Monitoring wildfire smoke dispersion using concentrations of PM10 and PM2.5

Abstract

Portugal as a member of the European Union (EU) follows the air quality legislation regulated by Directive 2008/50/EC that states that “Member States shall ensure that timely information about actual or predicted exceedances of alert thresholds, and any information threshold is provided to the public”. The southern European countries have biogenic contributions, mainly resulting from the long-distance transport of particulate matter, originated in arid regions, and from large forest fires, whose pollutants loads cannot be imposed limits. These sources lead to an additional contribution of atmospheric pollutants mainly during the spring and summer seasons. Climate change is increasing the vulnerability of the environment to extreme events, such as those that enhance the likelihood of forest fires. Due to the forecast increase in temperatures and decrease in precipitation and humidity it is likely there will be an increase in the frequency and intensity of forest fires in certain regions of the Mediterranean basin as is the case of Portugal. Monitoring the concentration of particles matters in the network of air quality stations can be a way of detecting the dispersion of smoke and giving possible alerts to affected populations. The analysis of PM10 and PM2.5 of the air quality network stations, from June to October, in the period 2010-2020, showed very high concentration values coincident with the years and periods of forest fires. This data was analysed for all stations by considering the average value of the concentration in the 2 days before the day of registration of fires with burned area, daily and by district, greater than 500ha, on the day of the fire occurrence and on two days later. Back trajectory analysis was performed using HYSPLIT Model to confirm the origin of the particles observed at the stations. It was observed that the percentage increase in the daily average value of PM10 concentration, between one day before and the day of the event, is variable but with 26% as an average value for the wildfire above 500ha district daily value, in the period 2010-2020 . This difference depends on the proximity of the forest fire, the area burned as an indication of its strength, the stability of the atmosphere and the height of the boundary layer. Dust episodes with origin in the deserts in Northern Africa often affect the southern regions of Portugal which leads to an increase the concentration of PM10 and PM2.5 particles. The days with high values of PM10 particles (and PM2.5), in the summer months, were studied in order to distinguish the days that could be related to dust particles, in general larger particles, from those which referred to smoke particles from forest fires, usually smaller particles. For this purpose, the relationship between the concentrations of PM2.5 in relation to PM10, in the data coincidence period, was considered. For the large fires of 2017, a more detailed assessment of these characteristics was made by studying the optical thickness of the aerosols and the Angstrom exponent of the VIIRS radiometer on board the Suomi NPP satellite.