When to plant a fall garden

How this calculation works

Every date here is your first frost date minus four things added together. Days to maturity from the seed packet. The length of the harvest window, because a crop you pick over three weeks has to be ready three weeks before you want the last of it. The fall factor. And a frost allowance that can be positive or negative depending on the crop.

The fall factor is 14days, and it is the part most people leave out. Catalogue maturity figures are measured in spring and summer conditions — lengthening days, warming soil. A fall crop grows through the opposite: every day shorter and cooler than the last. The same lettuce that hits 45 days in May takes closer to 60 in September. Extension guidance and the Johnny's backdating formula both add about two weeks to cover it.

The frost allowance runs both directions, which is why two crops with identical maturity dates can have planting deadlines a month apart. Bush beans are tender: one frost ends the plant, so the harvest has to be finished two weeks before the frost date to leave any margin. Carrots and kale are hardy: they keep producing two weeks past it, and taste better for having been frosted. That is a 28-day swing between two rows planted side by side.

Frost is not what ends the fall garden

This is the part almost every fall planting chart gets wrong by omission. Below roughly ten hours of daylight, most vegetables stop growing — not because of cold, but because there is not enough light to run photosynthesis at a rate that produces net growth. Eliot Coleman named the stretch when daylight sits under that threshold the Persephone period, and in the northern half of the country it is real and it is long.

At 44°N — Minneapolis, Portland, Maine — daylight drops under ten hours in early November and does not come back until early February. That is roughly three months in which a plant holds its size and does very little else. At 33°N it is a few weeks. Below about 29.5°N it never happens at all.

The practical consequence is a trap: you can have a mild November, no frost in the forecast, and a bed of half-grown cabbage that is simply never going to finish. The calculator flags this when it applies, because the fix is a planting date, not a row cover. Row cover holds warmth. Nothing you can buy adds daylight.

The working rule from winter growers is to get a crop to about three-quarters of its mature size before the period opens. What is standing when the light fails will hold through winter and can be harvested as you need it — it just will not get any bigger until the light returns. Which means a fall garden in the north is planned around a date in early November that has nothing to do with temperature.

August soil is too hot for the seeds you need to sow in it

The second failure mode is at the other end. Lettuce seed goes into thermal dormancy above about 80°F of soil temperature — it will not germinate, and it does not rot either, it just sits there. Spinach behaves similarly. In much of the country the calculated sowing date for these crops lands in the first half of August, when bare soil in full sun runs well past 80°F through the afternoon.

So the date the calculator gives you is correct and the sowing still fails. The workarounds are all about temperature at seed depth: sow in the evening and water with cold water, shade the bed with a board or a scrap of shade cloth until the seedlings appear, or start the flat somewhere cool and transplant. Pre-sprouting the seed on damp paper towel in the refrigerator for a few days is the most reliable of the lot for lettuce and spinach.

This is also why the fall garden favours transplants more than the spring one does. Four weeks in a flat under shade sidesteps the worst of it, and a transplanted crop occupies the bed for four fewer weeks — useful when the bed still has tomatoes in it.

Where the dates come from

Days to maturity are catalogue medians. Varieties within a crop vary by two or three weeks, and the figures are honest averages rather than the best case a breeder can quote — check your own packet and adjust if it differs by more than about ten days.

Day length is computed from your latitude by solar declination, not looked up in a table, so any latitude gives a real answer. It agrees with published sunrise tables to within about ten minutes, which is far tighter than the precision of anything else on this page. Note that day length depends only on latitude — Sacramento and Philadelphia are on nearly the same line, and get the same light on the same date despite very different winters.

Frost tolerance is grouped into three classes rather than graded per crop. Tender crops die at 32°F, semi-hardy take a light frost and stop at about 28°F, hardy crops survive into the low twenties and several improve for it. The frost sweetening is real chemistry: cold triggers starch-to-sugar conversion as an antifreeze response, which is why a November carrot beats a September one.

What this calculator can't tell you

The single largest source of error is your frost date itself. Published averages are just that — averages, typically meaning a 50% chance the frost arrives by then. Half of all years it comes earlier. If you want to be more sure than a coin flip, work from a date a week or two before the published average, or accept that one year in four the last crop will not finish.

Your own ground varies more than the map does. A south-facing slope, a wall that holds heat, or a bed at the top of a rise can run two weeks later than the low spot at the bottom of the same garden where cold air pools. Frost is heavier in a hollow. Nobody can calculate this for you; two seasons of paying attention will tell you more than any regional dataset.

Nor does this account for what is already in the bed. The real fall planting constraint is usually that the tomatoes are not out yet, and successional planning — knowing in July which bed frees up in August — matters more than any single date. Nor does it cover pest timing: fall brassicas often escape the cabbage worms that ruin the spring crop, which is a genuine argument for growing them in fall regardless of the tighter schedule.

Where fall planting fits

What you harvest in October has to go somewhere. The hardy roots on this list are exactly what a root cellar is for, and the greens and beans that will not keep fresh are candidates for the freezer. Both want sizing before the harvest arrives, not after.

One crop on the fall calendar is not on this list because it works differently: garlic goes in around the first frost and comes out the following July. The garlic planting calculator handles that one. And the same weeks are when the woodpile needs to be finished. Everything you work out collects on your my homestead page.