Traditional, unconditional, last-frost date
I had planted a few cold-hardy vegetables in my garden weeks prior to last weekend’s deep freeze. I put in some snow peas, potatoes, beets, garlic, onions.
It got down below 20F briefly on one of those nights. I can now say that all of those appear to have survived, with just a bit of TLC. That was in the form of capping the bed with radiant barrier, then adding a piece of plastic for air-tightness. (See Post G21-018, or my just-prior garden posts.)
It’s no surprise that we had a freeze. Our nominal “last frost date” is somewhere around April 22,so these plants were in the ground almost two months ahead of that. Instead, the interesting thing is that I had two weeks’ warning that the freeze would occur. The fourteen-day forecast accurately predicted that there would be a freeze that weekend, although the original forecasts understated the depth of that freeze.
This leads me to ponder the implications of reasonably-accurate long range weather forecasting and our “last-frost” dates. Folklore guidelines (“plant peas on St. Patrick’s Day”) and science-based “last frost date” guidelines predate the era of supercomputers that make long-range forecasting possible. Weather is still chaotic in the mathematical sense, and so not predictable at very long intervals, but we now have two-weeks-ahead temperature forecasts that are reasonably accurate.
I already rang the changes on this once, in post G21-005, Your 70th percentile last frost date is actually your 90th percentile last frost date. What you typically see cited as your “last frost date” is the date on which, historically, frost only occurred after that date around 30 percent of the time. But that’s an unconditional probability, as if you would plant on that date regardless. If, by contrast, you check your 14-day forecast on that date, and refrain from planting if frost is in the forecast, then you’ll convert that to a 90th percentile last-frost date. That conditional probability — chance of frost after that date, conditional on a frost-free 14-day forecast — gives you a much higher chance of avoiding a freeze after that date.
The upshot is that a reasonable prediction of the two weeks following the “last frost date” shifts the odds attached to that date considerably. It’s actually a lot safer to plant frost-sensitive plants on that date, in the modern world, than it was in the era when no forecasts extended more than three days. As long as you make that decision conditional on the extended forecast, and you’re smart enough not to plant if it looks like frost any time in the next two weeks.
At present, we’re creeping up on 14 days prior to our April 22 “last frost date”. And I’m pondering — just as an exercise in probability and statistics — whether that same math works 14 days in advance of the date.
And I’m pretty sure it does. If the 14-day forecast were completely accurate, then the conditional 70th percentile last frost date in this area would be April 9th. No frost in the forecast through April 22 would mean that the conditional odds of frost occurring after April 9 would be the same as the unconditional odds after April 22.
That is, April 9 is our conditional 70th percentile last frost date. If we have a decidedly frost-free 14 day forecast at that point, planting on that date bears the same risk of frost damage as planting blindly on April 22.
The only uncertainty there is in how accurate the 14-day forecast actually is, for daily low temperature.
Weather forecasts seem to be one of the few true ephemera of the digital age. They are published, and then they are replaced with the next day’s forecast. Nobody cares about yesterday’s forecast, other than those who have some deep professional interest in forecast accuracy. Accordingly, where you can look up the actual weather 14 days ago, I haven’t yet located a database that lets me look up the actual weather forecast 14 days ago.
So that’s going to have to remain an unknown, for the time being, unless I want to try to compile the data, for my location, day-by-day, myself. Or if I can find existing research that addresses this exact question of predicting a frost. So I’ll just have to leave that as saying that if the 14-day forecast shows lows that are well above freezing, then you can probably move your traditional (unconditional) 70th percentile last-frost date up by two weeks.
But is this just the second-biggest waste of time in the U.S.?
The second-biggest waste of time in the U.S.A. is doing something really well that doesn’t need to be done at all. (I heard that in a time-use seminar I attended decades ago.)
In the fall, frost protection has some clear advantages. The plants are already grown, the produce is already ripening. Protection from an unexpected early frost is a matter of saving garden produce that would otherwise be lost.
But as I hustle about protecting my plants in the spring, it invites the obvious question: Just how much am I gaining by planting these crops early? And to that, I will add not just planting early, but the whole process of starting seeds indoors, regardless of the planting date.
In reality, is this really just an example of the second-biggest waste of time in the U.S.?
Ultimately, while some plants may grow in the cold, they tend to grow slowly. At some level, that’s just basic chemistry. The rate at which a typical chemical reaction proceeds roughly doubles with every 10 degrees C of temperature increase. Sure, plants will develop enzymes to speed those processes in colder temperatures. But it doesn’t take a genius to notice that while they will grow, they sure won’t grow very fast.
What prompts this is my peas, which are now all of about 2″ high. And it’s getting on close to a month after they went into the ground. Is that head start worth it, compared to simply waiting for the nominal last-frost date and planting them then?
In short, I’m beginning to suspect that my current setup — plant early, provide frost protection, but no greenhouse — might just be the least efficient of all possible worlds. All the hassle of early planting, and (almost) none of the benefit.
Without a greenhouse structure (or poly tunnel, or similar) to warm the daytime air and soil temperatures, it seems like most of what I’ve done is to induce my plants to try to grow under inhospitable conditions. And they are responding accordingly.
Back when I was a low-effort gardener, I seldom mucked around with any type of early planting. I’d start seeds a couple of weeks before I planted them, just to be able to have a tiny visible plant to stick in the ground. (And so, have better chance of survival for (say) tomato plants.) But my opinion then was that the gains from very early planting were minimal. Give it a couple of weeks, and the (e.g.) peas planted later in the year will have effectively caught up with those planted earlier.
As a result, I’m now wondering whether I’ve been taking all this early-planting advice from people who do early planting and have some type of greenhouse arrangement on top of those early plantings. From what I’m observing so far, that would make a lot more sense than just sticking plants in the ground and protecting them from freezing as needed.
When I briefly Google for this topic, all I see is people touting the benefits of early planting. In effect, a series of statements that you’ll get more out of your garden if you do it. I’m not seeing any quantification of just how much more you get, from early planting alone (i.e., with frost protection but not a greenhouse or poly tunnel).
So, before I get any further caught up in this effort to see just how much I can push that last-frost date, and just how well I can protect those tender plants from frost, it seems like I need to assess the cost/benefit tradeoff.
I’ve proven that I can plant well in advance of that last-frost date. I can do that very well, thank you. But should I do that? I don’t think I’ve really answered that question. And, in particular, should I do that without some sort of setup to warm the daytime air and soil temperatures?
Maybe early planting without a greenhouse really is just the gardening equivalent of the second-biggest waste of time in the U.S.A. Clearly, that needs to be the next thing I test. For that, I need some sort of cheap, safe, low-effort greenhouse or poly tunnel. One that minimizes the chances that I’m going to bake my plants to death.
So that’s the next thing on the agenda. Replant what’s in my garden, one month after the original planting date. And work up a greenhouse covering that, as a lazy gardener, I can live with.
