Frost dates in Phoenix, and what to protect

The claim that “it never freezes in Phoenix” is not a myth in the way most gardening myths are. It is half true, and the half that is false is the half that costs you a tree.

Here is the true half. The 1991–2020 climate normals for Phoenix Sky Harbor give the city 0.4 nights below freezing in January, 0.1 in February, 0.1 in December, and nothing in the other nine months. A normal year contains 0.6 freezing nights. The annual figure in the same dataset says the same thing from another direction. And there is a stronger version of it: across the whole normal year there is no day whose normal low is anywhere near freezing. The coldest normal night of the year is 20 December, at 44.4 °F.

Here is the half that is false. The record at the same station, running from June 1933, contains 511 nights at or below 32 °F. Not in an exceptional decade. In the record.

Myth: the normals tell you the frost risk

The normals are a thirty-year average, and they are the right instrument for “what is a typical July”. They are the wrong instrument for “will my bougainvillea survive this winter”, because they average the bad winters with the mild ones and report the middle. A 0.4-night January is not a January that has four-tenths of a freeze in it. It is a January that usually has none and occasionally has several.

The record answers the other question. The chart above is the coldest night of each month in ninety-three years of observation, and it says three things plainly.

First, freezing happens in five months and only five: November, December, January, February and March. April through October have never produced a freezing night at this station, and the coldest April and October nights both stopped at 34 °F. If you are planting something marginal, that is the shape of the safe window.

Second, the cold is not evenly spread. January alone accounts for 212 of the 511 freezing nights, and December 140. Those two months carry 352 of them, nearly seventy per cent, in a stretch of sixty-two days.

Third, and this is the part the normals hide completely, it has been genuinely cold. The coldest January night in the record is 17.1 °F. February and December have both been to 21.9 °F, March and November to 25 °F. A 17 °F night is not a light frost; it is the kind of night that kills a young citrus tree to the ground and takes a bougainvillea with it. The Frost & Freeze Date Explorer is built on this record rather than on the normals for exactly that reason, and citrus in the low desert is the article that puts these nights against a specific tree.

Myth: one frost date covers the Valley

The extension’s frost handout gives a range rather than a date, and the range is wide: first frost anywhere from late November in Buckeye to mid-December in central Phoenix, and last frost from early February to early April.

Two things move a date across that span, and both are local. Elevation, because cold air settles, so a property at the base of a slope is a colder site than one a few hundred feet up and the same distance from the freeway. And urban density, because a built-up neighbourhood holds heat overnight in the way a new subdivision at the edge of the pavement does not. This is also why the same handout recommends planting the most cold-sensitive plants against a south- or west-facing wall, or near rock or paving that stores the day’s heat and gives it back after dark.

The practical version: your frost date is your address’s frost date, and there is no published table that can give it to you. The extension’s range is the best available starting point; a note of the mornings your own thermometer goes below 32 °F is what narrows it down.

Myth: cover it with plastic

The extension’s guidance is cloth, and the reasoning is worth knowing because it makes the method obvious.

A cover is not a blanket that warms anything. It is a lid on the heat the soil has stored all day and is now radiating out. That heat needs somewhere to collect, which is why the cloth goes all the way to the ground rather than being tied around the trunk like a scarf, and why a frame that holds the fabric off the leaves is better than fabric resting on them.

Plastic does not work for the same reason: it traps moisture against the leaf rather than trapping warm air around it. Cloth of the right weight protects to roughly 30 °F, and some types to 20 °F. It should come off in the morning once the temperature underneath the cloth has reached about 50 °F, because a plant held warm and dark through the day is being pushed into growth in the wrong month.

The same handout lists what testing has not supported, and it is a longer list than most garden advice admits: smudge pots, large fires, mulch used in the hope of warming soil, and chemical sprays. The first two are also obvious fire hazards in a dry winter, and worth ruling out on those grounds alone.

Myth: water well before a freeze and leave it at that

Watering has a place here, but a specific and limited one. The extension’s guidance is a trickle of water at the base of the plant running overnight, alongside a 100-watt outdoor bulb under a cover as an alternative, and then a limit that gets left out of most retellings: no more than three nights in a row.

The three-night limit is the honest part of that advice. Running water for a week is not more protection; the benefit is a small amount of heat released as the water freezes, and it is not a solution you can leave on indefinitely without soaking the soil and damaging the root zone.

Before any of that, though, there is the cheapest protection available: a plant that went into the cold snap well watered. Damage happens when ice crystals form inside the leaf and pull water out of the tissue. A plant already drought-stressed has less to lose. In a desert yard, where every plant is on the edge of dry by design, that is not a trivial point.

Myth: prune the damage off as soon as you see it

This is the piece of advice that costs people a second year rather than a first night.

Both the extension’s frost handout and its December gardening guide say the same thing: wait. Do not prune until the threat of freezing has passed and, preferably, until you can see new growth and can tell dead wood from live wood. There are three reasons and all three are practical. Damaged tissue still insulates what is underneath it. Dieback continues past the night of the freeze, so a branch that looks finished in the first week of January may not be. And a plant that has lost its top often regrows from wood that looked dead, from a base that looked dead, or from seed it dropped before it froze.

Cut it back in spring and you have lost nothing. Cut it back in January and you have removed the plant’s own protection at the point in the year when it most needs it.

What all of this adds up to

The Valley’s frost story is a story about frequency, not about possibility. Freezing here is uncommon (0.6 nights in a normal year) and when it comes it comes hard, in a short window, with 511 nights of record behind it and a coldest night of 17.1 °F.

That combination is exactly why preparation matters more here than the numbers suggest. A grower in a cold climate runs the same routine every winter and it becomes habit. A grower in Phoenix runs it once every three or four years, in the dark, having forgotten where the cloth is. The extension’s advice is written for the second person: have the cover, know the range of dates, and put the pruning shears down until spring.