GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Strategies for a seasonal thermal energy storage (STES) cooler

sgbotsford | Posted in Green Building Techniques on

STES = seasonal thermal energy storage.

I need a cold room for storing seedlings and bare root trees for 2-3 months of the year, from roughly the end of April to the middle of June. The temperature needs to be between 0.5 and 3. Humidity needs to be at 90% The cost of a conventional cooler system is unworkable, both because of the machinery, but also that I would need to upgrade my electrical service, as well as upgrade the power distribution line (700 feet) from the transformer to the proposed location.

The location is outside of Edmonton, Alberta, Canada. Spring melt starts in mid March. Snow is gone by mid April, although spring snows are frequent. Leaf day (when the poplars break bud) is usually the 10th of May plus or minus a week.

At present I’m considering a 24 x 24 foot space packed gravel floor, strawbale walls, clay/lime render with a vapour retardant paint on the outside. The rule of thumb I’ve gotten from the SB groups is that the warm side render should have 1/5 to 1/10 the permeability of the cold side.

A: It’s not clear to me if this is valid for a cold interior. The dew point is still inside the wall, but this is true for a warm interior too. How do bale walls dry out?

Cooth would be stored by digging a 16 foot wide 3 foot deep slope sided hole in the middle of the floor, lining with a pond liner, and filled with water. In winter the door would be left open, and a conventional 3000 cfm window fan used to move cold air into the room.

The pond would be decked, and the space above covered with 1″ thick OSB/Foam/OSB panels. These panels would be removed and stacked during the winter, and put in place in summer. (Yes, this means that I have storage at two levels. Most of this is boxes moved by hand. If it gets to the point where carts are needed, I can put permanent deck on the rest of the floor later.

We have an annual freezing index of 1750 C degree days. Ice thickness on lakes is approximated by t= c*sqrt(F) where t is the ice thickness in cm, F is the freezing index and c is a constant ranging from about 1.5 for a somewhat sheltered snow covered lake to as high as 2.7 for open lakes in windy areas. If I use 2, I get an ice thickness of between 80 and 90 cm.

If this fails to materialize, my backup plan is to put 1000 feet of plastic pipe in the bottom of the pond, a similar length on the north roof and circulate antifreeze. (Break pipe up into ~100 foot sections, and manifold)

B: Is this a workable strategy? In particular the ground temperature the first year is going to be its usual 10 C. I would like to avoid insulating the bottom of the pond, thinking that after a few years operation, I will have the local ground temperature close to freezing, allowing my ice to last longer. It also makes for simpler construction. Common frost depths here are 4-6 feet, and under driveways we put the waterline 8 feet down. My expectation is that the first year I would have shallower frost depth under the pond due to the delay of the pond freezing.

In operation the pond would be separated from the room by the insulated panels on the deck. Since it is below, there is little tendency for coolth to escape. A thermostat controls a louvered fan set into the deck. A ceiling fan keeps the air moving.

Foundation would be a rubble/ river run cobble filled trench, insulated with foam, but it’s not clear how much, or where to put it. Conventionally it would be on the outside to both keep the interior warm, and to keep the frost from heaving the foundation. Here, in operation, the inside is likely to have deeper frost than the outside, since there is no insulating layer of snow. At this point I’m thinking in terms of 2″ type II EPS that goes down 1 foot and sideways 2 feet so that there is in effect a 6 foot wide chunk of ground that the foundation sits in.

C: How should I detail the foundation.

The roof stumps me. Various sources says that since cooth doesn’t rise, the ceiling space doesn’t need to be any better insulated that the walls. The vapour barrier should be on the warm side, but it isn’t practical to get a continuous vapour barrier in between the trusses. I suppose in theory I could use A frame trusses and get a clear space across the roof, but I’m not confident in getting it right at the edges.

My present thought is to apply 6-8″ of type II EPS in two layers to the underside of the rafters, sealing the joints with spray can expanding foam. 6″ gives R24. Add another r12 of cellulose on top.

D: Is there a better way to do the ceiling insulation?

Finally: The humidity requirement creates an interesting situation: The vapor gradient in use is from the inside out, while the temp gradient is from the outside in. Very high humidity inside. Outside typical daytime values in spring run 30 to 60%

E: Does this affect the design?

F: what else have I overlooked.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. GBA Editor
    Martin Holladay | | #1

    When you wrote, "The temperature needs to be between 0.5 and 3," did you mean, "The temperature needs to be between 0.5°C and 3°C"?

    If so, that means that you are aiming for 33°F and 37°F -- right?

  2. GBA Editor
    Martin Holladay | | #2

    Depending on how crucial the upper temperature limit is -- and for the time being, I'm assuming that the upper temperature limit is 37°F -- you may be overthinking this. Perhaps all you need is an old-fashioned dirt-floored cellar (that is, a room that is mostly underground).

    I have such a cellar under my house in northern Vermont. In mid-April, the cellar temperature hovers around 33°F. It changes temperature very slowly. In mid-June, it's a little warmer, but not much.

  3. charlie_sullivan | | #3

    I agree that a simpler design could probably work. Insulating the pond and the actively bringing cold to and from it seems harder than just not insulating it. If the pond doesn't have enough ice forming, jugs or barrels could be added. And putting it below grade could make sense, although maybe having access without stairs is important?

    As for roof vs. wall insulation, one of the reasons that roof insulation is often spec'ed to be thicker than wall insulation is that the incremental cost to pile cellulose a little deeper in the attic is tiny compared to the cost to make walls thicker (once you include all the fussy details involved. It's more about bang for the buck than "need". Other differences between roofs and walls include solar heating of the roof surface and radiant cooling of the roof surface. A certified cool roof material on the surface facing the sky will help you on both counts, but isn't necessary, particularly if you use thick insulation.

    As for drying of the wall, it's tricky, because your cold side is also high humidity, so you don't really have a good drying plan either way. The conventional advice for that situation would be to use moisture tolerant materials, like EPS. Straw might be risky, but maybe that risk is acceptable. The damp parts might stay cool enough year round to avoid too much risk of mold or rot, and maybe it's OK if it doesn't last 100 years if you don't spend much money on building it. And at least nobody would be living in, so health risks would be low.

    But it does look like the dew point outside would be significantly higher than inside from late May through June, so a vapor control layer at the outside surface makes sense. If you keep it closed up during the summer to try to continue keeping it cool inside, you'll want that vapor control layer then too.

  4. sgbotsford | | #4

    The temps are in C, yes. The ideal temperature for seedling storage once they have thawed is just above freezing. You don't want the roots freeze (typically about -2) nor do you want them to be warm enough for chemistry to start up. The ideal would be 0.5C 24/7. In practice 2 degrees C is accepted, and brief excursions to 4 (door opening, sticking thermostats...) aren't the end of the world.

    It's not a sharp edged transition. Longer storage periods at higher temps => increased mortality and decreased vigor. Some species are more tolerant than others. Some mortality depends on fall conditioning.


    One of the other options I was looking at was using a larger building with a partition wall, and plastic water drums. A rain barrel on the ground is earth coupled -- the bottom freezes last, and the bottom is bulged. I am running an experiment in an insulated shed (R12) with barrels up on blocks so the bottom can freeze too, and with 3-4 inches of styrofoam peanuts floating on top. I think this will result in more uniform freezing, with the final bulge being mostly at the top instead of the bottom.

    Good point on bang for buck regarding ceilings.

    I don't know how I would do temp control in a root cellar. Our ground temp of 10 C is too warm for seedlings, and at the season's start it would be well below freezing.

  5. sgbotsford | | #5

    Progress report.

    I'm doing a trial run, using an old boiler shed. 2x4 construction, 10 feet x 10 feet, fiberglass batts, on a concrete slab.

    The ratty shed has been patched, cleaned, old flue and air intakes stuffed with insulation. 9 barrels 3/4 full of water have been moved in, supported off the slab with chunks of pallets. A window fan is on the floor keeping the air moving. 6 of the barrels have a 6-8" high cone of styrofoam peanuts floating on the water surface.

    I have a wireless thermometer in the shed, at roughly barrel top height, and one on the porch. Several times a day I record both temps.

    After 6 days of -6 to -18 C temps, one peanutless barrel has an inch of ice on the top surface. The peanut treated barrels have not yet started to freeze on top.

    In general the temp inside the shed is 1/10 of the outside temp. E.g if it's -15 outside, it's -1.5 C inside. There is a delay of a few hours.

    I did some rough calculations, using info on this site (thanks!) and my nominal R12 2x4 wall shed probably has an overall building envelope insulation value of between R3 and R4. AFAICT the concrete slab is just a slab on grade with no insulation. This makes it the dominant heat transport. OTOH adding baled straw around the foundation in spring will reduce the warmup then.

    Answering the thought of a root cellar in more detail:
    Ground temperatures to very close approximation are the same as shallow well water temperatures. Here that is 10 C or about 50 F. Potato cellars here are installed with two thermostats: One is set to bring in cold air from outside if the interior temp gets above 40 F, the other set to turn on a heater if the temp gets below 34 F. By the time I need coolth, outside temps are too high to be effective, although I could probably get an extra week by pulling in cold air just before dawn.

    I've heart of one guy who has a root cellar that he floods over the winter, buillding up 2 feet of ice. In spring he insulates the floor with straw, and brings his trees in. I don't have running water on site once there is frost in the ground. (My water system is on the surface, and is drained in fall. I'm not going to dig 2000 feet of 8 foot trench through poplar bush to have year round water.)

    I don't know how he controls his temp. I would expect it to be too cold in early spring, and develop extreme stratification in late spring, as surface ground temps increase. Perhaps he uses fans.

  6. charlie_sullivan | | #6

    Thanks for the update. A few thoughts:
    1) Stuffing holes with insulation might or might not air seal them--fiberglass insulation doesn't work well to block air flow, but foam insulation board would, especially with the edges caulked, sealed with canned foam, or good tape.
    2) If your goal is to freeze the water in the barrels, wouldn't you want to blow air in from outside, rather than just circulating air inside?
    3) Spaced cooled with ice, without salt on the ice, never get below freezing as long as the outdoor temperature is above freezing. The surface of the ice stays right at 0 C, and the room will stay a little above that temperature. It's self regulating, automatically aiming for the temperature range you want.

  7. sgbotsford | | #7

    All too true about the fiberglass. unfortunately there are still boiler pipes also in the way, putting a foam board would be a PITA. The fiberglass is iin bread bags.

    The long range plan is to freeze the barrels. But I have January and February. Right now I'm playing. By observing the system behaviour while the system is cooling, I get a better understanding of it. This may help come spring.

    We will likely have 2 weeks sometime this winter when temperatures drop below -30. The issue in spring will be how much below freezing the mass of ice, the slab, and the soil below the slab are. From -30 to 0 is about 1/4 of the coolth as ice to liquid. Hence keeping the seedlings physically above the ice. Run a fan to mix the air as needed.

  8. charlie_sullivan | | #8

    Fiberglass in bread bags sounds pretty effective, actually.

    And now I understand your experiments a little better. It sounds like ideally you would want the barrels to just finish freezing as you get to the beginning of your storage season, so they never get down to - 30. But the effort to try to control that might not be worth it.

  9. sgbotsford | | #9

    Between the barrels and 3 sets of shelves, the floor is not clear. But I think I can lay down styrofoam and OSB on the center of the floor, and loose lay fiberglass bats where it's awkward. If I do this in spring then the coolth stored in the slab is throttled back.

    I also have some surplus insulated tarps --the kind used when doing concrete pours in winter -- probably R2-3 I can wrap the barrels with. Come spring I will work on getting the temp right before the trees come.

    The way the indoor temp follows the outdoor temp right now is telling me that the surface area of the barrels is likely limiting. 9 barrels have about 200 square feet of area. The building has 520 square feet (Including ceiling and floor) of area. If we consider that the barrels have an effective R1 because of two boundary layers, and that the fan is disrupting one, then the barrel is about R 1/2 200 sq ft of barrel at R1/2 is 400 BTU/hr/F. But the room this morning was 28F. So 1600 BTU/hr

    The slab if uninsualated is coupled to the ground. While the surface of the slab would cool quite quickly, earth runs about R1 per foot, possibly 1.5 since it won't be wet. So right now I may be getting some warming between 50 degree earth and 30 degree room. 100 square feet at R1 with a 20 degree differential could be adding 2000 btu/hr. The rest of the building, 420 square feet at R6 with a 30 degree differential is losing heat at 2100 BTU/hour.

    If that were all, then the shed temp would be half way between earth temp and outside temp. In the low 20's In fact it's running above that, in the upper 20's We're flirting with my uncertainties here: How well insulated is the shed really; as the barrels freeze, their effective R value goes up, in a hard to figure way, since they don't freeze at a uniform thickness on all sides; As the slab cools, the differential temperature of the top foot declines, so the effective R value of the slab increases.

    If this model is correct then the temperature inside the shed will follow the outside temperatures more closely as both the barrels and slab continue to cool. I expect that I'll see a temp about half way between freezing and the average outside temperature.

    Just did a quick spreadsheet: If I call the slab R2, the walls and ceiling R12, the stud bridging R4.5 (OSB on both sides) the barrels R1, the door R2, and 1 air change per hour it fits this morning's temps to within 100 BTU per hour. Admittedly: I've a surplus of variables, but I'm pleased that WAGs get an agreement to within 5%

    I may be out to lunch, but not totally out to lunch.

    (Interesting thought: Is the coolth stored in the ground by itself sufficient for my purposes?)

  10. sgbotsford | | #10

    Shed report:

    It's now been over 3 weeks since I started my STES experiment. During this time we have been having typical winter weather. -5 to -22 C for extremes, but mostly -8 to -14. During that time, shed temps have tracked outside temps fairly closely, allowing for a few hours delay. See attached graph. A rough calculation showed that with these temps, I should be running low on barrel heat (the water mostly frozen) by the end of the the third week. And the graph shows the inside temp starting to slide. Of course this could also be a matter of declining heat from the ground slab. Not sure how to tell these two effects apart.

    I find amazing, that it's actuallly working pretty much as predicted -- at least so far. Given how sloppy my approach was, I expected 50% errors between prediction and observation. From experiences with minimal effects in home made green houses, I thought that my water would freeze faster than expected, and that the temperatures inside the shed would be close to the average outside temp by the end of two weeks.

  11. charlie_sullivan | | #11

    Thanks for the report. It's amazing how well your very simple correlation works.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |