Some hydrologists learn a river by walking its bank. Satish Regonda learned his by disciplining its decades of noise into a forecast — teaching a statistical model to say, with honest uncertainty attached, what a river is likely to do next. It is a discipline built on humility as much as mathematics: a probabilistic forecast is, in essence, an admission that the future is not one number but a distribution of possibilities, and the skill lies in describing that distribution well. Regonda has spent more than two decades doing exactly that, first on rivers eight thousand miles from where he grew up, and now, back in Hyderabad, on the questions his own country is asking with new urgency.
His training reads like a itinerary through the world’s serious water schools. A civil engineering degree from Kakatiya Institute of Technology and Sciences in Telangana led him to the Indian Institute of Technology, Kanpur, for a master’s in hydraulics and water resources, and from there to the University of Colorado Boulder for a doctorate in water resources engineering, completed in 2006. It was at Boulder that he built a multi-model ensemble forecasting framework for the Gunnison River in Colorado, work that fed directly into a decision-support system used to weigh the practical benefits of forecasting itself, not an abstract exercise but a tool meant to be used by someone deciding how much water to release from a dam.
That practical instinct carried him into the United States’ own forecasting establishment. As a National Research Council postdoctoral associate and then a research scientist with NOAA’s National Weather Service, at the Office of Hydrologic Development in Silver Spring, Maryland, he spent the better part of seven years developing what became Hydrologic Model Output Statistics, or HMOS, a technique for turning a single, deterministic streamflow forecast into a full probabilistic one. He did not merely publish the idea; he built a prototype, installed it at a River Forecast Center, and wrote the manuals that taught working forecasters how to trust a number that came with a confidence interval attached to it, rather than a false certainty. Alongside that, he held an appointment as an Associate Research Scientist at Johns Hopkins University’s Department of Earth and Planetary Sciences, and consulted for the World Bank on hydroclimate questions, the kind of assignment that asks an American-trained specialist to make his statistics useful in places the models were never built for.
In 2017, that path brought him to the Indian Institute of Technology Hyderabad, and it is here that the arc of his career acquires a second, more personal chapter. He has served as Head of the Department of Climate Change, building its curriculum and its collaborations from the ground up; he now holds an Associate Professorship spanning Civil Engineering and Climate Change; and since 2023 he has chaired the institute’s Rural Development Center, a role that returns him, in spirit, to the villages his statistics were always meant to serve. It is a career that began by teaching an American river forecast center to speak in probabilities, and has arrived at a Hyderabad institute asking what those same tools of statistics, data analysis and hydroclimate modelling can do for a Ganga at flood stage, a Godavari in spate, or a district administration trying to decide, with the little lead time nature allows, whether to sound an alarm.
Few researchers working on Indian flood hydrology today carry into that work the operational memory of having built a forecasting system that a river authority actually ran. NSH Research Desk
