Simulation of Smoke-Haze Dispersion from Wildfires in South East Asia with a Lagrangian Particle Model

Abstract:

Biomass burning, often related to agricultural deforestation, not only affects local pollution levels but periodically deteriorates air quality in many South East Asian megacities due to the transboundary transport of smoke-haze. In June 2013, Singapore experienced the worst wildfire related air-pollution event on record following from the escalation of peatland fires in Sumatra. An extended dry period together with anomalous westerly winds resulted in severe and unhealthy pollution levels in Singapore that lasted for more than two weeks.
Reacting to this event, the Met Office and the Meteorological Service Singapore have explored how to adequately simulate haze-pollution dispersion, with the aim to provide a reliable operational forecast for Singapore. Simulations with the Lagrangian particle model NAME (Numerical Atmospheric-dispersion Modelling Environment), running on numerical weather prediction data from the Met Office and Meteorological Service Singapore and emission data derived from satellite observations of the fire radiative power, are validated against PM10 observations in South East Asia. Comparisons of simulated concentrations with hourly averages of PM10 measurements in Singapore show that the model captures well the severe smoke-haze event in June 2013 and a minor episode in March 2014. Different quantitative satellite-derived emissions have been tested, with one source demonstrating a consistent factor of two under-prediction for Singapore. Confidence in the skill of the model system has been substantiated by further comparisons with data from monitoring sites in Malaysia, Brunei and Thailand.
Following the validation study, operational smoke-haze pollution forecasts with NAME were launched in Singapore, in time for the 2014 fire season. Real-time bias correction and verification of this forecast will be discussed.