Landslide susceptibility map finds Eastern Ghats as risk-prone

Landslides are among the most devastating geohazards, causing significant loss of life and property. Economic losses from landslides are estimated to reach between 1%-2% of the gross national product in many developing countries.

Recently, researchers at the Department of Civil Engineering and Yardi School of Artificial Intelligence at the Indian Institute of Technology-Delhi have mapped out landslide zones across India, claiming an accuracy of 95.73%.

The newly developed India Landslide Susceptibility Map, available on open access, classifies 4.75% of India as very highly susceptible to landslides, using a five-point scale from very low to very high susceptibility. Landslide susceptibility represents the potential of slope failure for given geoenvironmental conditions.

Official data shows approximately 420,000 square kilometres, or 12.6% of the land area (excluding snow-covered regions) in India are prone to landslide hazards. Nearly 50% of this vulnerable area is located in the North East Himalayas, including the Darjeeling and Sikkim Himalayas. In this context, a tool to estimate landslide susceptibility is invaluable for effective landslide management.

The current study lists Sikkim, Uttarakhand, Arunachal Pradesh, Himachal Pradesh, Manipur, Nagaland, Goa, Mizoram, Meghalaya, and Kerala as the 10 Indian states most likely to witness landslides. It also identifies landslide-prone zones in the Eastern Ghats, which, so far, have not been included in government landslide records or India’s official landslide susceptibility map.

Globally, the number of fatalities owing to landslides until May 2024 have been higher than any previous recorded year, reports The Landslide Blog by Dave Petley, an earth scientist who studies landslides, and the vice chancellor of University of Hull in the UK.

Erratic weather events such as excessive rainfall and subsequent flooding have been causing increasing destruction across India in recent years.

As of June 16, at least nine people have been killed in landslides across Sikkim. The eastern Himalayas have witnessed several such incidents this year. In May, Cyclone Remal swept through Assam and Meghalaya, resulting in a landslide in the Karimganj district that killed five people. Similarly, a stone quarry collapse in Aizawl, Mizoram claimed the lives of 29 people.

The western Himalayas and the Western Ghats have also experienced similar catastrophic events. In March, snowfall-triggered landslides blocked the Gangotri National Highway and other routes in Uttarakhand, while rains caused a landslide on a state highway in Pulianmala, Kerala.

Better landslide management

A susceptibility map aids in streamlining landslide management by identifying the most hazardous areas. This particular map classifies the country’s topography into five classes of susceptibility – very low, low, medium, high, and very high – based on 16 landslide conditioning factors.

These factors include various aspects of slopes, soil composition, weather conditions, and proximity to urbanisation and rivers, among others. The authors of the study note that the previous maps on landslide susceptibility were limited by state borders, data deficiencies and low resolution.

Nirdesh Sharma, a PhD scholar at the Indian Institute of Technology, Delhi, and lead author of the study, highlights the limitations of previously existing landslide susceptibility maps, which fall broadly into two types, both of which have been addressed in the India Landslide Susceptibility Map.

He explains, “Some maps, like those developed by NASA, are global but at a low resolution of about one kilometre whereas this map is at 100 metres. They have limited data compared to ours. The second type are regional maps at a higher resolution of 30 metres, but they are also deficient on data. Our landslide map combines the strengths of both approaches. It is developed using extensive datasets from Indian agencies, allowing us to capture regional patterns better than previous studies.”

The study authors conducted two tests to verify their claimed accuracy of 95.73%. Sharma elaborates, “In the first test, we withheld a portion of the data to train our model, which was then tested on the remaining data to assess its performance. The second test involved a qualitative analysis using data from the global fatal landslide catalogue developed by a UK research group. Our model showed good correlation with that database as well.”

Closer look at the map

The India Landslide Susceptibility Map reveals that a total of 13.17% of India is prone to landslides to varying degrees, with 4.75% classified as very highly susceptible. The study employed big data rather than coarse resolution data, utilising 154,329 landslide points from the national inventory meticulously mapped by the Geological Survey of India and 489 points from a global landslide repository.

SP Pradhan, an assistant professor at Indian Institute of Technology, Roorkee, specialising in rock mechanics and slope stability, notes, “Intense rainfall over a short period of time will definitely lead to slope instability, and such extreme weather events are increasing in recent times. However, more research is needed to understand the frequency of landslides in relation to climate change.”

Among Indian states, Sikkim has the highest percentage of land susceptible to landslides, while Arunachal Pradesh has the largest area susceptible. Outside the Himalayan region, Kerala has the highest susceptible area.

The map highlights susceptibility in the Western Ghats (Maharashtra, Karnataka, and Gujarat) and Eastern Ghats (Odisha, Andhra Pradesh, and Tamil Nadu). Additionally, parts of Chhattisgarh, Rajasthan, Madhya Pradesh, Jharkhand, Assam, northern West Bengal, and Jammu and Kashmir also show scattered signs of landslide susceptibility.

Even while acknowledging the map’s thorough consideration of natural and physical parameters such as rainfall and slope, CP Rajendran, an adjunct professor at the National Institute of Advanced Studies, who specialises in geodynamics, climate, and environmental studies, highlights its limitation in addressing human-induced changes like dams and roads.

He says, “The map looks at natural and physical parameters like rainfall data, slope, etc., but not human-induced changes like dams and roads.” Rajendran expresses concern that the map may underestimate the true extent of landslide risks in India, particularly in regions increasingly affected by infrastructural projects.

Eastern ghats get a mention

The map aligns with a global inventory of fatal landslide occurrences from 2004 to 2016, which identifies the Eastern Ghats as a landslide-prone zone. The ILSM highlights the susceptibility of the Eastern Ghats to landslides, marking a significant update in understanding the region’s risks. The map underscores the necessity for further research and emphasises the need for an updated landslide inventory to comprehensively study landslide behaviour in the Eastern Ghats.

Pradhan explains that the causes of landslides vary across different regions such as the Himalayas, Western Ghats, and Eastern Ghats due to diverse factors. These include variations in geohydrology, geological characteristics (such as rock type and fractures in the rock mass), changes in land use and land cover, weather events, and slope angles.

“In the Eastern Ghats, landslides are often triggered by cyclones that bring heavy rainfall,” Pradhan notes. “In the Himalayas, on the other hand, the collision of the Indian plate with the Eurasian plate led to fracturing of rocks, which, coupled with intense weathering, increased the vulnerability of the rock mass to instability.”

Pradhan says that water is a primary factor contributing to landslides in all regions. “Rainfall infiltrates into rock and soil, significantly raising the likelihood of landslides,” he explains. He also points out that human activities, such as infrastructure development, exacerbate the susceptibility of these regions to landslides.

Landslide management in India

Rajendran emphasises that intensified rainfall across India is contributing to more frequent landslides. He points out that events like the glacial melt leading to the lake outburst in Sikkim are examples of how changing climate patterns are causing sudden flooding and eroding hill slopes, thereby triggering landslides.

“The frequency of floods is going up, which is eroding the base of hill slopes, and triggering more landslides. We already have data on landslides; deficiency isn’t the problem, but the authorities who are planning and permitting infrastructural projects in vulnerable areas are choosing to ignore it,” he says.

Highlighting the Char Dham road project in Uttarakhand as a case where government policies on infrastructure development in mountainous regions have been disregarded, he writes in a column published in 2021: “By considering this project, the government ignores its own policy framework that recommends ‘best practice’ norms for infrastructure expansion in mountain regions.”

Pradhan stresses the importance of comprehensive guidelines for human settlements, construction, and infrastructure development in hilly areas. He advocates for careful consideration of slope angles to avoid steep slopes, emphasising the critical role of effective drainage systems and regular slope maintenance in mitigating landslide risks.

Sharma says he believes that sustainable development and proactive measures against natural hazards should be integral to infrastructure planning. He suggests that governments should utilise susceptibility maps like the India Landslide Susceptibility Map before investing in new projects. He proposes that areas identified with high landslide probability should be developed with “appropriate engineering solutions such as rock bolting, shotcrete, improved drainage systems, and retaining walls to enhance safety and resilience”.

Experts agree that informed decision-making and adherence to best practices in infrastructure development are paramount to mitigating the impact of landslides in India’s vulnerable regions.

This article was first published on Mongabay.