Atmospheric pollution may be of both anthropogenic and natural origin (Levine, 1996; Finlayson and Pitts, 1999; IPCC, 2000; Delmas et al., 2005). Concerning natural pollution, most active volcanoes emit, among others, gases (eg. sulphur dioxide) and particles into the atmosphere during eruption events (Hobbs et al., 1991; Bhugwant et al., 2009). It is thus important to conduct the monitoring of these pollutants for active volcanoes all over the world, in order to take the adequate measures of air pollution (eg. to establish evacuation plans for the surrounding population) and also to study the long-term trends and effects caused by volcanic activity. In this sense, since a decade the World Health Organization (WHO), the European Community and the French Ministry of Environment (MEDDTL: Ministère de l’Ecologie, du Développement Durable, du Transport et du Logement) have established SO2 guideline levels, following the severity of the impact of this pollutant on human health and on the environment (WHO, 2005).
Previous experimental (ground-based and air-borne) studies indicate that major volcanic eruptions such as Mount St Helens (May 1980), El Chicon (March-April 1982) and Pinatubo (June 1991) injected large amounts of solid particles and volatile gases in the troposphere and the stratosphere, up to an altitude of 25 km high (Turco et al., 1993; Trepte et al., 1993; Krueger et al., 1995; Schneider et al., 1999). Another study showed, in a statistical approach, the global distribution of volcanic SO2 degassing during the last century (1900-2000) and further indicated that each eruption (even for non monitored ones) could affect the stratosphere, based on empirical observations (Hamler et al., 2002). More recently, the eruption of the Eyjafjallajökull volcano in Iceland, which began on April 14th 2010 and ended in October 2010, caused enormous disruption to air travel across western and northern Europe, due to huge amounts of ash and particles emitted in the atmosphere.
Volcanic eruptions may cause irreversible environmental and ecological impacts (burial by ash, mud, etc.), via lava or pyroclastic flows, dusts and ash falls and/or gaseous emissions, while they may also increase the economic burden in socio-economic sectors, by causing infrastructure and habitation damages (e.g. houses, buildings, roads, fields and forests covered with ash and/or lava) (Munich Re, 1998; Brosnan, 2000). It is now clearly established that the volcanic pollutants have damaging effects on both human health and ecology. The human health effects from air pollution vary in the degree of severity, covering a range of minor effects to serious illness such as impaired pulmonary function as well as premature death in certain cases (Brantley and Myers, 2000).
Volcanic eruptions mostly emit water vapour, sulfur dioxide (SO2), hydrochloric acid (HCl), hydrofluoric acid (HF) and carbon dioxide (CO2) (Tabazdeh et al., 1993; Evans and Staudacher, 2001; Kinoshita, 2003; Dubosclard et al., 2004). Once emitted in the atmosphere, they contribute to acid rain and may also affect the stratosphere (Galindo et al., 1998).
The atmospheric SO2 emitted during volcanic eruptions is chemically transformed into sulphate aerosols (H2SO3 and H2SO4) during transport (McKeen et al., 1984). These sulphate aerosols may remain in suspension in the atmosphere for several years mainly in the stratosphere, while in the troposphere they are washed out in between weeks (Graf et al., 1993). A recent epidemiological study showed that the SO2 emitted during the eruption of the Piton de la Fournaise volcano might have a potential sanitary impact on the surrounding population (Viane et al., 2009).
The mean residence time of ambient SO2 concentration is about 1 day, in absence of sink processes such as rainfall (Delmas et al., 2005). However, the public may be exposed to sulfur dioxide concentrations that are higher than typical outdoor air levels. The effects of exposure to any hazardous substance depend on the dose, the duration, the way one may be exposed, the personal traits and habits. Short-term exposures to high levels of sulfur dioxide can be life-threatening (ATSDR, 1999; WHO, 2005).
The study of gaseous and particulate emissions in the atmosphere induced by natural events such as volcanic eruptions may therefore help for a better comprehension of their impact, especially on the atmospheric chemistry and on human health.
However, up to now, very few studies based on atmospheric measurements have been undertaken with links to sources such as volcanic eruptions in the southern hemisphere, especially at Reunion island, where atmospheric measurements such as SO2 and particles are still scarce (Bhugwant and Brémaud, 2001; Bhugwant et al., 2002; Halmer, 2005; Bhugwant et al., 2009).
This chapter presents the SO2 concentration measurements, undertaken over Réunion island and in the vicinity of the Piton de La Fournaise volcano during and off eruption events from 2005 to 2010. This volcano is located in the south-eastern part of the Réunion island, a tiny island found in the south-western Indian Ocean (21 S; 55,5 E). It is surrounded by quite densely inhabited regions from the north-eastern to the southern part of it. This active volcano erupts regularly, at least once a year, with notable amount of degassing. However, due to the geographic configuration, coupled with the meteorological conditions which prevail regionally (easterlies) and locally (land/sea breezes) over the island, the volcanic plumes are regularly transported towards the inhabited regions located downwind and may cause sanitary and environmental impacts.
The SO2 concentration was measured continuously with an analyser in different parts of the island and with passive diffusion samplers in the vicinity of the Piton de La Fournaise volcano since 2005. The tremor activity of the Piton de La Fournaise volcano and the meteorological data (winds and rainfall) were measured at several locations over the island for each eruption. The analysis of these datasets have contributed towards explaining the spatial distribution and the temporal evolution of the volcanic SO2 emissions, mainly due to dynamical (transport) and deposition (wet/dry) processes.
2. Guideline values for the SO2 concentration levels
It has been shown that high SO2 concentration levels present in ambient air may cause considerable environmental and sanitary impacts (EPA, 1997). It is thus important to conduct the monitoring of this pollutant, in order to take the adequate measures of air pollution and also to study its long-term trends and effects. In this sense, since a decade the World Health Organization (WHO), the European Community and the French Ministry of Environment (MEDDTL : Ministère de l’Ecologie, du Développement Durable, du Transport et du Logement) have established guidelines for the SO2 concentration levels, following the severity of the impact of this pollutant on human health and the environment.
The French National air quality criteria is mainly based on the Decree n 2002-213 of 15th February 2002 and recently on the Decree n 2010-1250 of the 21st October 2010, concerning the air quality survey, in particular for SO2, and its effects on the human health and the environment, in the air quality objectives, in the alert thresholds and in the limit values (ADMINET, 2002, 2010).
The European Council Directive (1999) also defines the limit values for sulfur dioxide in the ambient air (see Table I). The objectives of this directive are, to establish limit values and alert thresholds for sulfur dioxide concentrations in ambient air intended to avoid, prevent and/or reduce harmful effects on human health as well as on the environment.
The Table I summarizes briefly the main SO2 guideline values (critical levels) for the human health and for the vegetation, in order to establish the regulation context of this study (Council Directive, 1999 ; ADMINET, 2002; WHO, 2005).
II. Alert threshold for sulfur dioxide
500 μg/m3 measured over three consecutive hours at locations representative of air quality over at least 100 km2 or an entire zone or agglomeration, whichever is the smaller.
3. Experimental set-ups and measurements
The ORA measured SO2 concentrations using passive diffusion samplers on 10 inhabited sites (1 to 10) located in the vicinity of the Piton de La Fournsaise volcano from mid 2005 to late 2007. During the April-2007 eruption however, SO2 concentrations were measured using passive diffusion samplers on 12 (11 to 22) additional inhabited sites located close to the Piton de La Fournaise volcano. Since 2009, 4 more sites (23 to 26) have been added to reinforce the SO2 survey in the vicinity of the volcano (Figure 1). All the selected sampling sites are located in inhabited areas, with almost no traffic circulation and at more than 1 to 2 km away from busy highways, in order to avoid local anthropogenic contamination.
3.1. Atmospheric measurements
3.1.1. SO2 measurement using passive diffusion samplers
Passive diffusion samplers (Ferm, 1991) were used at 10 sites located around the Piton de La Fournaise volcano for the measurement of the ambient SO2 concentration since September 2005. The measurements by passive diffusion samplers are based on the property of molecular diffusion of gases and species-specific collection on an impregnated filter specific to the SO2 pollutant measured. The passive diffusion samplers, which were exposed over a period of one week to 10 days, provided quantitative SO2 concentration measurements and the spatial distribution of this pollutant during and off volcanic eruptions of the Piton de La Fournaise volcano.
The uncertainty on the SO2 concentrations measured by passive samplers was evaluated during two inter-comparison phases in late 2005, with two series of SO2 tubes sampled on the sites. One of the series was analysed by the Joint Research Center (JRC) and the other one by the Gradko International Limited (Bhugwant et Siéja, 2005, 2006). This sampling technique has also been tested in different tropical and subtropical regions (Carmichaël et al., 2003; Ferm and Rodhe, 1997).
On the monitoring sites, the samplers are installed mesh side down in the underside of a plastic disc screwed at the top of a wooden pole and left in position for one week to 10 days (see Figure 2).
After they are taken off, the exposed samples are stored in a refrigerator, in order to stabilize the trapped air samples. Then, they are sent to Gradko International Limited for subsequent laboratory analysis, via Ion Chromatography, in order to calculate the SO2 concentration. The precision of the samplers, expressed as mean percentage difference between duplicates was found to be in the 10-25% range for SO2 (Ayers et al., 1998).
3.1.2. SO2 monitoring using SF-2000 analyser
The ORA also measured SO2 concentration continuously with automatic analysers in an atmospheric network composed of fixed stations at different parts over Réunion island (Figure 3). During the April-May 2007 eruption, an SO2 analyser SF-2000 (SERES) model was installed temporarily at two additional sites located in the south-western and southern parts of the island (see points 20: TR_1 and 22: TR_2 respectively, in Figure 3).
The instrument analyses ambient air by the ultraviolet (UV) fluorescence technique, at a time-base of 15 minutes, subsequently averaged to 1 hour, for analysis with other datasets such as tremor intensity, winds and rainfall.
Sulfur dioxide molecules are brought to fluorescence once irradiated by ultraviolet radiation in the 190-390 nm wavelength range.
This UV fluorescence is the maximum for a wavelength found in the 210-230 nm range produced by a UV radiation generator which uses a zinc vapour lamp. In the case of the SF-2000 analyser, the UV radiation of 215 nm wavelength produced by the zinc vapour lamp excites SO2 molecules contained in the sampling vat, as shown by the equation (1) :
The excited SO2* molecule regains its initial energetic state E by emitting a radiation of wavelength λ’, as indicated in the equation (2):
The re-emitted energy is lower than h, the excitation energy, and hence the wavelength of the UV fluorescence radiation λ’ (240-240 nm) is greater than that of the stimulation source wavelength λ (=215 nm). The fluorescence phenomenon stops when the excitation source is suppressed. The intensity of radiation of the sample is proportional to the SO2 concentration: [SO2] = k x IF, where k is the proportionality factor and IF is the measured intensity (Ruidavets et al., 2005).
The SO2 concentration measured by the SF-2000 analysers at the four locations situated in the north-western part of the island is carried out continuously since January 2001. However, only atmospheric data collected during summer 2005 and spring 2007 will be discussed in this work, as notable volcanic pollutants emitted during these eruption episodes were dispersed over different inhabited parts of the island.
3.2. Seismic network
The Observatoire Volcanologique du Piton de la Fournaise (OVPF) is in charge of the seismic and deformation network which has been implemented over the Piton de la Fournaise volcano since 1980. This network is composed, among others, of 25 seismic stations, with 1Hz and large band instruments (Aki and Ferrazzini, 2000; Staudacher at al., 2009). It records the seismic activity and deformation of Piton de la Fournaise 24 hours throughout a day and transfers data by radio in real time to the observatory. The seismic stations closer to the volcano enclosure (Staudacher at al., 2009) and to the eruption site are best representative of the eruptive tremor, and have been studied in particular for the April-May 2007 eruption (Bhugwant et al., 2005, 2009).
3.3. Meteorological data
The local meteorological parameters such as wind speed, wind direction, and rainfall are measured from the instruments onboard a tower at 10 m above the ground level, close to most of the sampling locations. These meteorological stations are managed by the French Meteorological Service (Météo-France, 2000). The wind fields calculated from the ECMWF (European Center for the Medium-Range Weather Forecast) data archives are also analysed (not shown) in order to study the dynamical processes prevailing on regional scale over Réunion Island (Bhugwant et al., 2009).
3.4. Tropospheric SO2 concentration variability measured from satellite data
The Ozone Monitoring Instrument (OMI) is dedicated to detect and measure volcanic eruptions degassing and anthropogenic pollution from space. The OMI is a hyperspectral UV-Visible spectrometer with a resolution of 13x24 km at nadir. In order to gain more information about the spatial distribution of the SO2 and the characteristics (composition, etc.) of the volcanic plumes, satellite data retrieved each day from the Aura OMI satellite following the eruptions were analysed (Carn et al., 2003, 2007; Yang et al., 2007; Bhugwant et al., 2009).
3.5. Field observations of the Piton de la Fournaise activity
It is now well established that the Piton de la Fournaise volcano is one of the most active volcanoes in the world, with one eruption every 8 months, over the last century. Since 1998, a particular intense volcanic activity is observed, with 2 to 4 eruptions per year. These eruptions are mainly basaltic, producing generally lava fountains of 50 to 100 m high and fluid lava flows of aa or pahoehoe type and of aphyric to olivine-rich composition. Most eruptions of Piton de la Fournaise last between 2 weeks and 1 month. However, the March 1998 eruption lasted 196 days. As the driving force of every eruption is the trapped gas, they always accompany each eruption. However, significant gas emissions are observed at Piton de la Fournaise during intense eruptions only (Bhugwant et al., 2001, 2009).
3.6. Sanitary and environmental survey
Following the recent population control, Réunion island holds a population of 730 000 inhabitants (INSEE, 2010). About 24% of the population of the island lives in the vicinity of the Piton de La Fournaise volcano and may thus be exposed to sanitary hazards during eruption episodes. Moreover, a large part region localised close to the Piton de La Fournaise is composed of agricultural lands and also forests with endemic species, which may be subject to environmental impacts (eg. acid rains) during eruptions.
4. Results & discussions
4.1. Spatial SO2 concentration distribution during late December 2005 eruption
The methodology applied to calculate the spatial distribution of the SO2 concentrations measured at different locations over Réunion island is the kriging interpolation method (Carletti et al., 2000).
The Figure 4 presents the GIS (geographic information system) plot of the spatial distribution of the mean SO2 concentration measured during an eruption event which occurred on late December 2005-early January 2006, i.e., from 27th December 2005 to 3rd January 2006. The mean SO2 concentration shown in this figure consists of the data obtained from both passive diffusion samplers and the automatic analysers on the atmospheric stations. It may first be observed that the highest SO2 concentration ( 70 μg/m3) is mostly confined on the southern and the south-eastern parts of the island. Then, it decreases rapidly after a few kilometres from the eruption vent. Importantly, it may be seen that the eastern to north-eastern part of the island is weakly concerned by the SO2 concentration emitted during this eruption event. The main causes are, among others, the ranges of mountains Cilaos, Mafate and Salazie (see Figure 2), which constitute a natural barrier for the transfer of atmospheric pollutants downwind, over the eastern to northern parts of the island. Moreover, due to the conjunction of the important relief of the island and the easterlies prevailing over the south-western Indian Ocean, the air masses originating from the south-eastern part of the island tend to by-pass it (Météo-France, 2000). Also, the geographic configuration of the island tends to generate microclimates over the island. Consequently, the windward side of the island is more windy with high rainfall levels throughout the year while the leeward side of it is drier with less winds. This parameter may also in part explain the low SO2 concentration observed on the eastern to northern sectors.
4.2. Spatial SO2 concentration distribution during early January 2006 eruption
The Figure 5 presents the plot of the spatial distribution of the mean SO2 concentration obtained by passive diffusion samplers and automatic analysers during an eruption event which occurred from late December 2005 to mid-January 2006, i.e., from 3rd January 2006 to 12th January 2006. The highest SO2 concentration ( 370 μg/m3) is observed on the northern to south-western parts of the Piton de la Fournaise volcano. Interestingly, it may be seen that the high SO2 levels are constrained in a mountainous sector with a very low density of population. The contaminated air masses follow gullies and rivers lane found in the vicinity of the volcano (on the north-western to southern part of it) to attain the littoral regions of the island. The comparison of the two figures (4 and 5) shows that, in spite of the relief and the meteorology, the source intensity is also a dominating factor which can explain the sulfur dioxide variability observed for these two eruptive episodes.
4.3. Spatial SO2 concentration distribution during mid-April 2007 eruption
The Figure 6 presents the plot of the spatial distribution of the mean SO2 concentration obtained by passive diffusion samplers and automatic analysers during an eruption event which occurred from early April 2007 to early May 2007, i.e., from 18th April 2007 to 26th April 2007. The highest SO2 concentration ( 190 μg/m3) is mainly concentrated on the northern to south-western parts of the Piton de la Fournaise volcano while moderate levels are also measured up to the southern to south-western part of the island.
Since the SO2 surveillance in the vicinity of this volcano, it may be seen that during the previous eruptions, regularly the volcanic plumes are preferentially transported to the southern to north-western parts of Réunion island.
4.4. Temporal evolution of SO2 concentration in 2006 and 2007
The Figure 7 presents the monthly mean SO2 concentration measured by automatic analysers in atmospheric stations at different parts of Réunion island in 2006 and 2007. In absence of any eruption, the monthly mean SO2 concentration exhibits low levels, varying in the 3-12 μg/m3 range at all the monitoring sites.
The highest SO2 concentration ( 50 μg/m3) is observed at the MOB station (which is closest to the Piton de la Fournaise volcano) while the other stations are also but less impacted by the SO2 emitted in April 2007. This assessment is coherent with the results presented in Figure 6.
Importantly, it may be seen that the atmospheric stations located in the north-western parts of the island show low values while the MOB station exhibit high values during each eruption event registered in 2006 and 2007. The results also suggest a regional impact of the contaminants emitted during the Piton de la Fournaise volcano eruption and the implications on human health for the population exposed to atmospheric pollution.
4.5. Winds spatial distribution over Réunion island
Dynamical processes such as winds (direction and speed) have notable influence on atmospheric constituent variability. It is thus important to study the winds regime prevailing over the island in order to assess their influence on the SO2 concentration variability. The Figure 8 presents the wind roses annual mean distribution over Réunion island, calculated from winds data collected over meteorological stations located in different parts of the island from 1976 to 1995 (Météo-France, 2000). It may be seen that the hilly landscape of the island plays an essential role in the winds distribution. On one hand, it contributes to strengthen the winds in certain sectors, and, on the other hand, to place certain regions under cover. The land/sea breezes as well as the slope breezes, the constituent of which is perpendicular in the coasts, are clearly evidenced on most of the meteorological stations. The local winds distribution induced by the important relief of the island, coupled with the easterlies prevailing regionally over the south-western Indian Ocean, notably contribute to the dispersion of the SO2 over the island during and off eruption events.
4.6. Spatial distribution of rainfall in early winter (April)
In order to gain more information about the spatial distribution and the temporal evolution of the SO2 concentration; additional meteorological data collected over the island was also analysed. Figure 9 presents the spatial distribution of the monthly mean rainfall measured in April (from 1998 to 2008) over Réunion island. It may be seen that the maximum rainfall occurs in the southern to north-eastern regions, with a maximum rainfall of 1 200 mm in the East of the island. A comparison with the 4 to 6 suggest that this parameter is the main factor which may explain the lower SO2 levels observed in the south-eastern to the northern regions, especially during eruption events.
In this chapter the spatial distribution of SO2 concentration and its temporal evolution during several eruptions of the Piton de La Fournaise volcano has been analysed, in conjunction with other parameters, such as seismic and meteorological data.
A good correlation has been established between the seismic variability and the SO2 concentration variation in the vicinity of the volcano (Bhugwant et al., 2009).
The analysis of SO2 data collected in parallel with meteorological parameters show that the rainfall influences notably the SO2 spatial variability in particular over the eastern to northern regions of Réunion island, via scavenging processes.
The conjunction of the important relief of the island coupled with the winds (easterlies and land/sea breezes) also contribute to the spatial distribution and the temporal evolution of the SO2 concentration measured at different parts of it.
Importantly, the results of several years of SO2 measurements also show that the QO (Quality Objective) as well as the LV (Limit values for the protection of the ecosystems), based on an annual average, were not exceeded over the island during the 2005-2010 period. However, the SRI (Recommendation and information threshold) as well as the SA (alert threshold) was exceeded in some inhabited regions close to the volcano during previous eruptions, in particular during the April-May 2007 eruption (Bhugwant et al., 2009).
Recently, a preliminary epidemiological study showed that the SO2 emitted during the eruption of the Piton de la Fournaise volcano might have a potential sanitary impact on the surrounding population. Consequently, this investigation should be carried out in order to establish quantitatively the links between atmospheric pollution induced by eruption events and sanitary impacts.
This preliminary work may be of interest in particular to epidemiologists (eg. to establish the sanitary impact on the surrounding population) and to decision-makers (eg. for crisis managements during eruption episodes).
A close collaboration between the various supervisory boards (ORA, OVPF and Météo-France) may help to establish a decision-making tool to the decision-makers, in particular, to draw up evacuation plans in case of pollution alerts. The recent tools deployed, i.e., the atmospheric data collected from the satellite since a few decades is also a supplement to the ground-based measurements, for the monitoring of atmospheric plumes during future eruptions of this active volcano.
The present work is supported in part by the French ‘Ministère de l’Outre-Mer’ (MOM) Program. We acknowledge the French ‘Minsitry of Environment’ for the financial support of the ORA atmospheric network.
We also gratefully acknowledge the ORA staff for the technical support and for the atmospheric data sampling. We also thank the OVPF staff for preparing the tremor data. Météo-France is also gratefully acknowledged for providing the meteorological data.