Comparison of measured and simulated mean radiant temperature. Case study in Lisbon (Portugal)

Résumé 0

Mean radiant temperature (MRT) is one of the most relevant human bioclimatic indices, particularly suitable for assessing the influence of climatic parameters on thermal comfort outdoors. MRT can be calculated either from physical measurements, carried out using a globe thermometer or a pyranometer and a pyrgeometer, in order to quantify short and long wave radiation fluxes, or through computer simulation. The first method, although accurate, entails the measurement of radiant fluxes from six directions, therefore it is time consuming, complex and it requires the use of expensive equipment. The second method is based on RayMan, ENVI-met and SOLWEIG computational models, often employed in urban climatological research. The present research deals with the comparison of MRT data obtained by measurement and computational simulation in Campo de Ourique, a densely populated city quarter of Lisbon. The measurements were carried out in a park and in the surrounding canyon streets during four summer days in 2006. The three computational models present some differences regarding their viability in estimating MRT compared to measurement: in case of direct solar exposure RayMan and SOLWEIG provided particularly good approximation (with an average difference of 6 and 9%, respectively), whereas the average overestimation of MRT obtained by ENVI-met is around 20%. However, in shaded locations ENVI-met provided the most acceptable agreement between measured and simulated MRT; in contrast RayMan and SOLWEIG overestimated MRT by 15% on average.