Ice Station Thuban

A temporary enclosure made entirely of water

Similar projects worth following
This is a water hack. When the temperatures are cold enough, it is possible to construct a geodesic dome made from water ice. The material is easy to move in liquid form to nearby the construction location, by hose or bucket, and then it is just a matter of time before usable panels result from the phase change.

Unlike an igloo, which is made from air-filled and lightweight snow, this dome is made from ice, heavy, structural and able to support large loads, including its own weight. And since it is much more dense than snow, it can last a lot longer than an igloo, when below-freezing temperatures are sustained. It can also tolerate fairly large excursions into above-freezing ambient temperature, due to the large thermal mass of the ice. A little melting and re-freezing actually causes the panels to weld together, making the entire structure stronger.

This project was attempted to understand the practical difficulties with building such a structure, and to begin to understand some of the issues involved with scaling a structure like this to larger sizes.

In a seasonal situation (meaning, a location with both above- and below-freezing temperatures during the year), there is a certain minimum wall thickness that can be built, such that the warm weather cannot melt the structure entirely, before freezing weather returns. This project does not test that. That minimum thickness is too large to be practical in my location.

In an environment such as the Antarctic, where freezing temperatures can be sustained, melting is not the enemy - though a dome may be slowly eradicated by wear from wind and through sublimation. This project does not test that, either.

Thuban is a star in the constellation Draco, used in ancient times as the North star. It would have been fun if tolerances allowed this structure to be aligned with the motion of this star, but that was not the case. The name was chosen to point out that this is not Ice Station Thula.

The resulting dome was about nineteen feet in diameter, nine feet in height.  The triangular panels were approximately three feet on a side, and weighed about sixty pounds each. Approximately two tons of water were self-suspended above the ground with no other materials present (other than water impurities and tiny windblown objects - no reinforcing structure). The dome was assembled over a period of three weeks when temperatures almost remained below freezing (there were a few excursions into above-freezing temperatures, no problem). It lasted about three days above freezing before the top collapsed in, and melted entirely over a period of another three weeks.

This is an ongoing project, because there is a version 2. To be continued...

  • 1 × Ball of Whacks (Optional) a small model to keep track of the various triangular faces
  • 1200 × Gallons of water Construction material
  • 32 × Degrees Fahrenheit (or less) Construction material solidification agent
  • 5 × 10' x 20' Plastic sheeting, 3 mil Triangular panel form waterproofing
  • 108 × 0.25" x 4.5" x 48" Wooden boards Triangular panel form structure

View all 9 components

  • Dome sweet dome

    Kenji Larsen06/03/2014 at 04:52 0 comments

    I realized I basically did the whole build log while writing up the instructions, so check out the instructions. Some more pictures there too.

    So for this log I will write what I learned, pros and cons:


    • The triangular panels are very heavy. Smaller panels could have resulted with a higher order tesselation, but it would have increased the number of unique shapes to keep track of and orient. The tradeoff would have resulted in longer management time, and may have made it impossible to complete in the timeframe available. Managing fewer, but larger and heavier panels may have been faster, but was certainly not easy, and was actually dangerous. The panels are slippery, because they are made of ice, and they are pointy, being triangles. They do drop, and they do drop on your foot. I would say there was approximately a 25% casualty rate on these triangles just from slip and fall.
    • The wooden panels were thin, because I was looking for the least expensive solution for these forms.That in itself was not a con, but the thin wood flexed outward from the weight of the water. This means the edges were not quite straight, and therefore two adjacent triangles always had a gap of an inch or two somewhere along the seam. A mortar made of snow and water worked quite well to fill these gaps, but it did lead to imprecision in placement, further leading to spherical aberration errors that required constant correction.
    • The large panels each had a large thermal mass, making them take a very long time to freeze, even when the air temperature was very cold. Scooping out an air channel in the snow under the center of the form to create airflow and potential heat transfer may have helped a few percent.
    • Rhombic triacontahedron tessellation required that each triangle be placed in the correct orientation every time. Two shapes were mirror images, so it could get confusing, and I did make some mistakes I had to correct.
    • Plastic sheeting would develop pinholes. That would cause water to drain out faster than it could freeze, but too slow to notice the leak. Easily fixed with more plastic sheeting, but detecting this problem took hours, resulting in huge delays in the manufacture of a single panel. Significant when a single panel is such a large percentage of the total hemispherical area, and you are racing the weather.
    • Oh my aching back.
    • Oh my freezing hands.


    • The triangular panels were large, so when one was placed, it took up a large solid angle comprising the hemisphere. The placement task could be done quite quickly.
    • Rhombic triacontahedron tessellation provided a low number of unique triangular shapes, while providing a high degree of spherical approximation.
    • Sound effects: inside the dome was highly echoic and various focal points for sound existed. Two people talking inside the dome could find certain spots with an amplification effect.
    • Light effects: looked cool!
    • Unique - only one on the block!

    Here are some pretty pictures from after the melt.

View project log

  • 1
    Step 1

    DO NOT attempt this build. It is actually pretty dangerous, with large (sixty-pound), near-frictionless, three-pointed weights suspended above your head, aiming for your feet. They can crush you or spear you or cleave you. Once fully completed it is safe, but only temporarily. Do not attempt this on any property where you cannot 100% control access. Do not attempt this if there is any possibility of unauthorized access. The dome will eventually melt, and will collapse, and you cannot predict when. It weighs enough to be potentially fatal. It is advisable to intentionally destroy the dome once above-freezing weather is expected to stay.

    If you disregard this warning, then plan ahead. The preparation will take some time but is necessary to deploy as fast as possible once an expected run of below-freezing weather starts.

  • 2
    Step 2

    Calculate the size and shape of the triangular panels. The dimensions will depend on the size of the dome and the geodesic basis shape. The dome here was based on a second order tessellation of the rhombic triacontahedron. (Many geodesic domes are based on the icosahedron, but I think that the triacontahedron gives you the best bang for the buck on low number of unique triangular shapes and high spherical approximation.)

  • 3
    Step 3

    Obtain wooden planks (they don’t have to be wood but it is inexpensive and easy to cut). Make edge cross lap joints so that they form the triangles calculated in the previous step. Note that you can angle the joints as I did so that the edges of the resulting panels bevel according to the spherical radial direction, but it is not critical to do so. The approximated sphere is determined by the edge lengths, not the bevel angle.

View all 17 instructions

Enjoy this project?



Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates