Keywords

1 Introduction

Landslides could be utterly devasting natural hazards and they have a significant impact on the economy of Sri Lanka and create several severe problems, such as loss of human life, damage to property, and damage to the natural environment. One such catastrophic rain-induced landslide happened in the Athwelthota area in Kalutara District, Sri Lanka around 05.00 a.m. on 26 May 2017 causing nine deaths and destroying four houses completely. The landslide is situated on the right-hand side (RHS) upper slope of the 49-kilometer post of Thiruwanaketiya-Agalawatta road (B421), in Palindanuwara Divisional Secretariat, Kalutara District. The geographical position of the site is 6.543586°N, 80.283905°E. The Athwelthota area is known for its hilly terrain and is prone to landslides, especially during the rainy season.

2 Objectives

Studying rain-induced rapid and long-traveling landslides is significantly important for Sri Lanka in many contexts such as hazard assessment in climatic change, development of early warning systems, and infrastructure planning. Centering the primary objective of development of early warning technology to manage recurring landslide events, the National Building Research Organization (NBRO) and the International Consortium on Landslides (ICL) have initiated a 5-year research project called “Development of Early Warning Technology for Rain-induced Rapid and Long-travelling Landslides” (Project RRLL) in collaboration with the framework of SATREPS (Konagai et al. 2021).

The Athwelthota landslide has been selected as one of the pilot sites for field studies under Project RRLL. This study attempts to understand the relationship between structural geological features and geomorphological features in and around the Athwelthota landslide area to comprehend the causes of landslides. This study focuses on determining the contribution of in-situ features to the occurrence of landslides. However, equal attention shall be paid to external features such as rainfall infiltration and slopes’ in-situ conditions for landslide initiation (Figs. 1 and 2).

Fig. 1
A Google earth map of the Athwelthota landslide. It presents the aerial view of a densely vegetated mountainous landscape sloping steeply into a narrow valley.

Athwelthota landslide before failure (Source: Google Earth, Feb.2017)

Fig. 2
A Google earth map of the Athwelthota landslide. It presents the aerial view of a densely vegetated mountainous landscape sloping steeply into a narrow valley. A small steeply sloped area of cleared vegetation is outlined and labeled, boundary of landslide occurred on 2017.

Athwelthota landslide after the failure (Source: Google Earth, Dec.2017)

3 Methodology

3.1 Field Survey

An extensive field survey was undertaken covering the landslide area to investigate the crown area, main escarpment, landslide boundaries, variation of slope angle, bedrock exposures, and their properties. In this assessment, more attention has been paid not only to hydrogeological conditions but also to the general geology of the area and soil condition.

3.2 Seismic Survey

Seismic surveys were undertaken along two traverses, as shown in Fig. 3. These two traverses were determined based on the possibility of future landslides that could happen with torrential rainfall conditions. Further to this, the results of the seismic survey can be used as primary information to determine the borehole locations.

Fig. 3
A satellite map of the borehole locations and seismic traverse at Althwelthota landslide area. It includes rain gauge 1 to the north of the landslide boundary, 2 on its southeast, and seismic traverse 1 and 2 extending vertically between rain gauge 1 and 2, and 2 and borehole 4, in order.

Borehole locations and seismic traverse at Athwelthota landslide area

The collected subsurface data through borehole drilling were used to validate and refine the seismic survey data. This calibration ensures that the seismic traverse profile accurately represents the subsurface conditions and assists in interpreting seismic data for further analysis or exploration.

3.3 Geotechnical Investigation

Geotechnical investigations including borehole drilling were commenced on 01st December 2022 and completed by 20th March 2023. Borehole investigations were carried out at BH-01, BH-02, BH-03, and BH-04 locations (Table 1). Rock coring was done using a double tube core barrel having a 54 mm core diameter. Soil strata changing depths were recorded by carefully analyzing soil and rock samples. Both disturbed and undisturbed soil samples were collected and sealed as soon as possible at the same time/day to preserve their natural moisture content. Upon the completion of the boreholes, the borehole inclinometers, borehole extensometers, and water level meters were installed to monitor the movements and fluctuation of the groundwater table (Table 1).

Table 1 Summary of boreholes and instruments

4 Results and Analysis

4.1 Evaluation of the Geological Condition

The structural geology of a landslide refers to the geological features and structural characteristics that influence the occurrence and behavior of the landslide. These features can include the types and orientations of rock layers, fractures, faults, and folds in the area. The site itself is composed of Charnockitic biotite gneiss and Biotite gneissic rock. The site is located close to a shear zone trending 133 degrees (Fig. 4). The studied slope exists on the Northern escarp slope of the ridge. Hence, the unstable soil and rocks that exist on the slope move in the northeast direction. Moderate to highly weathered Garnetiferrous hornblende biotite gneiss with spheroidal weathering could be observed. According to field observations, the mafic content (Mica and Hornblende) of the rock is high, and it gradually increases.

Fig. 4
A geology map of the project area with a steeply sloped Athwelthota landslide. It includes Ver-J 2 and 70-J 1 to the east of the landslide boundary and 65- F to the southwest.

Geology map of the project area

Bedrock is dipping towards the southwest direction with an angle of 80 degrees. The bedrock of the region could be identified as highly jointed and highly fractured. Three major joint sets can be identified at the site as shown in Fig. 4. Table 2 summarises the discontinuities observed in the area.

Table 2 Summary of structural geological data

The bedrock of the crown region is highly deformed where the micro folds can be observed, and the slickensides of the rock suggest a shearing event. Pegmatitic intrusions, consisting of 1–10 cm thick quartzite bands were also observed in the upper region. Interbanded clay and gravely sand soil layers could be found in the completely weathered bedrock with slippery conditions. The landslide has been initiated as two separate branches. The main landslide and the branch originated from the left side of the landside. They followed the existing main and minor valleys, where those valleys had been formed due to two separate springs coming through the geological discontinuities. The water flowing through these discontinuities recharges the landslide body directly. Hence this will increase the groundwater table and accelerate the failure.

The orientation of the foliation is not facilitating the failure as it is oriented into the slope. Hence, an escarpment slope is created. In this case, the foliation is dipping at 60 to 65 degrees and oriented at 242 degrees. Hence, breaking along the foliation planes is not possible. Joint set 1 is dipping at 70 degrees and oriented in a 160-degree direction, which also doesn’t support the slope failure. Joint Set 2 and 3 are vertically oriented joints and their intersections could lead to direct toppling (33%) and oblique toppling (33%) as shown in the kinematic analysis results (Fig. 5). For this analysis friction angle was assumed to be 34 degrees.

Fig. 5
4 polar diagrams of slope failures. a to d present planar sliding, flexural toppling, wedge sliding, and direct toppling in order. a, b, and c have density concentration at the central friction zone and d has a few scattered in the margin along the northeast, northwest, southeast, and southwest.

(a) Planar sliding (b) Flexural toppling (c) Wedge sliding (d) Direct toppling

4.2 Comparison of Daily Precipitation

The study area receives heavy rainfall during the Southwest Monsoon, which occurs from May to September 2017. This period is associated with inter-monsoonal rains and is characterized by intense showers and thunderstorms. The daily rainfall recorded from the Baduruliya weather station during 2016 and 2017 are shown in Figs. 6 and 7 respectively. The compared rainfall data indicates that the area has received much more intense rain in year 2017 than year 2016.

Fig. 6
A bar graph of fluctuating trend for the daily rainfall during the year 2016 from 1 January to 26 December. Rainfall peaks on 24 May and has little to nil on several days including 17 January, 21 March, 13 September, and 10 December.

Daily rainfall during the year 2016

Fig. 7
A bar graph of fluctuating trend for the daily rainfall in 2017 from 1 January to 27 December. Rainfall peaks to 240 millimeters on 25 May.

Daily rainfall during the year 2017

Attributed to the geological discontinuities, the landslide mass could have saturated not only from the top-down wetting front but also from the bottom-up wetting front. These integrated conditions could have created favorable conditions for landslide initiation in May 2017 with 245 mm daily precipitation as highlighted by the red circle in Fig. 7. This idea can be further verified by referring to the continuous stream flow that appeared after the landslide. Therefore, it is important to analyze the adjacent soil slopes which are highly susceptible to future landslides.

4.3 Subsurface Stratification of the Soil Remained at the Right-Hand Side of the Slope

As described in Sect. 3.2, two seismic surveys were conducted to obtain more information required to portray the subsurface conditions of potential landslide masses. Traverse 1, extends around 300 m whereas traverse 2 covers about 100 m. Based on traverse 1 and by combining the BH-01 data the following description can be made.

The topmost part of the soil profile consists of loose colluvium soil, which is dark brown to yellowish brown, gravelly silty fine to medium sand with weathered rock fragments (Colluvium). Underlying the topsoil is a very dense, yellowish brown, sandy silty gravel. Sand is angular fine to coarse-grained, and the gravel is fine to medium-grained. Even though the soil thickness at the borehole location is 1.45 m, extending about 20 m in both the upper slope and down slope area as visualized in Fig. 8. The soil overburden thickness of BH-03 is 9 meters. Beyond that, it transitions into highly fractured boulders and bedrock (Fig. 9).

Fig. 8
An assumed subsurface profile of elevation versus distance along the Seismic traverse 1. It has a declining slope with loose debris on top followed by soil, dense soil, weathered rock, and fresh rock in order beneath. It includes maximum fresh rock from (240, 15) and (280, 5), approximately.

Subsurface profile along the Seismic traverse 1

The bedrock is garnet biotite gneiss, moderately weak, light grey, metasedimentary, moderately to highly weathered, and highly fractured.

Fig. 9
An assumed subsurface profile of elevation versus distance along the Seismic traverse 1. It has a declining slope with loose debris on top followed by soil, dense soil, weathered rock, and fresh rock in order beneath. It includes maximum loose debris in borehole 05 at (160, 55), approximately.

Subsurface profile along the Seismic traverse 2 (BH-03)

The geotechnical investigation at borehole location 3 revealed the following soil and rock conditions;

0.00 m to 0.9 m

This layer consists of loose, light brown, slightly gravelly silty sand. The sand particles range from fine to coarse in size, and gravel is also present. The presence of organic matter suggests that this layer is of residual origin.

0.90 m to 2.85 m

In this layer, the soil is described as medium dense to dense, with a black color mottled with reddish brown and light brown. It consists of silty fine to medium sand and is also of residual origin.

2.85 m to 4.80 m

This layer is similar to the previous layer, with medium-dense soil that is black-mottled with reddish brown and yellowish brown. The soil composition is silty fine to medium sand and is also of residual origin.

4.80 m to 6.00 m

This layer is described as very dense and consists of yellowish brown and light grey silty fine to medium sand.

6.00 m to 10.50 m

This interval corresponds to a highly fractured in-situ boulder. This fact indicates the presence of a large rock mass that has undergone significant fracturing. 10.50 m to 12.00 m: The final layer consists of dense, completely weathered rock. This indicates that the rock has undergone significant weathering processes, which can alter its strength, stability, and other engineering properties. The specific characteristics of the completely weathered rock would need to be evaluated to assess its suitability for other geotechnical considerations.

Bedrock Description

The identified bedrock at this location is described as light grey, highly to moderately weathered, and highly fractured biotite gneiss rock. Biotite gneiss is a type of metamorphic rock composed primarily of biotite minerals and characterized by a banded or foliated structure. The rock’s high degree of weathering and fracturing can significantly impact its engineering properties, such as strength and stability, and reduce the shear resistance.

4.4 Hydrogeological Condition

The highly jointed and moderately to highly weathered bedrock beneath the soil overburden in the landslide area plays a significant dual role in hydrogeology and slope instability due to the influence of joints and fractures on water flow and the mechanical stability of the rock mass.

The fractured rock at the site, characterized by three sets of joints and weak foliation planes, forms an intricate network of interconnected hydrogeological pathways. These pathways act as conduits for the movement of water, leading to crucial implications for groundwater flow dynamics.

The specific catchment area of the landslide spans 5 ha, featuring a gentle slope covered by a dense forest and a loose, organically rich topsoil. This topography facilitates the infiltration of surface water into the subsurface, effectively recharging the groundwater system. The fractures within the rock significantly enhance the permeability and porosity of the rock mass in contrast to unfractured rock formations. This enhanced permeability allows water to flow more freely through the rock, leading to increased potential for groundwater storage and movement. Water can find its avenue through these fractures more rapidly compared to its movement through intact rock or even soil overburden.

The highly jointed rock acts as a natural aquifer, holding water close to the unstable soil slope. This stored water is gradually released through natural springs into the existing valleys on the site. Throughout the year, both the mainstream and its branch stream, following the existing valleys, consistently discharges water. During the dry season, the discharge rate is approximately 10 liters per minute, while in the wet season, it rises more than 100 liters per minute.

Fractures serve as conduits for increased water infiltration into the rock mass. However, this process can also contribute to the weakening of the rock structure as water erodes surfaces along the fractures and raises the pore water pressure within these fractures. The groundwater monitoring records at location BH-01 reveal a notably swift reaction to the rainfall experienced in the region (Fig. 10).

Fig. 10
A bar cum line graph of groundwater levels and rainfall at Rain gauge 1 from 26 April 2023, 12 p m to 8 September 2023, 12 a m. Former has ascending peaks till 6 March and descending ones after. Latter peaks on 5 March and tends to decline after.

Groundwater monitoring records at BH01 location and rainfall records at Rain gauge 1

5 Conclusions

The intense rainfall recorded on May 25th, 2017, likely played a critical role in triggering the Athwelthota landslide. However, complex unique hydrogeological conditions associated with the local arena may also be vital to clearly understand landslide initiation. Therefore, a detailed examination was conducted to understand the relationship among hydrogeology, structural geology, and geomorphology of the Athwelthota landslide. The following conclusions can be drawn based on the study:

The Athwelthota landslide area is composed of Charnockitic biotite gneiss and Biotite gneissic rock. The site is between two shear zones trending towards 255 degrees and dipping about 60 degrees to the southwest direction. The bedrock is highly jointed and exhibits signs of deformation, including micro folds and slickensides. Pegmatitic intrusions and inter-banded clay and gravely sand soil layers are also present.

These geological features contribute to the instability of the slope.

The landslide area is influenced by groundwater flow through geological discontinuities. Two separate springs in the region recharge the landslide body directly, increasing the groundwater table and potentially accelerating the failure. During the Southwest Monsoon, heavy rain saturates the soil and geological discontinuities, further contributing to instability. It can be said that the Athwelthota landslide was influenced by not only rainfall infiltration but also excessive subsurface drainage. Understanding these relationships is crucial for landslide hazard assessment, early warning systems, and infrastructure planning in landslide-prone regions.