If you guessed that the U.S. has an average of about 9 new COVID-19 cases per 100K per day, AND that this is more-or-less unchanged from a week ago, then you’re today’s big winner.
Data source for this and other graphs of new case counts: Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 3/31/2022, from https://github.com/nytimes/covid-19-data.” The NY Times U.S. tracking page 3may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html
Maybe our response shouldn’t be binary
The U.K. seems to have made a secondary peak now. Note that their current rate is almost 20x the U.S. new case rate. It’s not that Omicron hit them particularly harder than it hit the U.S. Our peak rate was almost 250 new cases per 100K per day. It’s that it never really went away there. And now it’s back, as son-of-Omicron (BA.2).
Source: Johns Hopkins data via Google search
And yet, the U.K.’s response to COVID-19 — and our response to COVID-19 — seems to be an either-or, yes-no approach. Either yes, we’re in the middle of a pandemic, and precautions should be taken. Or no, we’re over it, go back to exactly what you were doing before this all happened. (Which is pretty much where the U.K is right now.)
And maybe that’s not the brightest approach. Maybe some modest permanent changes might be more efficient. Because maybe, just maybe, what we’re looking at for the forseeable future is a permanent change in the “disease environment”, for want of a better term.
Today’s Guardian has an insightful article about this. Really, it’s about what “endemic” is supposed to mean. As opposed to what they are actually experiencing in Great Britain.
I think it’s well worth the five minute read. Because, as I’ve said here before, I can’t quite get my mind around just how, exactly, Omicron/son-of-Omicrion is supposed to be come “endemic”, in the sense that the common cold is endemic. And that article explains that quite clearly.
Omicron/son-of-Omicron is ridiculously easy to spread, compared to other endemic diseases. The R-nought for BA.2 is estimated to be somewhere around 22, versus a typical value of 1.75 or so for seasonal flu. But immunity seems to fade rapidly, re-infections are now common, and the existing vaccines are mediocre, at best, at preventing any new infection. So, just like the flu, just because you had it last year doesn’t mean you can’t get it again this year. Flu shot or no flu shot.
When scientists use the term endemic, they mean something that a) is present in the population, and b) doesn’t flare up into huge outbreaks. We also think of it as implying something relatively mild, but the Guardian points out that (e.g.) tuberculosis and malaria are endemic in much of the world, and those most certainly kill a lot of people.
Well, take a look at Great Britain. Official policy now treats BA.2/son-of-Omicrion as if it’s endemic, but it currently meets none of the criteria. The U.K. hospital system is once again under strain from the volume of admissions for COVID.
So if I had one takeaway from the Guardian article, it’s that maybe we’re just not quite getting our minds straight about this. Everybody wants to return to the pre-COVID world. But that world no longer exists.
Right now, seasonality is in our favor. BA.2 or not, this is the time of year when conditions favor a reduction in spread of most or all airborne viral diseases. It’s the end of flu season. All other things equal, it ought to be the end of COVID season.
If we manage to get through BA.2, and into the summer, are we really just going to declare victory and pretend that this is all behind us, and nothing has changed? Yeah, probably, I’d guess we’re going to try to do that.
But I can’t quite get my mind around what a nice, politely-behaved endemic Omicron is supposed to look like. As the Guaradian points out, that’s probably a myth. It certainly looks like a myth for Great Britain, right now.
Basing policy on myth is generally not a good idea. But what set of rationally-thought-through permanent changes are we planning to implement? Sure looks like none, to me. E.g., at what new-case level would a Federal mask mandate on public transportation be re-instated (assuming it’s ever lifted)? Nobody can even ask that question without getting slapped down.
Which means that, in effect, we’re hoping there will never be another outbreak. Which really means that, with each new outbreak, we’ll just be winging it again.
Right now, COVID-19 presents less risk of hospitalization and death, for a vaccinated and boostered individual, than seasonal flu does. So, right now, nobody needs to think about the most ration response in the event of another outbreak. Which means that right now would be the right time to have some rational discussion about some forward-looking public health policy in this area.
Next time, maybe it would be better if our public health show was more scripted, and less improv.
The U.S. still stands at 9 new COVID-19 cases per 100K population per day, roughly unchanged over the past seven days. That’s not due to a uniformly unchanging situation across the country, but to offsetting effects. Rapid new-case increases in the Northeast are being offset by equally rapid continuing declines in the Mountain and Pacific regions.
In my last experiment, I showed how well a Ball (mason) jar worked as frost protection. In the coldest part of the night, the inside of the jar stayed 10 degrees F warmer than the outside. I thought that was exceptional performance for a lightweight uninsulated glass container. My explanation is that the glass traps long-wave infrared. And so, this works for the same reason that my radiant-barrier frost protection works. It prevents the garden bed from radiating heat energy off into space.
Long-wave infrared absorption would explain why glass worked well but polyethylene sheet was a near-total failure. A sheet of ordinary window glass will absorb about 86% of long-wave infrared, and reflect the rest. Polyethylene, by contrast, was reported to be almost completely transparent to infrared.
Accordingly, where a glass jar works well as a garden cloche, I figured that a plastic jar would not. And that’s what I tested last night.
Never let facts get in the way of a good argument.
There’s just one problem: Different plastics have different infrared absorption spectra. And it took me a while to track that down.
Using Wein’s Law, the spectrum of radiation emitted by my 50 F garden subsoil would peak somewhere around:
10 microns (micrometers) wavelength
10,000 nanometers wavelength
1000 waves per centimeter.
Those are three ways of saying the exact same thing.
So I wanted to find out how different plastics behaved with respect to long-wave radiation somewhere in that vicinity. That’s where most of the power from the upwelling long-wave radiation from the garden bed will be concentrated.
I never did find exactly the data that I wanted. But I came close. And, as it turns out, polyethylene’s absolute transparency in that region of the spectrum is the exception among plastics, not the rule.
The chart below show the absorbance spectra of various common plastics, with the long-wave infrared region highlighted. Note that the line for polyethylene is almost completely flat in that region. It absorbs almost no long-wave infrared. But PETE plastic, just below that, in fact absorbs infrared strongly right at the frequency where infrared from the soil will have its peak — wave number of 1000.
The upshot is that when I condemned all plastic for this use, I was too hasty. Avoid polyethylene, for sure. But, assuming the glass choche works as I have described it, PETE plastic ought to work reasonably well. Not as well as glass, but certainly not as poorly as polyethylene.
As an odd little footnote, Mylar plastic — the kind used to make space blankets — is the same stuff as PET/PETE plastic — polyethylene terephthalate.
Results
Below is a photo of a quart Ball jar (right) and the thick-walled PETE jar that I’m going to test. That was as close as I could get to the same size and shape as the Ball jar. FWIW, the PETE jar originally held salad dressing. You can see that it’s much thicker than (e.g.) a typical disposable water bottle or soda bottle.
When I tested that last night — two temperature loggers on a raised garden bed, one covered with the PETE bottle, one un-covered — sure enough, PETE works pretty well. But not as well as glass.
At the very coldest part of the night, the PETE jar provided between 4 and 6 degrees F of protection, or about half the maximum protection observed for the glass jar.
The lesson here is that when I condemned all plastics for use in frost protection, I was too hasty. Polyethylene sheet is a terrible choice, from the standpoint of trapping long-wave infrared. But PETE’s OK. Not quite as good as glass, but pretty close.
Source: Data from William and Mary COVID dashboard, Virginia data from Virginia Department of Health file of case counts by age group.
W&M just announced a mask-optional policy for the Williamsburg campus. It may be worthwhile to continue to track the weekly update, even though nothing much is happening now.
As you can see above, there was a slight uptick in cases for both the William and Mary campus and for the 18-24 age group in Virginia as a whole.
The U.S still stands at about 9 new COVID-19 cases per 100K population per day, unchanged from seven days ago.
Take that with a grain of salt, as an increasing number of states seem to be reporting their new case data at more-or-less random intervals.
Data source for this and other graphs of new case counts: Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 3/29/2022, from https://github.com/nytimes/covid-19-data.” The NY Times U.S. tracking page 3may be found at https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html
Separately, the more-infectious son-of-Omicron strain (BA.2) is now the dominant strain in the U.S. Which is great news, given that this has occured and we still aren’t seeing a European-style uptick in cases.Per the U.S. CDC’s COVID data tracker, BA.2 accounted for about 55% of new cases, for the week ending 3/26/2022.
If there’s any link between BA.2 and the regional patterns of increase and decrease shown on the chart above, it’s not obvious at a glance.
Source: CDC COVID data tracker
Cases are rising rapidly in the Northeast, which has a high proportion of BA.2. Cases are falling rapidly in the Pacific region, which also has a high proportion of BA.2. Whatever the impact of BA.2 is in the U.S., it doesn’t appear to be a prime driver of new case growth so far.
Edit 11/10/2023: On re-reading this, I think it’s wrong. The estimate for the energy radiating into the opening of the mason jar seems right, but the analysis fails to account for the energy radiating out of the opening of the jar. The net radiant energy input is much smaller than what I calculate below.
So now, the excellent performance of the mason-jar cloche is a bit of a mystery.
Sure, it works in practice. But does it work in theory?
My prior gardening post demonstrated that a standard Ball jar (mason jar) provides excellent frost protection, if used in a garden bed with relatively warm soil below the surface.
An inverted mason jar gives six or seven degrees Fahrenheit of frost protection. I never would have guessed that.
I had to do this experiment twice, because I didn’t believe the results the first time. For two nights running, I left a pair of temperature loggers (digital recording thermometers) on a raised garden bed, one beneath an inverted wide-mouth quart Ball (mason) jar, one in the open.
In both cases, the goofy little mason jar provided 6 to 7 degrees of warmth. That wasn’t quite enough to prevent frost inside the jar this morning (the temperature dipped below 20F around dawn). But that is, nevertheless, an impressive amount of frost protection from a device that has, for all intents and purposes, almost zero value as either insulation or thermal mass. Continue reading G22-006: Mason jar as frost protection — a winner!
Bottom line: To protect a raised garden bed from frost, use a space blanket. Or, better yet, use a piece of construction radiant barrier. Same idea, but much tougher material. Alternatively, use glass jars to cover individual plants.
I tested these methods, using temperature data loggers, and you can get anything from 5 to 10F of frost protection from them. And all of them work the same way, by trapping the heat energy from the (relatively) warm underlying soil, that would otherwise be radiated away into the air.
By contrast, a lot of advice you’ll get on the internet — and a lot of frost-protection products offered for sale — don’t do much. In this post, I get around to testing (and flunking) the use of a sheet of polyethylene plastic as frost protection.
Free advice is worth what you pay for it. If you’re lucky.
The internet offers plenty of advice on protecting your vegetable plants from frost.
Some of that is useful, but much of it is wrong. Or, if you wish to be more charitable, much of it is almost, but not quite, completely ineffective. It’s folklore that somebody read somewhere, and passed it along without testing it to see how well it works.
Is anyone surprised?
That includes the commonly-cited advice to cover your tender plants with floating row cover, or to cover them with a frame draped with plastic sheeting (e.g., polyethylene sheet) when there is threat of frost.
Neither of which works worth a damn, by the way. Better than nothing, but not by much. Which I will show, below, using temperature data loggers to test that empirically.
In this post, I’m going to finish what I started last year, regarding the use of radiant barrier material for frost protection. Unlike a house (say), on a cold, clear, still night, a garden bed loses far more energy from radiation that from conduction. Once you’ve established a pocket of still air over your bed (and so, stopped losses via convection), you get far more bang-for-the-buck by preventing heat losses via radiation than you would by focusing solely on preventing losses to the atmosphere via conduction. Hence the use of radiant barrier material.
For those of us whose understanding of heat losses comes from insulating our homes, this may seem counterintuitive. Just keep re-reading the “Unlike a house” line above, until it sinks in. For a garden bed, the relative importance of energy loss via radiation and conduction is flip-flopped, compared to a house.
The practical upshot of that is that if all that stands between your plants and a frosty death is a sheet of plastic, consider tossing a space blanket over that. That thin layer of radiant barrier material will substantially add to the minimal frost protection offered by a plastic sheet alone.
To be clear, I didn’t think up that advice. I got the idea from a document that ultimately comes from the Colorado State University extension service (.pdf). That’s seriously useful advice from people in a seriously cold climate.
Edit: Separately, a mason jar (or any glass jar) also provides excellent frost protection for individual plants. For the same reason — it blocks long-wave infrared radiation. I tested this in Post #G22-006.
This weekend’s problem
Given how quickly the weather can turn hot in this area, I decided to get the earliest start possible on some cool-weather crops. Soil temperatures in my raised beds were at or near 50F a few weeks ago, plenty warm enough for some cold-weather crops. So I planted some peas, beets, turnips, and potatoes, even though we were more than a month away from our “last frost date”.
These have now produced nice green shoots. And, of course, now we’re going to get a frost. And yet, I’m not worried.
It’s not like I didn’t have plenty of warning. Two weeks back, the long-term weather forecast for my area gave a warning of likely frost this weekend. True to the forecast, the National Weather Service is now calling for successive nighttime lows of 27F and 26F in Northern Virginia, before returning to above-freezing temperatures.
I’m not worried, because I have my frost protection already worked out. Tested as to performance (Post G21-018). I should get adequate protection for this freeze event just by covering my raised beds with cheap, sturdy radiant barrier material. (It’s the same concept as a space blanket, but a lot tougher.) But if I’m feeling particularly paranoid, I might throw a few tens of watts worth of electric night lights under that cover, just for insurance, left over from a prior experiment (Post #1412, heated faucet cover). An idea which I also stole from the U. Colorado extension service document cited above.
In this the rest of this post, I’m going to recap the radiant barrier method of frost protection, and then push a bit beyond that. At some point, I’ll come up with a raised bed cover that provides both insulation (against loss of heat energy via conduction) and radiant barrier (against loss of heat energy by radiation). For now, though, I’m mostly going to walk through why radiant barrier is far better than either plastic sheeting or floating row cover, for providing frost protection to your sensitive plants.
Part 1: The tin-foil-hat gardener.
Above, you see one of my raised garden beds with pieces of woven polyethylene radiant barrier clipped to the top.
I went through the some of the science and math behind the use of a radiant barrier for garden beds in Post G21-015. Briefly:
While a traditional tin-foil hat serves to block alien mind control rays and other forms of incoming radiation, garden radiant barrier seeks to block outgoing radiation. In particular, it serves to prevent heat energy in the garden bed from radiating away into space. On a cold, clear night, those radiant heat losses would be large, and preventing them from occurring keeps the bed warmer than it would be, without the tin-foil-hat. So let me start this off by telling you all that you really need to know about radiant barrier.
Radiant barrier greatly reduces heat losses through radiation IF AND ONLY IF at least one clean side of the barrier faces an air gap of at least an inch.
(Or if it faces a vacuum, if you happen to be NASA. Hence the origin of the term “space blanket” for one commonly-available radiant barrier material.)
In particular, it works just as well if either side faces an air gap. Either the one facing the warm object, or the one facing the cold exterior. Which means that your intuition that this works “like a mirror for heat” is only partly right. That’s OK. Just follow the rule and it doesn’t matter if your intuition fails you. This is physics, and nothing changes just because you can’t quite get your mind around it. That’s just the way it is.
For you who need a more practical example, consider Reflectix water heater wrap, sold as “radiant barrier insulation”. The shiny outside serves as a barrier to the emission of infrared energy. Touch it and you’ll feel that it’s warm. And yet, it’s doing its job. It’s warm precisely because it’s difficult for heat energy to radiate away from that shiny, “low-emissivity” surface.
For those of you whose bent is more historical than scientific, look up pre-technology ice making in India and Iran. Under the right circumstances, upward radiative losses from a pan of water can result in the production of (some) ice even when the ambient air temperature is above freezing.
Note that the rule is very specific: This shiny material must face an air gap on at least one side.
Get it dirty, and performance degrades. Why? Because dirt is perfectly capable of radiating heat away. Transfer the heat from the aluminum to the dirt via conduction, and off it goes.
Sandwich it between two other layers and it does nothing. Why? The other materials don’t prevent radiation. They are perfectly capable of radiating heat. Conduct heat through this material, radiate it away using some other material, and the heat is gone.
Use it as a ground sheet for your sleeping bag, ditto. No air gap. Lay it on the ground and cover it with snow, likewise (I think it would be no better than having a simple plastic sheet under the snow.)
You get the drift. One clean side must be open to at least an inch of air, NASA excepted.
Choice of material
Now that I have that out of the way — one clean side must face some sort of air gap — let’s proceed.
What you see in use above is a tough woven polyethelene material, with an aluminized coating on both sides. It’s used in building construction. Here’s an example from Amazon, 62 feet of 4′ wide material, for $40 (reference). At 20 linear feet to cover a 4’x16′ bed, that works out to be about $13 to cover one 4’x16′ bed.
This stuff is very tough and reasonably cheap. Given that I use it just a few days a year, I am sure it will last decades.
There are plenty of easily-obtained alternatives for covering your beds with radiant barrier material. Really, all you need is a thin sheet of something that’s extremely shiny. Think “space blanket”.
But I think they all have significant cost or performance drawbacks relative to woven-polyethelene radiant barrier sold for building construction.
Space blankets work fine (I tested that), but they tear incredibly easily, particularly when clamped in place on a windy day. In the long run, they’ll cost more than woven poly radiant barrier. A typical thickness for space blankets is 12 microns, or about 0.5 mils (thousandths of an inch) thick.
You can buy large sheets of aluminized mylar, anything from “space blankets in bulk” to sturdier, thicker material. Here’s some that’s 2 mils thick, and costs somewhat more than half what the woven radiant barrier costs (reference). The thicker material is tough, but my experience is that it doesn’t like to make the sharp-radius curves required to clamp it to a flat surface. I suspect (but don’t know) that this wouldn’t last as woven radiant barrier.
Reflectix or equivalent should work well, but looks like it’s almost 10 times as expensive as woven poly radiant barrier. Reflectix is a tough bubble-wrap type material, shiny on both sides. Presumably that would give both radiant barrier and some modest insulation against conductive losses through the top of the bed.
Performance
Last year I did a series of experiments to show that this works. I had two cheap temperature loggers — digital recording thermometers. I left half the raised bed uncovered, and logged the overnight temperature of the uncovered and covered portions of the bed.
Here are the results:
A layer of floating row cover does nothing whatsoever. This was thin row cover, but this is also just about exactly what I expected.
A covering of literal space blanket (alone) raised the overnight temperature by about 5 degrees F:
Putting some gallon jugs to warm up during the day, and then covering with space blankets at night, raised the overnight temperatures by 10 degrees F. This had about one gallon per 8 square feet, and the gallons were warmed to about 70F by the end of the day. The results below were completely consistent with the thermal energy stored in the water, relative to the energy storage (and poor conductivity) of the underlying soil bed alone.
Doing the same, but using that heavy-duty woven polyethylene radiant barrier instead of flimsy space blankets raised the beds an estimated 12 degrees F overnight. (Estimated because the battery on the control temperature logger died overnight.)
And now, a new test. You’ll often get advice to protect your plants from freezing with some sort of frame (e.g., “hoop house”) covered by a piece of plastic sheeting. So let’s now test that.
(And, to be honest, maybe the good results above were purely from covering the bed with some type of plastic sheet. Maybe the radiant barrier property of that sheet is just a red herring. My earlier calculations say that can’t plausibly be true, but that still ought to be tested empirically.)
It’s not entirely implausible that a plastic sheet would provide some protection. A single layer of polyethylene sheet should add a roughly R 0.85 (U.S. units, not comparable to S.I. R-values). But that plastic sheet is essentially transparent to infrared, and so provides no radiant barrier whatsoever.
I addressed this question, in theory, in Post #G21-015. At that time, all things considered, I figured I would need something like 60 gallon jugs of 70 F water, to keep the garden bed 10 degrees F warmer than ambient, on a cool night, with plastic sheet alone.. Roughly speaking, a gallon jug every square foot.
Here’s a picture of the setup. I have one data logger under that tightly-tucked plastic sheet. Most of the sheet sits well away from the bed. And a second data logger in the uncovered portion of the bed.
And below you see the results. The side of the bed tightly covered in plastic sheet was less than 1 degree F warmer than the un-covered side.
Summary and further reflection in the performance of unheated plastic greenhouses.
The upshot of this is that at least two common bits of internet advice for cheap frost protection seem to be more-or-less worthless when actually put to controlled trial. Neither floating row cover nor a sealed air space covered with polyethylene plastic was able to achieve even 1 degree F of warming, on average, of the course of a night.
So, despite all the customer testimonials, I remain skeptical of the idea that you’ll be able to save your plants from a hard frost something like this:
Source: Amazon.
Or something like this:
Source: Amazon.
On the latter, real growers understand that. Go on YouTube, and you can find plenty of seemingly reputable gardeners who have measured the impact of their unheated greenhouses and will tell you that the nighttime air temperature in the greenhouse is essentially no different from that of the outside air. (e.g., just over four minutes into this episode of Gardener Scott):
If you have enough thermal mass inside the greenhouse, and the area of the walls is small enough relative to the enclosed volume (i.e., the greenhouse is large), and you have (say) a double-walled greenhouse with two layers of plastic and so much higher insulating value for the walls, then you might get several degrees of nighttime warming.
But a two-foot-tall hoop house covered by a single layer of plastic isn’t going to do much. At least, that’s what my experiment suggests.
There may be some anti-frost advantage to the hoop house, in that the higher daytime temperatures — when the solar energy does drive the interior temperature well above ambient — might heat the soil enough to provide additional energy release from that warmer soil at night. So maybe a tiny hoop house, in place day and night, might serve as enough of a solar energy storehouse that there is some material frost protection.
That said, I doubt it. My soil temperatures right now in that bed are around 50 F, but the air above the bed was essentially no different from the ambient air temperature. I can’t believe that another 10F or so in soil temperature would turn that situation around.
That said, I haven’t tested that, and probably never will. I don’t use cold frames or hoop houses because, in my experience, as an inattentive gardener, I always end up cooking my plants, one way or the other. All it takes is one unseasonably sunny and warm day, and a lack of attention, and the season’s growth can be baked to death in an afternoon.
Finally, one way in which a small single-wall hoop house can provide significant protection is in snow. If it gets covered in snow, the snow layer acts as insulation. This is no different from unprotected plants, who are more likely to survive cold temperatures if buried in snow than if exposed to the air above the snow.
I don’t deny that plastic-covered hoop houses have value. They speed growth by raising daytime temperatures well above those of the ambient air. Within reason, higher temperatures mean faster growth, all other things equal.
But if you’re going to spend $40 on a small plastic hoop house, spend an extra $2 for a space blanket to cover it in case of unexpected frost. The single wall of plastic, by itself, seems to be good for maybe 1 degree F of frost protection. Add a space blanket radiant barrier should boost that to at least 5 degrees F of protection.
Extras for experts: Wall of water and glass cloche
There are at least two other commonly-suggested ways to protect plants from freezing (other than literally heating the space the plants are in).
One of those — the Wall O’ Water (r), has physics that are completely obvious.
But the most time-honored of all – the glass cloche — is a little harder to figure out.
First, it’s no secret how a “wall of water” device can protect plants from freezing. These provide a large thermal mass in the form of 2″ wide plastic tubes filled with water, forming a deep, narrow chamber that protects the plant from excessive heat losses from convection. That mass heats during the day, radiates at night, and keeps the interior of the structure above freezing even in the face of a hard freeze.
Source: Wall-o-water.com. Accept no substitutes
Figuring out why a cloche works is a little harder. Glass choches have been in use for so long — and would have been so incredibly expensive in the past — that I have no doubt that they work to some degree. It’s just not clear why they work.
A traditional cloche is a heavy piece of glass, roughly in the shape of a bell. And yet, I’m pretty sure that unlike the Wall O’ Water (r), these don’t work by thermal mass. Near as I can tell, a mid-sized garden cloche only weighs about 6 pounds. Moreover, the specific heat of common window class is only about 0.2 (reference). As a result, a typical garden cloche would only retain as much heat as roughly 1.2 lbs (or a little over one pint) of water.
N.B. specific heat is the amount of heat energy required to raise a given weight of a substance by a given temperature. In the U.S., God help is, that’s expressed as British Thermal Units required to raise one pound of a substance by one degree F (whilst proceeding at a speed no greater than 10 knots per fortnight). Under that system, water always has specific heat of 1.0, as that is precisely how the BTU is defined.
By contrast, the specification for the genuine Wall O’ Water (r) say that it takes more than 20 pounds of water to fill it (reference).
So, what appears me to be a typical glass cloche has only about 5 percent of the thermal mass of a Wall O’ Water. Moreover, people claim that you can use a mason jar as a cloche — just place it atop your plants if a nighttime frost is predicted. Surely a thin-walled mason jar has next-to-no thermal mass. A thin-walled mason jar surely has too little thermal mass to matter.
Instead, I wonder if the effectiveness of the glass choche is due to the spectral properties of ordinary glass. Glass is transparent to near-infrared — what you’d perceive as the warmth from a glowing filament, say. But glass is opaque to medium and far infrared — the type of heat radiation that would dominate the spectrum of infrared given off by soil.
Among other things, this is why infrared cameras (e.g., a FLIR camera) do not have glass lenses (reference). In the temperature ranges where a FLIR camera might be used — such as thermal imaging of a house to look for heat losses — glass is more-or-less completely opaque. In fact, FLIR cameras cannot see through ordinary window glass for the same reason, and by report, window glass appears both opaque and reflective when viewed by a FLIR camera (reference).
And so, in the end, I suspect that my sheets of radiant barrier and a glass cloche work by the same principal — they both reflect the long-wave IR radiation that is given off by warm ground. By keeping this energy from escaping, and maintaining a still air pocket, they keep everything inside that air pocket materially warmer than the outside air.
I guess, as I put the tin-foil hat on my beds for tonight’s freeze, that’ll be another thing to test. Two temperature data loggers side-by-side, one under a mason jar, one not. I’ll report that out tomorrow.
