Improved methods for estimating mean radiant temperature in hot and sunny outdoor settings
Jennifer K. Vanos,Konrad Rykaczewski,Ariane Middel,Daniel J. Vecellio,Robert D. Brown,Terry J. Gillespie
DOI: https://doi.org/10.1007/s00484-021-02131-y
IF: 3.738
2021-04-28
International Journal of Biometeorology
Abstract:Thermal comfort research has utilized various sensors and models to estimate the mean radiant temperature (MRT) experienced by a human, including the standard black globe thermometer (SGT), acrylic globe thermometers (AGT), and cylindrical radiation thermometers (CRT). Rather than directly measuring radiation, a temperature is measured in the center of these low-cost sensors that can be related to MRT after theoretically accounting for convection. However, these sensors have not been systematically tested under long-term hot and clear conditions. Further, under variable weather conditions, many issues can arise due to slow response times, shape, inaccuracies in material properties and assumptions, and color (albedo, emissivity) inconsistencies. Here, we assess the performance of MRT produced by various heat transfer models, with and without new average surface temperature (( {overline{T}}_{mathrm{s}} )) correction factors, using five instruments—the SGT (15 cm, black), tan and black CRTs, gray and black 38 mm AGTs—compared to 3D integral radiation measurements. Measurements were taken on an unobscured roof throughout summer-to-early-fall months in Tempe, Arizona, examining 58 full-sun days. Deviations without correcting for asymmetrical surface heating—found to be the main cause of errors—reached ± 15–20 °C MRT. By accounting for asymmetric heating through ( {overline{T}}_{mathrm{s}} ) calculations, new corrective algorithms were derived for the low-cost sensor models. Results show significant improvements in the estimated MRT error for each sensor (i.e., ∆MRTmodel − IRM) when applying the ( {overline{T}}_{mathrm{s}} ) corrections. The tan MRTCRT improved from 1.9 ± 6.2 to −0.1 ± 4.4 °C, while the gray AGT and SGT showed improvements from −1.6 ± 7.2 to −0.4 ± 6.3 °C and − 6.6 ± 6.4 to − 0.03 ± 5.7 °C, respectively. The new corrections also eliminated dependence on other meteorological factors (zenith, wind speed). From these results, we provide three simple equations for CRT, AGT, and SGT correction for future research use under warm-hot and clear conditions. This study is the most comprehensive empirical assessment of various low-cost instruments with broad applicability in urban climate and biometeorological research.