Validation of seasonal mean radiant temperature simulations in hot arid urban climates

We validated seasonal RayMan and ENVI-met mean radiant temperature (T_MRT) simulations to assess model performance in a sensitivity analysis from cold to extremely hot conditions. Human-biometeorological validation data were collected in Tempe, Arizona via transects during five field campaigns between 2014 and 2017. Transects were conducted across seven locations in two to three-hour intervals from 6:00 to 23:00 LST with a Kestrel meter and thermal camera (2014–2015) and the mobile instrument platform MaRTy (2017). Observations across diverse urban forms, sky view factors, and seasons covered a wide range of solar radiation regimes from a minimum T_MRT of 8.7 °C to a maximum of 84.9 °C. Both models produced large simulation errors across regimes with RMSE ranging from 8 °C to 12 °C (RayMan) and 11.2 °C to 16.1 °C (ENVI-met), exceeding a suggested T_MRT accuracy of ±5 °C for heat stress studies. RayMan model errors were largest for engineered enclosed spaces, complex urban forms, and extreme heat conditions. ENVI-met was unable to resolve intra-domain spatial variability of T_MRT and exhibited large errors with RMSE up to 25.5 °C for engineered shade. Both models failed to accurately simulate T_MRT for hot conditions. Errors varied seasonally with overestimated T_MRT in the summer and underestimated T_MRT in the winter and shoulder seasons. Results demonstrate that models should not be used under micrometeorological or morphological extremes without in-situ validation to quantify errors and assess directional bias due to model limitations.