Background

Gereb Segen dam is located 17 km southwest of Mekelle, capital city of Tigray regional state, northern Ethiopia and bounded by geographic coordinates 540,000 m to 545,000 m E and 1,479,700 m to 1,483,900 m N (UTM Zone: 37, Datum: Adindan) (Fig. 1). The embankment dam is 51 m high with a crest length of 850 m and reservoir capacity of 31 million m3 and is designed to provide water supply to Mekelle city. During the initial phase of construction, critical geological problems in relation to water tightness were identified. This research is aimed at addressing the geological challenges (water tightness and stability) of the dam site using geophysical and engineering geological investigations. The reservoir capacity of Gereb Segen dam is large compared to other micro-dams around the study area.

Fig. 1
figure 1

Location map of Gereb Segen dam site with reference to Ethiopia and Africa. The location map is extracted from SPOT imagery

Full size image

The variability in rainfall together with the extended duration of drought is threatening food and water security in Sub-Saharan Africa in general and East Africa in particular (De Hamer et al. 2008). Construction of dams in arid and semi-arid areas such as northern Ethiopia, where the main socioeconomic activity is rain-fed agriculture, is important and is directly linked to the erratic and variable rainfall (Yazew 2005; Walraevens et al. 2009). Northern Ethiopia, particularly the Tigray region, is often affected by drought largely as a result of uneven distribution of rainfall over many decades (Nedaw and Walraevens 2009; Berhane et al. 2013). One possible strategy for alleviating such a problem is through the use of various means of harvesting water for irrigation and drinking purposes. To this end, many water storage schemes have been undertaken by the Regional Water Bureau and Federal Ministry of Water Resources over the past 15 to 20 years through the construction of micro-dam reservoirs. However, many of the micro-dams have failed to store the expected amount of water. This is mainly due to lack of engineering geological, geophysical and geotechnical investigations prior to dam construction (Haregeweyn et al. 2005; Abdulkadir 2009; Berhane 2010a, b).

One of the challenges faced by designers and development planners is the absence of basic and systematic data related to ground conditions for structures such as dams. Engineering geological investigation plays a significant role in the design and execution of such civil engineering projects (Panda et al. 2014; Rahimi et al. 2014), as it aims to gain an understanding of the interrelationships between the geological environment and the engineering situation, the nature and relationships of the individual geological components, the active geodynamic processes, and the likely results of the changes being made (UNESCO 1976).

In the past two decades, more than 70 earthfill dams have been constructed in the Tigray regional state, mainly for irrigation purposes (Abay and Meisina 2015). Of these, nearly 60% of the dams located in the Mekelle Outlier are subject to serious leakage through the reservoir and dam foundation (Haregeweyn et al. 2006; Berhane et al. 2013). Half of the leaking dams rest on systematically jointed carbonate rocks (limestone and marl). According to Berhane et al. (2013), the limestone-marl intercalation unit with hydraulic conductivity of 10−6 to 1 m/s was found to be responsible for the excessive leakage of Hashenge and Arato micro-dams. According to Haregeweyn et al. (2006), ten micro-dams constructed on the shale-marl-limestone intercalation around Gereb Segen dam are characterized by very high leakage and become empty shortly after rainfall, as indicated in Table 1.

Table 1 Reservoir leakages in sedimentary basin of Mekelle Outlier (Haregeweyn et al. 2006)
Full size table

However, many dams have been built in carbonate foundations. Due to unknown karstic heterogeneity, many operating dams have leakage problems, e.g. Hales Bar Dam, USA (Lienhart 2013), Anchor Dam, USA (Jarvis 2003), Kalecik Dam, Turkey (Turkmen 2003), Mosul Dam, Iraq (Al-Saigh et al. 1994), Tangab Dam, Iran (Karimi et al. 2005; Mohammadi and Raeisi 2007) and Lar Dam (Uromeihy 2000). In carbonate terrane, karstification readily exploits the fractures and open structures developed as a result of valley stress relief and results in leakage of water from impoundment structures. Karst heterogeneity can be found on the surface (sinkholes) and in the subsurface (caves). The water leakage at Lar Dam is mostly due to the abundance of caves and conduits in the underlying limestone beds. Although many efforts have been made to prevent the leakage (mainly by grouting), none of these efforts have been successful (Uromeihy 2000). To investigate leakage problems, geotechnical (Bedmar and Araguas 2002; Milanovic 2004) and geophysical methods (Al-Fares 2011; Al-Saigh et al. 1994) must be applied.

Geological setting

Regional geology

The regional geology of northern Ethiopia has been studied at regional scale (e.g. Levitte 1970; Beyth 1971, 1972; Kazmin 1972; Garland 1980; Bosellini et al. 1997) and can be broadly divided into three major groups: the Neoproterozoic basement complex, Paleozoic-Mesozoic sedimentary sequences and Cenozoic trap volcanics (Levitte 1970). The Neoproterozoic basement complex of Mekelle area is the southern extension of the Arabian-Nubian Shield (ANS). It is composed of two major sequences (Beyth 1971): the Tsaliet Group, an older, mainly metavolcanic/volcaniclastic sequence; and the Tambien Group (Alene et al. 2006), a younger sediment-origin, low-grade metamorphic rock. Both groups are later intruded by syn- to post-tectonic intrusions, mainly granitic to granodioritic positions. The supracrustal metavolcanics/volcaniclastics contain intra-layer immature sediments mostly metamorphosed in the greenschist facies (Avigad et al. 2007). The magmatic-origin Tsaliet Group is composed mainly of intermediate to felsic and pyroclastics (welded tuffs, well-bedded lapilli tuffs and agglomerates; Beyth et al. 2003). The Tambien Group is exposed in many localities of the region. It is composed of mainly metasediments including slates, phyllites, graywackes, limestone/dolostone and diamectites/pebbly slates. Typical sections are defined in the Mai Kenetal and Negash synclinoria (Beyth et al. 2003). Late to post-tectonic granitoids, which are less deformed, frequently intrude the metasediments (i.e., May Kinetal and Yechilla intrusions) with the exception of the Negash granitoid mainly intruding the metavolcanics/volcaniclastics. At few locations, small stocks of porphyritic diorite intrude the upper parts of the sequence (Beyth et al. 2003 and references therein).

The sediments in the Mekelle Outlier began to be deposited around the Ordovician or Carboniferous and probably ended in the Lower Cretaceous before the eruption of the Trap volcanics (Beyth 1971). The sedimentary rocks unconformably overlie the Neoproterozoic basement complex and form a nearly circular outlier 8000 km2 in area around Mekelle. The Paleozoic sediments are the characteristic formation of northern Ethiopia, and can be subdivided in two unconformably bounded units, the Enticho sandstone and the Edaga Arbi glacials (Alemu et al. 2018). The Enticho sandstone comprises a cross-stratified quartz sandstone (quartz arenite; Avigad et al. 2007 and references therein) that locally contains varying size and composition conglomerate layers, and may be partially glacial in its upper part. Spatially, the rocks show wide thickness variation: from ~220 m to few meters. Unconformably overlying the Enticho sandstone and in some places the basement rocks, there are glaciogenic sediments (Edaga Arbi glacials) of predominantly dark-brown to light-red laminated/friable shale and siltstone which often contain dispersed pebbles or boulders, and minor diamictites (Bussert 2010).

The Mesozoic sedimentary rocks of the Mekelle basin form one of the best exposed basins in Ethiopia, constituting a nearly complete stratigraphic succession: from the lower Adigrat sandstone, through Antalo limestone to the Agula shale and Ambaradam sandstone. The Adigrat sandstone, lying unconformably on the Edaga Arbi glacial deposit, consists of up to 630 m of continental to shallow-marine siliciclastic sediment (Blanford 1869). It is fine–medium-grained, well-sorted and mature sandstone capped at each major deposition by highly lateritized clastic layers. Interpretation of a depositional environment for this unit is controversial as most previous studies (e.g., Assefa 1991; Bosellini et al. 1997) consider the Adigrat sandstone as fluviatile. However, Dawit (2010) reported that the depositional environment is characteristic of tide-dominated estuarine conditions. The Antalo Formation, a major stratigraphic unit of the Mekelle Outlier, is characterized by the intercalation of limestone, marl and shale units. Based on the field observations and dominancy of the layers, the Antalo Formation is grouped into three categories, namely as marl-limestone-shale, limestone-marl and shale-marl-limestone units. Some limestone layers have shown signatures of small-scale karstification. The Antalo Formation (limestone) is conformably overlain by the Agula Shale Formation (Bosellini et al. 1997) which is composed of shale with minor intercalations of marly limestone, mudstone and gypsum layers. At the central part of the outlier, the shale is intensively injected by sills and dikes (Fig. 2). The dikes and sills known as Mekelle dolerites are the youngest intrusive rocks. About 75% of the Mekelle Outlier is covered by the Antalo Formation and Agula shale with gypsum intercalations. The remaining part is covered by sandstones units (Adigrat and Ambaradam units).

Fig. 2
figure 2

Simplified geological map of central and eastern Tigray, northern Ethiopia. Circled numbers denote major WNW–ESE-trending fault zones: 1 = Wikro fault zone; 2 = Mekelle fault zone; 3 = Chelekot fault zone and 4 = Fursa Mariam fault zone. The yellow star indicates the site of the dam

Full size image

At the top of the Mesozoic succession, there is a detrital and mottled sandstone called Ambaradam sandstone (Estifanos et al. 2018) which consists of mudstone, siltstone and pebbly sandstone. After a major unconformity, the Mesozoic succession of northern Ethiopia is covered by a 2-km-thick pile of flood basalt (Hagos et al. 2010; Ukstins et al. 2002). Lavas of this province erupted ~30 Ma ago, during a short period of 1–2 Ma, to form a huge volcanic plateau (Hofmann et al. 1997; Kieffer et al. 2004). The Mekelle Outlier is bounded to the north and south by Wikro and Fursa Mariam extension faults, respectively. At its center, the Mekelle and Chelekot faults cut the outlier into three blocks (Fig. 2).

Geology of the dam site

The dam site is covered by four geological units, namely alluvial deposits, dolerite, micritic limestone and marl-limestone-shale intercalation units (Fig. 3) as outlined below. The local geological units are correlated with the regional geology of northern Ethiopia. The alluvial deposit is correlated with Quaternary deposit, dolerite with Mekelle dolerite, micritic limestone and marl-limestone-shale intercalation with the Antalo limestone formation regionally.

Fig. 3
figure 3

Geological map of Gereb Segen dam site and its surrounding area. Lithologic contacts were extracted from SPOT imagery

Full size image

Alluvial deposits cover the central part of the dam site in the valley floors along the main river and its tributaries (Fig. 3). This unit is composed of clay soil with units of gravel, with boulder-sized dolerite and limestone clasts. It has 10 m thickness at the center of the reservoir, but decreases towards the abutments. Dolerite is exposed as a sill in the right and left abutments, upstream and reservoir rim of the dam site. It is overlain by relatively jointed micritic limestone (Fig. 3). The thickness of the sill varies from 40 m to 50 m in the study area. The dolerite is highly weathered and shows exfoliation and spheroidal weathering (Fig. 4a). It shows dark-gray to black color and consists of medium- to fine-sized minerals.

Fig. 4
figure 4

a Spheroidal weathering of dolerite at the right abutment, b limestone outcrops, jointed in two sets at the right abutment of the dam, c marl-limestone-shale intercalation unit at the trench excavation of right abutment and d alluvial deposits overlying shale-marl unit

Full size image

A micritic limestone unit is found at the upper part of the left and right abutments and along the upper face of the reservoir rim. The field investigation showed that it is moderately to highly jointed (Fig. 4b). The aperture of the joints varies from 2 to 15 cm with an average joint spacing ranging from 50 to 125 cm without filling materials. Karstification in limestone is common in the downstream of the proposed dam site. A highly weathered marl-limestone-shale intercalation unit is also observed in both right and left abutments and downstream of the dam.

Materials and methods

Electrical resistivity imaging

Electrical resistivity method has advanced from conventional vertical sounding to 2D and 3D resistivity imaging, which can provide high-resolution electrical images (Colella et al. 2004). 2D electrical resistivity imaging was executed with the objective of determining the nature and thickness of the subsurface units and to detect geological weak zones like dykes, faults, cavities and fissures at the dam site. In the present study, 2D resistivity imaging was carried out using a Wenner array and ABEM Terrameter SAS 4000 (ABEM 1999, 2012). The data was processed using the RES2DINV package (GeoTomo Software, Loke 2001) to create a 2D resistivity model using the inverted apparent resistivity values (Griffiths and Barker 1993). The program makes use of the least-squares method based on a quasi-Newton optimization technique (Loke and Barker 1996; Dahlin 1996).

Four profiles (Fig. 5) of electrical resistivity imaging were selected at the dam site. Profile 1 (SW-NE) was conducted along the dam axis with 20 m electrode spacing covering a total length of 780 m (Fig. 4a). The maximum depth of investigation in this, and all other profiles, reached 25.5 m. Profile 2 (SE-NW) was oriented perpendicular to the dam axis covering a total length of 640 m with 20 m spacing between the electrodes (Fig. 5b). Profile 3 (NW-SE) was taken at the left reservoir rim of the rim site where lineaments were detected from satellite imagery. The electrode spacing was the same as that of profiles 1 and 2 with a total length of 680 m (Fig. 5c). Profile 4 (NW-SE) was taken at the right reservoir rim of the dam with 20 m spacing and a total length of 380 m (Fig. 5d).

Fig. 5
figure 5

Inverse model resistivity imaging sections of four profiles: a profile 1, b profile 2, c profile 3, and d profile 4

Full size image

Engineering geological investigation methods

The engineering geological investigations of the dam site include discontinuity surveying, core drilling and Lugeon testing. A number of field observations and discontinuity measurements were recorded at the dam site in accordance with the suggested method of the International Society of Rock Mechanics (ISRM 1981). The condition of major joint sets (orientation, spacing, persistence, roughness, aperture and filling) relative to an engineering structure largely controls the possibility of unstable conditions, excessive leakage and deformations (Maerz and Zhou 1999). A total of 117 discontinuities were recorded using a Brunton compass on the left and right banks. Discontinuity orientations were processed using a RockWorks15 linear rose diagram (RockWare Inc. 2009) (Fig. 6).

Ten borehole locations were selected based on the geological conditions, topography and geophysical survey of the dam site and drilled to a depth of 38–51 m, with a total core length of 451 m using NX wire line equipment (54.7 mm core diameter). The distribution of boreholes was: five at the dam axis, three at the reservoir, one at the spillway route and one at downstream part of the dam (Fig. 3). Rock quality designation (RQD, Deere 1968) and rock mass rating (RMR, Bieniawski 1989) were calculated from the drill cores to characterize rock mass quality at the dam site.

A total of 33 Lugeon tests were conducted in ten boreholes to assess the permeability of the rocks. The testing was carried out at intervals along the boreholes with the testing section sealed off by inflatable packers. The maximum test pressure was defined before commencing the test, and pressure stages were lowered after using maximum pressure (Quiñones-Rozo 2010; Öge 2017). The data obtained from boreholes were used directly to assess leakage risks, and compared indirectly with the data from the resistivity imaging profiles.

Results and discussion

2D electrical resistivity imaging

The inverse model resistivity profiles show variable resistivity values both laterally and vertically (Fig. 5). The obtained root mean square errors of the four profiles (1, 2, 3 and 4) are 12.2%, 5.5%, 9.4% and 6.3%, respectively. The 2D imaging section in Fig. 5a clearly depicts the lateral variations underlying the dam axis. The two abutments are characterized by resistivity values of 80 to 200 Ω.m, attributed to weathered dolerite and jointed black limestone. Near the central part of the dam axis (at around 430 m chainage) there are medium-resistivity (90 Ω.m) pockets at shallower depth, mainly attributed to alluvial boulders of the remnant buried channel. In the central part of the dam axis, the resistivity values of 36 Ω.m to 60 Ω.m are taken to represent interlayers of various resistivity values reflecting marl-limestone-shale intercalations. Figure 5a shows possible weak zones at chainages of 80 m and 560 m. Low resistivity values (15–23 Ω.m) at the bottom of the right abutment are identified as a shale-gypsum unit).

Fig. 6
figure 6

Rose diagram of dominant joints at the a right abutment and b left abutment of the dam site

Full size image

Profile 2 taken upstream-downstream (Fig. 5b) shows three major geological layers which correlate with the geological cross section (Fig. 7b) produced from the drilling tests. The top layer with low resistivity (13.5–20 Ω.m) is fine to coarse alluvial deposits underlain by a more resistive (22–30.4 Ω.m) layer of marl-limestone-shale. The third layer is more resistive with resistivity values ranging from 32 Ω.m to >42 Ω.m, representing a marly limestone unit. Towards the downstream, the resistivity values become high, while in the central part they are low, related to fine alluvial deposits to a depth of 10 m. The bottom part in the upstream of the dam recorded high resistivity (>45 Ω.m) which is interpreted as an air-filled karstified zone (Fig. 5b), as later confirmed by borehole drilling (Fig. 7c). It is common that cavities present at shallow depths above the water table in semi-arid areas that are filled by air (not by groundwater) produce high electrical resistivity values (Martínez-Pagán et al. 2013).

Fig. 7
figure 7

a Engineering geological map of the dam site, b geological cross sections along the dam axis (A-A’ section) and c upstream-downstream (perpendicular to dam axis) (B-B’ section). Electrical profiles (profiles 1, 2, 3 & 4) are indicated in red line

Full size image

The 2D resistivity imaging along the left reservoir periphery (Fig. 5c) near the spillway is characterized by variable resistivity values due to the presence of heterogeneous units such as weathered dolerite, fractured limestone and disturbed shale-marl-limestone intercalations. The thickness of the top layer shows medium resistivity (60–90 Ω.m) which increases from the northwest to southeast part of the left reservoir. But, its resistivity drops at chainage distance of 570 m onward towards the southeast corner. Figure 5d shows two layers along the right reservoir rim. The top thin layer is dominated by shale-mudstone towards the southeastern part and highly weathered dolerite in the northwestern part which exhibits low resistivity. The bottommost part of the NW direction is characterized by moderate to hard marly limestone.

Engineering geological investigations

Discontinuity surveys

Detailed discontinuity surveys were executed for the rocks outcropping in the dam site to examine the rock mass quality and their influence on the water tightness. Of the 117 discontinuity measurements recorded in the dam area, 28 were from the left abutment and 89 from the right abutment. Two sets of joints orienting NW-SE and NE-SW dominate on both the left and right abutments (Fig. 6). The NW-SE-trending joints are a potential route for reservoir water leakage as they are subparallel to the likely flow direction. The joints are mostly smooth, narrow apertures (2–15 cm), and joint spacing ranges from 50 cm to 125 cm. Moreover, some wedge failures were observed in both abutments and reservoir rim following the intersection of these joints.

Core drilling

Borehole sites were selected following the electrical resistivity imaging survey and surface geological mapping. From the borehole drilling, lithological cross sections (Fig. 7b) were prepared to understand the vertical and lateral variation of strata. Rock quality designation (RQD) (Deere 1968) is calculated for the rock units present in the dam site (Table 4). The RQD value for black limestone exposed at the rim of the reservoir is estimated from the surface measurements using a volumetric joint count (Palmstrom 1974), while the RQD of other rocks is estimated for the borehole core samples as shown in Table 2. Accordingly, the RQD values for the rock masses at both abutments are lower than at the central foundation. Thus, rocks at the two abutments are liable to exhibit excessive leakage and stability problems.

Table 2 Average RQD values obtained for core samples from the left abutment (LF), right abutment (RA) and central foundation (CF) of Gereb Segen dam site
Full size table

As shown in Fig. 7b, the top layer of the central foundation of the dam is covered by alluvial soils (clay, sand and gravel) and underlain by shale and marly limestone units. The weathered dolerite about 25–30 m thick, observed in both left and right abutments, is overlain by massive black limestone. A shale-gypsum unit is the bottommost unit ranging in thickness from 11 m in borehole BHDA-5 to 29 m in BHDA-2. A pure gypsum layer (0.5–1.5 m thick) is also detected in four boreholes (BHDA-2, BHDA-3, BHDA-4 and BHDA-5) (Fig. 7b). Occurrence of gypsum at shallow depth (9.5 m) along the central foundation in borehole BHDA-3 may interact with percolating water through joints and dissolve easily to form small caves and sinkholes. This can allow excessive leakage and cause differential settlement of the dam.

The drilling results show that cavities having sizes of 2.5 m (21.7–24.2 m depth) and 3 m (17.25–20.25 m depth) are present along the dam axis as shown in BHDA-1 and BHDA-4 (Fig. 7b). Similarly, 0.8 m and 2.7 m cavities were observed in boreholes BHRE-1 and BHRE-2, respectively, along the reservoir perpendicular to the dam axis (Fig. 7c). Cavities were well developed in marly limestone because it is susceptible to dissolution when it is in contact with water. This implies that the marly limestone rock in the dam site and reservoir area is potentially affected by cavity formation, which in turn could give rise to excessive leakage and dam instability.

Permeability testing

A total of 33 Lugeon tests were carried out, and Lugeon values for different rock units are presented in Table 3. Results revealed that permeability is high in marly limestone and dolerite units due to the presence of open fractures and a moderate degree of weathering. This may allow seepage through the abutments and central foundation of the dam. Rock masses having Lugeon values above 3 indicate a need for grouting treatment (Houlsby 1990). During the Lugeon test at the cavity section (e.g. BHDA-4), there was no permeability record as the pressure could not be developed due to complete water loss.

Table 3 Average Lugeon values for different rock masses at Gereb Segen dam site
Full size table

Shale-gypsum lenses and shale units are classified into very-low- and low-permeability zones, respectively. Though the permeability of shale-gypsum is low, the presence of very thin layers (3–10 cm) of gypsum may cause stability problems due to dissolution. Permeability tests were also conducted on alluvial soils, dominated by clay and silt. Falling-head permeability tests performed on fine-grained soils from three boreholes from the central foundation indicate low hydraulic conductivity (1.06 * 10−6 m/s). Hydraulic conductivity generally decreases with depth due to the effect of overburden (Lee and Farmer 1993; Nappi et al. 2005) as shown in Fig. 8a, although exceptions were noticed. In three boreholes (BHDA-1, BHDA-4 and BHDA-5), the hydraulic conductivity value decreases with increasing depth. But in two boreholes (BHDA-2 and BHDA-3), permeability is nearly constant with depth. From the Lugeon test results (Table 3), rocks near the central foundation, which is dominated by shale (2.9 uL) and shale-gypsum units (0.6 uL), show low and very low permeability, respectively. On the other hand, fractured marly limestone (56 uL) and weathered dolerite (65 uL) at the left and right abutments show high permeability (Fig. 8b), which could also be a possible source of leakage.

Fig. 8
figure 8

a Lugeon value versus depth for five boreholes drilled along dam axis. b Lugeon value contour map along the dam foundation

Full size image

Rock mass classification

Rock mass classification is an important geological factor for the design and construction of a dam. It not only evaluates the properties of rock masses, but also helps in determining the range of utilizable rock masses around the dam and the reasonable excavation depth of the dam foundation (Lashkaripour and Ghafoori 2002). Rock mass classification is calculated according to rock mass rating (RMR, Bieniawski 1989). The rating of each RMR parameter is summed to obtain the total RMR value, as presented in Table 4. The rocks are thus classified in the range of good rock quality (class II) to fair rock quality (class III). Hence, the rock masses can serve as a foundation for the earth dam. However, treatment is needed to reduce the permeability of rock masses as the seepage is expected to be high which could create foundation stability problems.

Table 4 RMR values of different rock masses at Gereb Segen dam site
Full size table

Conclusions

In this study, 2D electrical resistivity imaging and engineering geological investigations were conducted to examine the subsurface geological formations at the dam site. The site is underlain by alluvial deposits, dolerite, micritic limestone and marl-limestone-shale intercalation units. Four cavities (two at the reservoir, one at the left abutment and one at the right central foundation) ranging in size from 0.8 m to 3 m were found at the dam site within the marly limestone layer, forming a karstified zone at the dam site. The presence of the karstified zone at the reservoir and dam foundation will likely lead to significant leakage for Gereb Segen dam. From this finding, it is possible to deduce that the marly limestone unit is expected to have similar cavities around the reservoir which could be the source of excessive water loss through the reservoir and foundation of the dam. Furthermore, the presence of soluble gypsum layers at a shallow depth (9.5 m) in the dam foundation could further threaten the stability of the structure of the dam in the long term due to its potential for dissolution. Lugeon test results indicate that the two abutments were characterized by high-permeability zones (56–65 uL) due to the fractured marly limestone and weathered dolerite, while the central foundation is covered by low-permeability alluvial deposits underlain by an impervious shale unit. The results of the study show that the correlation between geophysical survey and subsurface investigation is satisfactory. For the karst zones at the foundation, a multiple-line grout curtain is recommended to reduce seepage losses, and grouting along the dam axis where jointed limestone is found. Additionally, blanketing of reservoir peripheries and the river bed is important.