Simulation of an extreme dust episode using WRF-CHEM based on optimal ensemble approach

Extreme dust episodes have been noticeably increasing in recent years. Dust burden during such events imparts several threats to the environment and human health. Therefore, forecast of such extreme events becomes of utmost importance to minimise the adverse impact on various socio-economy sectors. In the present study model experiments have been carried out for the simulation of an extreme dust episode that emanated over the western Indian region that coupled with previously transported dust from the West Asia region and resulted in the highly degraded air quality over the Indian region. For this purpose four ensemble members, using Weather Research and Forecasting model fully coupled with chemistry (WRF-CHEM), have been generated for identical model configuration but for the perturbed initial conditions. Model performance has been evaluated with respect to the available ground based measurements from Central Pollution Control Board (CPCB) and the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA2) data sets using standard statistical measures. Ensemble spread is noted amongst the four ensemble members generated based on 4 varying initial conditions, therefore it is suggested to use ensemble mean forecast for the dust storm monitoring and analysis. It is revealed from the analysis based on the categorical validation scheme that the WRF-CHEM simulated dust load is able to capture the locational accuracy with respect to the reanalysis/observational data sets over Indian subcontinent fairly well with 0% false alarm ratio, 78–92% probability of detection and 79–90% accuracy. It is noted that thick blanket of dust load enveloped Northern part of India subsequent to the dust storm which resulted in substantially dropped ground reaching solar radiation (~100–150 W/m²) and a consequential reduction in the surface temperature (~2 °C). Further to this enhanced temperature in association with reduced relative humidity from 800 to 600 hPa has also been noted in response to the enhanced dust loading in the model simulations. Results obtained in the present study suggests that WRF-CHEM is able to simulate both the direct and semi-direct effects of dust aerosols reasonably well. Present work has implications for improved prediction skills of WRF-CHEM in simulating extreme dust episodes and investigating the impact of dust aerosols on weather and climate.