Science & Tech · Science & Environment · 7 hrs ago
Researchers develop landslide warning system for Kerala’s Idukki district
Researchers from three universities developed a landslide warning system for Idukki, a district in Kerala’s Western Ghats in southern India.
The system combines satellite data, weather forecasts and readings from eight local weather stations to estimate rainfall and assess slope stability.
This matters because satellites can miss intense local downpours, while scattered weather stations cannot cover a whole district.
In a test of the 2018 landslides, calibration with rain-gauge readings raised detection sensitivity from 8.9% to 72.6%.
In a 2023 test on Munnar’s Gap Road, the system revised estimated daily rainfall from 20 millimetres to 63.9 millimetres after using ground readings.
The researchers have adapted the system to run automatically, giving disaster management authorities a tool to issue timely warnings and evacuations during monsoon seasons.
An international research team has developed an automated landslide early warning system for Idukki district in Kerala.
The system, called GLEWS, combines satellite data, weather forecasts and live rainfall readings from ground stations to issue color-coded public safety alerts.
Researchers tested it against landslide events in Idukki in 2018 and 2023.
In the 2018 test, calibrating satellite data with rain-gauge observations raised detection sensitivity from 8.9% to 72.6%.
The research team includes researchers from the University of Mississippi, Michigan Technological University and the University of Kerala.
- Who
- Researchers from the University of Mississippi, Michigan Technological University and the University of Kerala.
- What
- They developed GLEWS, an automated early warning system for landslides.
- When
- The system was reported on November 21, 2026. The study was published in a recent edition of Environmental and Engineering Geoscience.
- Where
- Idukki district, Kerala, India.
- Why
- Rainfall data gaps have hindered landslide prediction in mountainous terrain. GLEWS combines multiple data streams to support warnings.
This story does not have two clearly opposing sides.
The team tested GLEWS against floods and landslides in Idukki, finding calibrated data detected 72.6% of landslide conditions.
In a test of a landslide along Munnar’s Gap Road, calibrated rainfall estimates reached 63.9 mm, compared with 20 mm in raw satellite data.
The Hindu reported that the international team had developed the warning system.
- System
- Google Earth Engine-based Landslide Early Warning System (GLEWS)
- Ground stations
- Eight automatic weather stations across the region
- Terrain map
- 30-metre slope stability map
- High susceptibility threshold
- Factor of Safety rating of 1.5 or lower
- 2018 test sensitivity
- 72.6% with calibration, compared with 8.9% without it
- 2023 rainfall estimate
- 63.9 mm after calibration, compared with 20 mm in raw satellite data
1 source
What people are saying
Predicting landslides in mountainous terrain like Idukki has long been hindered by gaps in rainfall data.
To address the issue, GLEWS blends multiple data streams on the Google Earth Engine cloud platform.
In a second-phase test analysing the July 7, 2023 landslide along the Gap Road stretch of Munnar, raw satellite data recorded a modest 20 mm of daily rainfall.
After GLEWS applied its ground-gauge calibration, the estimated rainfall tripled to 63.9 mm, accurately reflecting the severity of the local downpour.