The Sikkim hydro engineer and the Teesta III reckoning
🏔️ Ashok Sharma, 32, works as a consulting hydro engineer in Gangtok, Sikkim. He was in the office on June 4, 2023, when an SMS alert came through: glacial lake outburst flood, Teesta III dam failed, 41 people dead downstream. No warning. No alert system. The moraine dam failed before dawn. Two hours and eleven minutes after the dam failed, people were dead. Ashok had never seen a discharge schedule like that. In May 2025, he was offered a lead consulting role on Teesta VI—the lower river, 66 MW facility. Worth ₹38 lakhs. Enough to run the office for a year. He signed. By March 2026, he had not read the glacial-risk data since the proposal was filed.
🚨 The problem
Three new proglacial lakes appeared on satellite imagery between March 15 and April 20, 2025—small, growing, in positions where breach would funnel into the same watershed that fed Lhonak Lake in 2023. A postdoctoral researcher published a preprint warning that new proglacial lake formation on the Teesta had accelerated in 2024–2025. Ashok saw the preprint circulated in an IEI email on April 28. He read the subject line. He did not read the paper. His DPR deadline was May 25. Updating the risk assessment would cost another ₹1.5 lakhs to the NIT Sikkim glaciologist and delay the DPR by two weeks. He did not update.
🚀 How GabFORGE helped
His junior colleague Tenzin brought a tablet into the office with an agent already installed. He asked it: "What has changed in glacial-risk data for the Teesta headwaters since March?" The agent listed the three new lakes, their growth rates, and flagged the IEI email from April 28 that Ashok had not opened. Ashok asked: "If all three lakes breach, what happens to my Teesta VI spillway design?" The agent replied: Your spillway is designed for 150 cumecs. Three lakes breaching = 500–800 cumecs additional discharge. You need to update the design.
🔍 Monitored the satellite data — tracked three new lakes and their growth rates. 💬 Flagged the research — surfaced the academic preprint Ashok had not read. 📞 Raised the design question — forced recalculation before submission.
Ashok called the NIT Sikkim glaciologist, who re-analyzed the risk as "high" for all three lakes and 70% probability of at least one breach by December 2025. He rewrote the DPR, added forty pages of detailed GLOF modeling, and resubmitted on June 18—three weeks late. NHPC approved the revised DPR on July 9 with a commendation for "enhanced risk-aware design." ✅
🇮🇳 Why this matters
Information fragmentation means satellite imagery lives in multiple systems, academic preprints arrive unvetted, and an engineer can be sleeping through a warning signal that arrived in an email subject line.
The long version contains [details from original article].