Machine Learning Sharpens Winter Storm Forecasts
A Penn State graduate student presented radar research aimed at improving snowfall and winter weather prediction.
Updated on Nov. 10, 2026 in Current Conditions

Live Poll
Do you trust your local forecasts to give enough warning before winter storms?
Song Zhang, a PhD student in Penn State's Department of Meteorology and Atmospheric Science, presented research on using machine learning with polarimetric radar to improve winter storm forecasts. The lecture was held in room 117 of the Earth and Engineering Sciences Building as part of the TRAIN seminar series.
Why it matters
Winter storms remain one of the most challenging forecast problems, and better estimates of snowfall and storm structure could help meteorologists anticipate hazardous conditions. State College residents and the Penn State community benefit when forecasting tools improve for regional winter weather.
The lecture drew over a dozen attendees in person and over Zoom to room 117 of the Earth and Engineering Sciences Building. Zhang's research links polarimetric radar snowfall estimates to real-time temperature changes by altitude, though the source gives no accuracy benchmarks.
The players
Song Zhang
Song Zhang is a PhD student in Penn State's Department of Meteorology and Atmospheric Science who has published several research papers on winter weather.
NASA
NASA is the U.S. federal space agency that sponsored the IMPACTS project under which part of Zhang's winter weather work was conducted.
The details
Zhang's work uses machine learning to estimate snowfall from polarimetric radar, which sends and receives horizontal and vertical pings to produce a 3D view of the atmosphere. Data assimilation connects those radar readings to real-time winter weather observations, and his research was part of NASA's IMPACTS project.
Timeline
January: a winter storm occurred that illustrates the forecasting challenge.
November 6, 2026: Zhang presented the machine learning forecasting lecture.
Seasonal Patterns
Zhang's work was part of NASA's IMPACTS project, a research program that studied winter storms and snowfall using airborne and ground instruments. The lecture extends that line of inquiry by applying machine learning to polarimetric radar data for winter forecasting.
Residents preparing for winter storms should monitor official National Weather Service forecasts and follow local advisories for snow and hazardous driving. Improved radar-based snowfall estimates may eventually sharpen warnings, but current guidance should guide planning.
The takeaway
Better winter forecasts depend on combining new sensing tools with data assimilation and machine learning, so storm predictions can improve for the communities they affect. Keep checking official forecasts as methods continue to be tested and refined.
Further reading
Explore more current conditions reporting for State College.
Source note: This article includes information reported by The Daily Collegian.
Live Poll
Do you trust your local forecasts to give enough warning before winter storms?










