighborhoods. This study develops a simple method to estimate Tc from Tr with the aid of the Temperatures of Urban Facets in 3D (TUF-3D) numerical model, which calculates 3-D sub-facet scale urban surface temperatures for a variety of surface geometries and properties, weather conditions and solar angles. The effects of geometric and meteorological characteristics – e.g., building planar area index (λp), wall facet area index (F), solar irradiance – on the difference between Tc and Tr were evaluated using the TUF-3D model. Results showed the effects of geometric and meteorological characteristics on the difference between Tc and Tr differ between daytime and nighttime. The study then sought to predict the relationship between Tr and Tc, using λp, F, and solar irradiance for daytime and only using λp and F for nighttime. Based on the simulated data from TUF-3D, the resulting relationships achieve a coefficient of determination (r2) of 0.97 and a RMSE of 1.5 K during daytime, with corresponding nighttime values of r2 = 0.98 and RMSE = 0.69 K. The relationships between Tr and Tc are evaluated using high resolution airborne thermal images of daytime urban scenes: r2 = 0.75 and RMSE = 1.09 K on August 6, 2013 at 12:40 pm; and r2 = 0.86 and RMSE = 1.86K on October 24, 2017 at 11:30 am. The new relationships were also applied to estimate Tc from Tr in Hong Kong retrieved from Landsat 5 Thematic Mapper (TM) and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). In the present climatic context, the difference between Tc and Tr can reach 10 K during daytime in summer, and 6 K during daytime in winter, with seasonal variation attributable to the variations in shortwave irradiance. The nighttime difference between Tc and Tr can also reach 2 K in both summer and spring seasons.

Original languageEnglish
Article number111540
JournalRemote Sensing of Environment
Publication statusPublished - Feb 2020


  • Remote sensing
  • Surface temperature
  • Thermal heterogeneity
  • Urban geometry

ASJC Scopus subject areas

  • Soil Science
  • Geology
  • Computers in Earth Sciences

Cite this