How Cold Is a Mountain Marathon Start? Why Past Weather Reads Warm

Past race-day weather at mountain starts reads warm, by about 9°F at 6 AM where we could check. What we measured, what the new note means, what to pack.

racecast.io·
WeatherMountain RacesMethodology

Open the weather history for the REVEL Big Cottonwood Marathon and it shows 57°F (14°C) at 6 AM, averaged over nine past races. A weather sensor near the start of that course measured 51°F (11°C) over the same mornings. At East Canyon the gap was larger: 63°F (17°C) in our history, 53°F (12°C) on the sensor.

The history is not made up. It is the readings of weather stations near the race. The trouble is where those stations are. At a race that starts high on a mountain, the nearest stations are often in the valley below, and a valley morning is warmer than the start line above it.

As of this week, 62 courses on racecast.io carry a one-line note under their weather history that says where the readings come from. This post explains what the note means, what we measured, and what to do with it if you are standing on a bus at 4 AM.

In short
The problemPast race-day weather comes from stations near the race. At mountain starts those stations are often thousands of feet lower, so the history reads warm.
How far offOn the courses we could check against a mountain sensor, the 6 AM history ran between 4°F and 15°F warm, about 9°F on average.
What changed62 courses now show a note under their weather history saying where the readings come from. No temperature on the site was changed.
Why no single numberCold air pools in valleys, and in some places the high ground is not colder than the low. The note gives no size in degrees.
Waterfronts tooBuffalo, Newport News and Surfside Beach have stations miles inland, so late-morning readings run warm.

How cold is the start of a mountain marathon?

Colder than our past race-day history says, at the courses where we could check. We compared the history to SNOTEL sensors, the snow-monitoring stations run by the USDA. Each sensor sat within about 5 miles (8 km) of the course point and within about 500 ft (150 m) of its height, and we adjusted its reading for the small height difference that was left. Here is the 6 AM hour, averaged over the past races we have on file:

RacePast racesHistory shows at 6 AMSensor measured at 6 AM
East Canyon Marathon9 (2018 to 2026)63°F (17°C)53°F (12°C)
REVEL Big Cottonwood Marathon9 (2017 to 2026)57°F (14°C)51°F (11°C)
REVEL Mt. Charleston Marathon8 (2017 to 2026)53°F (12°C)38°F (3°C)
Huntsville Utah Marathon3 (2024 to 2026)54°F (12°C)46°F (8°C)
Timberline Marathon5 (2021 to 2025)59°F (15°C)45°F (7°C)
Mountain Magic Trail Race5 (2019 to 2025)54°F (12°C)47°F (8°C)
High Uinta Half Marathon4 (2023 to 2026)59°F (15°C)55°F (13°C)

The gap runs from 4°F at High Uinta to 15°F at Mt. Charleston. It is not the same everywhere, and it is not the same every year. Across the seven courses in the table it averages about 9°F at 6 AM. A wider check of 13 mountain points on 71 race days, from 5 to 11 AM, came out the same: about 9°F too warm.

Few single mornings were close. Of the 43 race mornings behind the table, the history's 6 AM reading was within 5°F of the sensor on 10.

The wider check is below, one dot per race morning. The history read warmer than the sensor on 70 of the 71, and by 5°F or more on 57.

If you came here asking how cold the start of Big Cottonwood is: in past years the table says 57°F at 6 AM and a sensor near the start read 51°F. Plan for the colder number.

Why does past race-day weather read too warm at mountain starts?

Our history for a past race morning comes from a blend of the weather stations near the course. We do not publish a weather model's guess for those days. Where a station sits next to the course, that works very well. At ordinary races where we could check against an airport station within about 3 miles (5 km), the history differed from the airport's own readings by 0.24°F on average.

A mountain race breaks the assumption. Temperature falls as you climb, and the stations in the blend are often at airports and towns in the valley. On the mountain courses the note states the gap. On Big Cottonwood, the stations average more than 3,500 ft (1,100 m) below the start.

The table shows the valley's morning. The start line has its own, colder one.

Why not subtract a few degrees for every 1,000 feet?

The usual rule of thumb is that air cools about 3.5°F for every 1,000 ft you climb (about 6.5°C per kilometer). It is a rough average, and it is where most runners start. It is also not good enough to fix a history table, for three reasons we found.

  • The rate changes through the morning. The rate that would have matched our mountain sensors was about 2°F per 1,000 ft from 5 to 8 AM and close to 4°F per 1,000 ft by 11 AM. Cold air pools in valleys overnight, so at dawn the valley is not as much warmer than the mountain as the textbook says.
  • It changes from day to day. Between pairs of airports at different heights, the rate on a given morning ranged from the high ground being the warmer one to air cooling much faster than the average. A fixed rate is right on average and wrong on many individual mornings.
  • We do not know which height to subtract from. The history is a blend of several stations, and the weights are not published. Subtracting a flat amount for the height gap cut the typical error at our mountain points from about 9°F to between 5°F and 6.5°F, but it overshot at some courses and barely helped at others.

The first of those is the easiest to see. The line below is the rate that would have matched the sensors at each hour. It sits well under the rule of thumb until mid-morning.

Some places go the other way. At Death Valley the course sits below its stations, and the history ran 3 to 5°F too cold. In California's coastal hills, a nearby station showed the history too cold on one morning and too warm on another at Miwok. Marine air and fog set the temperature there more than height does.

That is why there is a separate kind of note for such courses, with no direction in it. And it is why none of the notes states a size in degrees.

What does the note under a race's weather history mean?

It sits under the averages in the weather history card on the race page. It tells you where the readings came from. It does not change them: the table shows the station readings as they are. There are four versions.

Stations well below the course (32 courses). Some part of the course averages 200 m (about 650 ft) or more above its stations, and the history ran 3°F or more warmer than a model-based weather archive at that point. For the start of Big Cottonwood the note reads:

Past race-day readings here come from weather stations that average more than 3,500 ft below the start. The course usually runs colder than shown.

Where the high part of the course is the point that passed, such as Pikes Peak and Leadville, the note says "below the high part of the course" and "the high part usually runs colder than shown".

A sensor beside the start (2 courses). Timberline and Cedar City Half do not pass the height test, but a sensor next to the start read colder than the history on 5 or more past mornings. The note reads: "Past race-day readings here come from weather stations away from the start. A weather sensor near the start has read colder than shown on past race mornings."

A different elevation, no direction (25 courses). These are valley floors, canyons, courses where only one test passed, the California coastal ranges, and two mountain starts (St. George and Big Sky) where a station at the start read about what the history shows. The note reads: "Past race-day readings here come from weather stations at a different elevation than the course. Conditions on the course can differ by several degrees." Death Valley, Escalante Canyons and Miwok carry this one.

A waterfront with far stations (3 courses). Covered in the next section.

Only some of these courses have a mountain sensor close enough to measure against. At the others the note rests on the height of the stations against the course, and nothing has been measured at the high points of Pikes Peak, Leadville or Hardrock. That is why the notes give no number.

Do waterfront races have the same problem?

A smaller version of it. At Buffalo, Newport News One City and Surfside Beach, the nearest station in the blend is 7 miles or more from the water at the first two and 14 miles or more at Surfside. Inland air warms faster than air over a lake or the ocean once the sun is up.

We checked against NOAA sensors on piers and sea walls. In the early morning (6 to 9 AM) the history was fine. Between 10 AM and 1 PM, at courses whose nearest station was 10 km (6 miles) or more away, it ran 3.6°F warm on average across 72 race days. The note for these three says that the waterfront usually runs a few degrees cooler than shown by late morning.

This matters for a slower marathoner, who is still out at noon. It matters little for a runner done by 9.

How many races does this affect?

62 of the roughly 1,500 races on the site have a note. At the rest, the stations are close enough to the course that we have no reason to flag them. Where we could check against an airport station within about 3 miles, the history was within a quarter of a degree of the airport's readings on average, and the morning's average was off by 3°F or more on only 1 of 287 race days.

The 62 are not all proven the same way. Our direct measurements are mostly in Utah and Nevada, with a smaller number elsewhere in the US. Nothing was measured abroad, and the notes there rest on station height alone.

This post is about the history only. The forecast is a separate source, and this study did not test how accurate it is at altitude.

What should you pack for a mountain start?

  • Treat the history's start-hour number as the valley's. If the race page shows 55°F at 6 AM and the note says the stations are far below the start, plan for a cooler start. On the seven courses where a sensor has shown it, the gap was 4 to 15°F.
  • Bring throwaway layers. A sweatshirt or trash bag you can drop at the start, and gloves. If you wait on a bus or in a corral for an hour, you stand still in the coldest part of the morning.
  • Dress the finish separately. At a course that finishes down in the valley, the temperatures near the end are closer to what the table shows.
  • Look at your own race's note. If it says the high part of the course runs colder, the cold stretch is up there, and that may not be the start line.

For the clothing side, see what to wear by temperature, how to run a cold marathon and the race day gear guide.

More guides

REVEL Big Cottonwood Marathon
See the weather history, the note and the race-day forecast
REVEL Mt. Charleston Marathon
See the weather history, the note and the race-day forecast