Sensitivity of online coupled model to extreme pollution event over a mega city Delhi

Abstract

Sensitivity of interactive Weather-Chemistry model has been examined to predict the air quality (1 and 3 days in advance) of Indian mega city Delhi during two identical extreme events of Diwali in 2012 and 2013. Analysis is conducted 3 days prior to 3 days later of Diwali day for both events to verify the rapid changes in fine particulate matter (PM2.5) due to widespread display of Diwali fireworks. The model successfully predicted the variability in PM2.5 during 2012 for the entire period of analysis with reasonable accuracy. Although model performed reasonably well until Diwali day in 2013 but it was unable to simulate rapid built up of PM2.5 (1500 μg−3 hourly average) during post Diwali day as it failed to capture unusual mid-night steep temperature gradient followed by a record lowering of boundary layer height. The predictability of the model and its limitation to micrometeorological processes are discussed in detail.