Aims of this doctoral thesis and organization of the manuscript
The first general objective of this thesis is to provide a robust selection of FT air masses at the PUY station using a combination different criteria. These criteria are based on LIDAR measurements using the wavelet covariance transform (WCT), BLH simulations using the European Center for Medium –Range Weather Forecasts (ECMWF) model, and in-situ measurements of BL tracers (NOx/CO and radon-222 (222Rn)).
This segregation of air masses in FT/BL will enable the realization of the second objective of this thesis, which is to establish the seasonal variation of aerosol size distribution and non-refractory-PM1 (NR-PM1) data at the PUY station with emphasis on the boundary layer/ free troposphere segregation.
For this I will use a long time series of in-situ measurement of physical (size distribution), and chemical (NR-PM1) properties which operate continuously at the PUY station over a 12-month period, during 2015-2016.
To answer these objectives, this thesis is based on the measurements that have been deployed at multiple sites including the Cézeaux, and the PUY in Clermont-Ferrand, as well as measurements made in the city of Clermont-Ferrand. In chapter 2 we have presented the main concepts that define aerosols, their properties and their characteristics, as well as their impacts. I will present in Chapter 3 the instrumentation put in place for the chemical and physical characterization of aerosols, and the methods used and / or developed to interpret these measurements and to answer the objectives of the thesis.
The rest of the manuscript will be devoted to the results, which will be articulated around three chapters. Chapters 4.1 and 4.2 present a detailed analysis of the physical and chemical composition of aerosol sources over 12 months period during 2015-2016. Chapter 4.3 will be dedicated to the study of the links between the chemical composition of PM1 at the high altitude PUY station and measurements made simultaneously in the city of Clermont-Ferrand.
Aerosol particles are an atmospheric component characterized by large diversity. In this part of the manuscript, we describe the properties of atmospheric aerosol. In particular, the focus will be on their microphysical and chemical properties, as well as describing how these properties can vary depending on their sources photochemical processing and evolution as they are transported through the atmosphere.
Aerosol particles are, by definition, all the liquid and / or solid particles suspended in the atmosphere. In general aerosol refers to a group of particles in either liquid or solid form and the term particle refers to an individual particulate species. In the thesis, these terms particle and aerosol will be used together to describe all suspended solid matter.
In order to distinguish the aerosol particles from the point of view of their source, there are two main classifications. The first proposes to classify the particles according to their origin, anthropogenic or natural/biogenic, whereas the second proposes to differentiate the particles from their mode of emission in the atmosphere, primary or secondary. It is finally possible to combine the two classifications for a more complete description, the primary and secondary particles being able to be of natural/biogenic or anthropogenic origin.
At a global scale, natural emissions account for between 70 and 90% of mass emitted particle fluxes (Delmas et al., 2005). Significant natural sources of particles include soil and rock debris (terrestrial dust), volcanic action, sea spray, biomass burning, and reactions between natural gaseous emissions. Table 2.1 presents a range of emission estimates of particles generated from natural and anthropogenic sources, on a global basis. Emissions of particulate matter attributable to the activities of humans arise primarily from four source categories: fuel combustion, industrial processes, nonindustrial fugitive sources (roadway dust from paved and unpaved roads, wind erosion of cropland, construction, etc.), and transportation sources (automobiles, etc.).
The following paragraphs will first describe the main natural and anthropogenic sources of aerosols, considering successively their primary and their precursor’s emissions.
Dust source in the atmosphere consist of two major sources – mineral and volcanic.
• Mineral sources arises from wind acting on soil particles. The major dust sources are located in arid and semiarid regions of the world, which cover about one-third of the global land area. The largest global source is the Sahara-Sahel region of northern Africa; central Asia is the second largest dust source. Estimates of global dust emissions are uncertain. Published global dust emission estimates since 2001 range from 1000 to 3000 Tg/yr (Table 2.1), and atmospheric burden estimates range from 8 to 36 Tg (Zender et al., 2004). The large uncertainty range of dust emission estimates is mainly a result of the complexity of the processes that raise dust into the atmosphere. The emission of dust is controlled by both the wind speed and the nature of the surface itself. In addition, the size range of the dust particles is a crucial factor in emissions estimates. It is generally thought that a threshold wind speed as a function of particle size is required to mobilize dust particles into the atmosphere. The average lifetime of dust particles in the atmosphere is about 2 weeks, during which dust can be transported thousands of kilometers. Saharan dust plumes frequently reach the Caribbean and Europe, and plumes of dust from Asia are detected on the west coast of North America.
• Volcanic sources have much more variable emissions. They depend on the frequency of direct volcanic eruptions in the atmosphere of volcanic ash, which consist mainly of pulverized rocks and minerals. These primary volcanic aerosols, usually greater than one micron in diameter, will have a short lifetime in the atmosphere, unless the ejection velocity allows them to reach the free troposphere and above. On the other hand, these eruptions are accompanied by gaseous releases rich in sulfur compounds such as SO2 and H2S, whose annual emissions range from 6 to 20 Tg of sulfur per year (Table 2.1). The oxidation in the atmosphere of these precursor gases leads to the formation of sulfate aerosols, which depend on the altitude at which they are emitted and therefore will have a lifetime of a few weeks (troposphere) to several months or even years (stratosphere).
Marine source is the most important natural source of aerosols, emitting between 1000 and 6000 Tg of sea spray per year (Table 2.1). These marine aerosols, of primary types, are produced by the mechanical action of the wind on the surface of the oceans (Woodcock et al., 1953). The emission of these aerosols thus depends mainly on the wind speeds, which above a certain threshold allow the tearing and suspension in the atmosphere of seawater droplets drying out to particles. The diameter of this marine spray can vary from a twenty nanometers to about ten micrometers. Marine sprays may contain a part of biological material and inorganic salts, more generally called sea salts (SS for sea salt). In terms of mass flux, most of the emissions of these marine salts are in the supermicron fraction (da> 1μm) (O’Dowd et al., 1997).
Biogenic sources are emissions that come from natural sources of particles, including pollens which correspond to coarse aerosols, with sizes range from 1 to 100 μm and spores which are emitted up to 30 Tg/yr (Table 2.1). Bacteria and viruses are also considered as biogenic particles, their diameter varying from a few nanometers to several microns for the largest bacteria (Jaenicke and Mathias-Maser, 1992). Marine sea-level emissions can also result in the suspension of organic material associated with marine biological activities, resulting in primary submicron organic aerosol emissions (Leck and Keith Bigg, 2008). Also a significant fraction of biogenic particles in forested areas are the source of precursor gas emissions such as volatile organic compounds (VOCs) as well as more moderate amounts of nitrogen oxides (NOx). The most Known VOC is the dimethyl sulfide (DMS) who’s mainly emitted by phytoplankton.
Natural Biomass Burning (BB) sources are resulting from the intentional burning of land which is the major source of combustion products to the atmosphere. Most of this burning occurs in the tropics, leading to emissions of gas precursors including VOC, NOx, SO2, and NH3 and primary aerosols including organic carbon compounds, mainly consisting of oxygen and hydrogen atoms, and black carbon (BC), mainly composed of carbon atoms. The quantity and type of emissions from a biomass fire depend not only on the type of vegetation but also on its moisture content, ambient temperature, humidity, and local wind speed. Seinfeld and Pandis, (2006) show an estimation of global emissions of trace gases from BB. They show that CH4 contribute 40 Tg/yr out of a total flux of 598 Tg/yr, CO contribute 700 Tg/yr out of a total source of 2780 Tg/yr. For NOx, biomass burning is estimated to contribute globally 7.1 Tg/yr, as compared to 33 Tg/yr from fossil-fuel burning.
Anthropogenic Biomass Burning sources such as wood-burning, represent emissions of 50 to 90 Tg/yr on a global scale (Table 2.1). In terms of primary aerosol emissions, the anthropogenic BB source appears to be superior to the natural source. In addition, these primary aerosols have predominantly submicron diameters (Rau, 1989) and a supermicron fraction (Park and Lee, 2003).
The chemical components present in particulate form in the atmosphere are extremely numerous and varied. These components are specific to the particles origin and formation process and, can evolve rapidly during atmospheric transport during condensation, coagulation, and photochemical processes. Understanding the chemical composition of aerosol particles is important since it can influence the optical properties of the particles (Maring et al., 2000; Pilinis et al., 1995), their health impact and their ability to form a cloud droplets. The most important chemical species in the submicron atmospheric aerosols are: ammonium, sulphates, nitrates, chlorides and organic matter (OM). Despite the challenges associated with in-situ aerosol chemical characterization, analytical techniques available today allow complex insights into the main particulate components properties (Putaud et al., 2010).
An example of this characterization is presented in Figure 2.1 This figure shows the annual average chemical composition from off-line filter measurements at 39 sites across Europe across different size ranges: PM2.5 and those associated with coarse particles (PMcoarse; PM2.5-10) (Putaud et al., 2010). The significant portion of unaccounted mass, referred as unacc, evidences the complexity of a total chemical characterization in Figure 2.1 (obtained as the difference between the gravimetrically measured aerosol mass concentration and the sum of the aerosol component concentrations) which is found in PM2.5-10 and PM2.5 fractions for most sites. This unacc mass fraction can result from analytical errors, a possibly systematic underestimation of the PM constituents whose concentrations are calculated from measured data (e.g. OM, carbonaceous matter (CM), and mineral dust (MD)), and aerosol-bound water (especially if mass concentrations are determined at RH >30%). Furthermore, these chemical compositions show significant variability according to the types of sites (urban, peri-urban, rural, regional background, etc.). These results also show differences in chemical composition according to the fractions in particle sizes. In Europe, PM2.5 is dominated by OM, with a significant fraction due to elemental carbon (EC) and by non-marine sulphate (nss-SO4). Nitrates (NO3) are species whose contributions appear to be roughly equivalent in PM2.5-10, PM2.5 and PM10, whereas MD and SS influence the PM2.5-10 and PM10, and are higher depending on their location (proximity to the Sahara or littorals, respectively).
There are, however, disadvantages in off-line filter measurements of aerosol chemistry: large uncertainties on the measurements and low time resolution.
Figure 2. 1. PM2.5 and PMcoarse (PM2.5-10) annual mean chemical composition in the 39 European sites of different typologies (Adapted from Putaud et al., (2010)). The sites are classified according to their geographical position (from left to right: north-west, south and central Europe). Pastel background colors indicate the site types (green: rural background, yellow: near-city, rose: urban background, grey: kerbside).
In recent years, instrumental advances in aerosol mass spectrometry (AMS) have allowed on-line and high-resolution temporal characterization of chemical composition and size distribution of non-refractory submicron aerosols (NR-PM1), which corresponds to species vaporizable at temperatures below 600 °C. This measurement technique therefore excludes MD, SS and EC, but on the other hand makes it possible to quantify the PM1 concentrations of ammonium, sulphates, nitrates, chlorides and organic matter. Furthermore, spectral analysis of different species allows insights into their origin, particularly primary or secondary organic aerosol. Although most AMSs do not measure refractory PM1 (or R-PM1), it should be noted that some versions of AMS make possible the more specific characterization of BC, SS or MD (Cross et al., 2009; Onasch et al., 2012). Several short term studies of AMS in the Northern Hemisphere (Jimenez et al., 2009; Ng et al., 2010; Zhang et al., 2005), shows that submicron particle mass is largely dominated by aged organic compounds. Petit et al., (2017) in urban sites in France, shows that all sites are characterized by a strong predominance of secondary pollution, and more particularly of ammonium nitrate, which accounted for more than 50% of submicron aerosols. Lanz et al., (2010) at an alpine region, show that the NR-PM1 aerosol particles are dominated by organics (36% to 81%). Other main constituents comprised ammonium (5–15%), nitrate (8–36%), sulfate (3– 26%), and chloride (0–5%). Freney et al., (2011) at PUY site, show highest nitrate and ammonium mass concentrations were measured during the winter and during periods when marine modified air masses were arriving at the site, whereas highest concentrations of organic particles were measured during the summer and during periods when continental air masses arrived at the site. Kiendler-Scharr et al., (2016) at urban and rural sites in Europe, show that organic nitrates contribute substantially to particulate nitrate and organic mass, and represent 34% to 44% of measured submicron aerosol nitrate and are found at all urban and rural site. Also, Zhang et al., (2007) identified chemical compositions and mean concentrations of NR-PM1 measured at 37 sites all over the world. These results show that, despite a certain variability in the chemical composition, the NR-PM1s are globally dominated by organic and sulphates species, with respective fractions of 45 and 32% on average over all 37 sites. The ACSM (Aerosol Chemical Speciation Monitor, (Ng et al., 2011)), recently developed for the routine and long-term chemical characterization of NR-PM1, is based on the same principle as AMS, but does not provide access to the size distribution. This instrument, deployed in our study for the chemical characterization of aerosols, will be described in more detail in chapter 3 of this manuscript.
Figure 2. 2. Annual average NR-PM1 relative chemical composition across Europe. Geographical locations are shown by coloured background panels. The type of site is indicated by coloured diamonds (Bressi et al., 2018).
In recent years, the deployment of mass spectrometers dedicated to aerosols during field campaigns conducted on different continents has made it possible to highlight geographical and seasonal trends in the chemical composition of NR-PM1. Long term measurements are important to facilitate continuous control of air quality, monitoring of climate changes, to study the seasonal variation, for long range transport and to increase statistics on the measurements. Long-term ACSM data sets have recently been presented in the literature almost all over the world such as: 2 years near Paris (France) (Petit et al., 2015), 1.5 years in central Oklahoma (Parworth et al., 2015), 1 year in Zurich (Switzerland) (Canonaco et al., 2013, 2015), near Johannesburg (South Africa) (Tiitta et al., 2014; Vakkari et al., 2014), near Barcelona (Spain) (María Cruz Minguillón et al., 2015; Anna Ripoll et al., 2015), and in the Southeastern United States (Budisulistiorini et al., 2016). Also, Bressi et al., (2018) identified the relative chemical compositions of NR-PM1 measured at 21 sites across Europe, where the ACSM (19 sites) and AMS (2 sites) were implemented. These results, presented in Figure 2.2, show that, despite a certain variability in the chemical composition, the NR-PM1s are globally dominated by organic aerosol (OA) over all 21 sites. It is also observed that the OA contributions are similar at remote, coastal, regional, urban or industrial areas (between 50% and 56%). Slightly higher OA contributions are found in Northern Europe compared to Southern Europe or Mid-Latitude Europe. Sulphate and nitrate contributions largely differs on the European site (9-44% and 4-42%, respectively). On average, higher sulphate and nitrate contributions are found remote/coastal areas and urban/regional areas, respectively. Higher sulphate contributions can be associated with shipping emissions of SO2 influencing coastal sites (Viana et al., 2014) and relatively long (days) lifetime and in-cloud formation kinetics of sulphate (Stein and Lamb, 2002 and references therein). Higher nitrate contributions can be associated with Local anthropogenic emissions of NOx – mostly from road and non-road transport, energy transformation and industrial combustion in Europe (Pay et al., 2012).
These high-time resolution measurements of aerosol chemistry performed across Europe can help assess the efficiency of abatement policies implemented on short- and long-term durations (e.g. traffic rationing and sulphur emission reductions, respectively), and at the local and European-wide scales.
The diversity of aerosol particles sources, the physical, and chemical mechanisms of aerosol formation leads to a large variation in the aerosol size distributions and morphologies. Diameters of aerosol particles are distributed over several orders of magnitudes from a few nanometers up to tens of microns. The particle size distribution (PSD) can be expressed in number or mass concentration, as well as a function of surface of volume. The measurement and characterization of the PSD is essential in order to understand aerosol particles sources and to determine the impact of different atmospheric physical, chemical and radiative processes.
Figure 2.3, displays an example of different size modes (expressed as volume distributions) that can be observed in the atmosphere as well as the processes at the origin of their formations and their deposits. Diameters corresponding to PM1, PM2.5 and PM10 as well as those of ultrafine, fine and coarse particles have also been reported for comparison. This figure shows in particular that the total suspended particles is distributed around three modes.
The ultrafine particles, corresponding to all particles with diameters less than 0.1 μm, are mainly composed of secondary aerosols. These ultrafine particles are divided into two distinct modes: the nucleation mode and Aitken mode. These particles are produced by homogeneous and heterogeneous nucleation processes (Kulmala and Kerminen, 2008). They can form during natural gas-to particle condensation or during condensation of hot vapor in combustion processes. Nucleation mode particle diameters are less than 0.02 µm whereas, in the Aitken mode particles are defined within a diameter range between 0.02 and 0.1 μm.
The fine particles covers all particles with diameters less than 2.5 μm, indicating that it comprises all the ultrafine particles, those resulting from the condensation of gases on these ultrafine particles or their coagulation, as well as some primary particles that are directly formed in the submicron sizes. These phenomena lead to the formation of particles with diameters greater than 0.1 microns, which consequently belong to the accumulation mode. The upper limit of the accumulation mode is generally considered to be 1 μm (Oberdörster et al., 2005; J. H. Seinfeld and Pandis, 2006; Whitby et al., 1978) although other studies reported the value of 2 μm (Jacobson, 2005; Nazaroff, 2004).
Table of contents :
Chapter 1: General introduction and aims of this doctoral thesis
1.1. General Introduction
1.2. Aims of this doctoral thesis and organization of the manuscript
Chapter 2: State of the art
2.1. Aerosol particles
2.1.2. Chemical composition
2.1.3. Microphysical properties
2.2. Spatial distribution
Chapter 3: Sampling sites and analytical tools
3.1. Sampling sites
3.2. Analytical tools
3.2.1 In-situ measurements
126.96.36.199 Particle size distribution
188.8.131.52.1 Scanning Mobility Particle Sizer (SMPS)
184.108.40.206.2. Optical Particle Counter (OPC)
220.127.116.11. Equivalent Black carbon (EBC) concentrations
18.104.22.168.1. Multi Angle Absorption Photometer (MAAP)
22.214.171.124.2. Aethalometer-33 (AE33)
126.96.36.199. Gas concentrations
188.8.131.52.1. Carbon monoxide (CO)
184.108.40.206.2. Nitrogen oxides (NOX)
220.127.116.11. Non refractory – PM1 chemical concentrations
18.104.22.168.1. Time-of-Flight Aerosol Chemical Speciation Monitor (ToF- ACSM)
22.214.171.124.2. Positive Matrix Factorization (PMF)
3.2.2. Remote Sensing
126.96.36.199. Aerosol layer height
188.8.131.52. ECMWF-ERA-Interim and LACYTRAJ
Chapter 4: Results
4.1. Seasonal variation of aerosol size distribution data at the Puy de Dôme station with emphasis on the boundary layer/ free troposphere segregation (Farah et al., (2018) published in atmosphere)
4.1.1. Segregating between Boundary Layer (BL)/Aerosol Layer (AL) and FT air masses
184.108.40.206. Comparison between the Aerosol Layer Height from LIDAR Profiles and Boundary Layer Height Simulated with ECMWF
220.127.116.11. Radon-222 (222Rn)
18.104.22.168. Comparison of the Four Criteria
22.214.171.124. Classification of Air Masses by Combining Four Criteria
4.1.2. Comparisons of the Free Troposphere and Boundary Layer Aerosol Properties
4.1.3. Aerosol Properties in the Lower Free Troposphere as a Function of Air Mass Type and Age
4.2. One year of on-line chemistry measurements of the non-refractory submicron aerosol at the Puy-de-Dôme with an emphasis on air mass transport and free troposphere / boundary layer conditions (Farah et al., (2018) in prep.)
4.2.1. Time series and mass concentrations
4.2.2 Seasonal variations
4.2.3 Diurnal variations
4.2.4 Air mass speciation
4.2.5 FT/BL conditions
4.3. Understanding sources of aerosol particles and their exchanges between the high altitude PUY and its surrounds.
4.3.2 Characterization of meteorology, back-trajectories
4.3.3 Bulk PM1 chemical composition
126.96.36.199 PUY Free Tropospheric (FT) conditions
188.8.131.52.1 Time series and concentrations
184.108.40.206.2 Diurnal variations
220.127.116.11 PUY Boundary Layer (BL) conditions
18.104.22.168.1 Time series and concentrations
22.214.171.124.2 Diurnal variation
126.96.36.199.2 Statistical study for the separation of sources
188.8.131.52.2.1 ME-2 results
184.108.40.206.2.2 ME-2 factors diurnal variations
Chapter 5: Synthesis and perspectives