GRC Transactions, Vol. 37, 2013 Fluid Chemistry of Menengai Geothermal Wells, Kenya Sylvia Joan Malimo Geothermal Development Company Ltd, Nakuru, Kenya smalimo@gdc.co.ke Keywords Fluid chemistry, geothermometers, aquifer fluid, gas composition, geochemistry ABSTRACT Production of fluids from high-temperature (>200 C) geothermal reservoirs by deep drilling has provided extensive information on the origin and chemistry of these fluids, the associated hydrothermal alteration and various hydrological characteristics of such systems. Geothermal geochemistry is used to identify the origin of geothermal fluids and to quantify the processes that govern their compositions and the associated chemical and mineralogical transformations of the rocks with which the fluids interact. Six (6) geothermal wells from the Menengai Geothermal Field in Kenya have been used in this study to infer the reservoir fluids with respect to species composition. Steam fraction, is estimated to range from 0.1 0.5 of the total well discharge, with the fluids feeding the aquifer having a near neutral ph. Correlation of aqueous silica species with increasing temperature gives a linear pattern indicative of silica as a temperature indicator. Fluid gas concentrations show an increase with increasing aquifer temperature with the one phase wells having the highest aquifer concentration of CO 2, H 2 S and H 2. The CO 2 concentrations show a systematic correlation with respect to selected aquifer temperature, indicating that activity of dissolved CO 2 in the deep fluid is controlled by local equilibrium between geothermal solution and secondary minerals but this correlation can be disturbed by excessive CO 2 input due to magmatic intrusions in the roots of the geothermal system. the Oligocene (30 million years ago) and activity has continued to recent times. The last 2 million years saw the development of large shield volcanoes within the axis of the rift. These centres are the most important geothermal prospects within the rift given that the association between rifting and most of the occurrences of geothermal energy is mainly due to shallow magma chambers underneath the young volcanoes within the rift axis. Volcanic geothermal systems are commonly associated with ring structures or calderas. Kenya has 14 geothermal prospects of which three Olkaria, Menengai and Eburru - are proven geothermal fields. 1.0 Introduction High temperature geothermal systems are associated with central volcanoes and in Kenya, they are associated with the development of the Kenya Rift Valley (KRV). The Kenyan Rift Valley is a continental scale volcano-tectonic feature that stretches from northern to southern Africa. Development of the rift started during Figure 1. Map of the Kenyan Rift Valley showing the geothermal areas. 425
Menengai Geothermal Field (MGF) is located within an area characterized by a complex tectonic activity associated with a rift triple junction (Figure 1). This is at the zone where the Nyanza rift joins the main Kenya rift. A large area around the caldera is covered by mainly pyroclastics erupted from centers that are associated with the Menengai volcano. The caldera floor covers an area of about 88 km 2, and is partially covered by young rugged lava flows that are post caldera in age. The major structural systems in the area are the caldera, Molo tectonic axis and the Solai graben. Menengai caldera is elliptical with the major and minor axis measuring about 12 km and 8 km respectively. The ring structure has been disturbed by the Solai graben faults on the NE end and one fracture at the SSW of the caldera wall extending southwards. The Molo TVA/Ol rongai fracture system intersect Menengai caldera on the NNW part (Lagat et al., 2010). Most of the caldera infill lavas are from fissure eruptions that flowed out of the fracture openings (Figure 2). A total of twelve (12) geothermal wells have been drilled successfully in the MGF. Production of fluids from high-temperature (>200 C) geothermal reservoirs by deep drillings has provided extensive information on the origin and chemistry of these fluids, the associated hydrothermal alteration and various hydrological characteristics of such systems. Figure 2. Location of Menengai wells considered for the study. Geothermal fluids are also of interest as analogues to oreforming fluids and they provide important information to further understanding of fossil hydrothermal systems that have been exhumed by erosion. Geothermal geochemistry is used to identify the origin of geothermal fluids and to quantify the processes that govern their compositions and the associated chemical and mineralogical transformations of the rocks with which the fluids interact. The high-temperature, high-pressure, volatile-rich two phase fluids discharged from wet-steam wells in volcanic geothermal systems pose unique challenges in terms of chemical analysis and modelling of initial aquifer fluid compositions. Geochemical assessments of geothermal fluids have provided insights on current reservoir conditions in different parts of the world after prolonged production thus creating further understanding on the behaviour of wells in geothermal systems resulting in insights for formulating resource management (e.g. Angcoy, 2010; Arnórsson et al., 2010; Gudmundsson and Arnórsson, 2005; Karingithi et al, 2010; Scott, 2011). 1.1 Objective The main objective of this study is to characterize the fluid feeding the aquifers in the MGF and attempt to model the aquifer fluid of the Menengai geothermal field with respect to geochemistry as aided by discharge fluid (two-phase) analysis and solute/ gas geothermometers. 2.0 Location and Geological Setting The Menengai caldera is a shield volcano with a summit elevation of 2,278 m (7,474 ft) located at 0 12 0 S and 36 4 0 E. The massive Menengai shield volcano occupies the floor of the East African Rift. Construction of a 30 km 3 shield volcano beginning about 200,000 years ago was followed by the eruption of two voluminous ash-flow tuffs, each preceded by major pumice falls. The first took place about 29,000 years ago and produced a large caldera. The second major eruption, producing about 30 km 3 of compositionally zoned per-alkaline trachytic magma about 8,000 years ago, was associated with formation of the present-day elliptical 12 x 8 km summit caldera (Lagat et al., 2010). More than 70 post-caldera lava flows cover the caldera floor, the youngest of which may be only a few hundred years old. The caldera floor covers an area of about 88 km 2, and is partially covered by young rugged lava flows that are post caldera in age (Lagat et al, 2010). No historical eruptions are known from Menengai. Fumarolic activity is restricted to the caldera with a few fumaroles located to the north western part of the caldera (at Ol rongai area). The chronology of syn- and postcaldera events is based on correlation with dated fluctuations in the levels of nearby lakes, suggesting that the two ash-flows may have been erupted at about 29,000 and before 12,850 years ago. Lake sediments inside the caldera provide evidence for a late intra-caldera lake from about 10,300 to 8,300 years B.P. (Leat, 1984). The Menengai geothermal field is located within an area characterized by a complex tectonic activity associated with a rift triple junction. This is at the zone where the Nyanza rift joins the main Kenya rift. A large area around the caldera is covered by mainly pyroclastics erupted from centers associated with Menengai volcano. The major structural systems in the area are the Menengai caldera, Molo tecto-volcanic axis (Molo TVA) and the Solai graben. The ring structure has been disturbed by the Solai graben faults on the NE end and one fracture at the SSW of the caldera wall extending southwards. The Molo TVA / Ol rongai fracture system intersects Menengai caldera on the NNW part. Most of the caldera infill lavas are from fissure eruptions that flowed out of the fracture openings. 426
3.0 Methodology 3.1 Sampling and Analysis Sampling, sample treatment/preservation and chemical analysis of fluids from wellheads of geothermal boreholes require specific techniques in addition to normal procedures used for surface and non-thermal waters. These techniques ensure attaining representative and uncontaminated samples (considering the high temperatures of geothermal fluids and the effects of cooling or exposing the samples to the atmosphere) and are a first in the steps required to determine aquifer fluid composition. The subsequent steps require that initial aquifer fluid chemical compositions be calculated based on a model for boiling. The procedures are in principle similar to the details discussed by different authors (e.g. Angcoy, 2010; Arnórsson et al., 2010; and Karingithi et al., 2010) and in accordance with the laid down procedures by Armannsson and Olaffson (2006). The primary data for this study are obtained from chemical analysis of water and steam discharges from six wells of which four discharge two-phase: liquid and gas (i.e. MW-01, MW-03, MW-04 and MW-12) and two discharge one phase: superheated steam (MW-06 and MW-09) in the Menengai high-temperature geothermal field. The well samples were collected and analysed by the geochemistry staff of the Geothermal Development Company. A chromium steel Webre separator was used to collect water and steam samples from the pipeline close to the wellhead of each well. A cold water jacketed coil of stainless steel was attached to the separator by a teflon-coated steel tubing. Before sampling, some of the sample was pumped through the sampling/ filtration apparatus to clean the sampling line and remove any contaminants. Steam samples were collected into gas sampling bulbs, which had been evacuated in the laboratory and contained 20 to 50 ml of freshly prepared 40 % w/v NaOH solution. The strong base is used to capture the major non-condensable gases (CO 2 and H 2 S) while residual gases (H 2, CH 4, N 2 and O 2 ) occupy the head space. Samples for analysis of all components except for ph, CO 2 and H 2 S were filtered on site to prevent interaction with any suspended matter through 0.45 μm filter papers into low density polyethylene bottles using a polypropylene filter holder. For the determination of major cations, the samples were acidified with concentrated nitric acid, 1 ml into 100 ml of sample. A 100 ml sample was collected for sulphate analyses. To this sample 2 ml of 0.1 M zinc acetate solution was added to remove H 2 S as ZnS. The ZnS precipitate was filtered from the sample. Two 100 ml low density polyethylene bottles were used to collect samples for the determination of ph and CO 2. Samples for determination of Cl and F were not treated, except for filtration. The non-condensable gases in the headspace of the gas sampling bulb were analysed by gas chromatography. H 2 S was determined onsite titrimetrically using mercuric acetate and dithizone (Ármannsson and Ólafsson, 2006; 2007; Arnórsson, 2000). CO 2 concentration and ph in the liquid phase were determined in the laboratory immediately upon return from the field by potentiometric titration and a calibrated ph electrode respectively. SiO 2, SO 4 and B were analysed by the UV Vis spectrophotometric method, while the major aqueous components Na, K, Mg, Ca, F, Al and Fe were analysed by the ISE method. Cl was analysed titrimetrically while TDS was measured using a TDS meter. It can be summarised that the analyses of water samples are grouped into onsite/immediate analysis (ph, CO 2, H 2 S), and laboratory analysis of major elements and some minor elements. Thirty five (35) samples from the six wells have been used (selected for the requirements of this study). 4.0 Discussion 4.1 Aquifer Fluid Temperature Many chemical and isotopic geothermometers are used to estimate the aquifer temperatures beyond the zone of secondary processes like boiling, cooling and mixing on the basic assumptions that the sampled fluids are representative of the undisturbed aquifers where local equilibrium conditions are achieved. Actual downhole measurements may or may not agree with geothermometers. In this study, the selected aquifer temperature is derived from the geothermometer developed by Fournier and Potter (1982), for two-phase wells, and an average of the CO 2, H 2 S and H 2 gas geothermometers developed by Arnórsson and Gunnlaugsson (1985), for the one-phase wells. The temperature logs of the main aquifers generally compare well with the calculated geothermometers for each well s sample(s), considering that some of the wells seem to have more than one producing aquifer (GDC unpublished report, 2012) From these selected aquifer temperatures (T f ), the aquifer fluid compositions were calculated using the WATCH 2.4 program (Arnórsson et al., 1982), version 2.4 (Bjarnason, 2010). The wells display temperature ranges of 180 C to 330 C, with an average of ~250 C. 4.2 Aquifer Fluid Composition WATCH 2.4 (Arnórsson et al., 1982 and Bjarnason, 2010) was used to infer the initial aquifer fluids. WATCH 2.4 is suited to handle geochemical data from wet-steam wells. The program reads the chemical analyses of water and gas samples collected at the wellhead then computes for the chemical composition at the selected reference temperature (T f ) including ph, aqueous speciation, partial pressure of gases, redox potentials and activity products of mineral dissolution reactions. No mineral precipitation or dissolution is assumed as the fluid is modeled from the aquifer to the surface. Another assumption considered is that chemical reactions do not change the fluid component concentrations and that reactions of fluid with casing material and or wellhead installations do not occur. To model the aquifer composition of the selected wells the analytical data were input into WATCH 2.4, and then iterated until the total silica concentration in the liquid aquifer was consistent with T f. Figure 3 to Figure 5 show the concentrations of various parameters with respect to the selected reference temperature (T f ). Steam fraction, X d, v, estimated at point of collection (using the selected aquifer temperatures) range from 0.1 0.5 of the total well discharge (Figure 3). The fluids feeding the aquifer have a near neutral ph indicative of good chemistry for fluid utilization. Measured ph of discharged samples affects the calculated initial aquifer ph, which also affects the relative concentrations of the carbonate and sulphide-bearing species (Angcoy, 2010). For the wells considered in this study, there seems to be a decrease of CO 2 with increasing ph of the aquifer fluid. Correlation of aqueous 427
silica species with increasing temperature gives a linear pattern indicative of silica as a temperature indicator. Fluid gas concentrations show an increase with increasing aquifer temperature with the one phase wells (MW-06 and MW- 09 having the highest aquifer concentration of CO 2, H 2 S and H 2 (Figure 5). The CO 2 concentrations show a systematic correlation with respect to selected aquifer temperature, indicating that activity of dissolved CO 2 in the deep fluid is controlled by local equilibrium between geothermal solution and secondary (2 ) minerals but this correlation can be disturbed by excessive CO 2 input due to magmatic intrusions in the roots of the geothermal system. The wells display high concentration of H 2 gas indicative of equilibrium vapor in the aquifer. This might also be indicative of steam condensation of reservoir gases in an upper aquifer leading to increased non-reactive gases e.g. H 2, in the discharge. The gases discharged from MW-06 reservoir are highly variable. This might be attributed to the little time of discharge and therefore the well had not stabilized. It is however expected that the well discharges will stabilize with time and more so once the production stage embarks. 5.0 Conclusions From the discussions in Section 5.0, it can be concluded that a. Menengai reservoir temperatures are estimated to be >200 C. b. Fluids are of bicarbonate type. c. Aquifer fluids have a near neutral ph good chemistry for fluid utilization. CO 2 H 2 S Figure 3. Steam fraction (X d,v ) at the point of collection of Menengai well discharges. H 2 Figure 4. Aquifer silica concentration of the Menengai wells. Figure 5a, b, c. Aquifer gas (CO 2, H 2 S and H 2 ) concentration of the Menengai wells. 428
d. Close to equilibrium of calcite in the aquifer fluid. e. Steam fraction (X d,v ): 0.1 0.5 of the sampled discharge. References Ármannsson, H. and Ólafsson, M. 2006: Collection of geothermal fluids for chemical analysis. ÍSOR, Reykjavík, report ÍSOR-2006/016, 17 pp. Angcoy, E.C. Jr., 2010: Geochemical modelling of the high temperature Mahanagdong geothermalfield, Letye, Philippines. University of Iceland, MSc Thesis, UNU-GTP, Iceland, report 1, 71 pp. Arnórsson, S. (editor), 2000: Isotopic and chemical techniques in geothermal exploration, development and Use. IAEA. 362pp Arnórsson, S., Angcoy, Jr., E.C., Bjarnason, J.Ö., Giroud, N., Gunnarsson, I., Kaasalainen, H., Karingithi, C., and Stefánsson, A., 2010: Gas chemistry of volcanic geothermal systems. Proceedings World Geothermal Congress 2010, Bali, Indonesia. Arnórsson, S. and Gunnlaugsson, E., 1985: New gas geothermometers for geothermal exploration - calibration and application. Geochim Cosmochim Acta, 49, 1307-1325. Arnórsson, S., Sigurdsson, S., and Svavarsson, H., 1982: The chemistry of geothermal waters in Iceland. I. Calculations of aqueous speciation from 0 to 370 C. Geochim. Cosmochim. Acta, 46, 1513-1532. Bjarnason J.Ö., 2010: The speciation program WATCH, version 2.4, user s guide. Iceland water chemistry group, Reykjavik, 9 pp. Fournier, R.O., and Potter, R.W.II, 1982: A revised and expanded silica (quartz) geothermometers. Geotherm. Res. Council, Bull., 11, 3-9. Karingithi, C. W., Arnórsson, S., and Gronvold, K., 2010: Processes controlling aquifer fluid compositions in the Olkaria geothermal system, Kenya. J. Volcanol. Geoth. Res., 196, 57-76. Lagat, J., P. Mbia, C. Mutria, 2010. Menengai prospect: Investigations for its geothermal potential. GDC internal report (2010), 64 pp. Leat, P.T., 1984: Geological evolution of the trachytic volcano Menengai, Kenya Rift Valley. J Geol Soc London, 141: 1057-1069. Scott, S.W., 2011: Gas chemistry of the Hellisheidi geothermal field. University of Iceland, MSc Thesis. REYST report 08-2011, 69 pp. Appendix Table A1. Menengai well discharges. MW-06 and MW-09 have one-phase discharges. Well No. Sample ID Date of sampling WSP (bar-g) Water Sample (mg/kg) Gas sample (mmol/kg) CO 2 H 2 S B SiO 2 Na K Mg Ca F Cl SO 4 Fe TDS CO 2 H 2 S H 2 CH 4 N 2 O 2 MW-01 315 23-Oct-12 2.8 9.1/24.6 6323 39 1.0 401 3707 147 0.03 86 641 7270 2342 1.1 14.2 0.0 10.1 1.8 MW-01 207 18-May-12 2.1 8.9/22.9 7260 20 197 4577 267 1.98 119 789 154 8140 2827 3.6 9.6 2.9 7.0 0.8 MW-01 348 6-Dec-11 3.1 8.8/24.8 7920 12 342 4202 235 111 497 184 8010 4441 3.2 5.9 14.7 15.9 0.0 MW-01 340 29-Nov-11 2.4 9.0/24.4 7348 11 346 4042 280 110 533 192 8120 4289 3.6 4.6 16.1 70.9 14.2 MW-01 238 10-Oct-11 0.3 9.4/20.0 7656 8 1.7 289 4407 303 0.23 2.39 115 568 255 0.41 8470 2710 2.8 2.4 4.2 20.4 0.0 MW-01 205 28-Sep-11 0.2 9.2/20.0 7612 17 330 4436 298 123 497 243 8760 1564 11.0 30.9 2.4 4.3 0.0 MW-01 193 14-Sep-11 0.6 9.2/20.0 7128 13 309 4065 304 0.22 2.10 127 710 380 0.41 8890 1479 12.8 29.4 1.9 1.8 0.0 MW-01 186 5-Sep-11 0.3 9.3/20.0 7414 18 1.9 292 4424 276 0.23 2.38 109 639 245 0.46 7650 571 8.7 7.8 0.4 0.9 0.0 MW-01 180 15-Aug-11 0.4 9.2/20.0 7590 13 4.4 311 4427 303 0.21 2.32 120 639 242 0.52 7615 860 11.2 14.4 1.0 2.2 0.0 MW-01 172 4-Aug-11 0.2 9.0/20.0 8140 25 316 4760 298 124 568 247 7415 844 13.2 12.6 1.7 1.7 0.0 MW-01 160 22-Jul-11 0.2 9.0/20.0 7436 43 306 4509 323 106 675 212 7185 1054 12.7 17.7 2.2 3.3 0.0 MW-01 150 13-Jul-11 0.1 7.2/20.0 7040 53 315 4249 266 122 355 183 6420 847 6.4 11.7 1.2 1.6 0.0 MW-01 139 29-Jun-11 0.1 9.2/20.0 6820 38 3.9 311 4201 247 0.14 1.73 118 710 271 0.57 7300 537 12.5 12.3 1.1 1.4 0.0 MW-01 130 14-Jun-11 0.1 9.1/20.0 6820 37 334 4264 289 108 675 320 7285 728 12.2 18.6 1.7 3.9 0.0 MW-03 417 28-Feb-13 0.3 9.7/20.8 4370 10 0.4 199 3084 98 0.13 108 999 431 6670 1181 2.4 6.6 20.0 21.0 0.2 MW-03 324 14-Nov-12 0.3 6.9/24.5 4268 14 0.6 183 2903 79 0.09 85 813 344 5670 763 0.4 4.6 10.0 20.0 3.8 MW-04 195 7-May-12 2.1 9.3/21.1 5060 172 296 3201 192 0.08 1.56 151 1022 257 7.13 7240 1064 19.4 84.3 8.7 1.8 0.0 MW-04 120 5-Apr-12 2.8 9.3/20.1 4960 253 473 3242 131 0.09 2.59 141 1040 287 3.82 7010 1536 54.0 79.9 5.5 15.9 0.0 MW-04 1 3-Jan-12 1.4 9.2/24.1 5830 187 1.0 347 3973 100 0.00 157 1172 261 7340 1789 20.0 91.7 4.2 21.0 0.0 MW-04 363 29-Dec-11 1.4 9.3/21.5 6446 119 1.8 348 3975 133 3.66 166 1065 246 7100 1036 18.9 70.7 3.1 15.9 0.0 MW-04 317 14-Nov-11 0.1 9.2/21.5 4840 118 1.8 351 3719 141 0.00 155 888 448 7710 1247 31.1 63.0 3.1 17.2 0.0 MW-12 375 11-Feb-13 2.8 8.9/24.8 2517 122 0.8 310 1826 14 0.07 55 647 391 3870 1708 31.2 142.0 7.4 35.1 3.6 MW-06 206 18-May-12 0.7 1805 64.9 138.0 20.7 6.0 0.0 MW-06 194 4-May-12 1.7 1604 44.7 121.0 17.4 2.0 0.0 MW-06 181 27-Apr-12 2.8 1820 57.7 140.0 19.2 0.0 MW-06 153 18-Apr-12 3.1 1734 53.6 158.0 20.9 1.0 0.0 MW-06 140 13-Apr-12 3.8 1575 56.0 156.0 20.7 0.0 MW-06 123 10-Apr-12 2.8 1437 42.9 146.0 21.7 0.0 MW-06 121 5-Apr-12 3.1 1733 79.1 118.0 16.2 44.0 0.0 MW-06 111 28-Mar-12 2.8 1189 45.8 169.0 23.7 0.0 MW-06 76 16-Mar-12 3.8 1417 31.6 181.0 28.3 0.0 MW-06 57 9-Mar-12 2.8 1691 37.0 48.0 19.7 0.0 MW-06 54 5-Mar-12 3.6 1671 40.9 47.0 7.2 0.0 MW-09 328 29-Nov-12 4.0 612 30.1 51.0 0.3 31.0 0.7 MW-09 326 26-Nov-12 4.0 779 31.2 25.0 44.7 17.0 1.5 429
430