Why a Winter Greenhouse Needs Mass, Not Glazing

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

Badly, and the reason is not the snow. A pop-up greenhouse fails in winter because of what it does not have inside it, and the missing thing is weight. An unheated enclosure has no night-time temperature elevation at all, because elevation is heat going in divided by heat leaking out, and after dark the heat going in is zero. Whatever the glazing is made of, by dawn the inside is the same temperature as the outside unless something in there was storing heat.

Which turns the usual advice inside out. Double glazing, extra film layers and sealing the gaps all change how fast the inside cools, not where it ends up. The only thing that changes where it ends up is mass, and a light structure standing on paving with trays on wire shelves has essentially none.

The day is the part that works, and it is worth 17 kelvin

Start with the half of the cycle that does behave as people expect, because it also produces the most satisfying check on this whole page.

A Backyard Discovery Aggie cedar cold frame publishes an unusually complete set of figures: 48 inches long by 28.5 wide and 32.25 high, 94.79 pounds, a double-wall polycarbonate lid, and a wax hinge that its record says opens "when temperatures exceed 70F and closes it again as things cool".

Those dimensions give a lid and footprint of 0.8825 square metres and side walls of 3.1833. Put a representative loss coefficient on each surface, 2.9 watts per square metre per kelvin for three quarter inch cedar and 3.6 for double-wall polycarbonate, and the whole frame loses 12.41 watts for every kelvin it sits above the garden.

Now the sun. Clear winter midday on a horizontal surface is around 300 watts a square metre, and polycarbonate passes about 80 percent of it, so 211.8 watts arrives inside. Divide by 12.41 and the frame runs 17.07 kelvin above the outside air.

On a 5 degree Celsius winter day that is 22.1 inside, which is 71.8 degrees Fahrenheit. The hinge opens above 70. That threshold is not a round number somebody picked. It is within two degrees of where this frame's own published geometry puts it at midday in winter, which means the wax was chosen to open at the point this specific design starts cooking what is inside it. Its record gets the reasoning exactly right in words: "Seventy degrees Fahrenheit is a sensible threshold: warm enough that the frame genuinely needs venting, cool enough that it opens before damage rather than after." The arithmetic agrees and says where the number came from.

Then the sun goes, and the only question is the time constant

After sunset the frame has no heat source. So it cools from wherever it was, and the shape of the cooling is an exponential whose time constant is the heat stored inside divided by the rate it leaks.

What is stored inside? Not the air, which is negligible. The soil. A cold frame sitting on the ground is thermally welded to the top fifth of a metre or so of growing medium, and over the Aggie's 0.8825 square metre footprint that is 141 kilograms of moist compost at about 1.8 kilojoules per kilogram per kelvin, which is 254 kilojoules per kelvin. Add roughly half the structure's own 94.79 pounds and the store comes to about 280 kilojoules per kelvin.

Divide by the 12.41 watts per kelvin and the time constant is 22,564 seconds, which is 6.27 hours. A twelve hour winter night is 1.914 time constants, so 14.75 percent of whatever elevation the frame had at dusk is still there at dawn. From ten kelvin at dusk that is one and a half kelvin at sunrise.

One and a half kelvin is a real and useful number. It is the difference between frost and no frost on a marginal night, and it is the entire thermal case for an unheated frame after dark. It is also not warmth in any sense a gardener would recognise from the daytime figure, and it exists because of the soil rather than because of the lid.

Our own record credits the wrong term, by a factor of 8.7

Here is where the arithmetic disagrees with something we have published. The Aggie's performance text says: "Double-wall polycarbonate is the right glazing for season extension. The air gap between walls insulates, so the frame holds overnight warmth better than a single sheet of glass or film."

The first sentence is right. The second attributes the overnight behaviour to the glazing, and the glazing is the small term.

A Beipoo 2 tier wooden greenhouse cold frame is the perfect control, because it is the same idea built the other way. It publishes 35.43 inches long by 14.57 deep and 27.56 high, fir rather than cedar, high transparency PVC rather than double-wall polycarbonate, and, crucially, a cabinet on legs with removable slatted shelves. Its own overview states the design intent precisely: "rather than protecting a patch of soil, it holds trays and pots at two working heights".

Run both numbers. Area-weighted, the cedar frame loses 3.05 watts per square metre per kelvin and the fir cabinet loses 5.50, a factor of 1.80. That is the comparison both records make. Now the stores: 280 kilojoules per kelvin against perhaps 18 for a cabinet holding trays on slats, a factor of 15.6. That is the comparison neither record makes, and it is 8.7 times larger.

In time constants it is brutal. The cedar frame cools with a 6.27 hour constant. The two tier cabinet cools with a 1,324 second one, which is 22 minutes, so a twelve hour night is 32.6 time constants and the inside is at garden temperature before the washing up is done.

So the Beipoo's drawback, "PVC panels insulate less than double-wall polycarbonate, so overnight heat retention is weaker", is true and is pointing at the wrong thing. The retention is weaker because the frame is deliberately not touching the ground. The thing that makes it genuinely useful, two levels of working height in a 35 inch footprint, is the same thing that removes 94 percent of its thermal store. That is a real trade and a good one for propagation in spring. It is not a winter structure, and the glazing is not why.

What a cover actually does after dark, and it is radiation

If an unheated enclosure cannot raise the night temperature, why does covering plants visibly prevent frost? Because the thing it stops is not conduction. It is the view of the sky.

On a clear night the atmosphere radiates downward like a body roughly 25 kelvin colder than the air at ground level, and a leaf facing it radiates upward into that gap. Put the numbers in: a leaf at 273.15 kelvin with an emissivity of 0.95, against a sky radiating like 248.15, loses 95.61 watts a square metre and gets almost nothing back.

Nothing replaces that except convection from the surrounding air, so the leaf settles below air temperature by 95.61 divided by the convective coefficient. In dead still air that is about 9.56 kelvin, at a moderate coefficient 6.37, and in a light breeze 3.82. Which is why frost forms on a clear calm night when the thermometer in the shed reads three degrees above freezing, and why a windy night at the same temperature is safe.

Now put any cover over it, including a single sheet of film with no insulating value whatsoever. The leaf now exchanges radiation with the underside of the cover, which is close to air temperature, so the net goes to nearly zero and the depression disappears. A cover does not make a plant warm. It stops the plant being colder than the air it is standing in, and that is worth between about four and ten kelvin on exactly the nights that matter.

Which is the honest answer to the original question. A pop-up greenhouse in winter is worth real frost protection, that protection comes from blocking the sky rather than from holding heat, and it works the first night and every night without needing mass, glazing or a seal. What it will not do is give you a growing season.

The snow, which is the other half of the question

Take the Aggie's published weight seriously, because it is the only weight in this group and it makes the snow case computable.

Wet snow runs 300 to 400 kilograms a cubic metre once it has settled and taken rain. Twenty centimetres of it is 60 to 80 kilograms a square metre, which over the Aggie's 0.8825 square metre lid is 52.95 to 70.60 kilograms, or 589 to 785 pascals.

The frame itself weighs 94.79 pounds, which is 42.99 kilograms. So a single moderate snowfall can place more weight on the lid than the entire structure weighs, by up to 64 percent. On a cedar and steel frame with a double-wall polycarbonate panel that is survivable and the five year warranty its record publishes suggests the maker thinks so too. On a fibreglass-rod and film structure of a few kilograms it is the collapse load, and it arrives overnight with nobody watching.

The Beipoo cannot be checked, and that is the finding. It publishes three dimensions and no weight anywhere in its record, so its snow margin, its wind margin and its ballast are all uncomputable. Not one cold frame or greenhouse product we carry publishes a snow load rating of any kind.

The drawback that is actually the answer

The Aggie's record lists this as its first drawback: "94.79 pounds, so effectively a fixed installation."

For a winter question that is the most valuable property the product has, and it is doing two separate jobs at once. The same mass that makes it immovable is the mass that holds the dawn temperature up, and it is also the ballast. Work the wind: the broad side is 0.9988 square metres, and at 15 metres a second with a drag coefficient of 1.2 the force is 165 newtons, which is 16.84 kilograms force against 42.99 kilograms of frame. A ratio of 0.392 means it stays where it is with nothing holding it down. Our page on choosing a patio umbrella and anchoring it works the wind loading on a garden structure through properly, and this page takes its method rather than repeating it.

A light structure gets the opposite answer. The two tier cabinet presents 0.63 square metres and takes 10.62 kilograms force in the same wind, which on an unpublished weight is unknowable but is certainly a large fraction of it. Which is where a 10 pack of ZOOFOX heavy duty ground stakes earns its place in this answer rather than any greenhouse accessory: 15.8 inch galvanised steel at 0.47 inches thick, ribbed to resist pull-out, with a loop for a guy line. Its own record is honest about where it stops, saying "Big structures in storms need extra anchoring" and "No stake replaces proper anchoring of a big structure in a storm". For a pop-up frame the stakes are not an accessory. They are the substitute for the 94.79 pounds a light structure does not have, and they only replace the ballast half of it, never the thermal half.

What would actually make it a winter greenhouse

Go back to the equation at the top. Elevation is heat in over loss out, so if you want an elevation after dark you need heat in. Our roundup of greenhouse heaters for frost free winter growing is the category that supplies it, and the arithmetic says how much: to hold the Aggie 5 kelvin above a winter night takes 5 times 12.41, which is 62 watts, continuously, for twelve hours, which is 0.74 kilowatt hours a night.

Sixty two watts for a four by two foot frame. Scale that to a walk-in structure with ten times the envelope and it is 620 watts all night. That single division is why unheated winter growing is mostly a myth and heated winter growing is mostly a utility bill, and it needs no product data at all.

The only product in our catalogue that supplies heat to a growing space is a Kokopro seed starter kit with grow lights and a heat mat, which holds 70 to 85 Fahrenheit and measures 10.2 by 8.07 by 6.69 inches. It is also, in its own words, for "indoor use only and must not be immersed", so it is not a cold frame heater and should not be pressed into service as one. Our page on whether heat mats make seeds germinate faster works through what bottom heat does and does not do, and makes the distinction this page keeps leaning on: soil temperature and air temperature are different quantities, and it is the soil one that matters to a seed.

A VEVOR 12 inch greenhouse circulation fan rated 1800 CFM is worth naming for the opposite reason: it solves the day problem rather than the night one, and it solves it for a structure far larger than a frame. Eighteen hundred cubic feet a minute is 0.85 cubic metres a second, and the Aggie's entire internal volume is 0.72 cubic metres, so this fan would exchange the whole frame more than once a second. Its real use is the 17 kelvin daytime rise in a walk-in house, where the alternative to venting is the same cooking problem the wax hinge exists to prevent. Its record also already flags a published internal contradiction in its own drawbacks, that "The listing's size attribute conflicts with the stated fan diameter", which we reported before this page and which still stands. Our page on why a fan's CFM rating is not the breeze you feel covers what that airflow figure does and does not promise.

The honest limit, and the one change that works

Three things on this page are exact and they are the three that carry the argument. An enclosure with no heat input has no steady state temperature elevation. Stored heat decays exponentially with a time constant equal to the store divided by the loss. And a surface exchanging radiation with a body at its own temperature nets zero, which is why a cover removes the sky depression completely rather than partly.

Everything thermal is representative and has to be, because not one cold frame or greenhouse product we carry publishes a U value, an R value or a glazing thickness. The 2.9 for cedar, the 3.6 for double-wall polycarbonate, the 5.5 for the PVC cabinet, the 0.2 metres of coupled soil, the 300 watts a square metre of winter sun and the 25 kelvin sky depression are all assumptions. Halve the coupled soil depth and the cedar frame's time constant falls from 6.27 hours to 3.33 and its dawn residue from 14.75 percent to 2.7, which makes the gap against the tray-based frame wider rather than narrower. Double the glazing performance and the 1.80 factor becomes 1.3, which makes the mass term more dominant still. The conclusion is robust in the direction that matters and the specific hours are not.

What this page deliberately does not compute is the collapse load of any actual structure, because that needs a tube wall thickness, a material and a joint detail and not one listing publishes any of them. The snow weight is computable and the strength it is pushing against is not, which means the right output is a weight rather than a verdict.

Which leaves the practical answer, and it costs nothing. If the goal is frost protection, the cover does that on its first night and the material barely matters, so clear the snow off it before it accumulates and stake it against the wind. If the goal is a growing season, the glazing choice is the wrong lever entirely. Put the frame on soil rather than on paving, which is the single biggest change available and the one nobody sells, and if you want it to still be above ambient at dawn, fill the inside with something heavy and dark. A row of water-filled bottles inside the frame adds its own mass to the store at the top of this page, and water holds 4.186 kilojoules per kilogram per kelvin, which is more than twice what the soil does per kilogram. Twenty litres of water is 83.7 kilojoules per kelvin added to a 280 store, which lengthens the time constant by 30 percent and lifts the dawn residue from 14.75 percent to 21.2. That is a bigger overnight gain than any glazing upgrade in this category, and the bottles are already in your recycling.