Land cover aware temperature correction of bicycle transects: A case study of mapping the air temperature in two Belgian cities

A low-threshold methodology for mapping the nocturnal air temperature in an urbanized region based on bicycle measurements and a number of in-situ weather stations is presented. Firstly, two corrections are applied on the bike measurements: a thermal inertia correction to account for the finite reaction speed of the temperature sensor and a temperature decline correction to offset the background temperature evolution during the mobile measurements. The latter includes for the first time land cover information to account for its impact on the nocturnal cooling rate. Secondly, a linear statistical model is used to produce regional maps of the air temperature based on the relationship between the corrected temperature and land cover around the bicycle transect. Applying this methodology on the Belgian cities of Antwerp and Ghent demonstrates that this low-threshold approach allows for a comprehensive picture of the air temperature and that it has advantages compared to more traditional decline corrections. Further improvements could be made by employing more accurate temperature sensors, applying more complex statistical models and performing multiple measurement campaigns per evening.