The KASE Weather dashboard just got its biggest upgrade since launch. The old version answered one question: is the weather at Aspen bad right now? The new version answers the question you actually have: how likely is it that today falls apart? It reads the whole day the way a local who has watched a thousand Aspen storm cycles would. Here is what changed, and why it matters for anyone flying in or out of Aspen-Pitkin County Airport (ASE).
The old model was honest, but nearsighted
Version one of the dashboard did one thing well. It took the current weather report at the airport, compared visibility, cloud ceiling, and wind against the operating limits of the jets that serve Aspen, and showed you a green, yellow, or red light. If the weather at the airport was bad, it told you so.
The problem is that Aspen flights do not fail for one reason. They fail for many. A flight can cancel under a bluebird sky because the plane never left Denver. It can cancel because a storm two states away broke the airline's schedule the night before. The old model could not see any of that. It watched one airport, one weather report, one moment in time. If you want the deeper story on all the ways a flight here goes sideways, we wrote about it in why Aspen flights get delayed.
The new model watches the whole system
Version two predicts the chance that today becomes a major delay day: a day where a third or more of Aspen's flights end up cancelled, diverted, or badly delayed. To do that, it pulls in signals the old model never touched:
- Better weather forecasting. Instead of one forecast, it uses guidance from a blend of dozens of National Weather Service models, including the odds of low clouds, snow amounts, and wind gusts hour by hour.
- Denver, plus six more airports. Nearly every Aspen flight touches a big hub: Denver, Los Angeles, Dallas, Chicago, Houston, San Francisco, or Atlanta. Together those seven cover about 96% of Aspen's commercial flights. The model checks the forecast at every one of them, because your Aspen flight cannot arrive if it never takes off from the other end.
- FAA operations. Ground stops, ground delay programs, and deicing events, at Aspen and at the hubs. These are air traffic decisions that pure weather data never shows.
- Yesterday's results. How did the airport actually do yesterday? Disruption at Aspen runs in streaks. Storm cycles last days, and broken airline schedules take time to heal. A rough yesterday makes a rough today much more likely, and it turned out to be the single strongest signal in the whole model.
The days that fooled the old dashboard
Testing surfaced a pattern we started calling the bluebird meltdown. The sky over Aspen is perfect. The old dashboard glows green all day. And yet a third of the flights collapse, because Denver is buried in snow, or thunderstorms are rolling through Dallas and Chicago, and the planes that were supposed to ping-pong in and out of Aspen are stuck somewhere else.
These were the old model's worst misses, and they are exactly the days the new model was built to catch. When the network is broken, the new dashboard says so, even when the local weather looks like a postcard. And if you end up on the wrong side of one of those days anyway, keep our guide on what to do when your Aspen flight diverts handy.
One surprise: gusts matter more than anyone thought
When we let the data speak, one finding stood out. Among all the pure weather signals, forecast wind gusts were the best single predictor of a bad day at ASE. Not ceiling. Not visibility. Gusts. The old model ignored gusts entirely; it only looked at steady wind speed. The new model treats the day's strongest forecast gust as a first-class input, and it earns that place in the results.
How we know it works
We did not just swap models and hope. We rebuilt two and a half years of history, day by day, using the US Department of Transportation's per-flight records for every Aspen arrival and departure, the same public data behind our Aspen delay and diversion statistics. Then we ran a fair fight: both models made their call each morning using only the information that existed at that moment.
The new model was trained on two winters and then tested blind on a third winter it had never seen. On that unseen winter, it caught 89% of the major delay days. The old model caught 46%. The new model nearly doubled the catch rate and cried wolf less often while doing it.
Honesty matters here, so two caveats. First, no forecast catches everything. A handful of meltdowns are airline-internal, invisible to any morning forecast; the dashboard's live FAA status strip is there to catch those as they develop during the day. Second, the model keeps learning. Every season adds new storm days to test against, and we will keep publishing how it performs.
What you'll see on the dashboard
The upgrade shows up on the live dashboard in three ways:
- A real probability. Instead of only a color, you get a number: the chance that today becomes a major delay day, updated all day long.
- The reasons why. When risk is elevated, the dashboard tells you what is driving it: strong gusts forecast, snow at the hubs, a rough day yesterday, a Denver ground stop.
- A live status strip. The moment the FAA issues a ground stop or delay program that touches Aspen or its hubs, it appears at the top of the page.
None of this changes how you should think about timing your trip. Mornings still beat afternoons, and some months are simply safer bets than others; our month-by-month guide to flying into Aspen still applies. What changes is how early and how clearly you see a bad day coming.
What's next
The model improves every season as new data arrives. The bigger step comes with KASE Weather Premium, launching summer 2026: instant email and text alerts the moment your travel day's risk level changes, so you hear about trouble before the gate agent does. You can join the Premium list here.
Until then, the smarter dashboard is free, live now, and watching a lot more than the sky over Aspen.