Mean radiant temperature (MRT) is a significant variable for outdoor thermal comfort studies. Two measurement-based methods can estimate MRT, one is globe thermometer – cheap, easily-applied but relatively inaccurate, another is integral radiation measurement method (also known as the six-directional method) - accurate but expensive. Due to low-cost and convenience, the globe thermometer has been widely used. Previous studies have improved its estimation accuracy by recalibrating the convection coefficients in the ISO method. Thus, it is pending to cross-compare the performance of these recalibrated methods.

This study aims to investigate the transferability of the recalibrated methods for estimating MRT in outdoor environment. First, field measurement was conducted in a subtropical city, Hong Kong. MRT was obtained through two methods: globe thermometer and integral radiation method. Second, the existing recalibrated convection coefficients were summarized, and the localized convection coefficient was recalibrated. Third, all recalibrated methods were compared for their performance. The impacts of measurement locations, devices, analysis time intervals were examined.

The results showed that the newly recalibrated method achieved the lowest estimation errors (RMSE = 3.84 °C). Other recalibrated methods presented higher RMSE (3.84–17.52 °C), similar as conventional ISO method (7.91 °C). Especially for open spaces, the coefficients from other cities should be cautiously applied when the accuracy requirement is less than ±2 °C. Kestrel and Grey globe are more recommended in subtropical cities. This study shed light on the application of globe thermometer for outdoor environment, and emphasized the necessity in recalibrating the convection coefficients locally.