Every autumn, as the Northern Hemisphere approaches winter, Judah Cohen begins to piece together the complex puzzle of the universe. Cohen, a research scientist in MIT’s Department of Civil and Environmental Engineering (CEE), has spent decades studying how Arctic conditions overshadow winter weather across Europe, Asia and North America. His research dates back to his postdoctoral work with Bacardi and Stockholm Water Foundations Professor Dara Entekhabi who looked at snow cover in the Siberian region and its connection to winter forecasting.
Cohen’s vision for the winter of 2025–26 highlights the season with indicators from the Arctic using a new generation of artificial intelligence tools that help develop a full picture of the atmosphere.
Looking beyond traditional weather drivers
Winter forecasts rely heavily on the El Niño–Southern Oscillation (ENSO), which is a tropical Pacific Ocean and wind pattern that influences global weather. However, Cohen notes that ENSO is very weak this year.
“The weaker ENSO is, the more important the climate indicators from the Arctic become,” Cohen said.
Cohen monitors high-latitude anomalies in his subseasonal forecasts, such as October snow cover in Siberia, early-season temperature changes, Arctic sea ice extent, and polar vortex stability. “These indicators can tell an incredibly detailed story about the coming winter,” he said.
One of Cohen’s consistent data predictions is the October weather in Siberia. This year, while the Northern Hemisphere experienced an unusually warm October, Siberia was colder than usual with early snowfall. “The cold temperatures associated with early snow cover tend to strengthen the formation of cold air that may spread to Europe and North America later,” says Cohen—weather conditions historically associated with more frequent colds later in winter.
Warm sea temperatures in the Barents–Kara Sea and the “easterly” phase of the quasi-biennial oscillation also suggest a weakening of the polar vortex in early winter. If this disturbance is accompanied by higher conditions in December, it leads to lower-than-normal temperatures throughout Eurasia and North America at the beginning of the season.
AI for sub-seasonal forecasting
While AI weather models have made impressive strides in short-range (one to 10-day) forecasts, these advances have not been effective over longer periods. Off-season forecasting spanning two to six weeks remains one of the industry’s most difficult challenges.
That gap is why this year could be a turning point in the subseasonal weather forecast. A team of researchers working with Cohen won first place for the fall season in the 2025 AI WeatherQuest subseasonal forecasting competition, hosted by the European Center for Medium-Range Weather Forecasts (ECMWF). The challenge tests how well AI models capture temperature patterns over multiple weeks, when predictability has historically been limited.
The winning combination model for machine learning pattern recognition for Arctic diagnostics Cohen has refined over decades. The system has shown significant advantages in multi-week forecasting, surpassing the best AI and statistical bases.
“If this level of performance holds over multiple seasons, it could represent a real step forward in off-season forecasting,” Cohen said.
The model also found a possible mid-December cold snap on the US East Coast much earlier than usual, weeks before such symptoms appear. This prediction was widely disseminated in the media in real time. If confirmed, Cohen explains, it will show how combining Arctic indicators with AI can increase lead time for impactful weather forecasting.
He adds: “Falancing a dangerous event three to four weeks in advance can be a difficult time. It will give utilities, transportation systems and community organizations more time to prepare.”
What is possible this winter
Cohen’s model shows a greater chance of colder-than-normal conditions in parts of Eurasia and central North America later in the winter, with stronger anomalies likely in the middle of the season.
“It’s still early, and patterns can change,” Cohen said. “But the ingredients for a very cold winter pattern are there.”
As Arctic warming accelerates, its impact on winter behavior becomes more apparent, making it even more important to understand these connections that shape energy, transportation, and public safety. Cohen’s work shows that the Arctic holds untapped sub-seasonal predictive power, and AI may help open it up to time frames that have long been challenging for traditional models.
In November, Cohen even emerged as a clue The Washington Post crossword, a small sign of how widely his research has permeated public debates about winter weather.
“For me, the Arctic has always been a place to watch,” he says. “Now AI is giving us new ways to interpret its signals.”
Cohen will continue to update his opinion throughout the season on his blog.
