GRC Transactions, Vol. 42, 2018

Early Stage 3D Model Construction for Well Planning: A Case Study from Barrier Volcanic Complex, Kenya.

Abraham., S1, Kimani., F1, Njau, K1, Baxter., C2, O’Brien., J3 1Olsuswa Energy, Mayfox House, P.O Box 14991 – 00800, Garden Road off Riverside Drive, Nairobi, Kenya 2Seequent UK Limited, Building 1, Chalfont Park, Gerrard’s Cross, Buckinghamshire SL9 0BG, United Kingdom 3Seequent Limited, 20 Moorhouse Ave, Addington, Christchurch 8011, New Zealand

Keywords Leapfrog, Olsuswa Energy, Barrier Volcanic Complex, Conceptual Model, Well.

ABSTRACT

Integration of surface data and inferred subsurface structure is a crucial part of geothermal resource delineation prior to beginning drilling campaigns. Inputs at this stage often include geological maps and cross sections, geochemical data, lidar or land survey data, and geophysical data. The integration of this early information informs the conceptual model and therefor drilling targets. Often this process is completed in a 2D environment where sometimes spatial context can be left behind. This paper highlights how incorporating surface exploration data into a 3D environment at the earliest stage can shed light on resource uncertainty and key features at the Barrier Volcanic Complex in Kenya. The Barrier Volcanic Complex (BVC) is a shield volcano located in Turkana, Kenya. The volcano Last erupted in 1921 and is the northernmost geothermal prospect in Kenya, lying in the Gregory Rift Valley at 2° 20’N, 36° 37’E. The volcanic complex forms a natural dam between Lake Turkana and the Suguta Valley. North to South topographic profiles indicate that this 20km long and 15 km wide, E-W trending ridge is a broad symmetrical feature with gently sloping flanks. BVC is a composite structure composed of four distinct volcanic centers namely; Likaiu East, Likaiu West, Kakorinya, and Kang’olenyang’. In the build up to the 3D and spatial integration processes, Olsuswa utilised data from earlier geoscientific works by the British Geological Survey (BGS) and Kenya’s Geothermal Development Company (GDC). This was in addition to data from Olsuswa’s detailed surface exploration program. A three dimensional representation of the system has helped with Abraham et al.

communicating with stakeholders and delineating key subsurface elements key to planning the initial exploration drilling strategy.

1. Introduction The Barrier Volcanic Complex is located in Turkana, Kenya and lies in the Gregory Rift Valley at 2° 20’N, 36° 37’E (Figure 1). Barrier volcanic complex is the northernmost geothermal prospect in Kenya at the southern shores of Lake Turkana. It is a complex composite of four volcanoes, namely; Kakorinya, Kang’olenyang’, Likaiu West and Likaiu East.

Figure 1: Map showing the location of the Barrier Geothermal Field.

Except in Kakorinya’s main caldera formation where pyroclastic deposits constitute the youngest rock types, a suite of basalt, hawaiite, mugearite, and trachyte lava flows constitute the recent lithologic formations in the other three volcanic centers. The project site covers the central area of the BVC surrounding the Kakorinya volcano. This area is transacted by a series of curvilinear N-NE trending normal faults which extend in an en echelon right-stepping fashion across the volcano. Geothermal surface manifestations in the complex are fumaroles, hydrothermally altered grounds, hot springs, silica sinter, and geothermal grass. Estimated gas geothermometry temperatures have given mean subsurface temperatures of > 281°C. Abraham et al.

This paper aims to investigate how incorporating surface exploration data into a 3D environment at an early stage. Historically collected data by BGS and GDC as well as local understanding of the area were used to infer the sub surface lithological 3D model in Leapfrog Geothermal. The model has helped delineating key subsurface elements key to planning the initial exploration drilling strategy.

2. Geothermal System in BVC Field 2.1 Geological Setting The geology of the BVC complex is comprised of a wide spectrum of lava types including basanite, basalt, hawaiite, mugearite, benmorite, trachyte and phonolite. Trachytic pyroclastic deposits cover much of the western slopes of Kakorinya and the summit area of Likaiu West. The oldest exposed rocks of the BVC are massively faulted porphyritic olivine basalts. These Pliocene foundation rocks were dated at about 4.53 Ma (Dunkley et al, 1993) and are well exposed in the adjacent rift margins. They cover most of the east and are called the Parkati Basalts. They occur in the faulted ground far east of Lake Logipi and around Latar, Southeast of Lake Turkana. The oldest lavas on the west are Lotikipi Basalts dated to be between 4.0-1.86Ma, (Dunkley et al, 1993). As observed by Dunkley et al. (1993), trachytic volcanism constructed the centres of Kang’olenyang’ and Likaiu East and major trachytes lavas are exposed within the inner trough. The youngest trachytes form the domes to the west of Kakorinya caldera around the caldera rim and these trachytes are dated to be about 0.05 Ma (Dunkley et al, 1993). Upper trachytes of 0.09 Ma (Dunkley et al, 1993) are exposed in the east of the caldera running in a N-S direction, with a minor outcrop of this formation being noted in the north west with a strip in the southwest. Older trachytes are exposed closer to the flanks in the east and are dated to be about 1.37 Ma (Dunkley et al, 1993). Pyroclastics cover most parts of the western side of Kakorinya while spots of alluvial sediments are scattered within this area. Some alluvials are found on top of cones inferring that the lake levels were much higher than they are at present. The youngest pyroclastic deposits on Kakorinya are airfall pumice lapilli and are best exposed in a thick wedge, which infills the western dipping slope between the caldera rim and the outer ring fractures. These deposits bank against and mantle the ring fracture escarpments and the pre-caldera domes. They are cut by the caldera wall in the west and bury the northern wall. This relationship indicates that the eruption of these trachytic tuff was broadly contemporaneous with the caldera collapse. Lacustrine sediments provide evidence of the existence of former Lake Suguta, which infilled the inner trough northwards from Emuruangogolak. The latest eruption in the area is historic and occurred in 1921. The erupted material was scoria basalts of Teleki Cone, which is still fresh and unvegetated (Figure 2).

Abraham et al.

Figure 2: Geological map of the Barrier Volcanic Complex and adjacent areas.

2.2 Structural Setting

Studies of the structural setting of the Barrier volcanic field and associated Kakorinya volcano were presented by Dunkley et al. (1993). These structures are part of the rift floor structural system. Structurally, the BVC is characterized by faults, steep ridges and four eruption centres. The general trend of the faults in the area are N-S and NNE-SSW, which is consistent with the regional stress of the area. The mostly faulted areas are to the east, from the south eastern tip of Lake Turkana where dense faults trending NNE-SSW intersect with the eastern rim of Kakorinya Caldera. Southeast of the rim, the faulting takes a sudden turn to the south. The western half of the complex is less intensely faulted, with the faults trending NNE-SSW just southwest of Kakorinya and NNW- SSE around Kang’olenyang’ volcano.

Abraham et al.

The heat source is postulated to be bound by the caldera system at Kakorinya. This is further supported by the higher frequency of surface manifestations around this caldera than around the other three volcanic centres within the area. 2.3 Surface Manifestations

Dunkley et al. (1993) delineated the thermal manifestations in the area. They include hotsprings, fumaroles, altered grounds, silica sinters and Fimbristylis exilis (“Geothermal grass”). Geothermal manifestations generally occur along a series of NNE-trending faults and fissures on the caldera floor and also around the caldera walls in the east and southeast (Figure 3). The hottest and most vigorous activity is associated with the caldera ring fractures and trachytic lava domes in the west of the caldera. The maximum recorded fumarole temperature is 96.4°C and occurs within the caldera floor (GDC, 2011). Outside the caldera there are a few fumaroles which exhibit low temperatures of up to 78.2°C (GDC, 2011). Hot springs frequently occur along the northern shores of Lake Logipi and around the eroded tuff cone of Naperito, although some of the hot springs occur under the lake (Figure 3). A maximum temperature of 70.0°C has been recorded for the hot springs (Dunkley et al, 1993). Silica sinters are common at many of the geothermal areas on Kakorinya and indicates former hot spring activity. The most spectacular development of sinter occurs on the trachyte lava domes cut by southwest wall of the caldera. They have sub-vertical dips, strike 008-028° and are parallel to a series of open fissures and faults. The veins extend up to the top surface of the lava domes where mounds of botryoidally sinter occur. Outside the caldera chalcedonic silica veins occur at several localities where the ring fracture meets a northeast-trending fault at weak fumaroles and steaming ground. Thermal indicators extracted from remote sensing data inform the PC4 image indicate the presence of hydrothermal alteration minerals southeast of Kakorinya caldera and in the Kang’olenyang’ area (Mutua et al. 2011). Thermal infrared imagery also shows thermal areas within Kakorinya area.

3. Historical Work Two historical studies have been used by Olsuswa to define the conceptual model and delineate the investigatory area. The BGS in 1993 conducted reconnaissance surveys which revealed the occurrence of strong surface manifestations. Subsequently during 2011, GDC carried out a preliminary reconnaissance surface exploration studies to establish the geothermal potential of the prospect (GDC, 2011). GDC used several geoscientific methods during the preliminary field surveys including, geophysical methods, utilizing resistivity techniques using Transient Electromagnetic (TEM) and Magnetotellurics (MT).

Abraham et al.

Figure 3: Location of geothermal activity on the summit of Kakorinya (Dunkley et al. 1993)

Geochemical techniques included the collection of gas and steam condensate samples from fumaroles to determine the nature of the geothermal reservoir. The results of the preliminary geoscientific surveys by GDC indicate the presence of a geothermal resource under the BVC. The heat source is understood to still be active as indicated by recent lavas which erupted in 1921. The heat source is associated with shallow magmatic intrusives beneath the volcanic complex. Estimated gas geothermometry has provided mean subsurface temperatures of over 281ºC (GDC, 2011).

4. Method for Data Visualisation and Constructing a 3D Model None of the geothermal areas surrounding the Kakorinya volcano have been explored for their geothermal potential. While historical surface data studies have been conducted there has been very little interpretation done on the subsurface. In order to construct the 3D model many inferred interpretations had to be made and in some instances surfaces were assumed vertical. This puts a lot of uncertainty into the interpreted output of the 3D model. Abraham et al.

The 3D conceptual model was built using Leapfrog Geothermal software. Leapfrog Geothermal uses an implicit method of modelling which utilises radial basis function (RBF) algorithms to extrapolate the surfaces (Alcaraz et al. 2011). The following highlights the workflow used to build the 3D conceptual BVC model in Leapfrog Geothermal. 1. Importing data into Leapfrog a. LiDAR, geological, surface and 1D resistivity maps were imported into Leapfrog Geothermal. 2. Define Topography a. LiDAR point data was used to construct a surface from a Delaunay triangulation calculation. 3. Determining the boundary area a. Given Olsuswa is at an early stage of exploration, the model boundary was defined by two general areas of interest. To the south it contains the Kakorinya caldera where the dominant surface manifestations lie and to the north, the boundary incorporates von Hahnels Bay and the Recent Phonolites. 4. Building the fault structure a. Four fault surfaces are perceived as being high importance. These surfaces were created by digitising using the surface data and due to uncertainty inferred to be vertical. 5. Building Lithological contact surfaces a. Aside from the geological map no other external data has been collected to define the lithological units. GIS lines were digitised in Leapfrog Geothermal tracing the contacts at the map and then adjusted using polylines and structural data to adjust the contacts to be more realistic. 6. Creating an MT model a. A geophysical survey was conducted by Geothermal Development Corporation. The 10ohm contour was digitised on the imported 1D inversion maps and interpolated to create a 3D surface.

5. Conceptual 3D model 5.1 Lithological Model Subsurface information was inferred from the surface geological map, unit descriptions and local knowledge of the area. To construct the lithological model GIS lines were traced along the map to define the contacts of the different lithologies at surface while Leapfrog Geothermal polyline and structural data editing tools were used to define the geometry of the formation to be geologically reasonable. Abraham et al.

Figure 4: Map and angled view of Leapfrog Geothermal 3D model showing inferred sub surface geometry.

While the Lower Trachytes have few scattered outcrops to the northern part of the model, the Upper Trachyte Lavas are well featured. Being the oldest and most abundant rock they were defined in the model as a volume only leaving their contacts to be defined by younger lithological surfaces. Consistent with a thickness of more than 100m of welded and non-welded pyroclastic flow deposits, breccias and air-fall pumice lapilli tuffs. The Lower Pyroclastic, through cross bedding show low angle cross-stratification (Dunkley et al. 1993) and course debris flow breccias occur near the base. Conversely the youngest pyroclastic deposits are air fall pumice lapilli tuffs. Exposed in the upper part of the south-west wall of the caldera, the Trachyte Lava Domes typically have carapaces of glassy, vesicular breccia (Dunkley et al. 1993). Intruding through the older rocks typically aligned with the zone of NE-trending faults, these are inferred to be vertical as many are bounded by faulting. To the north-west of the model, thin lava flows of basalt and basanite sit above the Upper Trachytes (Dunkley et al. 1993). Both the Lower and Upper Basalts were modelled to have the same shape with a thickness of about 5m as defined by (Dunkley et al. 1993) and referencing the contours of the topography surface. The Lower Basalts lie under the Upper Basalts in places. The most recent lava flow activities also pose an interest as it is inferred to be a potential source for the Barrier Geothermal System. 5.2 Structural Model The Kakorinya has two outer caldera ring faults. The inner ring structure was incorporated into the geological model to separate it into two fault blocks. This allowed the pre-dated formations to be truncated by the caldera ring.

Abraham et al.

Figure 5: Leapfrog Geothermal 3D model of topography showing geology map and fault structure and major fault surfaces.

Prior to the formation of the caldera an important phase of faulting influenced the Kakorinya volcano. This consists of a series of NNE-trending normal faults. Two dominant faults which link into the Southern part of the caldera and show associated Hot Grounds/Fumarolic Activity (Mutua et al.) were modelled as potential drilling targets. GIS lines were drawn to trace the map and the surfaces projected vertically down to define the faults. 5.3 Resistivity Model GDC conducted a magnetotellurics survey in 2011 of 18 MT and 8 TEM stations and estimated a high potential resource area to be roughly 160 km2 delineated by an area of resistivity of less than 30 ohm meters (Gichira et al, 2011). Olsuswa Energy’s concessional license widely overlaps the high potential area delineated by the MT. For the initial conceptual model, the 1D inversions were imported into Leapfrog Geothermal and the 10 ohm contour was digitised to define its volume. The low resolution of the stations did mean the 1D inversions did not align in 3D and some polylines had to be inferred or ignored to assist in creating a reasonable shape. The model shows a zone of shallow low resistivity which aligns to the two main NE trending faults. Abraham et al.

6. Implementation of Model for Well Planning Purposes Successful drilling of geothermal wells is best achieved through a comprehensive and up to date conceptual model of the system under scientific study. Historical reconnaissance surveys have so far focused on the interpretation of the surface. The inferred 3D model, although contains many uncertainties, has provided an initial subsurface context. By incorporating the 3 planned exploration wells into the model at an early stage, as the conceptual model updates and with new collected data it is possible to quickly update and refine the model to gain more certainty on its well paths and well target locations. Olsuswa aims to carry out a detailed surface study with proposed geological, geochemical, geophysical, structural analysis and seismic. This data will be reviewed and incorporated into the current 3D model to dynamically update and increase confidence in the conceptual model.

By having a history of the 3D conceptual model integrated with data collected, Olsuswa can use this to showcase to its stake holders the results of consecutive studies and refine the well targets for the 3 exploration wells. The drilling prognosis tool in Leapfrog Geothermal will also be applied to maximise the pre planning of drill engineering needs. Abraham et al.

7. Conclusion The BVC has no prior drilling conducted. This paper demonstrates how a 3D conceptual model, constructed in Leapfrog Geothermal, can be constructed at the very early stages of exploration by integrating historical studies of surface data to infer the subsurface. Although this model has many uncertainty to it, by combining all data into the one platform it helped with communicating with stakeholders and delineating key subsurface elements key to planning the initial exploration drilling strategy.

References

Alcaraz, S., Lane, R., Spragg, K., Milicich, S., Sepulveda, F., Bignall, G., (2011) “3D Geological Modelling Using New Leapfrog Geothermal Modelling Software”, Proceedings 36th Workshop on Geothermal Reservoir Engineering, Stanford, California, January 31 – February 2, 2011. SGP-TR-191 Dunkley P.N, Smith M, Allen D J, Darling W G., (1993) “Geothermal activity and geology of the northern sector of the Kenya Rift Valley”, Kenya Ministry of Energy, British Geological Survey Research Report SC/93/1., 1-180 Geothermal Development Company (GDC), (2011) “Remote sensing application in geothermal exploration, case study of barrier volcanic complex”, Geothermal Development Company, Internal report 943-48 Geothermal Development Company Ltd (GDC), (2011) “A Geothermal Resource Assessment Project report for the Barrier Volcanic Complex geothermal prospect”. Internal Geothermal Development Company (GDC) Report. 1-46 Gichira J., Simiyu C., Kangogo D., Yussuf Noor Y., Mwakirani R. and Wamalwa A. (2011) “Resistivity Structure of Barrier Geothermal Prospect.” p. 7 Mutua J., Mibei G., (2011) “Remote Sensing Application in Geothermal Exploration: Case Study of Barrier Volcanic Complex, Kenya”, GRC Transactions, Vol. 35