B.2 Harmonization of source apportionment and priorization of pollution sources
1. Sampling campaigns carried out between January 2013 and March 2014
The official sampling campaign started on January 14th, 2013 and ended one year later. The sampling sites are described at (http://airuse.eu/stations/). The duration of each sampling was 24 hours and the frequency of each sampling was one sample every three days. To study more in detail the Saharan dust intrusions, at the beginning of April it was decided to take one sample every day when models forecast the arrival of air masses from the Saharan region. On April 10th the Athens group started to send to all groups Hysplit, Skiron and NMMB/BSC-Dust forecasts, every 3 days, so it was possible to change in advance the time schedule of the sampling. The used collection substrates were Teflon and/or quartz filters as they are adequate both for weighing and also chemical analysis. No significant malfunctioning was reported during the campaign. A summary of the filters collected during the “standard” plus the supplementary sampling days is reported in the following table:
From 01/07/2013 till 15/09/2013 a Mini-AMS (Aerosol Mass Spectrometer) was used at the Barcelona site. AMS can provide quantitative information on the chemical composition of the sub-micrometric non-refractory fraction of aerosol particles with a high size and time resolution. The Mini-AMS will improve source characterisation and apportionment by distinguishing natural from anthropogenic sources. For shorter periods the aerosol was collected also with the Florence Streaker sampler which collects PM samples (PM2.5-10 (coarse fraction) and PM2.5 ( fine fraction)) with a 1 hour time resolution.
2. Analysis of the collected samples and advancement level of the chemical analysis
The analysis which have been performed are:
• Major elements and trace elements by a combination of methods PIXE (at Florence laboratory), ICP-MS (Inductively Coupled Plasma Mass Spectrometry, at Barcelona, Florence), ICP-AES (Inductively Coupled Plasma Atomic Emission Spectroscopy, at Barcelona and Florence), depending on the available techniques from each participant; by XRF (X-Ray Fluorescence) with polarized primary X-Ray beam and several secondary target, after an intercomparison with PIXE, only for Milan samples.
• Water-soluble ions by IC (Ion Cromatography) at Barcelona, Athens, Florence and Milan.
• Elemental and organic carbon by thermal/optical analysis at all the laboratories; carbonates on a subgroup of filters has been determined by the Porto group.
- Levoglucosan on a subset of PM2.5 filters from all the sampling sites by ARPA Lombardia
The complete datasets will be ready within the end of August 2014, before the expected date (31/12/2014).
PIXE analysis has been performed on the streaker samples uniformly collected during the week (i.e. with no problem during sampling) or sampled during the more interesting periods.
3. The advancement level of the statistical analysis and the results of the source apportionment
One of the aim of the action B2 is that the overall source apportionment strategy, from the sampling and analysis of the PM to the model application and the interpretation of the results, should be optimized and harmonized. In September 2013 a meeting has been organized in Florence to discuss and decide the receptor model and the protocol to be used. PMF (Positive Matrix Factorization) was chosen as receptor model because its widespread use by research groups in the world. Two programs (EPA3.0 and ME-2) have been compared on an old dataset (from the PATOS campaign in Tuscany) and they produced similar results for both the identified sources and their impact on PM10 mass. Furthermore, we have investigated in which way it is possible to combine the information obtained by the streaker sampler with 1-h resolution with the daily concentration data. During the meeting in Athens in March 2014 we decided that the new version of the EPA program will be used; the reason is that it is freely available on EPA web-site and it is surely more user-friendly respect to other programs; therefore it will be easier the for all the stakeholders to use it.
PMF analysis will be performed on the final, complete dataset, i.e. when all the chemical analysis will be finished. To harmonize the collection of the data a common general datasheet has been prepared. The final 4 datasheet from all the sites will be ready within the end of August. In September all the data will be validated and prepared in the proper format for PMF analysis.
During one week workshop in Barcelona from October 13th to October 17th representatives from all the participating groups applied the new version of the EPA program to the concentration data from all the 4 sites, following the protocol suggested in the JRC reference Report “European Guide on Air Pollution Source Apportionment with Receptor Models”. In such a way the harmonization in the use of the PMF receptor model is assured.
4. Activities connecting Action B2 with other AIRUSE Actions
The final results of action B2 will be important for actions B3 (Determination of the impact of natural sources (African dust, marine aerosol)), B4 (Determination of the impact of biomass burning), B5 (Determination of the impact of industrial sources), B6 (Determination of the impact of traffic related sources).
The streaker sampler was used to collect samples that were analysed together with the ones used in action B2 in the framework of action B7 (Testing of air mitigation measures – Development of air mitigation strategies) to test the effectiveness of CMA to reduce dust re-suspension. We collected for 5 weeks in parallel in two different sites (C.le Industria and C.le Valencia) during April-May and 2 weeks in August (at Palau Royal). The results have been used within action B7.
Three technical reports have been produced and may be seen bellow.
The training material prepared may be seen bellow:
Short course – sampling of atmospheric aerosols and analysis of their carbonaceous content (Aveiro)
In the frame of the project AIRUSE, the University of Aveiro ministered one short course on sampling of atmospheric aerosols and analysis of their carbonaceous content. The topic is related to Action B.2. (Harmonization of source apportionment and prioritisation of pollution sources) of AIRUSE. Within this Action, the Portuguese institution is carrying out a sampling campaign of PM2.5 and PM10 in Oporto. After detailed chemical analysis of the samples, source receptor models will be applied in order to estimate the contributions of the main sources. This short course, of practical character, addressed the following aspects: aerosol measurement overview, training in PM2.5/PM10 sampling (high and low volume), calibration of equipment, training in gravimetric measurements, training in organic and elemental carbon determination by a thermal-optical technique, training in carbonate analysis. The details of this course, agenda, and lectures are given in a supplementary pdf file.
Training activity on state-of-the-art receptor modelling (Florence)
From 9th to 12th September 2013, the AIRUSE training activity on state-of-the-art receptor modelling was carried out in Florence with the supervision of Dr. Fulvio Amato (IDAEA-CSIC). Participants from INFN, CSIC and University of Milan participated in the four-days activities which included:
- Overview of available programs for receptor modelling (PMF2, EPA PMF, ME-2)
- Training on Mulitlinear Engine scripting for constrained Positive Matrix Factorization
- Training on the implementation of a priori knowledge into the US EPA PMF model
- Intercomparison exercise
Training activity on road dust sampling device (Athens)
In August 2013 Angeliki Karanasiou from the coordinator IDAEA-CSIC visited the associated partner “NCSR Demokritos” in Athens, for a one day training of the technical staff on the road dust sampling device, designed and developed by the coordinator IDAEA-CSIC.
Demonstration of the device and sampling tests took place at a street near “Demokritos” and 3 of its members assisted: Konstantinos Eleftheriadis; Evangelia Diapouli and Stergios Vratolis.
Road Dust Sampling Campaign in Porto
The campaign was carried out, between the 9th of May and the 1st of July 2014, with the collaboration of the Traffic Division of the City Council, as well of the Municipal Police Municipality, a road dust sampling campaign was conducted on several streets of Porto and in a road tunnel (Avenida da Liberdade). The sampling device developed by Fulvio Amato was used to collect PM10 samples. These filters have been analysed by a thermal-optical method to obtain their carbonaceous content and by inductively coupled plasma techniques to get the elemental speciation. The “dust chemical profiles” will be useful to better estimate the contribution of resuspension to ambient PM10 levels.