In June of 2016, I got a call from Dave Mills of Abt Associates in Boulder, Colo. He and his team, all skiers, were in a late draft of a white paper commissioned by the Environmental Protection Agency entitled: Projected Climate Change Impacts on Skiing and Snowmobiling: A Case Study of the United States. That study was published in 2017, but not widely circulated. (The Climate Change Division at the EPA is keeping a low profile.) The study is both sobering and reassuring.
If future weather could affect your business, you should download the report at www.sciencedirect.com by searching “ski industry.” It’s tough reading in more ways than one, but if you don’t do it, you’re an ostrich.
By way of an introduction to the report, here’s my take on it.
The Future According to Abt
Abt has modeled the financial impact of climate change on downhill skiing, XC skiing, and snowmobiling at two future dates—2050 and 2090—and with two different rates of warming temperatures: one based on the world’s current rate of carbon output, the other based on reduced carbon output. It took into account several variables: rising temperatures, possible changes in natural snowfall and melt rates, population growth and shifts, and the impact of snowmaking. Of these, the most accurate, and for many resorts the most important, projections concern future temperatures. More on that soon.
The good news is that, given current snowmaking technology, a majority of existing ski areas will have sufficiently cold temperatures to maintain current opening dates and remain viable at least through 2050. The caveat: that means adopting increasing levels of snowmaking automation, to take advantage of the shorter snowmaking windows we can expect to see over the next 30-plus years.
The bad news is that, at our current rate of human-related carbon output, by 2090 winters could well become too short and warm to support the winter resort business anywhere but the central Rockies and High Sierras. While climate models cannot reliably project snowfall that far into the future, and the melt rate of machinemade snow is completely outside the realm of climate modeling, the projected rise in temperatures, especially east of the Rockies, will eventually make skiing and riding untenable, even with the best of current or envisioned snowmaking technology.
EPA has commissioned and published several similar studies, regarding human health, agriculture, forestry, water resources, and more. Winter sports made the list because someone at the EPA thought skiers might care enough to take personal action if they came to realize winter sports might not be there for the grandkids and beyond.
Abt’s premise is that climate change will lead to shorter winters, and that it can predict the extent of this change. The researchers, Dave Mills, Heather Hosterman, and Cameron Wobus, feel they can project length of season, based on knowing past snowfall and melt rates, and projecting these for the future. To this, they added ski-industry intel regarding snowmaking. There is some wiggle room regarding future projections, of course, but as we’ll see, the researchers accounted for different scenarios.
To orchestrate big data on climate and existing models to prepare their report, Abt tweaked the data to more accurately reflect the impact of climate change on 247 individual ski areas for which they knew the names, coordinates, base and summit elevations, and exposure to the sun of the individual trails.
profiling Winter Resorts
Unlike the climate change articles we can read in popular media every day, Abt was not looking to make projections for a continent, a state, or even a region. It was looking directly at 247 of our resorts from NSAA’s six reporting regions. Odds are, your ski area is a dot on their maps, two of which we reproduce here.
Abt used a well-vetted model, “Utah Energy Balance” (UEB) model, for how snowpack accumulates, melts, and its Snow Water Equivalent (SWE). In the West, water is the lifeblood, and a lot of it comes from melted snowpack. Water content of snowpack and rate of melt off has been studied for a long time.
Abt used the UEB because it was as reliable as any and had “high computational efficiency.” That was important, because it modeled 247 individual ski resorts, hour by hour, 20 years in the past with regression, and 72 years into the future. The team ran 300,000 model simulations.
Snowpack models like UEB need to be fed with weather data sets that include snowfall, temperature, humidity, hours and intensity of sunlight, and many other variables. For this, Abt used the North American Land Data Assimilation System (NLDAS-2). This uses NOAA weather data collection sites and other systems, including satellite assessments of snowpack, to predict the weather for almost any location. It does so by triangulating data and compensating for elevation and other site-specific variables. This is the same method used by weather services to provide, say, a 10-day forecast for a town of 98 people, where no one owns an accurate thermometer.
Abt further refined the accuracy of the NLDAS-2 data set for ski resorts. Using a combination of SNOTEL sites, which measure precipitation, temperature and several other weather variables, as well as other sources, it checked the model against observations from specific resorts. The model held up.
How Good Is the Model?
SnowConsult was invited to help Abt add snowmaking into its assessment of season length, which we’ll get to shortly. I have confidence in the model, particularly with respect to temperature, because Abt’s model mimics my actual hourly 10-year data from 30 clients across the country. Therefore, I believe this study gives us a very valuable tool for predicting what temperatures, on average, snowmaking will have to cope with in the future.
In my own work, I deal only with hours and temperatures where snowmaking is possible. I ignore natural precipitation, because it is not predictable. Forecasts for future precipitation in the EPA model were not weighted very much in its calculations, because there is no consensus among the models on snowfall. Globally, a warmer world will be wetter, but with huge regional differences in amount and form, unpredictable in the long term. Nonetheless, the “length of season” projections based on natural snowfall alone are interesting, if only because they represent one possible future consideration.
The questions are: how much warming will we see in the future? And can anyone use that data to predict future length of seasons? To help answer those questions, the EPA model has to be force-fed with projected temperature and precipitation data. Of the thousands of peer-reviewed scientific climate studies, 99 percent predict a warmer climate in the future. But they vary as to how fast the warming will be, what kind of precipitation could accompany it, etc.
To arrive at a sort of midrange projection, Abt modeled five representative climate studies under two different scenarios: one in which the world constrains carbon output to the atmosphere at “RCP4.5,” roughly in line with the Paris Climate Accord limits, and a second, “RCP8.5,” in which we continue pretty much on our current course.
At RCP (Representative Concentration Pathways) 4.5, Abt projects winters get shorter in 2050 by a few days at many, but not all, of the 247 resorts studied. Shorter yet by 2090, except for five locations. A-Basin, you can stop reading now.
According to the EPA report, at RCP8.5, winters could become considerably shorter for many resorts in the East and Midwest as soon as 2050. And by 2090 for most in the Midwest, East, and coastal Far West, skiing on snow, even machinemade as we know it, would not be economically viable. Skiing could continue in the Central Rockies and High Sierras for a 100-day season, and still longer at lottery-winner A-Basin.
Personally, I don’t share that dire assessment, even though I absolutely believe the temperature predictions. This is the best climate study ever done for our industry, but it doesn’t take into account next-gen snowmaking. Also missing is a recognition of the durability and resistance to melt of machinemade snowpack. No model accounts for that.
Will It Be Cold Enough?
Abt knew the baseline season length for those 247 resorts, thanks to NSAA data. But that baseline length of season very much includes the contribution of snowmaking. On the EPA map, a Southeast dot had a 90-day baseline season in the study. Without snowmaking, it would be perhaps 10 days.
Given the variable of snowmaking, the real value in the EPA study lies in what it says about future temperatures. The study also illustrates how future temperatures could affect the length of season compared to the baseline (i.e., today) at many of those 247 resorts, without a change in snowmaking.
One obstacle to knowing the impact of snowmaking in any scientific study is that there is no real data on how snowmaking impacts snowpack or season length. The UBE model of snowpack melt considers water content to range from 10 percent to 15 percent snow water equivalent. That’s fine for natural snow in the West, but our machinemade stuff is 40 percent-plus. No existing melt model can account for that, or machinemade mixed with natural snow, the most common snowpack at resorts.
The 450-Hour Threshhold
So instead, SnowConsult said, “Show me a ski area that has 450 hours under 28° before Christmas, and I will guarantee you a viable ski area.” I chose those parameters because even my least efficient snowmaking customers can open 100 percent of their snowmaking terrain if they have 450 hours before Christmas. The real future goal for my clients will be to open 100 percent of that terrain in 150 hours or less. Some have achieved this already.
In one day, Abt was able to estimate the dates by which each of the 247 ski areas would reach the 450 hour total, and to predict how that date would change over the next 32 and 72 years, for each ski area. The results are totally believable—Abt’s model for current dates is confirmed by data from my clients. For example, according to the EPA, the Holiday Valley dot can expect to reach 450 snowmaking hours between Dec. 15 and Dec. 31 at present. An 11 year history for Holiday Valley shows 471 hours by Dec. 31. Hence my confidence in the study’s temperature data.
Curious about how your future snowmaking hours might look? Abt can provide your resort’s projections for a reasonable fee (think $5,000 to $10,000).
Approximately 70 percent of modeled downhill skiing sites can reach 450 hours of snowmaking and be 100 percent open by Dec. 15 currently, with their existing systems. However, this percentage declines under each of the future scenarios by nearly half under both RCPs if systems don’t update—and by that I mean automate. Fortunately, the technology to maintain our current opening dates until 2050 exists today.
Before we all sigh in relief, there’s a caveat: EPA counted every single hour below 28ºF wet bulb, so those 450-hour mileposts are a little optimistic.
Can Snowmaking Compensate?
Under the “RCP4.5” scenario, the answer appears to be yes, until at least 2050, and for almost all resorts well beyond. But that’s only if each resort can adopt today’s cutting-edge snowmaking technology on a significant enough percentage of its terrain to be attractive. The Alps have pretty much done this already, as have some North American resorts.
Automated snowmaking systems will be essential, because the windows of snowmaking opportunity will get smaller, and the available hours will become less. This has been happening already, of course. My clients in the last 10 years have observed 30 percent fewer total snowmaking hours under 28ºF wet bulb. For clients with full automation, the diminished hours have not shortened their seasons at all. To the contrary, they are making 30 percent more snow, with 30 percent less available hours, and using 25 percent less kWh to do it. The most successful resorts use less than 3,000 kWh to make an acre foot of snow.
Top systems share three concepts:
• a high percentage of not just automated guns, but automated trails;
• a high percentage of fixed-location, tower-mounted stick or fan guns;
• process-controlled machine room resources.
But that’s not the norm. Many snowmaking systems are far from energy-efficient or ready for climate change. Some use more than 15,000 kWh to make an acre foot of snow.
The least efficient systems typically have a high percentage of first-generation manual air/water guns on hoses. It won’t do to simply run these systems harder. They are too inefficient to produce in the shorter windows of opportunity that are going to be the new norm.
Double Bottom Line
Automated, energy efficient snowmaking are not only our operational solution, they also reduce emissions. Some snowmaking systems account for 67 percent of the energy consumed by the resort during winter operations. I support all initiatives to save energy, from swapping out light bulbs to washable flatware vs. paper. But snowmaking is where the greatest strides can be made.
If the world won’t control carbon emissions, and we have to operate in a high-CO2, RCP8.5 scenario or worse, things get more difficult for all areas outside of the Central Rockies and the High Sierras, especially after 2050. We’d need some kind of technological breakthrough I don’t see coming. We already have “all temperature” snowmakers like the “Snowfactory,” but that is an energy-intensive solution suitable for special projects or intense urban markets, and not scalable for large resorts.
T0elluride CEO Bill Jensen infamously identified one-quarter of our areas as “sunset” resorts for lack of capital and market potential. Unabated climate change could be yet another nail in the coffin.
Lastly, the report tries to predict the financial impact that the shorter projected seasons will have on the industry as a whole. That includes possible changes in total visits nationwide, taking into account both climate change and population growth. These predictions are far less certain than for temperatures—perhaps enlisting more help from NSAA or other data-collection agencies would have helped. And by Abt’s own admission, predicting precipitation is more difficult than predicting temperatures.
That said, Abt’s conclusion is sobering. “Based on our modeling, the difference between RCP4.5 and RCP8.5 could represent the difference between preserving skiing and snowmobiling in the eastern half of the country and losing these activities almost completely by 2090,” the report states.
I’m not ready to accept as our fate a collapse in skier visits because of shorter seasons. Snowmaking can mitigate shorter seasons. In snowmaking we must trust.
And there’s reason to hope. In the report’s most optimistic projection, national visits actually rise, due to population growth and increased visits in the Rockies and Pacific West. Other regions, though, would experience declines.
Remember, the projections for season length and closing dates can be expanded by enhanced snowmaking. Still, the study highlights some worst-case scenarios that resorts should recognize, and act to overcome them.