At least one alternative must be selected

Real-time monitoring and Early Warning of Rainfall-induced Landslides

Please briefly describe your Water ChangeMaker journey

Landslides are one of the major natural catastrophes claiming the lives of many in India. According to a news article by Deccan Herald, “An international database of fatal landslides, prepared by the University of Sheffield, England shows that with close to 11,000 deaths due to landslides in 12 years, India tops a global list of nearly 56,000 casualties from 4,800 landslides around the world between 2004 and 2016.”The Himalayas, the Northeastern hill ranges, the Western Ghats, the Nilgiris, etc. are among the major landslide-prone regions of the country. Majority of these landslides are triggered mainly due to the rainfall received by the country during the monsoon season. Therefore, the problem was to detect such disasters ahead of time so that necessary precautionary and mitigation measures can be taken. Hence, our initiative was to develop systems to monitor such rainfall-triggered landslides and provide real-time warnings for saving the lives of the affected communities.

Please describe the change that your initiative created and how was it achieved

Through our initiative, we have conceived, designed, developed, and deployed a complete and integrated system, which is able to monitor the hydrologic, geologic, and meteorological parameters that could lead to a landslide and thus, detect the possibility of occurrence of such an event ahead of time. A momentum of change was made possible not only by leveraging the latest trends in technology such as IoT, Machine Learning & Artificial Intelligence, edge computing etc. but also by bringing in a participatory approach through community involvement. Multi-level, multi-phase intervention strategies were designed based on the communities’ socio-economic, and political background. Change agents had to be identified from the community, who will act as ambassadors for the change management and development of sustainable pathways for the prospective change. A Wireless Sensor Network (WSN) based monitoring system started functioning in the Munnar region of Western Ghats, South India in 2009, which later evolved into an IoT based system, comprising of a multi-level warning framework. An IoT system for monitoring and early warning rainfall-induced landslides was operational in Chandmari, Sikkim (North-east India) since 2015. Both these systems are continuously collecting real-time sensor data, which is being analyzed at the data management center located in our University. The system in Munnar was successfully validated by delivering early warnings in 2009, 2011, 2013, 2018, and 2019, thus saving human life in the deployment area.

How did your initiative help build resilience to climate change?

Climate change is contributing to increase in extreme weather conditions leading to an increase in frequency of landslides, and its unpredictability. The recent Kerala floods and landslides during 2018 & 2019 are an example of that. Hence it’s necessary to build models that could predict the possibility of occurrence of such disastrous events. In this regard, we have developed rainfall threshold models as part of our landslide early warning system. For instance, in our Sikkim deployment, these models were developed at two different spatial scales – i) site-specific model for Gangtok area of Sikkim, and ii) regional model for entire Sikkim area. We have also developed pore-water pressure forecasting and threshold models, which will estimate the anticipated pore-water pressure based on the rainfall forecast, and deliver alerts if it has crossed the threshold. The criticality based on the alert level obtained from these models gives us an idea of whether a landslide occurrence is imminent or not.

What water-related decisions did your initiative influence or improve?

In our experience, especially in our deployment in the Western Ghats, we determined through both historical data and our own observations, that rainfall induced landslides are usually an ‘upstream’ event for floods. Slides of landmass, especially in the hilly regions end up blocking the path followed by a natural flood flow, resulting in further build up of water and causing gigantic floods subsequently. This was clearly seen during the Kerala floods in 2018 and 2019. Hence our Landslide early warning system feeds into flood monitoring and warning systems already existent. We are also fine tuning our models to reflect this effect.

What were some of the challenges faced and how were they overcome?

One of the key aspects of this initiative was that it was thoroughly an interdisciplinary project involving multiple domains such as geology, geotechnology, meteorology, hydrology, sensor system, signal processing system, power system, communication & networking system, software development, etc. Hence, it was a challenging task to bring together these different components to build an integrated system, capable of achieving the end result of delivering timely, reliable and accurate warnings. Another challenge faced was the variability in the terrain and environment between the multiple deployment sites. Hence, the system design has to be enhanced accordingly to cater to the new requirements and challenges set forth by each study area. Ours being a community participatory system, meant there were further challenges in getting buy in from the local community and create a sense of ownership for them. This was a huge practical challenge.

In your view: Will the change that was created by your initiative continue?

The changes achieved by this initiative will definitely endure as we had brought in a sustainable solution to the problem that we addressed. The flexibility and ubiquitous nature of the system that was achieved through the adoption of IoT system is capable of supporting real world multiple monitoring applications for multiple sites using multi-communication networks. Thus, this system can even be extended to monitor multiple applications such as flood monitoring, drought monitoring, water quality monitoring etc. One of the major risk that we can foresee is the impact to the deployment when an actual landslide occurs in the specific locations where the devices are deployed.

What did you learn during the initiative or after? And is it possible that others could learn from you?

Several key learnings were achieved through the implementation of this project, from understanding the complex phenomena behind landslide triggering to how an effective warning system can be devised through the interlinking of science, technology, policies, and community. More than a decade long of IoT sensor big data such rain gauge data, moisture data, pore-pressure data, etc. was obtained from our real-time, real-world deployment, which was processed and analyzed in a spatio-temporal manner using techniques like Machine Learning & Artificial Intelligence, for comprehending the intricate process behind how rainfall is leading to landslides. At the decision-making level, we were able to learn how administrators & policy makers need to be engaged and empowered for dynamic change management, and how communities need to be trained in understanding different levels of risk communication. It was a combination of our experiences and knowledge acquired from research, publications, etc. which made it possible to bring such a change.