A parameterisation for urban surface exchange in mesoscale models is presented, which accounts for the effects of urban buildings on airflow without explicitly resolving them. The parameterisation considers horizontal and vertical surfaces (roof, canyon floor, and walls) and their impact on wind speed, temperature, and turbulent kinetic energy. It also accounts for radiation effects from urban canyons and modifies turbulent length scales in the TKE equation. The parameterisation is tested in a city model over flat terrain, showing improved performance compared to traditional methods that only adjust roughness length. The new parameterisation significantly affects air pollutant dispersion in urban areas. Urban air quality is complex due to multiple spatial and temporal scales. Pollutants are emitted at the urban scale and dispersed at the meso scale. Numerical models are used to study air pollution, requiring meteorological variables like wind, temperature, and humidity. These models must represent both urban and meso scales, with grid resolutions typically between hundreds of metres and a few kilometres. Urban structures cannot be resolved in detail, so their effects must be parameterised. Key urban effects on airflow include intense shear layers, wake diffusion, building drag, and differential heating/cooling. Traditional methods using Monin-Obukov similarity theory fail to capture turbulent structures in the urban roughness sublayer. This paper improves RSL representation, enhancing flow and dispersion characteristics. Surface temperature in standard models does not account for shadowing and radiation trapping effects. Recent efforts have aimed to improve urban surface representation in mesoscale models, with reviews available in Brown (2000). General attempts have focused on parameterising urban surfaces to better represent their impact on airflow and pollutant dispersion.A parameterisation for urban surface exchange in mesoscale models is presented, which accounts for the effects of urban buildings on airflow without explicitly resolving them. The parameterisation considers horizontal and vertical surfaces (roof, canyon floor, and walls) and their impact on wind speed, temperature, and turbulent kinetic energy. It also accounts for radiation effects from urban canyons and modifies turbulent length scales in the TKE equation. The parameterisation is tested in a city model over flat terrain, showing improved performance compared to traditional methods that only adjust roughness length. The new parameterisation significantly affects air pollutant dispersion in urban areas. Urban air quality is complex due to multiple spatial and temporal scales. Pollutants are emitted at the urban scale and dispersed at the meso scale. Numerical models are used to study air pollution, requiring meteorological variables like wind, temperature, and humidity. These models must represent both urban and meso scales, with grid resolutions typically between hundreds of metres and a few kilometres. Urban structures cannot be resolved in detail, so their effects must be parameterised. Key urban effects on airflow include intense shear layers, wake diffusion, building drag, and differential heating/cooling. Traditional methods using Monin-Obukov similarity theory fail to capture turbulent structures in the urban roughness sublayer. This paper improves RSL representation, enhancing flow and dispersion characteristics. Surface temperature in standard models does not account for shadowing and radiation trapping effects. Recent efforts have aimed to improve urban surface representation in mesoscale models, with reviews available in Brown (2000). General attempts have focused on parameterising urban surfaces to better represent their impact on airflow and pollutant dispersion.