October in Reno: the exact week to shut a swamp cooler down, from the NOAA freeze tables

September 19, 2026

If you want one answer: the week of September 28 through October 4 is when a swamp cooler in Reno, Sparks or Spanish Springs should be drained and shut down. That is not a rule of thumb. NOAA’s 1991-2020 climate normals publish first-freeze probability dates for the Reno airport and for the cooperative stations around the valley, and that week sits just ahead of the first date at which a freezing night stops being unlikely. Everything else here is how to move that week for where you actually live, and what has to happen inside it.

The published dates

Here is what the 1991-2020 normals give for Reno-Tahoe International Airport, at about 4,505 feet. Each row is the date by which there is a 10, 50 or 90 percent chance of having already hit that temperature in the fall:

Three numbers, three different jobs.

36 degrees is the warning shot, and its 50 percent date is October 7. Nothing in a cooler freezes at 36. But the airport reading is taken in a shelter roughly five feet above the ground over a standard surface, and that is not where your cooler lives. A flat or low-slope roof radiating heat to a clear Nevada sky on a calm night gets colder than the official number, and so does an exposed quarter-inch line running across it. When the airport reads 36, parts of the system have already been at freezing.

32 degrees is the real deadline, and the date that matters is October 5, not October 19. October 5 is the 10 percent date: one fall in ten has already had a freezing night by then. A cooler drained after that is rolling dice. October 19 is the coin flip, and November 2 is the near-certainty, which is the date for anyone who wants to be told the cooler has definitely taken a freeze.

28 degrees is what actually breaks things, at a 50 percent date of November 1. A brief touch of 32 may do nothing. A sustained 28 degree night is what expands trapped water hard enough to split copper and crack a polymer pump housing.

The other number worth knowing is the count: the airport averages 111.6 nights a year at or below 32 degrees. This is not a marginal freeze climate where a cooler might get away with it.

Why the week is the end of September

A homeowner reading that table could reasonably conclude they have until the middle of October. Two things argue against it.

The first is the gap between a thermometer and a roof, described above. Practically, subtract a few degrees from anything you read in a forecast.

The second is scheduling. Everyone in the Truckee Meadows reads the same ten-day forecast on the same evening, and when the first low-30s night appears in it, the phone does not stop. A shutdown booked in the last week of September is a scheduled stop at $100 to $200. A shutdown booked the afternoon a hard freeze is forecast is a scramble, and the ones that do not get done in time become April repairs.

Adjusting for where you live

The airport is one station. The valley spans about 350 feet of elevation between its lowest and highest towns in this service area, and the published cooperative stations do not agree with the airport.

Read the third column honestly. Only four of those rows have a published 50 percent date in the normals: the Reno airport at October 19, Sparks at October 5, Carson City at October 1 and the Stead station at October 7. Minden’s freeze normals are missing from the dataset entirely, and there is no published station for Sun Valley, Spanish Springs, Fernley or Dayton in it. The windows in the last column for those stops are our scheduling rule, based on elevation and on which published station each sits nearest. They are not NOAA dates and we are not going to dress them up as any.

Two things in that table are worth pausing on. Sparks, the lowest-elevation stop with published normals, freezes two weeks earlier than the airport at the 50 percent mark, and Carson City earlier still. Elevation is a guide, not a law: cold air drains and pools overnight, and a low spot in a valley can beat a ridge to the first freeze. If your street sits in a hollow, go a week earlier than this table says.

What has to happen in that week, and what can wait

Not every step is equally urgent, and it is worth knowing which is which if you are running out of daylight.

Before the first freezing night, in this order:

1. Shut the water off at the source. The hose bib or saddle tap that feeds the cooler, not just the float.
2. Drain the supply line. This is the one that costs money. The quarter-inch copper or poly line running up the wall and across the roof is what splits, and closing the float without emptying the line leaves it full. Open it at both ends and confirm it runs dry.
3. Drain the pan. Pull the plug or open the drain fitting, then sponge the rest out. Standing water freezes, expands and works the pan seams and the drain fitting open. While it is wet, scrape the loose scale out; it comes off far easier now than in April.
4. Pull the aspen pads. They are finished after a season, and left in place they hold damp against the pan all winter.

These can slip a week if they have to: fitting the cover, sealing the duct damper, and covering the interior ceiling grille. They matter for the heating bill and for keeping the Washoe Zephyr’s grit out over the winter, but they will not crack anything.

What it costs to get it wrong

A split supply line is $100 to $400 depending on the run, and it rarely travels alone: a frozen float is $100 to $250 and a cracked pump housing is $150 to $400 installed. That is how a missed October turns into a $300 to $600 April visit, against $100 to $200 for the fall winterization done on schedule. Booked with the spring start-up the pair is $200 to $400. Every one of those numbers is on the pricing page, and the pad and service calculator will show you the gap between a drained cooler and an undrained one on your own unit.

The worse outcome is the pan. A galvanized pan that was sound in September can be weeping by spring after standing water froze in it, and a pan is not a part you replace, it is the cooler. On the decades-old units still working on the post-war ranches north of the Sparks tracks and through the mid-century southwest, one missed October is how a repairable cooler becomes a $1,500 to $4,000 replacement.

If you want to keep cooling into October

Some years you do, and September can justify it: the normal September high is 83.8 degrees with 7.5 days at 90 or hotter. The move is to watch the Reno forecast every evening from about September 25, and the afternoon a night in the low 30s appears in the seven-day, do steps one through four. The cover and the damper can follow on the weekend.

October is a different argument. The normal October high drops to 70.4 with a normal low of 39.7, and the month averages 0.2 days at 90 or above. There is very little cooling left to buy and a real line to lose.

And if the cooler quits in that last warm week rather than shutting down gracefully, the warm air page has the five-minute check, and the repair cost calculator gives you a range before you decide whether it is worth fixing this late in the season.