Atmospheric rivers do not dance to the tune of El Niño, a meteorologist told an audience at the ILS Convergence 2026 conference.

F. Martin Ralph, founder and director for the Centre for Western Weather and Water Extremes in San Diego, California, explained that ARs — long narrow bands of concentrated water vapour in the sky — can disrupt the precipitation patterns people might expect from El Niño or La Niña.

He said El Niño can bring more ordinary storms, while an atmospheric river or series of them, can drive much more significant flooding.

He pointed to the 2015-16 El Niño weather phenomenon dubbed “Godzilla”. It prompted expectations of major impacts, including very heavy rain in Southern California, but the region was actually quite dry until March.

“Three ARs came in and got the rain season up to normal, so it was a big dud for precipitation in Southern California compared to what was expected,” Dr Ralph said.

He contrasted it with the highly impactful 1982 to 1883 and 1997 to 1998 El Niño seasons.

“We have learnt our lesson to be a little more cautious about guaranteeing big impacts, but we need to be preparing for them,” Dr Ralph said. “ARs are really the disrupter of the patterns we are expecting to see. If an AR or an AR family hits during an El Niño year, that is when we are going to see really significant flood damages and the like.”

Dr Ralph’s team has created a regional weather model to forecast ARs over the western United States during winter. The model combines detailed physics, large ensembles and deep learning.

It runs at 200 members — an individual weather simulation, or forecast run, that is part of a larger ensemble.

“That is compared with 30 members for the United States model and 50 members for the European model,” Dr Ralph said. “The larger ensemble helps researchers better characterise the range of possible outcomes, including extreme events.”

At the conference held at the Hamilton Princess and Beach Club, he discussed efforts to extend forecast skill further into the future.

He and his team and other experts are trying to improve week-two forecasts by collecting observations from several atmospheric rivers across the Pacific and Atlantic.

Better measurements could help reduce errors in the storms’ predicted position, strength and structure.

“When we get all three of them pinned down, we have a real chance to move the needle on week-two forecast skill,” he said.