PhD Candidate, University of Colorado Boulder
Building a real-time operational early warning system for riverine flooding: river stage forecasting and flood inundation mapping at multiple-day lead times, with the Brahmaputra River Basin as the primary testbed.
The lens sweeps across the basin behind this page, zooming in on where calm flow gives way to flood risk.
Stage and streamflow forecasting via hierarchical Bayesian networks (BHNM), and space-time precipitation post-processing (HBPP), in R and Stan.
Gauge-informed reconstruction of continuous water surface elevation from sparse SWOT altimetry (BRACE), across the Brahmaputra, Willamette, and CONUS-scale networks.
Bayesian Causal Networks linking Brahmaputra streamflow and Colorado River drought to ENSO, IOD, and MJO drivers.
Ensemble HEC-RAS 2D simulations driven by posterior stage samples, producing exceedance probability flood maps.
Unless noted otherwise, the work below is part of my Ph.D. research at the University of Colorado Boulder.
This entire page's backdrop is the live simulation: watch the water surface elevation climb on the clock below, then click the button to see what changes when a warning goes out ahead of time.
Applying BRACE across 250+ USGS gauges to benchmark WSE against OPERA DSWx-HLS and SWOT width products.
Jointly modeling Sentinel-2-derived river width and SWOT width at matched Willamette nodes.
Diagnosing atmospheric and oceanic precursors behind the 25-year Colorado River Basin drought.
Extending the published BCN paper with Exploratory Graph Analysis and correlation networks.
Causal drivers of All India Summer Monsoon Precipitation using graph-based Bayesian networks, feeding an upcoming AMS 2027 abstract.
AGU Fall Meeting 2026: Primary convener, Session H096, "Land-Ocean-Atmosphere Drivers of Hydroclimatic Extremes: Prediction and Early Warning in the South Asian Monsoon," with co-conveners Balaji Rajagopalan, Manabendra Saharia, and Priyam Deka. Invited speaker: Dr. Madhavan Nair Rajeevan.