What will El Niño bring during the 2023-24 ski season?

When contemplating what this year’s ski season will look like, Joel Gratz says one thing is for certain: There will be snow.

Gratz, the founding meteorologist of snow forecasting and weather website OpenSnow, hosted a community Q&A at The Westin Riverfront in Avon on Thursday, Oct. 12, amid a crowd of Broncos fans and a fresh dusting of snow on Beaver Creek.

Gratz grew up skiing at Shawnee Mountain in the Poconos of Pennsylvania, and pursued a career in meteorology after a career in ski racing didn’t pan out.

---

---

“I was not a good ski racer because I didn’t like to take chances, but I really like skiing,” he said.

Gratz came to Colorado for grad school and began forecasting powder days for him and his friends. From there, he started an email list — Colorado Powder Forecast — then a blog, and ultimately, OpenSnow was started.

---

---

Now, OpenSnow has 10 full-time employees and provides forecasts, snow reports and travel advice around the world on its website and mobile app.

Thursday’s evening brought tidbits and trivia including the snowiest spots in Colorado — Wolf Creek, Buffalo Pass and the West Elk Mountains — and the best wind direction for snow at Vail and Beaver Creek. In case you’re wondering, it’s northwest for Vail, and west for Beaver Creek.

However, the discussion also delved into broader topics, including everything from predicting this year’s snow season, when to call in sick for a powder day (and the accuracy of weather forecasting), machine learning, climate change and more.

El Niño and La Niña

Right now, as the first snowfall hits the slopes and the resorts begin firing up their snow guns, everyone wants to know what the ski season will bring. And one topic always dominates the conversation: Is it an El Niño or La Niña year?

Why?

“Because it’s the only thing we have to talk about that gives us a shred of understanding of what the season might be like; that’s why,” Gratz said.

Last season, which was a La Niña year, saw at or above the median snowpack compared to the last 35 years in Colorado.

This year, however, marks the return of El Niño.

Simply put, El Niño refers to “warmer water in the Pacific Ocean,” more specifically one spot in the ocean that is along the equator, Gratz explained. In this region, there are a lot of thunderstorms, which throw moisture and air from near the ground into the atmosphere.

“When we get warmer or cooler than average water temperatures, it changes thunderstorm patterns. The thunderstorm patterns change how much stuff goes up in the atmosphere, which then changes how the winds blow and the storm tracks,” he added.

So what’s the correlation between El Niño or La Niña and snowpack?

“There is no correlation between El Niño and snowfall across Colorado except for this: the stronger the El Niño, or the stronger the La Niña, the better chance we have of at least an average season,” Gratz said. “There’s a lot of variability in the middle.”

And this year, he added, it’s looking like a moderate to strong El Niño.

Historically, El Niño has brought higher-than-average snowfall in October, November, March and April and lower-than-average snowfall in December, January and February, Gratz said. However, is this what this year will bring?

“I wish I knew,” he said.

The accuracy of forecasting

This predicament perhaps demonstrates the near-impossibility of attempting to forecast what a ski season will bring. Gratz compared seasonal forecasting to Google Maps: sometimes it tells you a drive will take 12 minutes and it’s “almost right all of the time.”

“But sometimes in life, we just don’t have the right answer. And seasonal forecasting is one of those times,” he added. “Seasonal forecasting is not that good.”

There’s one other tidbit that also comes into play: “20% of the storms create about 80% of the snow,” Gratz said.

“So even if you roughly forecasted the storm track for the season, if you get a few blockbuster storms, it totally messes up the numbers compared to what you’d expect,” he added.

However, forecasting — and the accuracy of these forecasts — has only gotten better over the last 50 years. Gratz showed a graph showing the accuracy of forecasts over the past five decades, which demonstrated a trend of “up and to the right,” just below a “perfect forecast.”

“We’re not there and we’re probably not going to get there. But look at this progress, right? Slow and steady, getting there,” Gratz said.

The best forecasts take into account numerous weather models: “The more models you average together, the better chance you have at an accurate forecast,” he noted.

However, there’s a time limit for accuracy and predictability. When looking at how many days out you can forecast, perhaps unsurprisingly, the forecast one day out is “pretty darn good,” Gratz said.

This accuracy “degrades over time” and once it reaches beyond nine or 10 days out, “there’s very little accuracy in the forecast.” This is when it hits about 60% accuracy at this time, which Gratz noted is what statisticians deem “not very useful anymore.”

“From a powder chasing perspective, here’s how I think about this,” Gratz said. “About a week out, I’m thinking about it, but I’m not making plans. Three to five days out, I’m making plans that are cancelable. One or two days out, we’re totally locked in with the option to chase (powder) somewhere else if you wake up and something goes wrong.”

The future of forecasting

As forecasting continues to improve, the advent of machine learning and artificial intelligence in meteorology will have an impact on accuracy.

“It’s not going to revolutionize something tomorrow, but we’re on the cusp,” Gratz said.

Gratz explained that today, weather is predicted using physics-based weather models: “Even on supercomputers, those weather models take one to 12 hours, depending on the resolution, to run.”

The main thing AI would revolutionize is this speed: “These new AI machine learning models take seconds to minutes to effectively do the same thing,” Gratz said, adding that the “machine learning weather models look at the last 80 years of weather data, figure out the relationship, and then it’s just pattern recognition.”

So, it will bring speed, but also accuracy; already these machine-learning models “can out-forecast our best modeling,” Gratz said.

Climate change

And as meteorologists look to the future, climate change is certainly a topic of concern. However, when asked whether climate change will impact the accuracy of forecasting, Gratz’s answer was a simple no.

“Climate change actually does not change the physics of the atmosphere or the accuracy of the forecast models,” Gratz said.

How climate change will impact ski seasons and powder days was a slightly more complex answer.

Using graphs showing the precipitation in Colorado over the last 120 years, Gratz showed that while the annual amount of precipitation in the state fluctuates year over year, there is no overall trend. Looking at temperatures in Colorado during the same time period, there is an overall trend: up. This, he demonstrated, is the same for precipitation and temperature of the Colorado River Basin.

This doesn’t mean that climate change isn’t a problem, just that precipitation isn’t changing, Gratz noted. For one, “with warmer temperatures and the same amount of precipitation, you get drier soils” and therefore, a “moisture deficit,” he added.  

In terms of ski season, the warmer temperatures could certainly mean less snow at lower elevations, less snow and more rain on the shoulder seasons and earlier snowmelt, he added.

“But, if this winter, we happen not to get a lot of snow, don’t blame climate change; there is no long-term trend in precipitation,” Gratz said.

Changing the climate

And as the climate changes around us, Thursday’s event also delved into one method of humans attempting to alter the climate: cloud seeding. This technology refers to a weather modification technique that can increase a cloud’s ability to produce precipitation.

The technique only works for certain storms, the ones with the right temperature, wind direction and amount of moisture, Gratz said.

“Is cloud seeding a real thing? Yes. Does cloud seeding work a little bit? Yes, it does. How much does it work? Studies range from 2% to 15%,” Gratz said.

With this range, Gratz commented that it’s unlikely to bring about more powder days in the future.

Using 10% as an average — even though he indicated it’s likely more 2% to 5% — of how much cloud seeding increases snowfall, he noted: “If Vail gets 10 inches of snow, maybe cloud seeding is responsible for, in the best case scenario, one inch. That’s not going to change the powder day.”

“One inch of difference is what aspect you’re skiing on, trees or not, or wind blowing things in,” Gratz said.

That doesn’t mean the technique isn’t useful; this 10% increase over the course of a season could add a “reasonable amount of water to add to the snowpack,” he noted, adding that it’s good from a water perspective.

“It’s just not really going to change your skiing situation and your powder situation for a particular day,” Gratz said.