Climate Software Lab (CSL) is the climate technology arm of GPS Renewables, a full-stack clean fuels company. Composed of AI scientists and technology experts who leverage AI, machine learning and Information Technology, the lab works to create software solutions for climate adaptation and climate transformation. Additionally, the CSL team is integral in digitising GPS Renewables' internal workflows, developing real-time monitoring dashboards for operational KPIs, streamlining project procurement, and managing our HSE performance. Each of these interventions is highly scalable and improves efficiency. 

This is an INDIAai exclusive. CSL's Chief Sustainability Officer, Nipun OS, and Head of Technology, Arun M, speak about their newly launched AI-based Landslide Risk Assessment model, which can map landslide risks. The tool covers susceptible areas in the Indian subcontinent, where nearly 40 lakh people live. Read to find out how AI plays a significant role in Landslide Risk Assessment and what makes it relevant in India. 

What makes Climate Software Lab's (CSL) newly launched AI-based Landslide Risk Assessment model relevant today? 

Landslides pose a significant threat to parts of India, including critical, but already fragile ecosystems such as the Himalayas, North-Eastern hill ranges, Western and Eastern Ghats, Nilgiris, and Vindhyas.  

Unlike other natural disasters, it is difficult to predict when and where landslides will occur, and policy planners, scientists, and administrators are faced with the difficult task of formulating effective mitigation strategies.  

During our research, we found that many states in India depend on old and low-resolution susceptibility maps to draw up landslide prediction models.  

CSL's tool addresses this gap by providing high-quality risk visualisations that would enable policymakers to develop local-level disaster response strategies. The free and open-access tool makes climate-related data more accessible to more people. 

There are numerous AI-based landslide assessment models available in the market. What makes this model different? 

First, CSL's model is free and open source. 

It maps an area's susceptibility using a host of local factors, including tree height and its rate of change, soil clay content, and human modification. It also uses the Geological Survey of India's (GSI) dataset, which contains information about past landslide locations.  

It is based on NASA's existing Landslide Hazard Assessment for Situational Awareness (LHASA) model, but it is tailored to Indian conditions and provides better spatial resolution, enabling policymakers and administrators to plan local interventions for vulnerable populations and important assets. 

What is the Climate Software Lab? Can you explain the company's AI journey? 

The Climate Software Lab (CSL) is the climate technology arm of GPS Renewables, a full-stack clean fuels company.  

Within CSL, a team of data scientists and climate experts harnesses the potential of artificial intelligence (AI), machine learning (ML), geospatial systems, and data intelligence tools to develop software solutions for climate action and adaptation. 

The lab was born from our understanding of the urgent need for climate change mitigation and adaptation and our passion for using machine learning for the greater good. The idea is to develop a suite of tech-enabled open-source climate solutions that benefit all.   

Additionally, the CSL team plays an integral role in digitising and streamlining GPS Renewables' internal workflows by developing real-time monitoring dashboards for operational KPIs and project procurement and managing our HSE performance. Each intervention is highly scalable and improves the company's overall efficiency.

How do you think we can leverage AI in disaster management? 

With the increasing availability of high-resolution satellite imagery, machine learning systems can analyse vast datasets to predict and issue early warnings for natural disasters like landslides, floods, wildfires, and other threats.  

AI can further optimise emergency response by analysing real-time data on infrastructure damage and resource needs derived from these detailed satellite images.

This allows for faster deployment of personnel and supplies, saving lives and minimising losses. 

In short, AI can help us move from reactive to proactive disaster management, significantly improving preparedness and response. 

Are there any other AI models CSL developed that are significant in disaster management? 

The team is working on making the Landslide tool even more dynamic by including rain metrics, as rainfall is often a trigger event for landslides. The team is also developing tools to improve our in-house capacity to conduct preliminary environmental assessments. 

What are the company's plans under the pipeline? 

The CSL team plans to develop tools to monitor rainfall and visualise changing patterns. This is important considering that the nature of rainfall events is changing drastically, and these changing patterns also have implications for landslide prediction in near real-time.  

Over the past two years, the team has worked on developing an in-house technology solution called # Carbon. This solution combines machine learning and blockchain technologies to eliminate data tampering from sensors to storage in all our projects. 

The team is also developing tools to improve our in-house capacity to conduct preliminary environmental assessments to meet our ESG targets.