Thirteen percent of the state's land area is extremely susceptible to landslides triggered by extreme annual weather events.
Scene of the Pettimudi landlside, one of the biggest in the recent history of Kerala, that occurred on 6 August, 2020. (KA Shaji/South First)
The plantation workers did not find anything abnormal when incessant rains lashed Pettimudi, where they lived, some 30 km from Munnar on 6 August, 2020. Such torrential rains were normal in the area, and the workers of Kanan Devan plantations went about their tasks as usual.
For the workers, the day was long. The lights in the row houses — or layams — on the hill slope went off one by one as families retired for the day, oblivious to a huge boulder struggling to hold itself against the gushing water on the hilltop.
The boulder surrendered around 10.45 pm with a loud noise, and it rolled down to the Pettimudi river with a splash. A copious amount of mud, slush, debris, and water followed, levelling the once bustling settlement — along with the sleeping residents — in a flash.
At least 66 people were killed in the incident.
The search and rescue began in full swing the next day. Lily, a Belgian Malinois police dog sniffed out three bodies, while three other local canines, Rose, Tiger, and Kuvi refused to leave or accept food the rescuers offered.
Kuvi, an Indie pariah, soon became news as it frantically sniffed around the washed-away land, often pawing at places, as if searching for someone. The rescuers later learnt that Kuvi’s human family had been missing.
On day eight, a famished Kuvi led a police K9 trainer, Ajith Madhavan, through the thick fog, to what she had been searching for. She had found the now putrefied body of her beloved playmate, two-year-old Dhanushka, about four kilometres from the landslide site.
Kuvi’s was a story of unflinching loyalty and canine intelligence that caught global attention.
Landslides in Kerala have now — albeit tragically — earned the dubious distinction of a never-ending story, like the proverbial Sword of Damocles hanging over the heads of several hamlets in the high ranges.
Several human factors, including deforestation, and indiscriminate unscientific constructions, are fuelling the natural calamity now visible across Kerala.
Now, an artificial intelligence (AI)-enabled study by the Kerala University of Fisheries and Ocean Sciences (KUFOS) in association with Michigan Technological University (MTU) and the Pune-based Indian Institute of Tropical Meteorology (IITM) has indicated that things might get worse.
The study used data available between 1990 and 2022 and found that around 13 percent of the state’s land area is extremely susceptible to landslides, triggered by extreme annual weather events, such as flash floods and cloud bursts.
Among Kerala’s 14 districts, Idukki, Palakkad, Malappuram, Pathanamthitta, and Wayanad are highly vulnerable to high-magnitude landslides, which occur during the peak of the monsoon season.
With the help of the two external agencies, KUFOS prepared an AI-based landslide susceptibility map of Kerala, utilising deep learning technology, which provided an overall susceptibility overview of the state. The comprehensive research revealed a confluence of environmental stressors exacerbating the state’s vulnerability and it urged the government and stakeholders to initiate urgent mitigation efforts.
The data from over 32 years indicated a 3.46 percent increase in extreme landslide susceptibility zones, post the 2018 deluge.
The study also urged the government and policymakers to be accountable.
The reasons cited for the escalation of landslides included first-order stream disturbances or disrupting streams at source, slope cutting for road construction, and unscientific land-use practices.
The KUFOS researchers are now developing a mobile app for landslide mitigation, and it can coordinate diverse expertise and action plans. The app would be ready within six months. The study has also identified slopes within the 10° to 40° range as highly prone to landslides.
Speaking to South First, Girish Gopinath, who heads the Department of Climate Variability and Aquatic Ecosystems at KUFOS, said the study, Redefining Landslide Susceptibility Under Extreme Rainfall Events Using Deep Learning, was published in the latest issue of the science journal, Geomorphology.
“The high-range regions of the state, with approximately 31 percent of areas elevated above 600 m, fall under the extreme landslide susceptibility zones. This underscores the urgency of safeguarding not only populated areas but also ecologically sensitive, high-altitude regions,” Gopinath observed.
The researchers collected 3,575 landslide samples from extremely susceptible zones and analysed them using the AI tool.
“The landslide susceptibility map was prepared from the data generated using AI-aided analysis,” Gopinath said. The previous year, the state witnessed flash floods and landslides in many unanticipated locations in different districts.
Several areas in the hilly regions received over 15 cm of rain in a short span of three days, as if in a cloudburst, causing fresh landslides last year.
According to a reply by Union Minister for Science and Technology Jitendra Singh to a question in the Lok Sabha on July 28 last, Kerala accounted for 2,239 of the 3,782 “significant landslides” recorded in the country between 2015 and 2022. Kerala accounts for hardly 1.18 percent of the country’s total land area.
The minister said 59.2 percent of all reported landslides in India are happening in Kerala, despite its smaller size. According to S Sreekumar, a disaster risk consultant with the Kerala Institute of Local Administration, the state witnessed only 295 landslide-related casualties in 55 years between 1961 and 2016.
In the six years since, 374 people were killed in landslides, according to data from the Kerala State Disaster Management Authority (KSDMA).
“Ever since the devastating floods of 2018, the state has witnessed an alarming increase in the extent of deaths and damages,” Sreekumar told South First.
Across the state, the monsoon calendar has been witnessing a discernible change. As a result, more instances of high-intensity cloud bursts are concentrated over smaller areas that trigger massive landslides.
According to climate experts, the state has failed to identify vulnerable areas at the micro-level proactively. Instead of merely acting to mitigate the impact of rain-related disasters, they said the government must immediately map micro-level vulnerability across the state. They expect KUFOS’s proposed app to be a significant step in that direction.
Land and disaster management expert Dr KG Thara believed that the state’s local bodies should immediately be entrusted with comprehensive, ground-level vulnerability mapping. She recommended even ward-level mapping, as the situation was becoming more alarming by the day.
The experts also called for a massive restructuring in the nature and composition of the KSDMA. In an RTI reply in April last, the state’s nodal agency for disaster management said it lacked details on previous landslides.
“This itself points to the precarious situation in the state. The KSDMA must be revamped with experts who can act swiftly on the prevailing challenges,” SP Ravi of the River Research Centre at Chalakudy told South First.
Ravi said the mapping would help people understand how to use the sloped terrain more sustainably to prevent disasters.
According to Dr KV Thomas, a former National Centre for Earth Science Studies scientist, rubber monocropping, prevalent in the state’s high-range areas, has been significantly contributing to landslides.
“Rubber trees are incapable of holding the soil during incessant rains. There is a need for planting creepers intercropping in rubber plantations to avoid soil slipping,” Thomas told South First.
He said that rubber, along with stone quarry operations and unscientific construction activities, aggravate the climate crisis in the Western Ghats.
Sreekumar said construction work has continued unabated, even in vulnerable areas, with the tacit support of authorities. He said road and culvert construction must consider possible challenges from flash floods and landslides. The “unscientific construction mechanism” in the state has been adding to the lethal problem.
Thomas also felt that field surveys and mappings would help facilitate evolving mitigation measures, which CK Vishnudas, the Hume Centre for Ecology and Wildlife Biology director, also stressed.
The mapping can focus on changes in land-use patterns, mainly in the hills. Vishnudas said people in highly susceptible areas should be shifted to safer grounds during the rains, and farmers should be given real-time weather data.
“The prevalence of tropical humid climate across the region enhances chemical weathering, forming thick soil stratum (rich in clay) over the crystallines,” Gopinath explained.
“This unconformity between the crystalline rocks and the overburden forms the slip plane for most rainfall-triggered landslides in Kerala,” he added.
“Rainwater trapping pits dug up along the slopes to recharge groundwater resources trigger landslides during heavy rain. Human activities in the periphery of Wayand and Idukki high ranges are a major factor contributing to landslides,” Gopinath further said.
The KUFOS research has collected 3,575 landslide samples for AI-aided analysis. About eight percent (1,400 sq. km.) of the area in the Western Ghats in Kerala has been classified as a critical zone for mass movements such as debris flow, landslides, slumps, and rock falls.
“Only local panchayats can make a difference,” Sreekumar opined. “Every ward must have an evacuation response team (ERT). There should be more local resource groups; they should be clear about their mission. If all these could function efficiently, we can reduce fatalities,” he stressed.
During a disaster, the first 60 minutes is the golden hour. How much we can do in that one hour and how fast we can move people to relief camps are all important. Convincing people to move to relief camps is a major task because many will be reluctant. A local ERT will be more effective in such cases,” he added.
May 07, 2024
May 06, 2024
May 05, 2024
May 05, 2024
May 05, 2024
May 05, 2024
