Chief of disaster risk reduction
Sanjay K Srivastava is presently the Chief ofDisaster Risk Reduction at UN Economic and SocialCommission for Asia and the Pacific (ESCAP).He leads and guides ESCAP’s analytical policy research on disaster risk reduction and climate change adaptation that feeds into the UN intergovernmental meetings, negotiations, and agreements as well as supports to development of national DRR strategies, National Adaptation Plan etc.
Prior to the present position, he was the ESCAP Regional Adviser on Disaster Risk Reduction; Head of SAAARC Disaster Management Centre – New Delhi; Scientist/Engineer and Deputy Project Director of Disaster Management Support Programme at Indian Space Research Organisation (ISRO) at ISRO HQ Bangalore since 1991.
He has received 2 nd ESCAP innovation awards (2022 and 2024) and is the recipient of ISRO’s Team excellence award in 2008-09. He has more than 150 publications to his credit. He is the lead author of the UNESCAP’s Asia-Pacific Disaster Report, WMO’s State of Climate in Asia and State of Climate Southwest Pacific.
Academic qualification: Ph D (Remote Sensing) from IARI, New Delhi 1989; Master of Science (Physics) from Patna University (1984) and Bachelor of Science (Physics Honours) from Patna University 1980. University of California, Berkely Haas (US) – Executive Education – Certificate course on Digital Transformation: Leading People, Data and Technology (2023)
More heat in the ocean intensifies cyclonic risk
The planet is surpassing past heat records at an alarming rate. The year 2024 was the hottest on record, continuing a decade-long trend of unprecedented global temperatures. All the past ten years now rank among the hottest ever recorded. This manifests in terms of more heat in the coupled ocean and atmosphere systems that supercharge tropical cyclones in the Bay of Bengal and Arabian Sea. Further, the TCs that intensify rapidly are difficult to predict. This can lead to forecast errors. Rapid intensification, curvature, newer and complex tracks result in unchecked casualties, damage and losses. Cyclone Titli east coast of India, 2018, after the landfall, triggered landslides and flooding that killed more than 50 people, despite a very precise early warning.
At the 52nd Session of WMO/ESCAP Panel on Tropical Cyclone, AI to scale early warnings for All was discussed. The UN initiative on Early Warnings for All EW4All by 2027 advances but vulnerable countries continue to suffer disproportionately. Two important challenges have emerged. One, even the countries with the Multi-hazard Early Warning System MHEWS capacities, there are gaps that exist among the pillars: risk knowledge, forecasting and detection, warning dissemination, and timely response actions. Two, whilst floods and storms have traditionally been associated with the worst losses, extreme heat has emerged as a major killer.
EW4All: Can Artificial Intelligence be a game changer?
Data revolution: The 4 pillars of EW4All are interdependent and data intensive. The transformative potential of AI lies filling in critical gaps that exist in data, modeling and scenario visualization. The RI of TCs remains one of the most challenging phenomena to forecast because of its unpredictable and destructive nature. Traditional forecasting methods, such as numerical weather prediction and statistical approaches, often fail to consider the complex environmental and structural factors driving RI, while AI has improved RI predictions
Technology trends: The enabling machine and deep learning technologies that create data for AI-enabled EW4All are evolving rapidly. These technologies include segmentation algorithms, intelligent satellites to detect image changes, impact scenarios, drones mapping and sensing and multi-lingual natural language processing. The trend is moving towards drone-enabled communication, fully automated drone swarms with connectivity, digital twin technology, trustworthy AI and crowdsourcing, and automated image analysis.
5 Ways AI can scale: AI can significantly enhance disaster risk knowledge for EWS especially in data-scarce regions. FloodSENS, for example, has created an algorithm that efficiently reconstructs flooded areas under partial cloud cover in optical satellite images, using Machine Learning and auxiliary derivative layers from digital elevation models, and water flow algorithms
AI advances predictive analytics, real-time data assessments, platforms that consolidate and disseminate severe weather information. On effective communication, AI optimizes alerts, tailor messages through various channels, translate warnings into multiple languages to facilitate
Source: UNU EHS 2024
actionable warnings. AI’s ability to simulate different emergency scenarios offers valuable insights for preparedness and response through real-time assessments . Google Research, for example, uses AI to accurately predict riverine flooding and help protect livelihoods in over 80 countries up to 7 days in advance, including in data scarce and vulnerable areas.
Challenges: It is important to recognize that AI is a complementary tool for gathering insights to ensure that innovation is people-centred and equitable. The challenges include, first, AI tools should meet certain standards to not only improve data management and curation but to address bias, security, ethics and other important considerations. Transparent algorithms, accountability measures and bias mitigation strategies are essential to ensure an ethical application of AI to leverage EW4All. Building trust in AI-generated warnings is still a challenge. False alarms or incorrect data can erode confidence, validation of AI-driven alerts is necessary to maintain trust. Furthermore, AI-based solutions require substantial resources that are particularly missing in data-scarce regions.
Leveraging Regional Cooperation
The Regional Specialized Meteorological Centre RSMC at the India Meteorological Department, under the PTC framework, monitors cyclones across the Northern Indian Ocean—from Oman to Myanmar. Working closely with its 13 PTC member states, the RSMC facilitates the delivery of early warnings for transboundary tropical cyclones. As members modernize their MHEWS, collaborative research is underway to integrate AI and machine learning methods into forecasting. A notable innovation is the use of AI to monitor and predict Tropical Cyclone Heat Potential TCHP a critical factor in enhancing EW4All within the specific context of tropical cyclones.
The Regional Specialized Meteorological Centre RSMC at the India Meteorological Department, under the PTC framework, monitors cyclones across the Northern Indian Ocean—from Oman to Myanmar. Working closely with its 13 PTC member states, the RSMC facilitates the delivery of early warnings for transboundary tropical cyclones. As members modernize their MHEWS, collaborative research is underway to integrate AI and machine learning methods into forecasting. A notable innovation is the use of AI to monitor and predict Tropical Cyclone Heat Potential TCHP a critical factor in enhancing EW4All within the specific context of tropical cyclones.