Fourier transform spectroscopy of the gas plume of Masaya volcano, Nicaragua

We have now obtained field observations of volcanic gases with a portable fourier transform infrared (FTIR) spectrometer at a number of volcanoes since our first measurements at Mount Etna in 1994. In addition to Etna, prior measurements with this instrumentation were made at Vulcano, Stromboli and Soufriere Hills volcano. There were several reasons for our interest in working at Masaya volcano. Its intermittent lava lake activity, strong degassing (up to 50 kg s-1 of sulfur dioxide, comparable to Mount Etna), and ease of access (the volcano is enclosed in a national park and the active summit crater can be reached and partly circumnavigated by road) make it an obvious laboratory volcano. In addition, Masaya is one of the few known volcanoes responsible for basaltic plinian activity (20 and 6.5 ka BP), and its subdued topography (about 600 m above sea-level) and persistently strong degassing (Figure 1) result in considerable damage downwind both to vegetation and human health (raised incidence of respiratory disease). The volcano is only 35 km from Managua. Information concerning the chemistry and dynamics of the gas phase in Masaya magma is therefore crucial in understanding the control of degassing on eruptive style, and assessment of the environmental impacts of the volcanic gases.

Figure 1. View of Santiago crater, Masaya.

Figure 2. Cartoon of field configurations for FTS:
(a) active measurements, (b) solar occultation.

1. Outline of field campaign and project management

Four members of our Fourier-transform spectroscopy (FTS) group were involved in the Masaya fieldwork: Clive Oppenheimer (Geography, Cambridge), Peter Francis (Earth Sciences, Open), Mike Burton (Geography, Cambridge - EC postdoc), and Lisa Boardman (Earth Sciences, Open - NERC PhD student). In addition, Matthew Watson (Geography, Cambridge - NERC PhD student) came out for 1 week with the NERC Equipment Pool for Spectroscopy’s (EPFS) new CIMEL sunphotometer, in order to measure optical depths of Masaya’s plume. We timed the field season so as to collaborate with colleagues (John Stix, Pierre Delmelle, Katie St. Amand) from the University of Montreal and another Open University student (Glyn Williams) who took two ultraviolet sensing correlation spectrometer (Cospec) instruments to Masaya (see section 2.2). We worked closely with the Istituto Nicaraguense de Estudios Territoriales (INETER) which provided logistical support including vehicles and drivers. Except for some difficulties importing the equipment into Nicaragua (which were quickly resolved through intervention of the British Embassy in Managua) the field operation ran very smoothly.

The first FTS data were obtained on 21 February 1998, and observations were sustained through to 25 March 1998. Lisa Boardman received full training on both operation of the FTIR spectrometer and retrieval procedures, and part of the dataset will be analysed in a chapter of her PhD dissertation. Spectral data were collected in several open-path configurations (Figure 2): active measurements were obtained by aligning spectrometer and silicon carbide lamp across the summit crater; passive observations utilised the Sun, or hot intracrater vent as infrared source. Solar occultation spectra of the plume were collected at three locations: at the summit, 15 km distant, 30 km distant. Of the order of 15 000 individual spectra were obtained representing by far the largest FTIR dataset we have yet obtained. In advance of the fieldtrip, considerable effort went into design of retrieval algorithms. At the time of writing, most of the active data have been processed but work is still required to complete analysis of the solar occultation data. We aim to complete processing of all data and have all results written up for publication within the next 3 months. This "final" report is therefore not the last word on the project.

2. Progress towards achieving original aims

The following sections link directly (though not in the same order) to the five points highlighted in the ‘Aims’ section of the original proposal.

2.1. Real-time retrievals

In an ideal arrangement, it would not be necessary to save single-beam spectra in the field for subsequent processing but to automate the retrieval process to run alongside data collection. This would not only save considerable labour but also cut down on disk space requirements (each single-beam spectrum requires of the order of 70-90 kb). We are not far off being able to achieve this in theoretical terms but would need a more powerful pc if it were going to be practical for field use. Therefore, at Masaya, we compromised by simplifying the retrieval process by analysing the ratios of each sigle-beam spectrum to a "reference" spectrum. This reference was chosen for each set of measurements as the spectrum in that set containing the least amount of volcanic gas. The ratio eliminates to a large degree spectral lines for the interfering background (non-volcanic) atmosphere. This procedure is only suitable for analysing the active FTS data, however, since the solar measurements are affected by changing air mass factor through time, and the generally high amounts of gas in all spectra.

The retrieval of volcanic gas composition from spectra is an inverse problem. In order to determine the state of the atmosphere (x) given a measurement (y) we generate a forward model that simulates the measured ratio spectrum as accurately as possible. If x is a vector containing the parameters to be fitted in, e.g. volume-mixing-ratio, frequency shift etc. and y is the measurement ratio spectrum, then F(x) is the simulated spectrum, and the aim of the retrieval is to determine the values of x when F(x)=y. The gas pressure and temperature (which were recorded simultaneously in the field) were used to generate an optical depth spectrum for the volcanic gases, SO2, HCl or HF. The forward model was calculated by scaling the optical depth to a requested volume-mixing-ratio for a given pathlength, taking the negative exponential to calculate transmittance, before convolving with the theoretical instrument line shape defined by the resolution, apodization and field of view of the instrument. A retrieval window was then selected for each of these gases and the parameters in x were adjusted until the optimum fit to y was achieved. This determined the quantity of each of the above three gases in each spectrum. The code, written in IDL, was used to analyse all active data at the end of each day. Given the relatively slow speed of our field computer this was an acceptable compromise.

We have also worked on an alternative retrieval scheme which does permit analysis of single-beam spectra. The main difference is that more than one gas is fitted simultaneously, and this allows solar spectra to be analysed. To retrieve the quantity of volcanic gas in a solar spectrum we simulate the background interfering gases and the 100% transmittance level to determine the pure volcanic gas spectrum, whereas in the active data we generally have access to a high quality background spectrum, more or less free of volcanic gas. Preliminary analyses are encouraging with single-beam retrieved ratios for HCl/SO2, etc., identical (within errors) to those obtained from ratio spectra. This methodology has also yielded our first FTS measurements of CO2/SO2, something we alluded to in the original proposal. This has provided the first estimate of the CO2 flux from Masaya volcano (see next section). Reliable estimates of CO2 degassing rates at individual volcanoes are of considerable interest because of the wider implications of volcanic degassing for the global carbon cycle.

The full complement of gases we have retrieved so far includes HF, HCl, SO2, and CO2. HBr, OCS, CO, H2S, and SiF4, are at or below detection limits. Given that we collected spectra through very dense fumes above the crater at times, the amounts of these other gases must be very low. For example, the HCl/HBr volume ratio must be at least 250 for us not to detect HBr spectral lines. We have recently purchased a gas calibration cell and tests are underway to determine the accuracy of our retrievals with respect to traceable standards.

2.2. Fluxes of gases

A significant benefit of coordinating fieldwork with Stix’ team was parallel use of FTS with the Cospec which yields estimates of emission rates of volcanic SO2. At Masaya the Cospec was operated from a vehicle travelling along the Pan American Highway 15 km downwind from the summit and approximately perpendicular to the plume direction. By integrating the vertical column amount of SO2 along the transect, and multiplying this value by wind speed, the emission rate of SO2 may be estimated. Multiple scattering effects in the plume can lead to overestimation of the SO2 burden but upwind cloud processing can deplete the plume of SO2 before measurement (Oppenheimer et al., 1998). The Montreal group is invetigating these processes but for the timebeing we take the Cospec results at face value in order to derive fluxes of other gas species, M, by multiplication of SO2 emission rate by gas mass ratios (M/SO2) determined by FTS. Table 1 shows the HCl emission rates calculated on a daily basis. The SO2 emission was found to be quite variable over the month of observations, dropping to 7.8 kg s-1 and reaching as high as 65 kg s-1. Taking the mean SO2 emission rate of 24 kg s-1 and multiplying by the mean HCl/SO2 mass ratio (0.28) suggests an average HCl emission rate of around 6.7 kg s-1. Table 2 indicates emission rates obtained for other gases based on the SO2 emission rate, and a comparison between Masaya’s degassing and a number of other basaltic volcanoes.

Figure 3. CO2 vs. SO2, 23 Feb.

Figure 4. HCl vs. SO2, 22 Feb-9 Mar.

Our estimate for Masaya’s emission of CO2 is (39 kg s-1), is based on the observed CO2/SO2 volume ratio of 2.36 (Figure 3). We are confident about the CO2 retrieval because the ordinate on the CO2 axis (Figure 3), when divided by the path length across the crater (512 m) yields a value of 369 ppm, consistent with the non-volcanic background level of CO2. The CO2 flux, which excludes any contribution from diffuse flank degassing, is high compared with some other basaltic volcanoes (Table 2). While Masaya’s emission rate of CO2 appears a factor of 10 less than that reported for Mount Etna (Allard et al., 1991), this latter value is considerably at odds with the long-term SO2 output (55 kg s-1) and directly measured CO2/SO2 ratio for Etna (around 1). Masaya is also a signficant emitter of halogens (Table 2).

Table 1. Measured SO2/HCl ratios, Cospec observations of SO2 emission rate, and corresponding HCl emission rates

Date

SO2/HCl volume ratio

standard deviation (1 s)

HCl/SO2 mass ratio

SO2 emission rate (kg s-1)

HCl emission rate (kg s-1)

22-Feb-98

2.29

0.18

0.25

21.7

5.4

23-Feb-98

1.82

0.19

0.31

41.9

13.0

24-Feb-98

2.22

0.20

0.26

17.6

4.6

25-Feb-98

2.57

0.54

0.22

12.8

2.8

27-Feb-98

2.07

0.33

0.27

nd

 

28-Feb-98

1.88

0.59

0.30

nd

 

1-Mar-98

1.87

0.25

0.30

64.5

19.4

2-Mar-98

1.94

0.14

0.29

22.7

6.6

3-Mar-98

1.96

0.34

0.29

7.8

2.3

5-Mar-98

1.95

0.22

0.29

nd

 

6-Mar-98

1.92

0.16

0.30

nd

 

9-Mar-98

1.81

0.30

0.31

nd

 

Table 2. Comparative emission rates of carbon, sulfur and halogens for Masaya and some other basaltic volcanoes

Volcano

CO2

(kg s-1)

SO2

(kg s-1)

HCl

(kg s-1)

HF

(kg s-1)

reference

Masaya, 1998

39

24

6.7

0.65

this work

Masaya, 1979-1985

 

14.7

9.6

0.18

Stoiber et al., 1986

Kilauea, 1995

2.8-3.5

1.6

0.05

0.03

Gerlach & Graeber, 1985; Gerlach et al., 1998

Mt. Etna, 1975-1997

400

66

8.2

2.2

Allard et al., 1991; Francis et al., 1998

Stromboli, 1997

 

9.25

6.2

 

Allard et al., 1994; our FTIR data

Erta ‘Ale, 1971-4

0.62

0.49

0.046

0.03

Le Guern et al., 1979

Poas, 1982

2.3-8.9

8.8

0.46

0.040

Casadevall et al., 1984

Erebus, 1986-91

 

2.4-8.2

2.2-4.2

1.3-1.9

Zreda-Gostynska et al., 1993

global volcanic

2500

590

13-350

1.9-190

Stoiber et al., 1987; Symonds et al., 1988; Gerlach, 1991

 

2.3. Temporal variability

An important aim was to examine the temporal variability of gas ratios. Surveillance of ratios, especially of redox pairs of gases (e.g., CO/CO2, SO2/H2S) can illuminate subsurface magmatic and hydrothermal processes, thereby supporting efforts to predict eruptions. Temporal evolution in the ratios of CO2/SO2, HCl/SO2 and HCl/HF in the gas phase has been used to infer changes in magmatic systems feeding volcanoes. Up to now, we have focused on variations in HCl/SO2 for the Masaya dataset but will be examining variation in other species (HCl/HF, CO2/SO2) soon. Figures 4 and 5 show all the retrievals for HCl/SO2 obtained from active sensing FTIR observation over a 2 week period. These indicate a small but significant fluctuation with a period of a few days in the first week and more stable ratios (slightly above 0.5 for molecular HCl/SO2) in the second week. Over much shorter timescales of a few minutes, we occasionally observe smaller amplitude fluctuations in HCl/SO2 ratios (Figure 6).

Figure 5. SO2/HCl ratio. Feb 22-Mar 9.

Figure 6. SO2/HCl ratio. Feb 23.

One explanation for HCl/SO2 variations in gas emissions is that HCl/SO2 increases with the degree of degassing from a magma body due to the greater solubility of HCl over SO2. In this case, HCl/SO2 variation is controlled by evolution of sulfur degassing, such that a stagnant batch of magma should degas increasingly Cl-rich fluids. This model also implies that while the absolute fluxes of SO2 should decline as the magma degasses, the HCl flux should remain more or less constant. At Etna, however, a three-fold increase in HCl emission rate has been observed following eruptions, indicating a substantial increase in Cl exsolution (Pennisi and Le Cloarec, 1998). The pressure-dependence of Cl solubility in basalt is not well-constrained but experimentation with more silicic melts suggests that Cl is more soluble at lower pressures. Pennisi and Le Cloarec (1998) proposed an alternative interpretation of variations in HCl/SO2 at Etna in which increases represent arrival of less-degassed magma batches to the near surface, while decreases result from preferential S-degassing of shallow erupting magma. By analogy, the longer period, larger amplitude events at Masaya may result from fresher, less degassed batches of magma ascending into a magma pond close to the surface. It would be interesting to see if there is any pattern to the increase and decay of HCl/SO2 (and other) ratios, for example, steeper increases in HCl/SO2 than decreases. We will be giving further attention to interpretations for these variations.

The short period variations are not always observed and could relate to timing of gas separation from magma at the surface of the magma pond (e.g., bubble bursts), and plume chemistry. If gas release to the atmosphere is episodic (timescales of minutes) then it is possible that small fluctuations could arise in observed HCl/SO2 ratios at the crater rim due to the greater solubility of HCl in aqueous aerosol. As a given batch of gas stagnates in the crater it might be expected to show decreasing HCl/SO2 as the HCl dissolves in airborne water droplets formed from condensation of volcanic steam. Passage of a new gas bubble across the magma-air interface would be seen shortly afterwards in rising HCl/SO2.

FTS offers the possibility of high temporal resolution monitoring of gas ratios - this simply has not been practical hitherto. It may be possible to deduce aspects of the replenishment, residence time and sizes of magma batches feeding the upper conduit system through interpretation of variations in a range of gas ratios.


Figure 7. Solar spectra collected at Santiago rim.

 
2.4. Solar occultation
We obtained spectral data by solar occultation on many days during the field campaign (Figure 7). At the time of writing, only one day of solar spectra has been analysed due to ongoing work with the retrieval algorithm. The preliminary results, however, are encouraging since the retrieved ratios for HCl/SO2 and HCl/HF are consistent with those obtained for the active data.

 

Figure 8. Retrievals for solar spectra at Masaya
and Etna, showing (a) HCl vs. SO2, and (b) HCl vs. HF.

Figure 8 shows retrieved amounts of these three gases compared
with our solar occultation data for Mount Etna. With respect to SO2,
Masaya is more halogen rich than Etna.

2.5. Plume chemistry

Solar spectra were obtained at three distances from the vent with the aim of comparing retrieved gas ratios (principally for HCl, SO2, and HF) effectively as a function of plume age. Broadly, we might expect to see HCl/SO2 drop with distance from the vent due to the greater solubility of HCl in aqueous aerosol. Additional experiments with the two Cospec instruments operated simultaneously at different distances downwind have revealed an apparent loss of SO2 over these distances, too, so any relative losses for HCl and HF could be corrected to yield total depletion rates. Knowledge of the lifetime of these different gas species will contribute to modelling both the environmental impacts of volcanic plumes, and their chemical and radiative effects in the troposphere. Completion of this aspect of the investigation is a priority but requires further testing of the single-beam retrieval methodology for the more dilute plumes at greater distances from source.

3. Other experiments

3.1. Sunphotometry

The NERC EPFS 8-channel sunphotometer was used at the summit of Masaya during the final week of the field campaign. Observations were obtained both through and outside the plume in order to characterise the background optical thickness of the atmosphere and thereby estimate the plume spectral aerosol optical thickness (AOT). The wavelength-dependence of the AOT can be used to model the aerosol size distribution. NERC student Matthew Watson is analysing these data as part of his PhD dissertation.

3.2. Lunar occultation

The spectrometer was deployed one night during a full moon at Masaya. Initial analysis, while showing considerably more scatter than solar occultation and active data, yields a similar volume ratio for HCl/SO2. We hoped that sufficiently accurate measurements of this ratio could be compared with daytime measurements to assess the importance of diurnal variations in plume chemistry (arising from air temperature, irradiance changes, etc) but this seems unlikely.

3.3. Hot vent spectra

On several days we collected FTIR spectra passively, using the hot vent (and source of the gas plume) on the crater floor as a source of infrared radiation. Observations of crater glow at night indicated that lava was probably present within a few tens of metres of the vent but just out of view of the crater rim. Inspection of these spectra, however, reveals a much more complicated retrieval problem because gases are seen in both emission and absorption, and the gas temperature must vary from near magmatic to ambient along the observation path. If we can establish a methodology for analysing these spectra (possibly by modelling a multilayer atmosphere along the path) observations into the crater could offer the best chances for obtaining spectra with gases such as OCS and HBr present above detection limits.

4. Dissemination of results

At present two papers are in press which arise from this proposal (both acknowledge this NERC grant and are appended):

·         Francis, P., Burton, M., Oppenheimer, C., Remote measurements of volcanic gas compositions by solar occultation spectroscopy, Nature, awaiting proofs. While this paper focuses on FTS observations at Mount Etna, it uses the retrieval algorithm developed under the Masaya proposal (which funded purchase of two IDL licences).

There is still much to be written up from this project, however, and we have a clear publication plan and division of labour amongst the members of the group:

·         Lisa Boardman: analysis, interpretation, and publication of the multitemporal changes in gas ratios observed in the active data. Also undertaking analyses of fluid inclusions in spatter erupted from Masaya in 1997. This work will also represent a part of her PhD dissertation.

5. Beneficiaries of research

We see two key benefits of developing FTS studies of volcanic gases in the field:

·         Application to geochemical surveillance of volcanoes. Our new NERC project on Soufriere Hills Volcano, Montserrat, has already revealed an order of magnitude decrease in HCl/SO2 ratios in the plume between 1996 and 1998, which may indicate an increasing meteoric (hydrothermal) signature. These results should assist in considering whether or not the eruption has ceased. Since we began our work with FTS in 1994, we have seen colleagues purchase the same spectrometer (at Cascades Volcano Observatory, and Los Alamos) for volcanological work. In addition, we know of two other volcanological groups seeking funds for purchase of an FTIR instrument. Ultimately, we hope that this will become a technique routinely employed by a number of volcano observatories in support of hazards assessment and eruption prediction. Long-term geochemical datasets compared with other geophysical parameters promise the greatest insights into magmatic degassing and hydrothermal systems. All datasets collected at Easter will be shared between the UK and Montreal groups and INETER. We hope to continue working with spectroscopic techniques for gas and aerosol analysis at Masaya.

6. Contribution to training

This project has contributed directly to the training of two NERC PhD students (Boardman and Watson) and one EC PDRA (Burton). All are being encouraged to lead authorship of at least one paper arising from the project.

7. References & other publications by our group on FTS

Allard, P. Carbonelle, J. Metrich, N. and Zettwoog, P. Eruptive and diffuse emissions of carbon dioxide from Etna volcano. Nature, 351, 38-391 (1991).

Allard, P., Carbonelle, J., Metrich, N., Loyer, H., and Zetwoog, P. Sulphur output and magma degassing budget of Stromboli volcano, Nature 368, 326-330 (1994).

Casadevall, T.J., et al. Sulfur dioxide and particles in quiescent volcanic plumes from Poas, Arenal, and Colima volcanoes, Costa Rica and Mexico. J. Geophys. Res 89, 9633-9641 (1984).

Francis P, Burton M, and Oppenheimer C, Remote measurements of volcanic gas compositions by solar FTIR spectroscopy, Nature in press (1998).

Francis P, Chaffin C, Maciejewski A, Oppenheimer C. Remote determination of SiF4 in volcanic plumes: a new tool for volcano monitoring, Geophys. Res. Lett. 23, 249-252 (1996).

Francis P, Maciejewski A, Oppenheimer C, Chaffin C, Caltabiano T. SO2:HCl ratios in the plumes from Mt. Etna and Vulcano determined by FTS, Geophys. Res. Lett. 22, 1717-1720 (1995).

Gerlach, T.M. Present-day CO2 emissions from volcanoes, EOS, Trans. Am. Geophys. Union 72, 249-.

Gerlach, T.M., and McGee, K.A. Rates of volcanic CO2 degassing from airborne determinations of SO2 emission rates and plume CO2/SO2. Geophys. Res. Lett. 25, 2675-2678 (1998).

Le Guern, F., Carbonelle, J., and Tazieff, H. Erta ‘Ale lava lake: heat and gas transfer to the atmosphere. J. Volcanol. Geotherm. Res. 6, 27-48 (1979).

Oppenheimer C, Francis P and Stix J. Depletion rates of SO2 in tropospheric volcanic plumes, Geophys. Res. Lett. 25, 2671-2674 (1998).

Oppenheimer C, Francis P, and Maciejewski A. Spectroscopic observation of HCl degassing from Soufriere Hills volcano, Montserrat, Geophys. Res. Lett., in press (1998).

Oppenheimer C, Francis P, and Maciejewski A, Volcanic gas measurements by helicopter-borne fourier transform spectroscopy, Int. J. Remote Sens., 19, 373-379 (1998)

Oppenheimer C, Francis P, Burton M, Maciejewski A, Boardman L, Remote measurement of volcanic gases by fourier transform infrared spectroscopy, Applied Physics, in press (1998).

Pennisi, M and LeCloarec, M. Variations in Cl, F and S in Mt Etna's plume (Italy) between 1992 and 1995. J. Geophys. Res. 103, 5061 - 5066 (1998).

Stoiber, R, Williams, S.N. and Huebert, B. Annual contribution of sulfur dioxide to the atmosphere by volcanoes. J. Volcanol. Geotherm. Res. 33, 1-8 (1987).

Stoiber, R.E., Williams, S and Huebert, B.J. Sulfur and halogen gases at Masaya caldera complex, Nicaragua: total flux and varations with time. J. Geophys. Res. 91, 12,215- 12,231 (1986).

Symonds, R.B., Rose, W.I., and Reed, M.H. Contribution of Cl- and F- bearing gases to the atmosphere by volcanoes. Nature, 334, 415-418 (1988).

Zreda-Gostyka, G and Kyle, P.R. and Finnegan, D.L. Chlorine, fluorine from Mt. Erebus, Antarctica, and estimated contributions to the Antarctic atmosphere. Geophys. Res. Lett. 20, 1959-1962 (1993).