The Northern Great Plains (NGP), spanning Montana, Wyoming, the Dakotas, and Nebraska, is more than just America’s agricultural heartland. This vast region produces two-thirds of U.S. wheat and over half its beef, while its unique "pothole" wetlands support 50% of North America’s waterfowl. But this ecological and economic powerhouse faces a hidden crisis: wildly unpredictable warm-season rains. Prasad Thota did his PhD research at the University of Colorado to solve this puzzle: why does this region swing between devastating droughts and catastrophic floods, and can we predict these extremes?

The Rainfall Variability Puzzle
Prasad’s work revealed a striking pattern: 65% of the NGP’s annual rain falls during just five warm months, mostly in intense bursts. This isn’t gentle drizzle, it’s make-or-break precipitation for farms, ecosystems, and communities. In 2011, floods ravaged the Missouri River basin; a year later, drought caused $12 billion in losses. Rural and Native American communities bear the brunt. Prasad traced these swings to distant ocean temperatures and atmospheric "teleconnections" – climate handshakes between the Pacific and the Plains. Using NOAA climate data, he identified a critical driver: a temperature gradient in the North Pacific Ocean. When warm waters near Alaska contrasted sharply with cooler waters southwest of them (i.e., strong gradient), moisture surged into the Plains via the "Great Plains Low-Level Jet." When this gradient weakened, droughts occurred.

From Climate Clues to Predictive Power
But how do you turn this insight into forecasts? Prasad built a semi-Bayesian hierarchical model, a sophisticated statistical framework that "learns" from history. He fed it predictors like Pacific Ocean temperatures and land moisture levels (using the Palmer Drought Index) 1-2 months before the rainy season. The model didn’t just predict seasonal totals; it estimated the risk of extreme 3-day downpours. Validated against 60 years of station data, his forecasts outperformed traditional methods, offering ranchers and water managers a crucial heads-up. "The North Pacific’s memory holds the key," Prasad explained. "Its patterns persist long enough to give us a forecasting window."

Climate Models vs. Reality
Finally, Prasad tested whether current state-of-the-art climate models (like CESM2) could replicate the patterns he identified. While these models captured some broad features, they often failed to reproduce key details such as the east-west rainfall gradient. Further, while observed NGP rainfall increased slightly since 1980 (especially eastward), these models generally project a drier future. He traced these shortcomings to an unrealistic warming pattern in the Pacific Ocean simulated by the model, which weakened the mechanisms that typically bring summer moisture to the region. "If models can’t capture current dynamics," Prasad cautioned, "their drought projections for farmers need careful interpretation."

A Toolkit for Resilience
Prasad’s dissertation delivers more than diagnostics, it offers solutions. His Bayesian framework could evolve into an operational forecast tool, providing early warnings for floods and droughts. Future work could refine it with real-time data or extend it to project climate impacts. For the Northern Great Plains, where a single rainstorm can make or break a season, this science isn’t abstract. It’s the foundation for adapting to an uncertain climate, protecting ecosystems, food security, and rural livelihoods. NC CASC Climate Scientist Imtiaz Rangwala who was a co-advisor on Prasad’s PhD notes: "Prasad’s work certainly shines light on factors affecting rainfall in the Northern Plains and ability to predict them months in advance."

Prasad Thota successfully defended his PhD in Civil Engineering in May 2025. His work was partially supported by the North Central Climate Adaptation Science Center and the CIRES Graduate Student Research Award Program. Prasad also developed various decision-support tools (R-Shiny Apps) for the NC CASC and the Western Water Assessment.