- Email: [email protected]
Sulfur gas emissions from geothermal power plants in Iceland
Geothermics 29 (2000) 525±538 www.elsevier.com/locate/geothermics
Sulfur gas emissions from geothermal power plants in Iceland Hrefna KristmannsdoÂttir a,*, MagnuÂs Sigurgeirsson a, HalldoÂr AÂrmannsson a, Hreinn Hjartarson b, MagnuÂs OÂlafsson a a Orkustofnun, GrensaÂsvegur 9, 108 ReykjavõÂk, Iceland The Icelandic Meteorological Institute, BuÂsta+avegi 9, 105 ReykjavõÂk, Iceland
b
Received 27 July 1998; accepted 23 April 1999
Abstract Sulfur gas (H2S and SO2) emissions from geothermal ®elds in Iceland have been studied as part of a project, aimed at enhancing environmental research concerning eects of geothermal development. Short-term measurements of the gases have been carried out in several high-temperature geothermal ®elds in Iceland. In four exploited ®elds, baseline values for the concentration of sulfur gases have been obtained by long-term measurements. The data strongly re¯ect the dependence of gas concentrations on climatic factors, especially precipitation. Interpretation of the data by air distribution modeling, and by simple experiment, indicate minor, or at least very slow conversion of H2S to SO2 at atmospheric conditions in Iceland. 7 2000 CNR. Published by Elsevier Science Ltd. All rights reserved. Keywords: Pollution; Atmosphere; Dispersion; Weather; Conversion; Iceland
1. Introduction As a consequence of the increased emphasis on the environmental impacts of energy production, a project was started in 1991 to study the environmental * Corresponding author. Tel.: +354-569-6000; fax: +354-568-8896. E-mail address: [email protected] (H. KristmannsdoÂttir). 0375-6505/00/$20.00 7 2000 CNR. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 3 7 5 - 6 5 0 5 ( 0 0 ) 0 0 0 2 0 - 1
526
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
impact of geothermal developments in Iceland. The project was initiated by Orkustofnun (The National Energy Authority, NEA), and was conducted with the cooperation of the main producers of high-temperature geothermal energy in Iceland. The project entailed, ®rstly, an assessment of the present status at the ®ve main sites of high-temperature geothermal production in Iceland, and secondly, the de®nition of several priority projects (KristmannsdoÂttir and AÂrmannsson, 1995) to be carried out within the scope of the project. One of the main eects of geothermal exploitation on the environment is the emission of gases with geothermal steam (Axtmann, 1975; AÂrmannsson and KristmannsdoÂttir, 1992). The greenhouse gases CO2 and CH4, together with sulfur gases, are of main concern in this respect. The sulfur gas emitted from geothermal plants is in the form of H2S. This gas is toxic in high concentrations and has a very unpleasant odour even in very low concentrations; its presence is often of great environmental concern. However, people are only in danger of being poisoned inside poorly-ventilated buildings, at plant or drill sites, or in enclosed topographic lows. The smell may be more unpleasant to people not used to it, whereas people accustomed to it may not notice it. The H2S gas may, however, be oxidized to SO2, causing acidi®cation of rain and soil, which is of great concern. If all the H2S gas from geothermal power plants in Iceland were to be quantitatively converted to SO2, at least some plants would be considered unacceptably polluting and the removal of H2S from steam would be made mandatory. Measurements of pH and sulfate concentrations in precipitation in the vicinity of the Olkaria power plant in Kenya (Muna and Ojambo, 1985) and of pH in the neighbourhood of the Svartsengi power plant in Iceland (Bjarnason, 1991), reveal neither pH changes nor SO2 addition. Thus, there is evidence that probably not all the H2S is converted to SO2, at least not immediately. This has been a matter of considerable discussion, as the sulfur chemistry is complicated and little research has been done on the conversion of geothermally emitted H2S to SO2 in the atmosphere. Brown and Webster (1994) claim that oxidation of H2S within aerosols is a slow process, but Cox and Sandalls (1974) concluded that photo-oxidation of H2S to SO2 is a major loss process for H2S in the atmosphere. H2S is highly soluble in water and will be eectively washed out during heavy rainfall. It has long been known by observations (Allen and Day, 1935; unpublished NEA data) at the ground surface in the geothermal ®elds and around fumaroles that some of the H2S is oxidized to sulfur, which accumulates near or within the geothermal ®eld. The solid sulfur precipitated will gradually react with the soil to form metal sulfates, such as gypsum, which may be bene®cial, rather than harmful, to the environment. Gypsum is commonly encountered as an alteration product around sulfur mounds (unpublished NEA data). One would thus expect that the local climate would be of great signi®cance for the course of oxidation of H2S in the atmosphere. In a dry and sunny climate, the oxidation might lead to a great amount of H2S being oxidized to SO2, whereas in a wet and cold climate a greater portion of the H2S would end up as solid sulfur. Sulfur gas emission is primarily a local pollution concern as requirements imposed by environmental authorities need to be met; however, there is also a
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
527
global pollution concern because of international conventions regarding emission of SO2 into the atmosphere. In this respect, the question of the possible conversion or rate of conversion of H2S to SO2 is of major concern. When the environmental research project was initiated at Orkustofnun, few data were available on the concentration and dispersion of H2S in the atmosphere within and outside the geothermal ®elds. One of the main tasks of the project was therefore to measure the concentration of sulfur gases in the atmosphere within and outside the sites of geothermal power plants. To study the conversion of H2S to SO2, similar monitoring stations were also established at some distance outside one of the developed geothermal ®elds. The concentration values obtained were used as a basis for modeling the dispersion of the gases in atmospheric air. Some simple experiments were also made to study the physical conditions and time needed to convert H2S to SO2. 2. Emission of H2S from Icelandic geothermal ®elds Geothermal steam is emitted from fumaroles in most high-temperature geothermal ®elds prior to development. After development, the produced ¯uid will in almost all cases greatly exceed the natural out¯ow from the ®eld. The produced ¯uid will be separated by boiling and most of the gases will remain in the steam phase. After condensation of the spent steam, either in cooling towers of the electrical power plants or in heat exchangers of the hot-water plants, the gases are either vented to the atmosphere, extracted or reinjected with the water. Geothermal ®elds produce steam with very dierent gas contents, but the type and design of a power plant also greatly in¯uence the quantity of gas that is released (Axtmann, 1975). During development, the reservoir pressure of a geothermal ®eld will generally decrease, leading to draw-down of the water table and subsequent increased degassing and formation of a steam zone. This will, in turn, increase the ¯ow of steam from fumaroles, and they will not dry out as usually happens to natural thermal springs when a ®eld is developed. The development of a geothermal ®eld is, therefore, likely to result in both a greatly increased steam ¯ow from fumaroles as well as produced steam. In Svartsengi ®eld there was no visible steam emission prior to development, but after 20 years of production the emission through fumaroles was estimated to be 15 t/year (Table 1). The total annual emission of H2S from all geothermal areas in Iceland (Fig. 1) is estimated to be about 13 kt, of which less than 7 kt is from developed ®elds. About 5 kt is emitted from produced steam in geothermal power plants and 1.5±2 kt from fumaroles in developed ®elds. The estimated out¯ow from all the exploited areas is shown in Table 1 and Fig. 2, and the estimated out¯ow from selected non-producing ®elds for comparison is given in Fig. 2. In the case of produced steam, the numbers are calculated from measured production and gas composition (Orkustofnun data) in 1996. In the case of fumarolic out¯ow, direct measurement is not always possible (KristmannsdoÂttir and AÂrmannsson, 1995). The steam out¯ow from fumaroles in Kra¯a and one other geothermal ®eld has
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
528
Table 1 Hydrogen sulphide (H2S) emission (in 1996) from exploited geothermal ®elds in Icelanda Geothermal H2S emission from ®eld fumaroles (t/year)
H2S emission from produced steam (t/year)
Total H2S emission % of (t/year) total
Reykjanes Svartsengi Nesjavellir NaÂmafjall Kra¯a Hverager+i Total
60 170 1880 1300 1600 40 5050
65 185 2020 1770 2300 180 6520
a
5 15 140 470 700 140 1470
1.0 2.8 31.0 27.1 35.3 2.8
Compiled by the authors.
been repeatedly assessed by visual comparison with the out¯ow from wells. For other ®elds, the out¯ow has been estimated by multiplying the area of geothermal activity extension by the mean fumarolic steam out¯ow per area unit in Kra¯a ®eld. Measurements of gas composition (Orkustofnun data) from each ®eld were used to calculate the H2S emission from the steam out¯ow estimated for each ®eld. In the few ®elds where there are no data on gas composition due to lack of distinct fumaroles (only steaming ground), data from Kra¯a were used for the assessment. There is major production from Kra¯a (AÂrmannsson et al., 1987), Svartsengi (BjoÈrnsson and SteingriÂmsson, 1992) and Nesjavellir (Gunnarsson et al., 1992) ®elds but production from Reykjanes, Hverager+i and NaÂmafjall is as yet limited. In NaÂmafjall, the concentration of H2S in the steam is very high, but Reykjanes and Svartsengi ®elds are brine ®elds with relatively low H2S concentrations in steam. In the Nesjavellir ®eld, there is both large steam production and a high concentration of H2S in the steam.
Fig. 1. Annual discharge of H2S from geothermal ®elds in Iceland.
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
529
3. Short-term measurements Preliminary measurements were made of the concentration of sulfur gases and mercury in the atmosphere within ten high-temperature geothermal ®elds in Iceland, both producing and non-producing ®elds (IÂvarsson et al., 1993). Point measurements of H2S were made several times in each ®eld during June±July 1993 (Appendix A). In all the ®elds both H2S and SO2 were subsequently collected for measurement at the site of highest H2S concentration, as described in Appendix A. The range of H2S concentrations was found to be from 0 to 400 mg/m3 for point measurements and 0 to 202 mg/m3 for measurement of gases collected on ®lters over a 24-h period. The variation in measured values at the same points could be in orders of magnitude from one time to another. The highest concentration was, as expected, found in Nesjavellir ®eld, and was about 40 times that in Svartsengi ®eld (point measurements). The range in SO2 concentration was from <0.1 to about 18 mg/m3, also with great variation from one time to another at the same place. The weather in Iceland, and therefore the wind direction, is changeable, so the large periodic variations observed were expected.
Fig. 2. Emission of H2S from some geothermal ®elds in Iceland. Numbers in brackets are emissions in t/year. The rectangular area shows the location of the map in Fig. 5.
530
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
4. Long-term measurements The results of the short-term measurements of the concentration of sulfur gases in the atmosphere showed that to get reliable baselines for the concentration level, measurements had to be carried out over several months. Gases were therefore measured for 4±6 months in four of the producing high-temperature geothermal ®elds: Svartsengi, Nesjavellir, Kra¯a and NaÂmafjall. At Svartsengi, Nesjavellir and Kra¯a, the sampling points were at the power plant sites, where the highest concentrations were found during the short-term measurements. At NaÂmafjall, where there is no large power plant, the sampling point was located close to the steam/water separator for the two main production wells supplying the diatomite plant; this is near the site of a planned future power plant. The long-term measurements were carried out over the years 1994±1996. The results show that the dependence of the H2S concentration on weather is signi®cant for the long-term measurements in respect to both wind conditions and precipitation (Sigurgeirsson et al., 1995; Sigurgeirsson and KristmannsdoÂttir, 1996a). A graphic demonstration of such dependence is dicult except for one factor at a time. Examples of the changes associated with wind direction are shown in Fig. 3, and for precipitation in Fig. 4. It appears that the concentration of H2S in Svartsengi in October was not greatly related to wind direction, but was highest when the wind was from either NE or NW (Fig. 3). As the wind is very changeable in Iceland, one would not expect the dependence on wind direction to be re¯ected in the mean concentration for each day. Such a correlation is indeed clear on some days, but not on others. There is, however, only one sampling site in each ®eld, and this may have biased the results.
Fig. 3. Measurements of H2S and SO2 in Svartsengi ®eld, during one month. Wind directions are shown to the right.
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
531
Fig. 4 illustrates that when there is heavy rain or snow, the concentration of H2S in the atmosphere at Kra¯a appears to be almost quantitatively eliminated, probably due to dissolution in the precipitation. This is supported by analysis (Gunnlaugsson, personal communication) of precipitation samples, which were collected over a period of several years at Nesjavellir geothermal ®eld. An increase in sulfate concentration has been clearly correlated to rainy days. No clear dependence of the SO2 concentration on weather conditions is observed and no direct correlation between the concentration of the two sulfur gases has been found either. Both the examples shown (Figs. 3 and 4) re¯ect these relations. The mean concentrations of H2S at each of the four geothermal ®elds are similar (Table 2). It is slightly higher for Nesjavellir than for the others, and slightly lower for Svartsengi than Kra¯a and NaÂmafjall. The mean concentration of SO2 varies much more. It is lowest in Svartsengi, and here the concentrations are not very dierent in summer and winter. At Nesjavellir, the mean concentration is higher than in Svartsengi. It is similar at both sites during the summer, but higher in Nesjavellir during the winter. In Kra¯a, the values are generally higher than in Svartsengi and Nesjavellir. In NaÂmafjall, the values are considerably higher than in the other ®elds. The burning of crude oil at a diatomite plant operating within NaÂmafjall ®eld causes the SO2 contamination at that site. This pollution may also aect the values from Kra¯a, which is only 11 km away from NaÂmafjall. Long-term measurements were also carried out at two sampling points: Korpa and IÂrafoss (as shown in Fig. 5), far from Nesjavellir ®eld (Sigurgeirsson and
Fig. 4. Measurements of H2S and SO2 in Kra¯a ®eld, over a one-month period. Weather conditions are indicated by the numbers to the right: 1±3 are days with no precipitation (0 is no precipitation and no sun; 1 sunny day; 2 drifting snow; 3 sandstorm; 4 foggy day; 5 little rain; 6 heavy rain; 7 heavy snow). These are standard descriptions given by the Icelandic Meteorological Oce.
9.5±15.2 6.6±11.6 9.6±12.1 10.5±11.4 0.4±2.0 0.3±1.0 n.m.
Range of concentration of H2S (monthly mean) 15.2 6.6 9.6 11.0 0.5 0.3 n.m.
9.5 10.5 11.3 10.5 1.2 0.5 n.m.
1.7 1.0 2.4 5.5 0.1 0.4 1.6
Monthly mean Monthly mean Mean of H2S in Aug./ of H2S in Dec./ concentration Sept. Jan. of SO2
1.0-2.9 0.8±1.3 2.1±2.8 2.5±8.3 1.0±0.2 0.3±0.5 0.4±2
Range of concentration of SO2 (monthly mean)
Concentrations of H2S and SO2 in mg/m3. For reference, results of measurements of SO2 from ReykjaviÂk are shown.
13 10 11 11 1.0 0.5 n.m.
Nesjavellir Svartsengi Kra¯a NaÂma®jall IÂrafoss Korpa Reykjavik a
Mean concentration of H2S
Location
1.0 1.0 2.4 3.0 0.1 0.5 n.m.
2.9 1.0 2.3 2.5 0.1 0.4 n.m.
Monthly mean Monthly mean of SO2 in Aug./ of SO2 in Dec./ Sept. Jan.
Table 2 The mean concentration of H2S and SO2 in the atmosphere over a period of 4±6 months within the geothermal ®elds, and a one-year period at IÂrafoss and Korpa (outside Nesjavellir ®eld)a
532 H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
533
KristmannsdoÂttir, 1996b) in order to study the dispersion and possible conversion of H2S to SO2 by time. The measurements were carried out continuously for 1 year at these points, and the mean concentrations over this time are shown in Table 2. In this table the annual mean concentrations of SO2 in ReykjaviÂk (BenjamiÂnsson, 1993) are also shown for comparison, but no measurements of H2S exist from the city or other locations in Iceland. The concentration of H2S is higher at IÂrafoss than at Korpa, but the reverse is true for SO2. This could mean that more H2S had been converted to SO2 on the way to Korpa: the distance from Nesjavellir to IÂrafoss is shorter than that from Nesjavellir to Korpa, and there are also high mountains blocking the way. The air transportation route from Nesjavellir to IÂrafoss would be expected to be much less in¯uenced by the landscape. One would expect much more dispersion and lower concentration of H2S gas at Korpa than at IÂrafoss. Korpa is nearer to ReykjavõÂ k than both Nesjavellir and IÂrafoss, the highway to North Iceland is nearby, and the main harbour is just across the bay. Pollution from transportation (ships and vehicles) and industry would thus be expected to in¯uence the concentration of SO2 at Korpa. Most of the H2S, both at Korpa and IÂrafoss, is expected to have originated from Nesjavellir.
Fig. 5. The location of the gas monitoring stations at IÂrafoss and Korpa, relative to Nesjavellir geothermal ®eld. The location of this map is shown in Fig. 2.
534
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
5. Experimental conversion of H2S to SO2 Simple experiments were carried out at Nesjavellir to study the physical conditions and time needed to convert H2S to SO2. Geothermal steam containing H2S gas was mixed with atmospheric air and held in a specially made airtight cylinder at ambient temperature for several months. The concentration of both H2S and SO2 was measured regularly. The cylinder was made of 3 mm thick plexiglass and had a total volume of about 25 l. It was kept inside an unheated hut and thus shielded from weather and direct sunshine, but at near-to-ambient temperature. As it is dicult and expensive to control the humidity and UVradiation in such an experiment, this simple setup was chosen for a start. Samples of air were drawn through an airtight valve and measured by DraÈger ampoules in ppm by volume. The main result of this simple experiment was that no traces of SO2 were ever measured in the air samples drawn from the cylinder. The original concentration of H2S was about 1000 ppm (vol) and it decreased to about 200 ppm in less than one month, after which it was stable for 3±4 months. The H2S was probably oxidized to sulfur, but this was not observed, nor was any attempt made to recover any chemicals from the cylinder after the experiments were stopped. The experiment is not strictly comparable with conditions in nature because no sunshine or change in humidity could in¯uence the reactions of H2S. However, in this dark, dry and relatively cold environment no conversion of H2S gas to SO2 was observed. 6. Discussion of results Concern about sulfur gas emissions during geothermal production is partly due to potential local pollution and partly to global environmental requirements regarding emission of SO2 into the atmosphere. Our measurements of the concentration of sulfur gases in air, within the geothermal ®elds, show great temporal variations. The concentration is, as would be expected, very much in¯uenced by both wind direction and precipitation. For the probable conversion of H2S to SO2, temperature sunlight and as well as precipitation, presumably play a role. To get any meaningful comparison of measurements from dierent geothermal ®elds, the mean concentrations must therefore be obtained over a long period of time during which weather conditions are observed and recorded. The mean concentration of atmospheric H2S, measured over several months, is surprisingly similar in all the production ®elds. In Nesjavellir, where it is highest, it is still of the same magnitude as in Svartsengi where the amount of H2S emission is 30 times less. The concentration of SO2 in air within the geothermal ®elds varies more than that of H2S, mainly due to the variable amounts of pollution from factories. High concentrations at NaÂmafjall, and probably also Kra¯a, are almost certainly caused by pollution from the burning of crude oil at the NaÂmafjall diatomite plant. The
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
535
concentration of SO2 in the air at Nesjavellir ®eld is almost the same as in the town of ReykjavõÂ k, whereas at Korpa the concentration is about one-fourth of the ReykjavõÂ k value. As Korpa is much nearer to ReykjavõÂ k than Nesjavellir, the pollution from the town is expected to be much more prominent there. Because there are no other sources of pollution near Nesjavellir, the presence of excess SO2 is probably due to conversion of some H2S to SO2. However, the SO2 concentration is still very low compared to the total amount of H2S emitted from wells and fumaroles. A similar argument applies to Svartsengi. The increased SO2 in Nesjavellir in the winter could thus be due to greatly increased hot water production during the cold wintertime. The dierence in hot water production in Svartsengi is much less between summer and winter as the water is sold at a ®xed price, and not by metering, as at Nesjavellir. How much of the H2S will eventually be converted to SO2 is still to some extent uncertain. Only a small fraction of the H2S appears to be converted to SO2 within the ®elds. This is found to be true even on days with almost no wind, when the emitted gas stays within the ®elds for some time. On days of high rainfall, H2S appears to be completely removed from the atmosphere. The monitoring stations at Korpa and IÂrafoss were established to obtain values for use as a basis to model the dispersion of the gases from Nesjavellir geothermal ®eld. Older, background data for the gases are very scarce. Reliable background values for H2S and SO2 in the atmosphere in Iceland, outside ReykjavõÂ k, are not available, but the concentration of SO2 normally appears to be below 0.1 mg/ m3and that of H2S less than 0.3 mg/m3 (detection limit for the human nose). No measurements exist for H2S in the air distant from the geothermal ®elds. The only existing background measurements of SO2 in atmospheric air, except for those of our project, are from ReykjavõÂ k. Unfortunately, the budget of our study did not allow us to run stations for background measurements. There appear to be some ambiguities in the results from Korpa and IÂrafoss. The H2S concentration is higher at IÂrafoss than at Korpa, as would have been expected. The SO2 concentration at IÂrafoss is almost at background level, whereas the concentration at Korpa is higher. This is probably due to pollution from ReykjaviÂk, but might also indicate some oxidation of H2S to SO2. The dispersion of the actual amount of gases discharged into the atmosphere from Nesjavellir power station was modeled by an air dispersion model (AFTOX, Trinity Consultants, 1997) for the observed wind speed and direction to points at the same distance as the monitoring stations at Korpa and IÂrafoss. The modeling work partially suered because there are few measurement stations available to provide data to stabilize the model, and only meagre background data for the gas concentration. Unfortunately, the weather recordings at Nesjavellir were not very reliable either, and therefore, weather recordings from a nearby recording station often had to be used. As the weather in Iceland is very changeable, this added to the uncertainty in the modelling. Firstly, we selected days with near constant wind direction and no precipitation; the calculated concentration of H2S, purely by air dispersion at such conditions, compared well with the one measured at Korpa and IÂrafoss, indicating that little of the H2S was converted to SO2. Then, results from
536
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
summer and winter days with no precipitation were compared in order to determine the possible eects of dierent temperature and/or sunlight conditions, but no dierences were detected between the modeling runs. However, dierences in prevailing wind directions between winter and summer seasons reduce the reliability of such direct comparisons. The concentration of H2S on rainy days was calculated by the model and the correlation between calculated and measured values was found to be worse than on dry days: the calculated concentrations on rainy days tended to be higher than those measured. This may be due to washing of H2S from the air, as has been observed and measured by Gunnlaugsson (personal communication) and IÂvarsson et. al (1993).
7. Conclusions . Short-term measurements of the atmospheric concentration of sulfur gases were successfully carried out in ten high-temperature geothermal ®elds in Iceland. . The concentrations of H2S and SO2 showed large temporal and spatial variations, mainly due to variations in wind direction and amount of precipitation. . Long-term measurements, made in four exploited ®elds and at two distant sites, generally con®rmed the short-term data. . Modelling of H2S data from Nesjavellir geothermal ®eld indicates minor, or at least very slow, conversion of H2S to SO2 for the atmospheric conditions in Iceland, suggesting that only a small fraction reacts within the geothermal ®elds or within a radius of 15±25 km. However, the modelling does not give very conclusive results, due to lack of measurements and other uncertainties. . Precipitation eectively washes H2S from air, so only a small fraction of the H2S discharged from the geothermal ®elds is believed to end up as SO2. . An experiment, in which a steam and air mixture was held in an airtight container, showed no conversion of H2S to SO2 over a period of several months. . There is a clear need for much more monitoring data, both from within and distant from the geothermal ®elds.
Acknowledgements The three main producers of high temperature geothermal ®elds; Landsvirkjun, Hitaveita ReykjavõÂ kur and Hitaveita Su+urnesja are duly thanked for support with the projectas well as the Icelandic Ministry of Environment. The head tecnician of Orkustofnuns geothermal laboratorium KristjaÂn H. Sigur+sson is thanked for excellent supervision of the chemical analysis and improvement of the methodology. Trevor Hunt is thanked for editorial suggestions which improved the paper
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
537
Appendix A. Methods of sulfur gas sampling and measurement The aim of the short-term measurements of sulfur gases was to obtain a rough estimate of the concentration levels of the gases, and to compare the precision, accuracy and usefulness of several dierent analytical methods. The short-term measurements were also used to select sites for long-term measurements in the developed ®elds. The short-term measurements were both point measurements and measurements based on sampling of gases over a 24-h period on wetted ®lters and in liquids. In the point measurements, the concentrations of H2S were measured on a grid extending over the area of visible activity (IÂvarsson et al., 1993). The number of points chosen (7±39) was based on the size of the ®eld. Each measurement was the mean of 5±15 readings, at about 2 min intervals. The measurements were made using a hand-held Jerome 621 meter, based on the measurement of H2S accumulated on a gold ®lm. The detection limit was 1 ppb (accuracy at 50 ppb is 22 ppb), and the range of the meter is 1±500 ppb. The results of the point measurements of H2S were used to select sites of the highest concentration for sampling gases over a 24±48 h period. Sampling was performed by pumping measured quantities of air through a H2O2 solution, and ®lters wetted in AgNO3 and KOH solutions. When collection in each place lasted for 2 days, the ®lters were changed after 24 h. The SO2 collected in the H2O2 solution was oxidized in the solution to SO4, which was measured by a Dionex ion chromatograph. The SO2 collected on the KOH wetted ®lter was oxidized to SO4 by soaking the ®lter in 20 ml of a 0.3% H2O2 solution. SO4 was then analysed by a Dionex ion chromatograph and calculated as mg SO2/ m3. The H2S collected on a ®lter wetted in 0.5 ml of a 2% AgNO3 solution was reacted to form Ag2S, which was dissolved by heating in concentrated HNO3. The Ag was then measured by Atomic Absorption Spectroscopy, and the concentration of the corresponding H2S calculated in mg/m3. Blanks were run for all the measurements and at each sampling point. For the long-term measurements, sampling sites in the developed ®elds were the same as for the short-term measurements. The sampling sites outside Nesjavellir ®eld were selected to be at the closest weather observation sites. Testing showed that the method based on sampling on ®lters wetted in KOH and AgNO3 solutions for the collection of SO2 and H2S, respectively, was the most reliable, and was therefore chosen for further work. The ®lters were changed every 24 h, and two to three blanks were run with each measured lot (15±20) of ®lters.
References Allen, E.T., Day, A.L., 1935. Hot springs of the Yellowstone National Park. Publication 466, Carnegie Institute, Washington, USA (525 pp). AÂrmannsson, H., KristmannsdoÂttir, H., 1992. Geothermal environmental impact. Geothermics 21, 869± 880.
538
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
AÂrmannsson, H., Gudmundsson, AÂ., SteingriÂmsson, B.S., 1987. Exploration and development of the Kra¯a geothermal area. JoÈkull 37, 13±30. Axtmann, R.C., 1975. Emission control of gas euents from geothermal power plants. Environmental Letters 8, 135±146. BenjamiÂnsson, J., 1993. Pollution of air in the town of ReykjaviÂk. A yearly report from the health authorities in ReykjaviÂk (in Icelandic). Bjarnason, J.OÈ., 1991. On the pH of precipitation in Svartsengi. Report JOÈB-91/02. Orkustofnun, ReykjaviÂk, Iceland (in Icelandic), 4 pp. BjoÈrnsson, G., SteingriÂmsson, B.S., 1992. Fifteen years of temperature and pressure monitoring in the Svartsengi high-temperature geothermal ®eld in SW-Iceland. Geothermal Resources Council Transactions 16, 627±634. Brown, K.L., Webster, J.G., 1994. H2S oxidation in aerosols. In: Proceedings 15th PNOC-EDC Geothermal Conference, 37±44. Cox, R.A., Sandalls, F.J., 1974. The photo-oxidation of hydrogen sulphide and dimethyl sulphide in air. Atmospheric Environment 8, 1269±1281. Gunnarsson, AÂ., SteingriÂmsson, B.S., Gunnlaugsson, E., Magnusson, J., Maack, R., 1992. Nesjavellir geothermal co-generation power plant. Geothermics 21, 559±583. IÂvarsson, G., Sigurgeirsson, M.AÂ., Gunnlaugsson, E., Sigur+sson, K.H., KristmannsdoÂttir, H., 1993. Measurements of gas in atmospheric air. The concentration of hydrogen sulphide, sulphur dioxide and mercury in high-temperature geothermal areas. A co-operative project of Orkustofnun and Hitaveita ReykjaviÂkur. Report OS-93074/JHD-16, Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 69 pp). KristmannsdoÂttir, H., AÂrmannsson, H., 1995. Environmental impact of geothermal utilization in Iceland. In: Proceedings World Geothermal Congress, 2731±2734. Muna, E.W., Ojambo, S.B. 1985. Possible in¯uence of gases emitted from Olkaria geothermal ®eld on the rainwater of its surroundings. Report No. GC/GEN/038. Kenya Power Company, Geothermal Project, Olkaria. Sigurgeirsson, M.AÂ., Sigur+sson, K.H., KristmannsdoÂttir, H., 1995. Measurements of sulphur gases in the atmosphere. The concentration of hydrogen sulphide and sulphur dioxide at Svartsengi and Kra¯a. Report OS-95025/JHD-18B. Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 34 pp). Sigurgeirsson, M.AÂ., KristmannsdoÂttir, H., 1996a. Measurements of sulphur gases in the atmosphere. The concentration of hydrogen sulphide and sulphur dioxide at Nesjavellir and in Bjarnar¯ag. Report OS-96020/JHD-10B. Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 69 pp). Sigurgeirsson, M.AÂ., KristmannsdoÂttir, H., 1996b. Measurements of sulphur gases in the atmosphere. The concentration of hydrogen sulphide and sulphur dioxide at Korpa and IÂrafoss. Report OS96020/JHD-11B Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 59 pp). Trinity Consultants, 1997. User Guide for BREEZE HAZ-AFTOX dispersion model. Trinity Consultants, USA (206 pp).
Sulfur gas emissions from geothermal power plants in Iceland Hrefna KristmannsdoÂttir a,*, MagnuÂs Sigurgeirsson a, HalldoÂr AÂrmannsson a, Hreinn Hjartarson b, MagnuÂs OÂlafsson a a Orkustofnun, GrensaÂsvegur 9, 108 ReykjavõÂk, Iceland The Icelandic Meteorological Institute, BuÂsta+avegi 9, 105 ReykjavõÂk, Iceland
b
Received 27 July 1998; accepted 23 April 1999
Abstract Sulfur gas (H2S and SO2) emissions from geothermal ®elds in Iceland have been studied as part of a project, aimed at enhancing environmental research concerning eects of geothermal development. Short-term measurements of the gases have been carried out in several high-temperature geothermal ®elds in Iceland. In four exploited ®elds, baseline values for the concentration of sulfur gases have been obtained by long-term measurements. The data strongly re¯ect the dependence of gas concentrations on climatic factors, especially precipitation. Interpretation of the data by air distribution modeling, and by simple experiment, indicate minor, or at least very slow conversion of H2S to SO2 at atmospheric conditions in Iceland. 7 2000 CNR. Published by Elsevier Science Ltd. All rights reserved. Keywords: Pollution; Atmosphere; Dispersion; Weather; Conversion; Iceland
1. Introduction As a consequence of the increased emphasis on the environmental impacts of energy production, a project was started in 1991 to study the environmental * Corresponding author. Tel.: +354-569-6000; fax: +354-568-8896. E-mail address: [email protected] (H. KristmannsdoÂttir). 0375-6505/00/$20.00 7 2000 CNR. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 3 7 5 - 6 5 0 5 ( 0 0 ) 0 0 0 2 0 - 1
526
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
impact of geothermal developments in Iceland. The project was initiated by Orkustofnun (The National Energy Authority, NEA), and was conducted with the cooperation of the main producers of high-temperature geothermal energy in Iceland. The project entailed, ®rstly, an assessment of the present status at the ®ve main sites of high-temperature geothermal production in Iceland, and secondly, the de®nition of several priority projects (KristmannsdoÂttir and AÂrmannsson, 1995) to be carried out within the scope of the project. One of the main eects of geothermal exploitation on the environment is the emission of gases with geothermal steam (Axtmann, 1975; AÂrmannsson and KristmannsdoÂttir, 1992). The greenhouse gases CO2 and CH4, together with sulfur gases, are of main concern in this respect. The sulfur gas emitted from geothermal plants is in the form of H2S. This gas is toxic in high concentrations and has a very unpleasant odour even in very low concentrations; its presence is often of great environmental concern. However, people are only in danger of being poisoned inside poorly-ventilated buildings, at plant or drill sites, or in enclosed topographic lows. The smell may be more unpleasant to people not used to it, whereas people accustomed to it may not notice it. The H2S gas may, however, be oxidized to SO2, causing acidi®cation of rain and soil, which is of great concern. If all the H2S gas from geothermal power plants in Iceland were to be quantitatively converted to SO2, at least some plants would be considered unacceptably polluting and the removal of H2S from steam would be made mandatory. Measurements of pH and sulfate concentrations in precipitation in the vicinity of the Olkaria power plant in Kenya (Muna and Ojambo, 1985) and of pH in the neighbourhood of the Svartsengi power plant in Iceland (Bjarnason, 1991), reveal neither pH changes nor SO2 addition. Thus, there is evidence that probably not all the H2S is converted to SO2, at least not immediately. This has been a matter of considerable discussion, as the sulfur chemistry is complicated and little research has been done on the conversion of geothermally emitted H2S to SO2 in the atmosphere. Brown and Webster (1994) claim that oxidation of H2S within aerosols is a slow process, but Cox and Sandalls (1974) concluded that photo-oxidation of H2S to SO2 is a major loss process for H2S in the atmosphere. H2S is highly soluble in water and will be eectively washed out during heavy rainfall. It has long been known by observations (Allen and Day, 1935; unpublished NEA data) at the ground surface in the geothermal ®elds and around fumaroles that some of the H2S is oxidized to sulfur, which accumulates near or within the geothermal ®eld. The solid sulfur precipitated will gradually react with the soil to form metal sulfates, such as gypsum, which may be bene®cial, rather than harmful, to the environment. Gypsum is commonly encountered as an alteration product around sulfur mounds (unpublished NEA data). One would thus expect that the local climate would be of great signi®cance for the course of oxidation of H2S in the atmosphere. In a dry and sunny climate, the oxidation might lead to a great amount of H2S being oxidized to SO2, whereas in a wet and cold climate a greater portion of the H2S would end up as solid sulfur. Sulfur gas emission is primarily a local pollution concern as requirements imposed by environmental authorities need to be met; however, there is also a
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
527
global pollution concern because of international conventions regarding emission of SO2 into the atmosphere. In this respect, the question of the possible conversion or rate of conversion of H2S to SO2 is of major concern. When the environmental research project was initiated at Orkustofnun, few data were available on the concentration and dispersion of H2S in the atmosphere within and outside the geothermal ®elds. One of the main tasks of the project was therefore to measure the concentration of sulfur gases in the atmosphere within and outside the sites of geothermal power plants. To study the conversion of H2S to SO2, similar monitoring stations were also established at some distance outside one of the developed geothermal ®elds. The concentration values obtained were used as a basis for modeling the dispersion of the gases in atmospheric air. Some simple experiments were also made to study the physical conditions and time needed to convert H2S to SO2. 2. Emission of H2S from Icelandic geothermal ®elds Geothermal steam is emitted from fumaroles in most high-temperature geothermal ®elds prior to development. After development, the produced ¯uid will in almost all cases greatly exceed the natural out¯ow from the ®eld. The produced ¯uid will be separated by boiling and most of the gases will remain in the steam phase. After condensation of the spent steam, either in cooling towers of the electrical power plants or in heat exchangers of the hot-water plants, the gases are either vented to the atmosphere, extracted or reinjected with the water. Geothermal ®elds produce steam with very dierent gas contents, but the type and design of a power plant also greatly in¯uence the quantity of gas that is released (Axtmann, 1975). During development, the reservoir pressure of a geothermal ®eld will generally decrease, leading to draw-down of the water table and subsequent increased degassing and formation of a steam zone. This will, in turn, increase the ¯ow of steam from fumaroles, and they will not dry out as usually happens to natural thermal springs when a ®eld is developed. The development of a geothermal ®eld is, therefore, likely to result in both a greatly increased steam ¯ow from fumaroles as well as produced steam. In Svartsengi ®eld there was no visible steam emission prior to development, but after 20 years of production the emission through fumaroles was estimated to be 15 t/year (Table 1). The total annual emission of H2S from all geothermal areas in Iceland (Fig. 1) is estimated to be about 13 kt, of which less than 7 kt is from developed ®elds. About 5 kt is emitted from produced steam in geothermal power plants and 1.5±2 kt from fumaroles in developed ®elds. The estimated out¯ow from all the exploited areas is shown in Table 1 and Fig. 2, and the estimated out¯ow from selected non-producing ®elds for comparison is given in Fig. 2. In the case of produced steam, the numbers are calculated from measured production and gas composition (Orkustofnun data) in 1996. In the case of fumarolic out¯ow, direct measurement is not always possible (KristmannsdoÂttir and AÂrmannsson, 1995). The steam out¯ow from fumaroles in Kra¯a and one other geothermal ®eld has
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
528
Table 1 Hydrogen sulphide (H2S) emission (in 1996) from exploited geothermal ®elds in Icelanda Geothermal H2S emission from ®eld fumaroles (t/year)
H2S emission from produced steam (t/year)
Total H2S emission % of (t/year) total
Reykjanes Svartsengi Nesjavellir NaÂmafjall Kra¯a Hverager+i Total
60 170 1880 1300 1600 40 5050
65 185 2020 1770 2300 180 6520
a
5 15 140 470 700 140 1470
1.0 2.8 31.0 27.1 35.3 2.8
Compiled by the authors.
been repeatedly assessed by visual comparison with the out¯ow from wells. For other ®elds, the out¯ow has been estimated by multiplying the area of geothermal activity extension by the mean fumarolic steam out¯ow per area unit in Kra¯a ®eld. Measurements of gas composition (Orkustofnun data) from each ®eld were used to calculate the H2S emission from the steam out¯ow estimated for each ®eld. In the few ®elds where there are no data on gas composition due to lack of distinct fumaroles (only steaming ground), data from Kra¯a were used for the assessment. There is major production from Kra¯a (AÂrmannsson et al., 1987), Svartsengi (BjoÈrnsson and SteingriÂmsson, 1992) and Nesjavellir (Gunnarsson et al., 1992) ®elds but production from Reykjanes, Hverager+i and NaÂmafjall is as yet limited. In NaÂmafjall, the concentration of H2S in the steam is very high, but Reykjanes and Svartsengi ®elds are brine ®elds with relatively low H2S concentrations in steam. In the Nesjavellir ®eld, there is both large steam production and a high concentration of H2S in the steam.
Fig. 1. Annual discharge of H2S from geothermal ®elds in Iceland.
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
529
3. Short-term measurements Preliminary measurements were made of the concentration of sulfur gases and mercury in the atmosphere within ten high-temperature geothermal ®elds in Iceland, both producing and non-producing ®elds (IÂvarsson et al., 1993). Point measurements of H2S were made several times in each ®eld during June±July 1993 (Appendix A). In all the ®elds both H2S and SO2 were subsequently collected for measurement at the site of highest H2S concentration, as described in Appendix A. The range of H2S concentrations was found to be from 0 to 400 mg/m3 for point measurements and 0 to 202 mg/m3 for measurement of gases collected on ®lters over a 24-h period. The variation in measured values at the same points could be in orders of magnitude from one time to another. The highest concentration was, as expected, found in Nesjavellir ®eld, and was about 40 times that in Svartsengi ®eld (point measurements). The range in SO2 concentration was from <0.1 to about 18 mg/m3, also with great variation from one time to another at the same place. The weather in Iceland, and therefore the wind direction, is changeable, so the large periodic variations observed were expected.
Fig. 2. Emission of H2S from some geothermal ®elds in Iceland. Numbers in brackets are emissions in t/year. The rectangular area shows the location of the map in Fig. 5.
530
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
4. Long-term measurements The results of the short-term measurements of the concentration of sulfur gases in the atmosphere showed that to get reliable baselines for the concentration level, measurements had to be carried out over several months. Gases were therefore measured for 4±6 months in four of the producing high-temperature geothermal ®elds: Svartsengi, Nesjavellir, Kra¯a and NaÂmafjall. At Svartsengi, Nesjavellir and Kra¯a, the sampling points were at the power plant sites, where the highest concentrations were found during the short-term measurements. At NaÂmafjall, where there is no large power plant, the sampling point was located close to the steam/water separator for the two main production wells supplying the diatomite plant; this is near the site of a planned future power plant. The long-term measurements were carried out over the years 1994±1996. The results show that the dependence of the H2S concentration on weather is signi®cant for the long-term measurements in respect to both wind conditions and precipitation (Sigurgeirsson et al., 1995; Sigurgeirsson and KristmannsdoÂttir, 1996a). A graphic demonstration of such dependence is dicult except for one factor at a time. Examples of the changes associated with wind direction are shown in Fig. 3, and for precipitation in Fig. 4. It appears that the concentration of H2S in Svartsengi in October was not greatly related to wind direction, but was highest when the wind was from either NE or NW (Fig. 3). As the wind is very changeable in Iceland, one would not expect the dependence on wind direction to be re¯ected in the mean concentration for each day. Such a correlation is indeed clear on some days, but not on others. There is, however, only one sampling site in each ®eld, and this may have biased the results.
Fig. 3. Measurements of H2S and SO2 in Svartsengi ®eld, during one month. Wind directions are shown to the right.
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
531
Fig. 4 illustrates that when there is heavy rain or snow, the concentration of H2S in the atmosphere at Kra¯a appears to be almost quantitatively eliminated, probably due to dissolution in the precipitation. This is supported by analysis (Gunnlaugsson, personal communication) of precipitation samples, which were collected over a period of several years at Nesjavellir geothermal ®eld. An increase in sulfate concentration has been clearly correlated to rainy days. No clear dependence of the SO2 concentration on weather conditions is observed and no direct correlation between the concentration of the two sulfur gases has been found either. Both the examples shown (Figs. 3 and 4) re¯ect these relations. The mean concentrations of H2S at each of the four geothermal ®elds are similar (Table 2). It is slightly higher for Nesjavellir than for the others, and slightly lower for Svartsengi than Kra¯a and NaÂmafjall. The mean concentration of SO2 varies much more. It is lowest in Svartsengi, and here the concentrations are not very dierent in summer and winter. At Nesjavellir, the mean concentration is higher than in Svartsengi. It is similar at both sites during the summer, but higher in Nesjavellir during the winter. In Kra¯a, the values are generally higher than in Svartsengi and Nesjavellir. In NaÂmafjall, the values are considerably higher than in the other ®elds. The burning of crude oil at a diatomite plant operating within NaÂmafjall ®eld causes the SO2 contamination at that site. This pollution may also aect the values from Kra¯a, which is only 11 km away from NaÂmafjall. Long-term measurements were also carried out at two sampling points: Korpa and IÂrafoss (as shown in Fig. 5), far from Nesjavellir ®eld (Sigurgeirsson and
Fig. 4. Measurements of H2S and SO2 in Kra¯a ®eld, over a one-month period. Weather conditions are indicated by the numbers to the right: 1±3 are days with no precipitation (0 is no precipitation and no sun; 1 sunny day; 2 drifting snow; 3 sandstorm; 4 foggy day; 5 little rain; 6 heavy rain; 7 heavy snow). These are standard descriptions given by the Icelandic Meteorological Oce.
9.5±15.2 6.6±11.6 9.6±12.1 10.5±11.4 0.4±2.0 0.3±1.0 n.m.
Range of concentration of H2S (monthly mean) 15.2 6.6 9.6 11.0 0.5 0.3 n.m.
9.5 10.5 11.3 10.5 1.2 0.5 n.m.
1.7 1.0 2.4 5.5 0.1 0.4 1.6
Monthly mean Monthly mean Mean of H2S in Aug./ of H2S in Dec./ concentration Sept. Jan. of SO2
1.0-2.9 0.8±1.3 2.1±2.8 2.5±8.3 1.0±0.2 0.3±0.5 0.4±2
Range of concentration of SO2 (monthly mean)
Concentrations of H2S and SO2 in mg/m3. For reference, results of measurements of SO2 from ReykjaviÂk are shown.
13 10 11 11 1.0 0.5 n.m.
Nesjavellir Svartsengi Kra¯a NaÂma®jall IÂrafoss Korpa Reykjavik a
Mean concentration of H2S
Location
1.0 1.0 2.4 3.0 0.1 0.5 n.m.
2.9 1.0 2.3 2.5 0.1 0.4 n.m.
Monthly mean Monthly mean of SO2 in Aug./ of SO2 in Dec./ Sept. Jan.
Table 2 The mean concentration of H2S and SO2 in the atmosphere over a period of 4±6 months within the geothermal ®elds, and a one-year period at IÂrafoss and Korpa (outside Nesjavellir ®eld)a
532 H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
533
KristmannsdoÂttir, 1996b) in order to study the dispersion and possible conversion of H2S to SO2 by time. The measurements were carried out continuously for 1 year at these points, and the mean concentrations over this time are shown in Table 2. In this table the annual mean concentrations of SO2 in ReykjaviÂk (BenjamiÂnsson, 1993) are also shown for comparison, but no measurements of H2S exist from the city or other locations in Iceland. The concentration of H2S is higher at IÂrafoss than at Korpa, but the reverse is true for SO2. This could mean that more H2S had been converted to SO2 on the way to Korpa: the distance from Nesjavellir to IÂrafoss is shorter than that from Nesjavellir to Korpa, and there are also high mountains blocking the way. The air transportation route from Nesjavellir to IÂrafoss would be expected to be much less in¯uenced by the landscape. One would expect much more dispersion and lower concentration of H2S gas at Korpa than at IÂrafoss. Korpa is nearer to ReykjavõÂ k than both Nesjavellir and IÂrafoss, the highway to North Iceland is nearby, and the main harbour is just across the bay. Pollution from transportation (ships and vehicles) and industry would thus be expected to in¯uence the concentration of SO2 at Korpa. Most of the H2S, both at Korpa and IÂrafoss, is expected to have originated from Nesjavellir.
Fig. 5. The location of the gas monitoring stations at IÂrafoss and Korpa, relative to Nesjavellir geothermal ®eld. The location of this map is shown in Fig. 2.
534
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
5. Experimental conversion of H2S to SO2 Simple experiments were carried out at Nesjavellir to study the physical conditions and time needed to convert H2S to SO2. Geothermal steam containing H2S gas was mixed with atmospheric air and held in a specially made airtight cylinder at ambient temperature for several months. The concentration of both H2S and SO2 was measured regularly. The cylinder was made of 3 mm thick plexiglass and had a total volume of about 25 l. It was kept inside an unheated hut and thus shielded from weather and direct sunshine, but at near-to-ambient temperature. As it is dicult and expensive to control the humidity and UVradiation in such an experiment, this simple setup was chosen for a start. Samples of air were drawn through an airtight valve and measured by DraÈger ampoules in ppm by volume. The main result of this simple experiment was that no traces of SO2 were ever measured in the air samples drawn from the cylinder. The original concentration of H2S was about 1000 ppm (vol) and it decreased to about 200 ppm in less than one month, after which it was stable for 3±4 months. The H2S was probably oxidized to sulfur, but this was not observed, nor was any attempt made to recover any chemicals from the cylinder after the experiments were stopped. The experiment is not strictly comparable with conditions in nature because no sunshine or change in humidity could in¯uence the reactions of H2S. However, in this dark, dry and relatively cold environment no conversion of H2S gas to SO2 was observed. 6. Discussion of results Concern about sulfur gas emissions during geothermal production is partly due to potential local pollution and partly to global environmental requirements regarding emission of SO2 into the atmosphere. Our measurements of the concentration of sulfur gases in air, within the geothermal ®elds, show great temporal variations. The concentration is, as would be expected, very much in¯uenced by both wind direction and precipitation. For the probable conversion of H2S to SO2, temperature sunlight and as well as precipitation, presumably play a role. To get any meaningful comparison of measurements from dierent geothermal ®elds, the mean concentrations must therefore be obtained over a long period of time during which weather conditions are observed and recorded. The mean concentration of atmospheric H2S, measured over several months, is surprisingly similar in all the production ®elds. In Nesjavellir, where it is highest, it is still of the same magnitude as in Svartsengi where the amount of H2S emission is 30 times less. The concentration of SO2 in air within the geothermal ®elds varies more than that of H2S, mainly due to the variable amounts of pollution from factories. High concentrations at NaÂmafjall, and probably also Kra¯a, are almost certainly caused by pollution from the burning of crude oil at the NaÂmafjall diatomite plant. The
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
535
concentration of SO2 in the air at Nesjavellir ®eld is almost the same as in the town of ReykjavõÂ k, whereas at Korpa the concentration is about one-fourth of the ReykjavõÂ k value. As Korpa is much nearer to ReykjavõÂ k than Nesjavellir, the pollution from the town is expected to be much more prominent there. Because there are no other sources of pollution near Nesjavellir, the presence of excess SO2 is probably due to conversion of some H2S to SO2. However, the SO2 concentration is still very low compared to the total amount of H2S emitted from wells and fumaroles. A similar argument applies to Svartsengi. The increased SO2 in Nesjavellir in the winter could thus be due to greatly increased hot water production during the cold wintertime. The dierence in hot water production in Svartsengi is much less between summer and winter as the water is sold at a ®xed price, and not by metering, as at Nesjavellir. How much of the H2S will eventually be converted to SO2 is still to some extent uncertain. Only a small fraction of the H2S appears to be converted to SO2 within the ®elds. This is found to be true even on days with almost no wind, when the emitted gas stays within the ®elds for some time. On days of high rainfall, H2S appears to be completely removed from the atmosphere. The monitoring stations at Korpa and IÂrafoss were established to obtain values for use as a basis to model the dispersion of the gases from Nesjavellir geothermal ®eld. Older, background data for the gases are very scarce. Reliable background values for H2S and SO2 in the atmosphere in Iceland, outside ReykjavõÂ k, are not available, but the concentration of SO2 normally appears to be below 0.1 mg/ m3and that of H2S less than 0.3 mg/m3 (detection limit for the human nose). No measurements exist for H2S in the air distant from the geothermal ®elds. The only existing background measurements of SO2 in atmospheric air, except for those of our project, are from ReykjavõÂ k. Unfortunately, the budget of our study did not allow us to run stations for background measurements. There appear to be some ambiguities in the results from Korpa and IÂrafoss. The H2S concentration is higher at IÂrafoss than at Korpa, as would have been expected. The SO2 concentration at IÂrafoss is almost at background level, whereas the concentration at Korpa is higher. This is probably due to pollution from ReykjaviÂk, but might also indicate some oxidation of H2S to SO2. The dispersion of the actual amount of gases discharged into the atmosphere from Nesjavellir power station was modeled by an air dispersion model (AFTOX, Trinity Consultants, 1997) for the observed wind speed and direction to points at the same distance as the monitoring stations at Korpa and IÂrafoss. The modeling work partially suered because there are few measurement stations available to provide data to stabilize the model, and only meagre background data for the gas concentration. Unfortunately, the weather recordings at Nesjavellir were not very reliable either, and therefore, weather recordings from a nearby recording station often had to be used. As the weather in Iceland is very changeable, this added to the uncertainty in the modelling. Firstly, we selected days with near constant wind direction and no precipitation; the calculated concentration of H2S, purely by air dispersion at such conditions, compared well with the one measured at Korpa and IÂrafoss, indicating that little of the H2S was converted to SO2. Then, results from
536
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
summer and winter days with no precipitation were compared in order to determine the possible eects of dierent temperature and/or sunlight conditions, but no dierences were detected between the modeling runs. However, dierences in prevailing wind directions between winter and summer seasons reduce the reliability of such direct comparisons. The concentration of H2S on rainy days was calculated by the model and the correlation between calculated and measured values was found to be worse than on dry days: the calculated concentrations on rainy days tended to be higher than those measured. This may be due to washing of H2S from the air, as has been observed and measured by Gunnlaugsson (personal communication) and IÂvarsson et. al (1993).
7. Conclusions . Short-term measurements of the atmospheric concentration of sulfur gases were successfully carried out in ten high-temperature geothermal ®elds in Iceland. . The concentrations of H2S and SO2 showed large temporal and spatial variations, mainly due to variations in wind direction and amount of precipitation. . Long-term measurements, made in four exploited ®elds and at two distant sites, generally con®rmed the short-term data. . Modelling of H2S data from Nesjavellir geothermal ®eld indicates minor, or at least very slow, conversion of H2S to SO2 for the atmospheric conditions in Iceland, suggesting that only a small fraction reacts within the geothermal ®elds or within a radius of 15±25 km. However, the modelling does not give very conclusive results, due to lack of measurements and other uncertainties. . Precipitation eectively washes H2S from air, so only a small fraction of the H2S discharged from the geothermal ®elds is believed to end up as SO2. . An experiment, in which a steam and air mixture was held in an airtight container, showed no conversion of H2S to SO2 over a period of several months. . There is a clear need for much more monitoring data, both from within and distant from the geothermal ®elds.
Acknowledgements The three main producers of high temperature geothermal ®elds; Landsvirkjun, Hitaveita ReykjavõÂ kur and Hitaveita Su+urnesja are duly thanked for support with the projectas well as the Icelandic Ministry of Environment. The head tecnician of Orkustofnuns geothermal laboratorium KristjaÂn H. Sigur+sson is thanked for excellent supervision of the chemical analysis and improvement of the methodology. Trevor Hunt is thanked for editorial suggestions which improved the paper
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
537
Appendix A. Methods of sulfur gas sampling and measurement The aim of the short-term measurements of sulfur gases was to obtain a rough estimate of the concentration levels of the gases, and to compare the precision, accuracy and usefulness of several dierent analytical methods. The short-term measurements were also used to select sites for long-term measurements in the developed ®elds. The short-term measurements were both point measurements and measurements based on sampling of gases over a 24-h period on wetted ®lters and in liquids. In the point measurements, the concentrations of H2S were measured on a grid extending over the area of visible activity (IÂvarsson et al., 1993). The number of points chosen (7±39) was based on the size of the ®eld. Each measurement was the mean of 5±15 readings, at about 2 min intervals. The measurements were made using a hand-held Jerome 621 meter, based on the measurement of H2S accumulated on a gold ®lm. The detection limit was 1 ppb (accuracy at 50 ppb is 22 ppb), and the range of the meter is 1±500 ppb. The results of the point measurements of H2S were used to select sites of the highest concentration for sampling gases over a 24±48 h period. Sampling was performed by pumping measured quantities of air through a H2O2 solution, and ®lters wetted in AgNO3 and KOH solutions. When collection in each place lasted for 2 days, the ®lters were changed after 24 h. The SO2 collected in the H2O2 solution was oxidized in the solution to SO4, which was measured by a Dionex ion chromatograph. The SO2 collected on the KOH wetted ®lter was oxidized to SO4 by soaking the ®lter in 20 ml of a 0.3% H2O2 solution. SO4 was then analysed by a Dionex ion chromatograph and calculated as mg SO2/ m3. The H2S collected on a ®lter wetted in 0.5 ml of a 2% AgNO3 solution was reacted to form Ag2S, which was dissolved by heating in concentrated HNO3. The Ag was then measured by Atomic Absorption Spectroscopy, and the concentration of the corresponding H2S calculated in mg/m3. Blanks were run for all the measurements and at each sampling point. For the long-term measurements, sampling sites in the developed ®elds were the same as for the short-term measurements. The sampling sites outside Nesjavellir ®eld were selected to be at the closest weather observation sites. Testing showed that the method based on sampling on ®lters wetted in KOH and AgNO3 solutions for the collection of SO2 and H2S, respectively, was the most reliable, and was therefore chosen for further work. The ®lters were changed every 24 h, and two to three blanks were run with each measured lot (15±20) of ®lters.
References Allen, E.T., Day, A.L., 1935. Hot springs of the Yellowstone National Park. Publication 466, Carnegie Institute, Washington, USA (525 pp). AÂrmannsson, H., KristmannsdoÂttir, H., 1992. Geothermal environmental impact. Geothermics 21, 869± 880.
538
H. KristmannsdoÂttir et al. / Geothermics 29 (2000) 525±538
AÂrmannsson, H., Gudmundsson, AÂ., SteingriÂmsson, B.S., 1987. Exploration and development of the Kra¯a geothermal area. JoÈkull 37, 13±30. Axtmann, R.C., 1975. Emission control of gas euents from geothermal power plants. Environmental Letters 8, 135±146. BenjamiÂnsson, J., 1993. Pollution of air in the town of ReykjaviÂk. A yearly report from the health authorities in ReykjaviÂk (in Icelandic). Bjarnason, J.OÈ., 1991. On the pH of precipitation in Svartsengi. Report JOÈB-91/02. Orkustofnun, ReykjaviÂk, Iceland (in Icelandic), 4 pp. BjoÈrnsson, G., SteingriÂmsson, B.S., 1992. Fifteen years of temperature and pressure monitoring in the Svartsengi high-temperature geothermal ®eld in SW-Iceland. Geothermal Resources Council Transactions 16, 627±634. Brown, K.L., Webster, J.G., 1994. H2S oxidation in aerosols. In: Proceedings 15th PNOC-EDC Geothermal Conference, 37±44. Cox, R.A., Sandalls, F.J., 1974. The photo-oxidation of hydrogen sulphide and dimethyl sulphide in air. Atmospheric Environment 8, 1269±1281. Gunnarsson, AÂ., SteingriÂmsson, B.S., Gunnlaugsson, E., Magnusson, J., Maack, R., 1992. Nesjavellir geothermal co-generation power plant. Geothermics 21, 559±583. IÂvarsson, G., Sigurgeirsson, M.AÂ., Gunnlaugsson, E., Sigur+sson, K.H., KristmannsdoÂttir, H., 1993. Measurements of gas in atmospheric air. The concentration of hydrogen sulphide, sulphur dioxide and mercury in high-temperature geothermal areas. A co-operative project of Orkustofnun and Hitaveita ReykjaviÂkur. Report OS-93074/JHD-16, Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 69 pp). KristmannsdoÂttir, H., AÂrmannsson, H., 1995. Environmental impact of geothermal utilization in Iceland. In: Proceedings World Geothermal Congress, 2731±2734. Muna, E.W., Ojambo, S.B. 1985. Possible in¯uence of gases emitted from Olkaria geothermal ®eld on the rainwater of its surroundings. Report No. GC/GEN/038. Kenya Power Company, Geothermal Project, Olkaria. Sigurgeirsson, M.AÂ., Sigur+sson, K.H., KristmannsdoÂttir, H., 1995. Measurements of sulphur gases in the atmosphere. The concentration of hydrogen sulphide and sulphur dioxide at Svartsengi and Kra¯a. Report OS-95025/JHD-18B. Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 34 pp). Sigurgeirsson, M.AÂ., KristmannsdoÂttir, H., 1996a. Measurements of sulphur gases in the atmosphere. The concentration of hydrogen sulphide and sulphur dioxide at Nesjavellir and in Bjarnar¯ag. Report OS-96020/JHD-10B. Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 69 pp). Sigurgeirsson, M.AÂ., KristmannsdoÂttir, H., 1996b. Measurements of sulphur gases in the atmosphere. The concentration of hydrogen sulphide and sulphur dioxide at Korpa and IÂrafoss. Report OS96020/JHD-11B Orkustofnun, ReykjaviÂk, Iceland (in Icelandic, 59 pp). Trinity Consultants, 1997. User Guide for BREEZE HAZ-AFTOX dispersion model. Trinity Consultants, USA (206 pp).