Surface and groundwater hydrochemistry in the mid-Gregory Rift , Kenya : first impressions and potential implications for geothermal systems

The University of Glasgow has a long tradition of scientific endeavour in the Gregory Rift Valley. This paper details some of the history and inspiration behind current hydrological efforts and details results from a 2016 field excursion to this region. A range of surface and ground waters were sampled and analysed for physical, chemical, and stable isotope composition as scoping investigation into geothermal-related hydrological systems. The results allow us to make some initial observations that will be followed up by additional multi-seasonal data collection. Our initial results show clear chemical and isotopic signals for river, lake, hot spring and Menengai geothermal well waters. 1 The Gregory Rift Valley and the University of Glasgow The Gregory Rift Valley (GRV) represents the section of the East African Rift System (EARS) that dissects Kenya and Tanzania [1]. This part of the EARS is named after John Walter Gregory, who carried out the first scientific expedition there in 1892-1893 [2], and who would later take up a 25-year tenure as the Chair of Geology at the University of Glasgow (UoG) from 1904 onwards [3]. Gregory is most famous for his recognition and description of a new type of topographic depression, which he recognised as a result of tectonic separation and subsidence, which he called a ‘Rift Valley’ [4]. In the years since, much work has been carried out to decipher the geological evolution of the EARS, and assess its geothermal power potential. Drilling investigations in the GRV started in 1956, 75 km NW of Nairobi, and by 1981 the Olkaria I power plant was producing 15 MW. Since then over 100 wells have been drilled, the number of power stations has grown to four and capacity has grown to 569 MW [5]. Further exploration and testing is underway to develop multi-MW projects at Eburu, Longonot, Suswa, and Menengai in order to exploit Kenya’s estimated reserves of up to 10 GW [6]. The UoG took a greater interest in Kenyan geothermal development in 2012 with the appointment of Prof. Paul Younger to the Rankine Chair of Energy Engineering. Younger was particularly enthusiastic about advancement of hydrogeological cycle knowledge in regions with high geothermal potential but significant hydrological vulnerabilities such as * Corresponding author:neil.burnside@glasgow.ac.uk © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). E3S Web of Conferences 98, 07004 (2019) https://doi.org/10.1051/e3sconf/20199807004 WRI-16


The Gregory Rift Valley and the University of Glasgow
The Gregory Rift Valley (GRV) represents the section of the East African Rift System (EARS) that dissects Kenya and Tanzania [1].This part of the EARS is named after John Walter Gregory, who carried out the first scientific expedition there in 1892-1893 [2], and who would later take up a 25-year tenure as the Chair of Geology at the University of Glasgow (UoG) from 1904 onwards [3].Gregory is most famous for his recognition and description of a new type of topographic depression, which he recognised as a result of tectonic separation and subsidence, which he called a 'Rift Valley' [4].
In the years since, much work has been carried out to decipher the geological evolution of the EARS, and assess its geothermal power potential.Drilling investigations in the GRV started in 1956, 75 km NW of Nairobi, and by 1981 the Olkaria I power plant was producing 15 MW.Since then over 100 wells have been drilled, the number of power stations has grown to four and capacity has grown to 569 MW [5].Further exploration and testing is underway to develop multi-MW projects at Eburu, Longonot, Suswa, and Menengai in order to exploit Kenya's estimated reserves of up to 10 GW [6].
The UoG took a greater interest in Kenyan geothermal development in 2012 with the appointment of Prof. Paul Younger to the Rankine Chair of Energy Engineering.Younger was particularly enthusiastic about advancement of hydrogeological cycle knowledge in regions with high geothermal potential but significant hydrological vulnerabilities such as the GRV, where structure-magmatism-water connections can have a significant impact on vulnerable populations.In his travels to the GRV and the Main Ethiopian Rift he observed anomalous, non-precipitation-driven lake flooding and related salinization of rivers that provide essential fresh water resources to local populations.He theorised that permeable structural links between deep geothermal waters and fresh surface water bodies could be playing a considerable role.Advised by Younger, a University of Glasgow team set out to investigate the geothermal systems within the GRV from 28 June to 5 July 2016.This paper details the results of this trip and sets a baseline for future hydrological efforts in the region.

Methodology and Results
Various surface and ground water samples were collected from across the GRV (Fig. 1).Prior to sampling, field parameters were measured on site.Field, sampling and analytical methodology are available in [7].In addition, two rain water samples were collected from the field trip accommodation at Maili Saba Camp (5 km NNE of Nakuru; Fig. 1).
Hydrochemical results showed distinct variations between lake, river, hot spring (HS) and Menengai geothermal well (MW) samples (Table 1; Fig. 2).Bogoria (pH ~ 10) and Elementaita (pH ~ 9.7) are soda lakes rich in blue-green algae and have geyser activity.Specific electrical conductivity (SEC) measurements from two Bogoria HSs (~ 6.6 mS/cm) and 16 HS pools on Elementaita (~ 4 mS/cm) -of which three were sampled (K35 to K37) -were considerably lower than lake water values (~ 44.7 and ~ 12.0 mS/cm respectively), so any geothermal impact assessment on these waters is challenging to assess.The high alkalinity of Lake Bogoria consumed all available acid during field titration, so the K18 value was used as a proxy for plotting of subsequent Bogoria samples (Fig. 2).Geothermal water temperatures where beyond limit of measuring equipment (> 75°C), measured pH 8.5 to 9.3, and have an enriched Na/Ca ratio in comparison to most surface waters, suggesting ion exchange with Na + bearing host rocks.Lakes Nakuru, Bogoria and Elementaita are hypersaline in nature and have equivalent, or greater, ionic values and Na/Ca ratios than geothermal waters due to evaporation driven solute concentration.The Hand O-isotopic data (Table 1; Fig. 3) illustrate that rain and river waters are predominately associated with the GMWL, with rain at isotopically enriched in D and 18O relative to river and geothermal waters, whilst lake and geothermal waters fall along the evaporative trend defined by the Kenya Meteoric Water Line (KMWL) calculated using January/February rainfall [9].Baringo HS2 (K9) seeps from unconsolidated sediments on the shores of the lake and represents a mixture of geothermal and lake waters.

Conclusion
Geothermal waters are of high enough concentration that any significant volumetric discharge to the surface would likely result in noticeable perturbations in fresh water (i.e.Baringo and Naivasha) lake salinity.Lakes Bogoria and Elementaita waters, however, have such high ionic strengths that even substantial ingress of geothermal fluids would be difficult to detect.Lake Nakuru has significant ionic enrichment versus Lakes Baringo and Naivasha; and is similar -in terms of chemical composition -to Menengai waters.Thermal waters and lakes show a clear difference in lighter (thermal) and heavier (lake) isotopic populations, bar one geothermal well and hot spring outlier.The hot spring (BAR HS2) is on the shore of Lake Baringo and has a similar isotopic signature, indicating extensive mixing of thermal and lake waters.The geothermal well sample (MW-09) plots close to isotopic values for Lake Nakuru.When coupled with ionic data, the initial isotopic evidence related to the Menengai Geothermal field and Lake Nakuru looks to be in support of Younger's lake level fluctuation hypothesis and suggests permeable connections between deep thermal and surface waters in this location.Evidence for significant interaction between thermal and lake waters at other sampling locations is inconclusive.A definitive assessment is difficult using this 2016 data set, which represents a single point in time and cannot account for any seasonal variation.However, the data presented here forms a powerful baseline that shows clear distinctions between deep thermal and surface waters that lays the foundations for future investigations.Further work is underway, including inter-seasonal data, to assess potential geothermal fluid flow pathways and the degree of interconnectivity with surface water bodies across the GRV.

Fig. 2 .
Fig. 2. Piper plot of anion and cation data for waters of the Gregory Rift Valley.