A customisable downscaling approach for local-scale meteorological and air pollution forecasting: Performance evaluation for a year of urban meteorological forecasts

In this paper we develop a customisable downscaling approach for local-scale air quality and meteorological forecasting applications, using The Air Pollution Model (TAPM) with the Conformal Cubic Atmospheric Model (CCAM). The CCAM–TAPM system allows users to define their own local pollution emissions, terrain and land-use data and then downscale the meteorological and air pollution forecast on PC hardware. To evaluate the meteorological component of the CCAM–TAPM system, we have produced 7-day forecasts dynamically downscaled to 3 km resolution, centred on Melbourne (Australia). We produce 365, 7-day forecasts with a forecast starting on each day during 2003, initialised with National Centre for Environmental Prediction (NCEP) 1° Global Forecasting System (GFS) analyses (i.e., no boundary conditions are needed). We estimate that CCAM–TAPM can simulate 10-m wind speeds and 2-m temperatures for up to two forecast days with errors within 10% of previous TAPM studies in which air pollution was successfully predicted.