Forecasting brown haze in Auckland.

Brown haze is a visual indicator of poor urban air quality. In Auckland, brown haze typically occurs on cold winter days during the weekday morning commute. There is large interannual variability in the number of brown haze events observed in the region. Recent research identified a clear link between brown haze and poor respiratory health outcomes in Auckland. This paper analyses the meteorological characteristics of 23 brown haze events in Auckland (as identified from camera images) and proposes a simple Numerical Weather Prediction (NWP) forecast scheme that doesn't explicitly model pollutants, but instead predicts weather conditions historically conducive to brown haze. Results showed a distinctive atmospheric profile associated with the development of brown haze conditions. A forecast scheme was designed to capture this specific meteorology, and was subsequently validated against camera imagery during a pilot phase lasting three months during the winter of 2017, both at the 30-hour and 54-hour forecast validity period. Results showed 97% forecast accuracy for both 30- and 54-hour forecasts, and a probability of detection of 89% and 78% respectively. One limitation of the current forecast scheme is that it does not forecast local emissions, and so does not differentiate between weekdays and weekends. Human assessment would still be required before forecasts could be released to the public. Another limitation is that validating the scheme is subjective (brown haze is a visual indicator). Future work will validate the forecast scheme against quantitative measurements of surface air pollution concentrations over a longer period (3 winters), since initial relationships between the forecast index and observed air pollution concentrations looked promising. [ABSTRACT FROM AUTHOR]