meteorological factors controlling soil gases and indoor ... · meteorological factors controlling...

S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 7 ( 2 0 0 9 ) 1 3 6 2 – 1 3 7 2

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ loca te / sc i to tenv

Meteorological factors controlling soil gases and indoor CO2

concentration: A permanent risk in degassing areas

Fátima Viveiros⁎, Teresa Ferreira, Catarina Silva, João L. GasparCentro de Vulcanologia e Avaliação de Riscos Geológicos, University of the Azores, Rua Mãe de Deus, 9500-801 Ponta Delgada,Azores, Portugal

A R T I C L E D A T A

⁎ Corresponding author. Tel.: +351 296 650147E-mail address: [email protected]

0048-9697/$ – see front matter © 2008 Elsevidoi:10.1016/j.scitotenv.2008.10.009

A B S T R A C T

Article history:Received 13 January 2008Received in revised form 3 July 2008Accepted 1 October 2008Available online 8 November 2008

Furnas volcano is one of the three quiescent central volcanoes of São Miguel Island (AzoresArchipelago, Portugal). Its present activity is marked by several degassing manifestations,including fumarolic fields, thermal and cold CO2 springs and soil diffuse degassing areas.One of themost important soil diffuse degassing areas extends below Furnas village, locatedinside the volcano caldera. A continuous gas geochemistry programme was started atFurnas volcano in October 2001 with the installation of a permanent soil CO2 efflux stationthat has coupled meteorological sensors to measure barometric pressure, rain, air and soiltemperature, air humidity, soil water content and wind speed and direction. Spike-likeoscillations are observed on the soil CO2 efflux time series and are correlated with lowbarometric pressure and heavy rainfall periods. Stepwise multiple regression analysis,applied to the time series obtained, verified that the meteorological variables explain 43.3%of the gas efflux variations. To assess the impact of these influences in inhabited zones amonitoring test was conducted in a Furnas village dwelling placed where soil CO2

concentration is higher than 25 vol.%. Indoor CO2 air concentration measurements at thefloor level reached values as higher as 20.8 vol.% during stormy weather periods. A similartest was performed in another degassing area, Mosteiros village, located on the flank of SeteCidades volcano (S. Miguel Island), showing the same kind of relation between indoor CO2

concentrations and barometric pressure. This work shows that meteorological conditionsalone increase the gas exposure risk for populations living in degassing areas.

© 2008 Elsevier B.V. All rights reserved.

Keywords:Carbon dioxideVolcanic gasesPublic healthHazardVolcanic environment

1. Introduction

Volcanic gases can pose a permanent threat to population notonly during eruptive unrest but also in quiescent phases ofvolcanoes (e.g. Blong, 1984; Williams-Jones and Rymer, 2000;Hansell and Oppenheimer, 2004). Soil diffuse degassing areasare potentially one of the main hazardous zones as the gasesare continuously released and can rise below edifices withoutbeing noticed by the population. Carbon dioxide (CO2) isamongst the most important diffused gases released by soildegassing (Allard et al., 1991; Baubronet al., 1991) and if presentat in high concentrations can become particularly dangerous

; fax: +351 296 650142.ov.pt (F. Viveiros).

er B.V. All rights reserved

for public health, since it prevents oxygen respiration. Carbondioxide is colourless, odourless, can silently affect populationand above 15 vol.% may cause death by asphyxia (NIOSH/OSHA, 1981; Blong, 1984; Le Guern et al., 1982; Wong, 1996;Williams-Jones and Rymer, 2000). Table 1 summarizes themain effects of CO2 exposure at different concentrations. Inaddition, since CO2 is denser than air, it can accumulatehazardously in poorly ventilated or low-lying zones.

In the last decades several hundred deaths occurred involcanic areas due to CO2 release (Hansell and Oppenheimer,2004; Weinstein and Cook, 2005; Hansell et al., 2006). Themostcatastrophic events were the Dieng Plateau (Indonesia) gas

.

mailto:[email protected]

http://dx.doi.org/10.1016/j.scitotenv.2008.10.009

Table 1 – Toxicology of carbon dioxide (based on NIOSH/OSHA, 1981; Le Guern et al., 1982; Wong, 1996; Williams-Jones andRymer, 2000)

CO2

concentration(%)

Exposuretime

Effects Occupationalexposure limit

Observations

0.033 – – – Atmospheric air0.5 – – 8 h/day in a

work environmentPEL — permissible exposurelimit

0.5–3.0 Several hours Difficulty in breathingIncrease of the cardiac rhythmHeadache

3.0 – – 15 min STEL — short-term exposurelimit

N15 min 100% breathing accelerationHeadacheMuscle weakness

5.0 – 300% breathing accelerationHeadacheMuscle weaknessDepressionSicknessVertigoNoise in the earsSomnolence

10.0 Severalminutes

Unconsciousness. Quick recovery when theperson is moved to a ventilated environment.

N15.0 – Unconsciousness and death.

1363S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 7 ( 2 0 0 9 ) 1 3 6 2 – 1 3 7 2

cloud emission, causing the death of at least 142 persons (LeGuern et al., 1982; Allard et al., 1989), and the gas release fromMonoun (1984) (Giggenbach et al., 1991) and Nyos lakes (1986)(Barberi et al., 1989; Baxter and Kapila, 1989; Oskarsson, 1990)in Cameroon, which caused the deaths of 39 and 1700 personsrespectively. Even in areas of quiescent volcanism, CO2

degassing may cause dangerous accidents to people andanimals, as it is the case of Central Italy (Rogie et al., 2000;Pizzino et al., 2002; Beaubien et al., 2003; Carapezza et al., 2003;Barberi et al., 2007; Costa et al., 2008).

In the Azores archipelago themost recent casualties causedby CO2 exposure occurred in 1992, with the death of two touristsinside the Furna do Enxofre lava cave (Graciosa Island), due toair CO2 concentrations higher than 15 vol.% (Gaspar et al., 1998).

Another hazardous situation took place in 1997, in RibeiraSeca village, on the northern flank of Fogo volcano (S. MiguelIsland), where some families were indefinitely removed fromtheir houses due to the high indoor CO2 concentrations (Gasparet al., 1998; Ferreira et al., 2005). More recently, similar problemswere identified in Mosteiros village (Sete Cidades volcano, S.Miguel Island) and in Praia do Almoxarife village (Faial Island).

This work aims to address the influence of meteorologicalvariations on the indoor CO2 concentration. Several studiesperformed indifferent degassing areas have already shown thatenvironmental variables may highly influence the gas releasedfrom soil (e.g.Klusman andWebster, 1981; Asher-Bolinder et al.,1991; Hinkle, 1991, 1994; Chiodini et al., 1998; Oskarsson et al.,1999; Diliberto et al., 2002; Granieri et al., 2003). This study isintended to understand if the meteorological variables thathighly control the soil diffuse degassing processes may alsoinfluence the indoorCO2 values. Recently, somestudies (Marley,2001; Rowe et al., 2002; Denman et al., 2007) reported seasonal

variations on the radioactive gas radon (222Rn), explainedmainly by meteorological influences and occupancy factors.This work shows, not only seasonal variations on the gasflux but, above all, focus that lethal indoor CO2 concentrationsmay be reached in only a few hours due to extreme weatherconditions, posingahiddenandpermanent risk forpopulations.In this particular case, it was also observed that the soil CO2

efflux station could work as an early-warning system since thesoil flux antecipates increases in the indoor CO2 concentrationsome hours prior to their occurrence.

2. Study area

The Azores Archipelago comprises nine volcanic islands(Fig. 1A), located where the Eurasian, American and Africanplatesmeet (Searle, 1980). Due to this complex tectonic settingseismic and volcanic activity are frequent in the archipelago.Since its settlement, in the 15th century, several volcaniceruptions and destructive earthquakes have been reportedcausing thousands of deaths and severe damages (e. g.Weston, 1964; Silveira et al., 2003). The islands of São Miguel,Terceira, Graciosa, Faial and Pico show important degassingareas, associated with the hydrothermal systems of someactive volcanic centres, including fumarolic fields, thermaland cold CO2 springs and soil diffuse degassing zones.

S. Miguel Island is formed by three quiescent centralvolcanoes, Sete Cidades, Fogo and Furnas (Fig. 1B). Furnas is atrachytic central volcano and started to form around100000 years BP. It comprises a summit depression 5×8 kmwide formed by two nested calderas controlled by NW–SE andNE–SW faults (Guest et al., 1999). Ten intracaldera explosive

Fig. 1 – (A) Location map of the Azores archipelago and (B) S. Miguel Island digital elevation model. Legend: (a) Sete Cidadesvolcano, (b) Fogo volcano and (c) Furnas volcano. The red square represents the area under study.

1364 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 7 ( 2 0 0 9 ) 1 3 6 2 – 1 3 7 2

trachytic eruptions occurred in this volcano in the last5000 years, two of which happened in historical times (1439–43; 1630) (Queiroz et al., 1995; Cole et al., 1995). At present,volcanic activity is marked by four main fumarolic fields, threeof them located inside the volcano caldera: (1) Furnas Lakefumarolic field; (2) Furnas Village fumarolic field and (3) Ribeirados Tambores fumarolic field. On the south flank of the volcanoit can be found Ribeira Quente fumarolic field with several gasemanations occurring along Ribeira Quente stream. In Furnas

Table 2 – Chemical composition of the main fumarolic fields, in

Fumarolic fields

Reference CO2 H2S

Caldeira Grande LG920128 98.54 0.787Lagoa das Furnas LG920128 99.08 0.387Ribeira dos Tambores FU91RT03 97.65 1.010

Soil diffuse

Soil gas Minimum Aver

Soil CO2 concentration (vol.%)b 0.0Soil temperature (°C)b 11.5 1Soil radon concentration (Bq m−3)c 45.9 670

Descriptive statistics of soil diffuse degassing at Furnas volcano accordin

village several hot springs andCO2 cold springs are also present.Fumarole outlet temperatures are close to the boiling point ofwater (95–100 °C). The main components of the fumarolicdischarges are water vapour (H2O) followed by carbon dioxide(CO2), hydrogen sulphide (H2S), hydrogen (H2), nitrogen (N2),oxygen (O2),methane (CH4) and argon (Ar) (Table 2) (Ferreira andOskarsson, 1999; Ferreira et al., 2005). CO2 is the main gasreleased by soil diffuse degassing at Furnas. Recent work (Silva,2006) measured values for the radioactive gas radon (222Rn) in

molar %, from the work of Ferreira and Oskarsson (1999)a

compositiona

CH4 H2 N2 O2 Ar

0.0047 0.236 0.383 0.0418 0.01080.0026 0.133 0.358 0.035 0.00850.0229 0.343 0.940 0.038⁎

degassing

age Maximum Number of measurements

5.2 94.6 11979.8 74.8 11972.0 11,0808.0 175

g to Sousa (2003)b and Silva (2006)c.


the main soil diffuse degassing area in Furnas caldera and wasfound to be mostly associated to CO2 anomalous areas.

Furnas village is locatedwithin the caldera of Furnas volcanoand, according to the last Portugal census of population andhousing (Census, 2001), has 1541 inhabitants living in 895dwellings. About 22% of the buildings have raised woodenfloors (Pomonis et al., 1999), which consist of old traditionalconstruction that creates a ventilated space mostly to reducethe moisture of the wood floor. By its characteristics, thisconstruction technique also prevents or retards the gas migra-tion into the houses.

Themeanannualprecipitation inFurnascaldera, basedon65climatological years, is 1992 mm according to Marques et al.(2007). The same authors refer that the monthly rainfall dis-tribution exhibits seasonal pattern, with significant differencesbetween the “rainy season” that extends between October andMarch, and the “dry season” with a minimum rainfall in July.

3. Previous works

Baubron et al. (1994) performed an initial soil CO2 concentra-tion map (scale 1/2000) of the inhabited area of Furnas villageshowing that it was built over a significant soil diffusedegassing area. A new survey, using the same methodology,was performed ten years later by Sousa (2003) covering amoreextended area (Fig. 2). The data obtained in those works showsoil CO2 concentration and soil temperature values higherthan 90 vol.% and 70 °C, respectively (Table 2). In the case of

Fig. 2 –Soil CO2 degassing map from Furnas vil

Furnas village, the soil CO2 degassing anomaly looks to haveremained remarkably stable between the two surveys. Thesestudies are important for seismo-volcanic monitoring as, infuture, alterations in the anomaly shapemaymean significantchanges in the deep system. Baxter et al. (1999) based on thework of Baubron et al. (1994) estimated that about one third ofthe houses at Furnas village were located in areas with highCO2 degassing (N1.5 vol.% in the soil). According to the sameauthors, it was not possible to associate loss of human lives tohigh indoor CO2 concentration in the Furnas area. Never-theless, cases of dizziness in high CO2 concentration areas inFurnas volcano have been described through time and can beread in historical accounts from the 16th century (Frutuoso,1522–1591).

During our field work, some people reported breathingacceleration and asphyxia symptoms, occasionally, insidetheir homes. In addition, deaths of little animals (chickens,birds, cats and mice) in some non-ventilated spaces were alsoreported. Some persons also mentioned the empirical asso-ciation between the gas increases inside houses and extremeweather conditions.

4. Methodology and results

4.1. Soil CO2 efflux permanent station

In the scope of the Azores seismo-volcanic monitoringprogramme a soil CO2 efflux permanent station (GFUR1) was

lage, according to the work of Sousa (2003).

Table 3 – Sensor type and descriptive statistics of the data acquired during the monitoring test at Furnas volcano

Variables Minimum Average Maximum Std. Dev. N Type of sensor

GFUR1 soil CO2 efflux (g m−2 d−1) 26.0 245.9 1313.3 87.0 1173 DRÄGER Polytron IR CO2

GFUR1 barometric pressure (hPa) 964.7 998.9 1012.3 7.1 1173 PT100B Pressure TransducerGFUR1 rainfall (mm h−1) 0.0 0.4 22.8 1.6 1173 Rain gauge LASTEMGFUR1 air relative humidity (%) 52.4 89.8 98.2 11.2 1173 TermohigrometerGFUR1 air temperature (°C) 1.2 11.3 20.6 3.3 1173GFUR1 wind speed (m s−1) 0.0 0.6 8.0 1.0 1173 LASTEM GonioanemometerGFUR1 soil temperature (°C) 12.0 13.8 15.6 0.7 1173 Pt 100 ThermometerGFUR1 soil water content (%) 22.3 27.5 48.7 2.7 1173 CS616 ReflectometerIndoor CO2 concentration (vol.%) 0.2 4.7 20.8 3.1 1011 ANAGAS IR CO2

GA barometric pressure (hPa) 948.0 983.5 995.0 7.4 1011 Pressure Transducer


installed at Furnas volcano in October 2001, next to the Furnasvillage fumarolic field (Fig. 2). This station performs measure-ments by the “time 0, depth 0” accumulation chambermethod(Chiodini et al., 1998) and the soil CO2 efflux value representsthe gas concentration increase inside a chamber during acertain interval of time, which was programmed to be 130 sfor GFUR1 station. The CO2 efflux unit is g m−2 d−1. Theaccumulation chamber and the station datalogger unit arecovered with a shelter. The automatic station (manufacturedby WestSystems, Italy) performs measurements on an hourlybasis and has coupled meteorological sensors. Namely it

Fig. 3 –Variation of the soil CO2 efflux and the barometric pressuMarch 2003.

acquires information about barometric pressure, air tempera-ture, air relative humidity, wind speed and direction, rainfall,soil water content and soil temperature.

The station has a local memory able to store up to 2048measurements, however every hour the data are transmittedto the Centre of Volcanology and Geological Risks Assessmentin the University of the Azores via GSM telemetry system.Sensor types, as well as the descriptive statistics of the mainstudied variables, are shown in Table 3.

Time series acquired by the different sensors are plottedin Fig. 3 and show that some meteorological variables

re (A)/rainfall (B) at GFUR1 between 3rd February and 24th

Fig. 4 –Soil CO2 efflux monthly average at GFUR1 between March 2002 and April 2006. The time series was smoothed based onthe compound data smoother defined in Velleman (1980).


influence the gas flux oscillations. From all the monitoredvariables, barometric pressure showed an evident negativecorrelation with the gas flux (Fig. 3A). Significant spike-likeincreases in the gas flux can also be observed concomitantlyto intense rainfall periods (Fig. 3B). Long-term changes areobserved in the soil CO2 efflux time series registered at GFUR1during 4 years (Fig. 4). Excluding the first year of dataacquisition, the data show that soil CO2 efflux values arehigher during winter months than in summermonths. Thesetendencies may be explained by the lower barometricpressure values and the intense rainfall events that aremore frequent in winter periods and may cause a globalincrease on the gas flux. In order to understand and quantifythe environmental influences on the gas flux, stepwisemultiple regression analysis (Draper and Smith, 1981) wasapplied to the data series obtained at the GFUR1 station, from3rd February to 24th March 2003, the period used to performthis monitoring test. Soil CO2 efflux was used as dependentvariable and the monitored meteorological variables weretested as the independent variables. All the monitoredvariables were tested, being included in the regressionmodel only the significant ones (based on the t testsignificance) and the variables that increase the adjusted2

more than a threshold of 1% (Draper and Smith, 1981; Freund

Table 4 – Stepwise multiple regression analysis for data obtain

Independent variable Coefficient B Standard error of B Co

Intercept 4283.43 310.70Rain (b8 mm h−1) 21.66 2.28Barometric pressure −4.28 0.33Rain (N8 mm h−1) 34.78 2.95Soil temperature 16.33 3.08Number of observations (N) 1173

and Wilson, 1998). This adjusted R2 value is a measure of theamount of variation about the mean explained by the fittedregression equation (Draper and Smith, 1981).

The statistical analysis confirmed most of the preliminaryobserved relations between the monitored variables andshowed that 43.3% of the soil CO2 efflux variation (adjustedR2 value, Table 4) can be explained by the influence of theenvironmental monitored variables. Rainfall, barometric pres-sure and soil temperature are the variables with statisticalmeaning to explain the gas oscillations (Pb0.01). Coefficientβ expresses the influence of each independent variable on theCO2 efflux (Draper and Smith, 1981). According to thecoefficient β sign, barometric pressure has a linear negativerelationship with the gas flux. By contrast, rainfall and soiltemperature have a positive correlation with the gas flux.Rainfall gives rise to different effects (Table 4) on the gas fluxdepending on the rain amount and, for this reason, the samephysical variable is split into two data series, one for valueslower than 8 mm h−1 and other one for values above thatthreshold. It is possible to observe that both data sets showa positive correlation with the gas flux, but the coefficientβ shows different weights. In fact, the increase in the soil CO2

efflux is more evident for high rainfall periods (N8 mm h−1)(observe on Table 4 the coefficient β value). Analyzing the

ed at GFUR1 between 3rd February and 24th March 2003

efficient β t test Signif. of t test Adjusted R2 increase (%)

13.79 0.000.26 9.52 0.00 28.8

−0.35 −12.94 0.00 6.90.31 11.80 0.00s 6.30.14 5.31 0.00 1.3

Adjusted R2 (%)= 43.3

Fig. 5 –Variation of indoor CO2 concentration, barometric pressure and rainfall during the monitoring test at Furnas village.


adjusted R2 changes, rainfall appears as the most influencingvariable. By contrast, soil temperature is the onewith the leastrelationship to gas flux.

4.2. Indoor CO2 measurements

Spike-like oscillations of the gas flux may be caused byvariations in meteorological factors, as demonstrated in theprevious section. In order to check the impact of thesemeteorological influences on indoor CO2 concentration, anexperiment was performed during February andMarch 2003 ina selected house at Furnas village (Fig. 2), during a period withno significant seismic activity. This house is built over an areawithout a significant thermal anomaly and with soil CO2

concentration ranging from 25 to 50 vol.% (Sousa, 2003). Thebuilding has two floors with a raised wooden ground floor.

Fig. 6 –Variation of soil CO2 efflux at GFUR1 and indoor CO2 conceMarch 2003.

The CO2 indoor values were registered with a portableinfrared CO2 concentration analyser equipped with a datalog-ger (ANAGAS™ CD 95 model from Geotechnical Instruments,United Kingdom) that also registers the barometric pressurevalues. CO2 concentrations are measured in vol.% from 0 to100, with 0.5% precision. The equipment was installed at theground floor, in a bedroom, and the sampling point was set atthe floor level.

Data was acquired every hour from 3rd February to 24thMarch 2003 and was programmed to be contemporaneouswith data acquired at GFUR1, located about 0.5 km far from theselected house (Fig. 2).

The CO2 concentration in the bedroom ranged between 0.2and 20.8 vol.% with an average of 4.7 vol.% (Table 3). At thesame time, the soil CO2 efflux values at GFUR1 changed from26.0 g m−2 d−1 to 1313.3 g m−2 d−1. During the monitored time,

ntration at the monitoring test site from 3rd February to 24th

Fig. 7 –Variation of indoor CO2 concentration and barometric pressure at Mosteiros house between April and December 2006.


95% of the indoor CO2 valueswere above 0.5 vol.%, which is thePermissibleExposure Limit (PEL) for a 8hworkingenvironmentperiod defined by the National Institute of Occupational Safetyand Health (NIOSH), the Occupational Safety and HealthAdministration (OSHA) in the United States (NIOSH/OSHA,1981) and the Commission of the European Communities(1991, 2006). The minimum value registered was 0.2 vol.%, sixtimes higher than the average atmospheric air CO2 concentra-tion (Table 1). The maximum value obtained, 20.8 vol.%, isabout forty times higher than to the PEL.

Fig. 5 shows that the higher CO2 concentration values arenegatively correlated with the barometric pressure and occursome hours after heavy rainfall periods. The influence of theother monitored variables on indoor gas concentration wasnot possible to establish.

Comparing the indoor CO2 concentration and the soil CO2

efflux values at GFUR1 (Fig. 6) it is possible to observe thatincreases in the soil gas flux values foresee for some hours(from 5 to 10 h) the indoor CO2 concentration increases.

Recently, a similar test was performed in a house locatedon Mosteiros village on the Sete Cidades volcano flank, in thewestern part of S. Miguel Island. The infrared detector wasinstalled in the 1st floor of a house located in an area wheresoil CO2 concentration values range between 5 and 10 vol.%.Data were acquired on an hourly basis between 1st Apriland 31st December 2006. The indoor CO2 concentrationranged between 0 and 17.4 vol.% with an average value of0.6 vol.%. It was also possible to observe significant increaseson the indoor CO2 concentration, which are coincident withdecreases in the barometric pressure (Fig. 7).

5. Discussion and conclusions

Stepwise multiple regression analysis seems appropriate tounderstand the relations between the different monitoredvariables. Based on this statistical analysis, it is also possibleto remove the meteorological influences from the gas flux.

The most influent meteorological variables on the gasflux oscillations at themonitoring sites at Furnas volcano arethe barometric pressure and the rainfall. The inverse cor-relation between barometric pressure and soil CO2 efflux hasalready been observed in other volcanic systems (e.g.Chiodini et al., 1998; Rogie et al., 2001; Granieri et al., 2003)and can be explained by the upward migration of gases fromdeeper layers due to a decrease in the barometric pressureand a driven back of the gas into the ground during baro-metric pressure increases, the so-called barometric pumpingeffect (Auer et al., 1996; Martinez and Nilson, 1999; Neeper,2001). Rainfall influence is split into two intervals positivelycorrelated with flux, but for rainfall amounts higher than8 mm h−1, the positive relationship is more significant. Thispositive effect on the gas flux may be partially related with acovering effect of these stations that allows the gas to escapethrough the dry area. During heavy rainfall periods, the soilaround the CO2 efflux station gets completely saturatedfilling all the pores with water and obstructing gas release tothe surface; thus the accumulation chamber site remainsdryer for a longer time favouring the gas flux below it. All thegas filling in the voids in the surrounding area is conveyed inthis place, causing the observed positive spike-like anoma-lies that are more significant for values higher than 8 mmh−1. The same phenomenon seems to occur in a broaderscale in the zone located below the dwellings, since theindoor CO2 concentration increase is observed some hoursafter rainfall episodes. This positive correlation between thesoil CO2 efflux and the rainfall is not observed in all thepermanent stations installed at S. Miguel Island (Viveiroset al., in press). Despite the covering effect of the stations,topographic effects and drainage area may also play animportant role. Soil temperature influence only appearswhen barometrtic pressure and rainfall are not important.This positive correlation may be potentially explained bythe drying of the superficial soil layers causing slightincreases in permeability and consequently in the soil CO2

efflux.

http://doi:1016/j.jvolgeores.2008.07.005

http://doi:1016/j.jvolgeores.2008.07.005


During extreme meteorological conditions, characterized bylow barometric pressure and intense rainfall, the indoor CO2

concentration at the Furnas house reaches the highest values.Despite the similar behaviour showed by soil CO2 efflux andindoor CO2 concentrations, the sharp increases do not occursimultaneouslywith indoorCO2 increasebeingdelayed for somehours (from 5 to 10 h) in relation to the soil CO2 efflux increase.This delay time response may be due to the presence of thevented space that exists below the house allowing the aerationand dilution of the gas prior to its ingress into the house.

The described indoor gas behaviour shows that the gashazard increases during extreme weather conditions. Inaddition, the ventilation of the houses is highly reducedduring these periods, once the house occupants tend to closedoors and windows.

Some spike-like changes detected by previous works(Baxter et al., 1999) in dwellings at Furnas village wereattributed to non-registered microseismicity, however con-sidering rain data from that time, those increases can beexplained by extreme meteorological conditions.

Several works (e.g. King, 1993; Salazar et al., 2002; Beaubienet al., 2003) related gas changes in degassing areas withseismic activity however, this work shows that increases inthe gas flux and concentration inside dwellings can occursolely induced by meteorological changes. During the periodunder analysis the seismo-volcanic activity both at Furnas andSete Cidades volcanoes was quite low, so it is considered thatthe meteorological influences were responsible for theobserved gas flux and concentrations oscillations.

The lethal CO2 concentration values obtained during themonitoring test indicate that, even though currently volcan-ism in the Azores is quiescent, the population is at permanentrisk from exposure to CO2. In addition, it is necessary toconsider that the sharp gas increases occur in a few hours andwhen it happens during the night the risk is even greatest. Thelethal values obtained in themonitoring tests atMosteiros andFurnas villages were registered on houses located in low andmiddle risk zones, respectively (according to the Baubron et al.,1994 classification) and not in the highest CO2 soil concentra-tion anomalous areas. Furthermore, the raisedwooden floor inthe Furnas village case should supposedly diminish thehouse's vulnerability to the volcanic gases, but it seems thatit only retards the gas increase in thehouse. This points to evenhigher indoor CO2 values at other more vulnerable housesplaced in higher CO2 soil concentration areas. Future worksshould extend these monitoring tests to other houses atFurnas village, in different anomalous degassing areas, inorder to choose the permanent station to be used as an early-warning system to hazardous situations.

In this work the gas studied was the CO2, one of the mostimportant in volcanic regions, however, there are other gasesreleased diffusely by the soil, including 222Rn and H2S that canalso be influenced by the same meteorological factors as CO2

with all the consequences for the public health that will imply.The H2S is other gas released in geothermal areas andconcentrations higher than 700 ppm may be lethal (Beaubienet al., 2003; Durand and Scott, 2005). No detailed H2S mappingis available for Furnas village; nevertheless some occasionalmeasurements were done along the main soil diffuse degas-sing area and was only detected near the fumaroles. Further

studies need to be carried out to understand the hazardassociated with this particular gas.

The obtained results and the experience fromother cases intheAzoresArchipelago, confirm that evenareaswithnovisiblegasmanifestations show lethal CO2 concentrations, leading toan ideal situation that should be to forbid building in certaindegassing areas. Soil diffuse degassing anomaly maps shouldconstitute a fundamental tool used by the land-useplanners inorder to select the proper areas to build.

Unfortunately there are already several populated areasbuilt over degassing areas and, even if we consider that themost advisable scenario should be removing the populationfrom these dangerous areas, we know that these drasticresolutions are unrealistic to apply. For this reason, at leastseveral mitigation measures should be taken into considera-tion and some construction techniques should be introduced.Not allowing the construction of basements or cellars,installation of natural ventilation systems and the use of animpermeable membrane are some of the possible rules thatshould be legislated for buildings located in degassing areas,similarly to the earthquake-resistant construction thatalready exists in active seismic areas.

Even though the house under study has a raised woodenfloor that could be considered as a preventive mitigationmeasure there is still gas ingress into it. It seems that thesystem may retard the entry of the gas into the house but itwill not solve the problem. As a cautionary note, efforts weremade in Rotorua, New Zealand, to mitigate gas flux intohouses, but the preventive measures failed after some years(Durand and Scott, 2005). So, we must insist that the referredmitigation measures may delay the problem, but do not solveit completely and that the best solution should be not allowingto build in degassing areas.

This work emphasizes the existing hidden permanent riskthat should be taken into account for public health riskassessment purposes in all degassing areas, since significantgas flux increases may occur even in quiescent periods ofvolcanic activity, due only to meteorological factors.

Acknowledgements

The authors would like to thank to the owners of the monitoredhouses for allowing to install the CO2 detectors that enabled thisstudy. We wish to acknowledge Dr. Catarina Goulart and Dr.JamesDay for reviewing the English. FátimaViveiros is supportedby a PhD grant from Fundo Regional da Ciência e Tecnologia,RegiãoAutónomadosAçores (M3.1.2/F/017/2007). Theauthorsarethankful toDr.GiovanniChiodini andananonymous reviewer fortheir commentsandsuggestions that improved thequality of thispaper.

R E F E R E N C E S

Allard P, Dajlevic D, Delarue C. Origin of carbon dioxide emanationfrom the 1979 Dieng eruption, Indonesia: implications for theorigin of the 1986 Nyos catastrophe. J Volcanol Geotherm Res1989;39:195–206.


Allard P, Carbonnelle J, Dajlevic D, Le Bronec J, Morel P, Robe MC,et al. Eruptive and diffuse emissions of CO2 from Mount Etna.Nature 1991;351:387–91.

Asher-Bolinder S, Owen DE, Schumann RR. A preliminaryevaluation of environmental factors influencing day-to-dayand seasonal soil–gas radon concentrations. Field Studies ofRadon in Rocks, Soils and Water, 1971. US Geol Surv Bull; 1991.p. 23–31.

Auer LH, Rosenberg ND, Birdsell KH, Whitney EM. The effects ofbarometric pumping oncontaminant transport. J ContamHydrol1996;24:145–66.

Barberi F, Chelini W, Marinelli G, Martini M. The gas cloud of LakeNyos (Cameroon, 1986): results of the Italian technical mission.J Volcanol Geotherm Res 1989;39:125–34.

Barberi F, Carapezza ML, Ranaldi M, Tarchini L. Gas blowout fromshallow boreholes at Fiumicino (Rome): induced hazard andevidence of deep CO2 degassing on the Tyrrhenian margin ofCentral Italy. J Volcanol Geotherm Res 2007;165:17–31.

Baubron JC, Allard P, Sabroux J, Tedesco D, Toutain JP. Soil gasemanations as precursory indicators of volcanic eruptions.J Geol Soc London 1991;148:571–6.

Baubron JC, Baxter P, Coutinho R, Allard P, Ferreira T, Gaspar JL.Gas hazard in dwellings (confined spaces or spaces belowground). City of Furnas (S. Miguel, Açores). 1/2000 Map. BRGM,France. 1994.

Baxter P, Kapila M. Acute health impact of the gas release at LakeNyos, Cameroon, 1986. J Volcanol GeothermRes 1989;39:265–75.

Baxter P, Baubron JC, Coutinho R. Health hazards and disasterpotential of ground gas emissions at Furnas Volcano, SãoMiguel, Azores. J Volcanol Geotherm Res 1999;92:95–106.

Beaubien SE, Ciotoli G, Lombardi S. Carbon dioxide and radon gashazard in the Alban Hills area (central Italy). J VolcanolGeotherm Res 2003;123:63–80.

Blong RJ. Volcanic hazards. A sourcebook on the effects oferuptions. Academic Press; 1984. 424pp.

Carapezza ML, Badalamenti B, Cavarra L, Scalzo A. Gas hazardassessment in a densely inhabited area of Colli Albani Volcano(Cava dei Selci, Roma). J Volcanol Geotherm Res 2003;123:81–94.

Census. Instituto Nacional de Estatística, Recenseamento Geral daPopulação e Habitação. Portugal; 2001 (in Portuguese).

Chiodini G, Cioni R, Guidi M, Raco B, Marini L. Soil CO2 fluxmeasurements in volcanic and geothermal areas. ApplGeochem 1998;13:543–52.

Cole P, Queiroz G, Wallenstein N, Gaspar JL, Duncan AM, Guest J.An historic subplinian/phreatomagmatic eruption: the 1630 ADeruption of Furnas Volcano, São Miguel, Azores. J VolcanolGeotherm Res 1995;9:17–135.

Commission of the European Communities. CommissionDirective 91/322/EEC of 29 May 1991 on establishing indicativelimit values by implementing Council Directive 80/1107/EECon the protection of workers from the risks related to exposureto chemical, physical and biological agents at work. Off JEuropean Union L 1991;177:22–4.

Commission of the European Communities. Commission Directive2006/15/EC of 7 February 2006 establishing a second list ofindicativeoccupational exposure limit values in implementationof Council Directive 98/24/EC and amending Directives 91/322/EEC and 2000/39/EC. Off J European Union L 2006;38:36–9.

Costa A, Chiodini G, Granieri D, Folch A, Hankin RKS, Caliro S, et al.A shallow layer model for heavy gas dispersion from naturalsources: application and hazard assessment at Caldara diManziana, Italy. Geochem Geophys Geosys 2008;9(Q03002).doi:10.1029/2007GC001762_________________________.

Denman AR, Crockett RGM, Groves-Kirkby CJ, Phillips PS, GillmoreGK, Woolridge AC. The value of seasonal correction factors inassessing the health risk from domestic radon—a case study inNorthamptonshire, UK. Environ Int 2007;33:34–44.

Diliberto IS, Gurrieri S, Valenza M. Relationships between diffuseCO2 emissions and volcanic activity on the island of Volcano

(Aeolian Islands, Italy) during the period 1984–1994. BullVolcanol 2002;64:219–28.

Draper N, Smith H. Applied regression analysis. 2nd ed. JohnWiley& Sons, Inc.; 1981. 709pp.

Durand M, Scott BJ. Geothermal ground gas emissions and indoorair pollution in Rotorua, New Zealand. Sci Total Environ2005;345:69–80.

Ferreira T, Oskarsson N. Chemistry and isotopic composition offumarole discharges of Furnas caldera. J Volcanol GeothermRes 1999;92:169–79.

Ferreira T, Gaspar JL, Viveiros F, Marcos M, Faria C, Sousa F.Monitoring of fumarole discharge and CO2 soil degassing in theAzores: contribution to volcanic surveillance and public healthrisk assessment. Ann Geophys 2005;48N(4/5):787–96.

Freund RJ, WilsonWJ. Regression analysis. Statistical modelling ofa response variable. USA: Academic Press; 1998. 444pp.

Frutuoso G. Livro Quarto das Saudades da Terra. In: Saudades daTerra VII, editor. Ponta Delgada, Instituto Cultural de PontaDelgada 2nd ed. 1522–1591; 1522–1591. (In Portuguese).

Gaspar JL, Ferreira T, Queiroz G, Pacheco JM. CO2 soil degassingand risk assessment in the inhabited area of Ribeira Seca(S. Miguel Island, Azores). EC Advanced Study Course 1998.Volcanic Hazard Assessment, Monitoring and Risk Mitigation.European Commission, Environment and Climate Programme(abstract); 1998.

Giggenbach WF, Sano Y, Schmincke HU. CO2 — rich gases fromLakes Nyos and Monoun, Cameroon; Laacher See, Germany;Dieng, Indonesia, and Mt. Gambier, Australia— variations on acommon theme. J Volcanol Geotherm Res 1991;45:311–23.

Granieri D, Chiodini G, Marzocchi W, Avino R. Continuousmonitoring of CO2 soil diffuse degassing at Phlegraean Fields(Italy): influence of environmental and volcanic parameters.Earth Planet Sci Lett 2003;212:167–79.

Guest J, Gaspar JL, Cole P, Queiroz G, Duncan AM, Wallenstein N,et al. Volcanic geology of Furnas Volcano, São Miguel, Azores.J Volcanol Geotherm Res 1999;92:1–29.

Hansell A, Oppenheimer C. Health hazards from volcanic gases:a systematic literature review. Arch Environ Health 2004;59(No.12):628–39.

Hansell A, Horwell C, Oppenheimer C. The health hazards ofvolcanoes and geothermal areas. Occup Environ Med2006;63:149–56.

Hinkle ME. Seasonal and geothermal production variations inconcentrations of He and CO2 in soil gases, Roosevelt HotSprings Known Geothermal Resource Area, Utah, U.S.A. ApplGeochem 1991;6:35–47.

Hinkle ME. Environmental conditions affecting concentrations ofHe, CO2, O2 and N2 in soil gases. Appl Geochem 1994;9:53–63.

King CY. Gas-geochemical approaches to earthquake prediction.Isotopic and geochemical precursors of earthquakes andvolcanic eruptions. Proceedings of an Advisory Group Meetingheld in Vienna, 9th to 12th September. IAEA; 1993. p. 22–37.

Klusman RW, Webster JD. Preliminary analysis of meteorologicaland seasonal influences on crustal gas emission relevant toearth prediction. Bull Seismol Soc Am 1981;71(1):211–22.

Le Guern F, Tazieff H, Faivre Pierret R. An example of healthhazard: people killed by gas during a phreatic eruption: DiengPlateau (Java, Indonesia), February 20th 1979. Bull Volcanol1982;45(2):153–6.

Marley F. Investigation of the influences of atmospheric conditionson the variability of radon and radon progeny in buildings.Atmos Environ 2001;35:5347–60.

Marques R, Zêzere J, Trigo R, Gaspar J, Trigo I. Rainfall patterns andcritical values associated with landslides in Povoação County(SãoMiguel Island,Azores): relationshipswith theNorthAtlanticOscillation. Hydrol Process 2007;22:478–94. doi:10.1002/hyp.6879____________________.

Martinez MJ, Nilson RH. Estimates of barometric pumping ofmoisture through unsaturated fractured rock. Transp PorousMedia 1999;36:85–119.

http://dx.doi.org/10.1029/2007GC001762

http://dx.doi.org/10.1002/hyp.6879


Neeper DA. A model of oscillatory transport in granular soils, withapplication to barometric pumping and earth tides. J ContamHydrol 2001;48:237–52.

NIOSH/OSHA. Occupational Health Guidelines for ChemicalHazards, DHHS (NIOSH) Publication No. 81-123. U.S.Government Printing Office; 1981.

Oskarsson N. Carbon dioxide bursts of Lake Nyos, Cameroon,modelled as periodic supersaturation in a countercurrentreactor. J Volcanol Geotherm Res 1990;42:307–18.

Oskarsson N, Pálsson K, Ólafsson H, Ferreira T. Experimentalmonitoring of carbon dioxide by low power IR-sensors: soildegassing in the Furnas Volcanic Centre, Azores. J VolcanolGeotherm Res 1999;92:181–93.

Pizzino L, Galli G, Mancini C, Quattrocchi F, Scarlato P. Natural gashazard (CO2, 222Rn) within a quiescent volcanic region and itsrelations with tectonics: the case of Ciampino–Marino Area,Alban Hills Volcano, Italy. Nat Hazards 2002;27:257–87.

Pomonis A, Spence R, Baxter P. Risk assessment of residentialbuildings for an eruption of Furnas Volcano, São Miguel, theAzores. J Volcanol Geotherm Res 1999;92:107–31.

Queiroz G, Gaspar JL, Cole P, Guest J, Wallenstein N, Duncan AM,et al. Erupções vulcânicas no Vale das Furnas (ilha de S.Miguel, Açores) na primeira metade do século XV. Açoreana1995;8(1):159–65 (In Portuguese).

Rogie JD, Kerrick DM, Chiodini G, Frondini F. Fluxmeasurements ofnonvolcanic CO2 emission from some vents in central Italy.J Geophys Res 2000;105B4:8435–45.

Rogie JD, Kerrick DM, Sorey ML, Chiodini G, Galloway DL.Dynamics of carbon dioxide emission at Mammoth Mountain,California. Earth Planet Sci Lett 2001;188:535–41.

Rowe JE, Kelly M, Price LE. Weather system scale variation inradon-222 concentration of indoor air. Sci Total Environ2002;284:157–66.

Salazar JML, Pérez NM, Hernández PA, Soriano T, Barahona F,Olmos R Cartagena R, et al. Precursory diffuse carbon dioxide

degassing signature related to a 5.1magnitude earthquake in ElSalvador, Central America. Earth Planet Sci Lett 2002;205:81–9.

Searle R. Tectonic pattern of the Azores spreading centre andtriple junction. Earth Planet Sci Lett 1980;51:415–34.

Silva C. Estudo da desgaseificação difusa no vulcão das Furnas(Ilha de S. Miguel): o caso do 222Rn. Master Thesis, University ofAzores; 2006. 167p. (In Portuguese).

Silveira D, Gaspar JL, Ferreira T, Queiroz G. Reassessment ofthe historical seismic activity with major impact in S.Miguel Island (Azores). Nat Hazards Earth Syst Sci2003;3:615–23.

Sousa F. Cartografia de anomalias geoquímicas associadas aprocessos de desgaseificação difusa no Vulcão das Furnas(S. Miguel, Açores): contribuição para a avaliação de riscosvulcânicos. Master Thesis, University of Azores; 2003. 106p.(In Portuguese).

Velleman PF. Definition and comparison of robust nonlinear datasmoothing algorithms. J Am Stat Assoc 1980;75:609–15.

Viveiros F, FerreiraT,CabralVieira J, SilvaC,Gaspar JL. Environmentalinfluences on soil CO2 degassing at Furnas and Fogo volcanoes(São Miguel Island, Azores archipelago). J Volcanol GeothermRes in press. doi:1016/j.jvolgeores.2008.07.005_______________________________.

Weinstein P, Cook A. Volcanic emissions and health. In: Selinus,editor. Essentials of Medical Geology — Impacts of the NaturalEnvironment on Public Health. Elsevier; 2005. p. 203–26.

Weston FS. List of recorded volcanic eruptions in the Azores withbrief reports. Bol Mus Lab Mineral Geol Fac CieÃnc Univ Lisb1964;10(1):3–18.

Williams-Jones G, Rymer H. Hazards of volcanic gases. In:Sigurdsson H, editor. Encyclopedia of Volcanoes. AmericanPress; 2000. p. 997–1004.

Wong KL. Carbon dioxide. Spacecraft Maximum AllowableConcentrations for Selected Contaminants, vol. 2. Natl AcadSci, U S A; 1996. p. 105–87.

meteorological factors controlling soil gases and indoor ... · meteorological factors controlling...

Documents