Every fall, because the Northern Hemisphere approaches winter, Judah Cohen begins to piece together a fancy atmospheric puzzle. Cohen, a research scientist in MIT's Department of Civil and Environmental Engineering (CEE), has spent many years studying how conditions within the Arctic set the tone for winter weather in Europe, Asia, and North America. His research stems from his postdoctoral work with Professor Dara Entekhabi of Bacardi and the Stockholm Water Foundations, where he checked out snow cover within the Siberian region and its relationship to winter forecasting.
Cohen's outlook for winter 2025-26 marks a season marked by indicators emerging from the Arctic using a brand new generation of artificial intelligence tools that help develop the complete atmospheric picture.
Look beyond the standard climate drivers
Winter forecasts rely heavily on El Niño Southern Oscillation (ENSO) diagnostics, that are the conditions of the tropical Pacific and atmosphere that influence weather all over the world. However, Cohen notes that ENSO is comparatively weak this yr.
“When ENSO is weak, climate indicators from the Arctic develop into particularly vital,” says Cohen.
Cohen monitors high-latitude diagnostics in his subseasonal forecasts, comparable to October snowpack in Siberia, early-season temperature changes, Arctic sea ice extent, and polar vortex stability. “These indicators can tell a surprisingly detailed story in regards to the upcoming winter,” he says.
One of Cohen's most consistent data predictions is October weather in Siberia. This yr, because the Northern Hemisphere experienced an unusually warm October, Siberia was colder than normal and snow fell early. “Cold temperatures coupled with early snow cover are likely to increase the formation of cold air masses that may later reach Europe and North America,” says Cohen — weather patterns historically related to more frequent cold spells later within the winter.
Warm ocean temperatures within the Barents-Kara Sea and an “eastern” phase of the quasi-biennial oscillation also suggest a possible weaker polar vortex in early winter. When this disturbance is combined with December surface conditions, it ends in lower than normal temperatures early within the season across parts of Eurasia and North America.
Subseasonal AI forecast
While AI weather models have made impressive progress, evident in forecasts for brief periods (one to 10 days), these advances will not be yet applicable to longer periods. Subseasonal forecasting over two to 6 weeks stays certainly one of the largest challenges in the sector.
This gap is why this yr could possibly be a turning point for subseasonal weather forecasting. A team of researchers led by Cohen won first place for the autumn season within the subseasonal prediction competition AI WeatherQuest 2025, hosted by the European Center for Medium-Range Weather Forecasts (ECMWF). The challenge assesses how well AI models capture temperature patterns over multiple weeks where forecasting has historically been limited.
The winning model combined machine learning pattern recognition with the identical Arctic diagnostics that Cohen has refined over many years. The system demonstrated significant progress in multi-week forecasts, outperforming leading AI and statistical baselines.
“If this level of performance is sustained over multiple seasons, it could represent an actual advance for off-season prediction,” says Cohen
The model also detected a possible mid-December cold surge on the US East Coast much sooner than usual, weeks before such signals typically occur. The forecast was widely publicized within the media in real time. If validated, Cohen explains, it might show how combining Arctic indicators with AI could extend the lead time for predicting influential weather.
“It could be a game-changer to report a possible extreme event three to 4 weeks prematurely,” he adds. “It would give utilities, transportation systems and public agencies more time to organize.”
Whatever this winter may bring
Cohen's model shows that parts of Eurasia and central North America usually tend to experience colder than normal temperatures later within the winter, with the strongest anomalies likely occurring in the course of the season.
“We are still within the early stages and patterns may change,” says Cohen. “But the ingredients for a colder winter are there.”
As Arctic warming accelerates, its impact on winter behavior is becoming increasingly clear. It is subsequently becoming increasingly vital to grasp these relationships for energy planning, transport and public safety. Cohen's work shows that the Arctic has untapped subseasonal forecasting capabilities, and AI could help unlock these on time frames which have long been difficult for traditional models.
In November, Cohen even emerged as an informant Crossword puzzlea small sign of how widely his research has found its way into the general public debate about winter weather.
“For me, the Arctic has all the time been the place to look at,” he says. “Now AI gives us latest ways to interpret its signals.”
Cohen will proceed to update his outlook because the season progresses Blog.
