In the shadow of the Himalayas, where rivers roar and mountains shift, a pioneering project quietly transforms how we forecast floods and landslides. Led by Amrita Vishwa Vidyapeetham, this initiative is setting a national precedent in climate resilience through state-of-the-art heterogeneous models, which include AI-enabled IoT-based wireless sensor network systems for surface and subsurface measurements, as well as data-driven and numerical models for impact-based forecasting and risk assessment of multi-hazards.
A team of three researchers from the Amrita Center for Wireless Networks and Applications—Mr. Nitin Kumar M, Ms. Hari Chandana Ekkirala, and Mr. Bichu B.K. Raman undertook a field mission to the regions of Gangtok, Mangan, Dikchu, and Chungthang in Sikkim under the guidance of Dr. Maneesha V Ramesh, Provost of Amrita University.
The primary objectives of the visit were to carry out a Differential GPS (DGPS) survey of surface cracks observed in the region and to measure the flow velocity of the Teesta River using hydrological instrumentation. At the core of this initiative lies a critical mission: to understand, simulate, and predict the complex cascade of events involving extreme rainfall events, landslides, and floods that lead to catastrophic disasters—before they unfold—enabling proactive risk mitigation and early warning.
The Regional Challenges
The Eastern Himalayan state of Sikkim faces mounting risks from natural hazards, including glacial lake outburst floods (GLOFs), flash floods, landslides, and land subsidence. These challenges are compounded by the region’s high-altitude terrain, fragile slopes, erratic precipitation patterns, seismic activity, and rapidly retreating glaciers.
In contrast, the East Sikkim district faces a different set of geohazards. The region, particularly around the capital, Gangtok, is affected by land subsidence, the formation of cracks on roads and infrastructure, and slow-moving landslides. These challenges are exacerbated by rapid urban expansion and intense rainfall.
Comprehensive Field Assessment in North Sikkim
The research team undertook surface-crack mapping and household-level surveys to understand the socio-economic and structural impacts of these movements. Gangtok, a hill station with a history of severe landslides, such as the 1997 cloudburst-induced landslide, continues to face elevated risk, making such localized studies essential for mitigation planning.
In response, a team from Amrita University undertook a comprehensive field assessment in North Sikkim to support the development of resilient multi-hazard early warning systems.
The evaluation included a technical survey of the affected hydropower dam to evaluate structural vulnerabilities and the potential for cascade failures.
In parallel, the team conducted stakeholder interviews with government officials, dam operators, and community leaders to identify key challenges and opportunities in early warning communication, preparedness, and disaster response.
An Acoustic Doppler Current Profiler (ADCP) was also utilised to measure river discharge and flow velocity, providing critical data for hydrodynamic modeling and post-event hazard reconstruction. These integrated efforts aim to build a more data-informed, site-specific, and community-centric approach to multi-hazard risk reduction in the Eastern Himalayas.
Researchers from Amrita have developed a real-time, data-driven, multi-hazard forecasting and early warning dissemination platform. This platform is not only dynamic and responsive but also impact-based, integrating real-time monitoring, modeling, and decision support tools to inform effective disaster response and preparedness.
A Real-Time, Data-Driven, Multi-Hazard Forecasting and Early Warning Platform for the Eastern Himalayas
At its core, the system operates at a granular spatial scale of approximately 5 km² across vulnerable regions of North Sikkim. This enables localized simulation of hydrological processes and hazard behavior, particularly for GLOFs, landslides, and floods.
The platform utilizes heterogeneous models to simulate key hydrometeorological variables, including rainfall, snowfall, evapotranspiration, soil moisture, groundwater infiltration, and discharge, thereby capturing the entire chain of water movement from canopy interception to subsurface storage.
Precipitation data is disaggregated into rainfall and snowfall to reflect high-altitude glacial dynamics. These inputs flow through a rainfall runoff model, generating real-time water accumulation estimates that enable the rapid identification of flood-prone or slope-unstable areas under various conditions.
Multi-Source Data Integration for Dynamic Risk Forecasting
The platform leverages a rich set of heterogeneous, multi-source datasets to support high-resolution, real-time forecasting. These include:
These inputs feed into an integrated modeling system that monitors triggering factors (e.g., rainfall, snowmelt) and causative factors (e.g., slope, drainage density, vegetation cover) to assess risk levels in each sub-basin dynamically.
Structural-Level Monitoring and In-Situ Instrumentation
In addition to regional modeling, the platform supports site-specific structural-scale monitoring. In high-risk areas experiencing land subsidence (the gradual caving in or sinking of an area of land) and widening cracks, tiltmeters and crackmeters will be installed on critical structures to monitor their movement.
These sensors measure, in real-time, the progressive widening or movement of cracks and structural tilt, enabling the early detection of instability and the prediction of failure mechanisms at the structural level.
A key innovation of the platform lies in its impact-based modeling, which goes beyond hazard forecasting to include mapping of elements at risk, such as hydropower dams, roads, bridges, schools, settlements, and agricultural areas.
The platform identifies which specific assets are at risk within a sub-basin and connects this information to nearby shelters, evacuation routes, and critical service infrastructure.
Hence, the system operates at four interconnected domains:
This multi-scalar approach ensures that decision-makers are informed not just about where risk exists, but how, why, and when it evolves over space and time.
Field Trials and Teesta’s Challenge
One of the model’s most critical variables—discharge (streamflow)—requires precise in-situ measurements. To address this, the team deployed an Acoustic Doppler Current Profiler (ADCP) at key river cross-sections in the Teesta.
But Teesta is no ordinary river.
Often dubbed one of India’s fastest-flowing rivers, Teesta posed extreme operational challenges. In the rugged terrain of Sikkim during the onset of the monsoon, researchers had to string ropes across the river and manually guide the ADCP from one bank to the other. On multiple occasions, the device toppled in the current, and ropes snapped under pressure. Undeterred, the team pushed on.
Local Army personnel stationed nearby took note. Intrigued by the scientific mission—and witnessing its potential firsthand—they lent their manpower and logistical support, helping the researchers safely complete multiple cross-sectional discharge measurements. The camaraderie extended beyond science, with shared meals in Army tents and an appreciation for the value of this collaborative work.
A Comprehensive Multi-Hazard Risk Management System
Ultimately, this platform supports all phases of the disaster management cycle—preparedness, mitigation, response, and recovery. It is not just a forecasting tool; it is a decision-support ecosystem that integrates real-time data, structural monitoring, in-situ instrumentation, machine learning tools, and geospatial intelligence.
By combining regional-scale modeling, site-specific sensor data, and structural risk assessment, it provides a scalable and replicable model for multi-hazard risk reduction—one that is grounded in both scientific rigor and field realities.
This system represents a critical step forward in creating resilient mountain communities in the Eastern Himalayas and can serve as a blueprint for hazard-prone regions globally.
A Commitment to Resilience
Amrita University’s project is an integral part of our Chancellor Mata Amritanandamayi Devi (Amma’s) vision of compassion-driven research that combines science, community engagement, and sustainable innovation. Through international and government partnerships, the university is redefining what it means to build climate resilience in India.
As Sikkim braces for future monsoons, it does so with a new ally—data.
“Science cannot stop the rain. But with the right tools, it can ensure the rain doesn’t turn into disaster.”
