Although the urban heat island (UHI) is well studied, the dynamic nature (i.e. with wind) receives little attention. The concept, urban heat advection (UHA), can warm air temperatures in surrounding rural areas. This
may lead to a misinterpretation of local climate and bias in long-term climate records. Using observational
analysis and numerical modelling this thesis investigates these limitations by spatially quantifying UHA. A
methodology to separate UHA from the background air temperature was applied to a high-density urban
observation network in the city of Birmingham, UK demonstrating mean downwind UHA of 0.4oC and up to
1.2oC at individual stations (wind speeds 2 – 3 m s-1). This UHA methodology was adapted to show that even
small urban areas (~1 km2) can produce a mean UHA of 0.6oC. TheWeather Research & Forecasting numerical
model was used to refine the UHA methodology (accounting for regional heat advection) and conduct semiidealised
simulations. Here, a square city with 16 km size produced UHA of 2.4oC at the city edge, with 0.5oC
warming extending 9 km downwind. A relationship was found between city size and UHA intensity, enabling
statistical scaling. This demonstrated an approach to estimate UHA without the need for computationally
expensive simulations.