Is It Possible to Build an Earthship in Moist, Freezing, Expansive Clay Soil?

We have beautiful, heavy, clay-rich soil. It’s great for cob (which explains the local proliferation of cob houses), great for earthen plasters and earthen floors but it seems to pose some challenges when it comes to underground construction such as Earthships. I hope in this post to outlines the challenges and what solutions I have come across to deal with if. If you’ve built in an Earthship with such soil then please stop by and share your experience with it. Midway into writing/editing this post I came across this thorough description of expansive soil and their potentially adverse effects on construction. From reading it and a few others resources I feel is it important to note that:

  • Almost all mentions of expasive soil issues are in relation to foundations. An Earthship has no foundations.
  • Almost all mentions of expensive soil relate to hard-concrete responding to intense uneven pressures. An Earthship is inherently a “softer” structure (then concrete) embedded in the earth. We intend to embrace that concept and even our floor will be a “soft” earthen floor and not a rigid concrete slab.
  • Expansive soils are not inherently a problem – fluctuations in their moisture content is a big problem. If moisture content is stabilized then the problem is largely diminished.
  • An Earthship is inherently a massive structure (even more so with a living roof we intend to add) and as such an Earthship is capable of “pushing back” against the forces of surrounding expansive soils.

In addition to all this we have had an opportunity to observe how things behave in real life which is an excellent teacher – especially as I am about to get into a lot of theoretical ideas. We live in a cob house that was built in 1934 and it is structurally sound. It has partial peripheral stone foundations and is holding up find sitting on expansive clay. Other houses in the village were built with without any foundations and have been standing for many years. Of course there are also decayed houses … but I cannot say what kind of role soil-expansion had to play in their history. Though they have not yet withstood the test of time, we have built two underground concrete boxes (with manhole access) for our water infrastructure. Their walls are ~10cm thick with rebar – and they have shown no signs of stress problems. We have had a few ditches open over recent months and they have seen many transitions from wet to dry and they also showed no structural decay.

Expansive Clay Soil

It took me time to understand what all the structural fuss and warnings are about “Expansive Clay Soil” and it all went back to understanding the structural qualities of the soil itself. The soil composition itself is considered clay-rich and my first misconception was that that meant it was mostly clay. This is incorrect … clay rich soil has a relatively small percentage of clay particles in it – typically ranging from 10% to 25%. While the clay is not a major quantitative element it is a dominant qualitative one.

Clay particles in clay-rich soil expand when they come in contact with water. As it absorbs water it becomes sealed and much less penetrable for water. This makes saturated clay-rich soil slow-percolating. We witnessed this clearly when we dug a small hole in the ground and filled it with water. The small pool stayed in place for quite some time, percolating into the ground very slowly.

This makes clay-soil a structural force to deal with when building an underground house. There are two phenomena that lend a hand to the expansive behavior of clay. The first was mentioned above – clay absorbs moisture and expands. So, for example, the fall season rains saturate the soil and the clay expands. Then winter brings into play the second phenomena – freezing. The saturated and already expanded clay soil is now exposed to freezing temperatures which cause even more expansion.

These forces are insignificant in a bucket of clay but can translate into potentially thousands of tons of force pushing up against a house that is buried in the ground. After carefully internalizing the “expansive” behavior of clay it seems to me that the problem is not the clay rich soil itself but it’s exposure to moisture.If it’s kept dry then clay rich soil is actually an excellent structural soil – it dries into a very solid earth – ideal for rammed earth tires … IF you keep it dry.

Implications for Earthships

I can identify numerous implications of working with/in rich-clay-soil when it comes to Earthships:

  • Construction work – wet clay soil becomes a heavy muddy substance very difficult to get around in let alone to work with.
  • Rammed Earth Tires – if a tire is packed with clay earth that is not dry (enough?) then when it will dry it out the earth in it will get compacted some more. That could be a structural nightmare.
  • Drainage – clay rich soil has no drainage – it saturates with water and seals itself. Period. Though I haven’t seen this acknowledged in any written materials (off and online) my impression is that that is actually a welcome feature to one half of the water drainage problem of a house – surface water (once the soil is saturated) will simply flow away – so all you have to do is divert it to make sure it flows where you want it to go (preferably away from the house). The other half of the water drainage problem comes from below … and that problem isn’t unique to clay-rich-soil. It simply means that you need to have good drainage beneath the floor and around the house.
  • Structural Pressure – now we that we have figured out that the earth around the house may be pushing up against the house with tremendous force – something needs to be done about it.
  • Insulation – wet earth sucks warmth out of the house. The Earthship is bermed with earth all around and so to maintain energy efficiency any contact with wet earth must be avoided or mitigated.

Possible Solutions

Tire Walls Inherit Strength

In a typical underground house the forces of the soil would be acting directly on standard earth-proofed walls (usually concrete). The first main difference about Earthships is that the walls are massive … twice the width of typical walls. The tightly packed tires offer much more structural resistance then their counterpart typical walls. In addition we are planning to have the internal walls also be tire walls for additional mass and structural support.  In addition, outer Earthship tire walls are designed to lean back into the surrounding earth which offers even more lateral strength. 

Use Dry Clay Soil

This is easier said then done (at least in the Romanian climate which can rain any time and for any duration of time):

  • The building site would have to be setup with a large area which is arranged to dry soil.
  • This area would have to be sheltered from the rain.
  • It would also need to have exposure to sun and open-air circulation to promote drying.
  • It would need to have a large surface area so that a substantial quantity of soil can be dried sufficiently for use in both tires (slow and continuous consumption of soil over a long period of time) and backfilling (rapid consumption of soil in a very short period of time).

I suppose that if the excavated earth was placed in a narrow (north-south) and long (east-west) mound and that if that mound was covered with a slightly elevated clear plastic cover while enabling comfortable access for both wheelbarrows and a tractor – that soil could be reasonably dried!?

This also means that the tire walls themselves need to be kept dry throughout the project. Since Romanian weather includes rain-showers throughout most of the year (except of course in the subzero temperatures of winter) – this means that there need to be plenty of cover materials on site and a quick response when rain showers do appear.

Bring in Alternate Soil

Though it can become a substantial expense it is possible to bring in sandy, good draining soil for both ramming tired and backfilling. The supply of soil can be regulated as needed so it can be kept reasonably dry. Also, since it doesn’t suffer from expansion it would be OK to use it when moist knowing it will eventually dry out.

This is something I would prefer to avoid because (a) it is costly and (b) it goes against the core idea of using local materials for construction.


The design of the Global Model Earthship introduces a perimeter wall of insulation and moisture barrier set about 1 meter away from the outside of the tire walls. This creates two distinct backfill areas: (1) between the tire walls and the insulation; (2) outside the insulation and moisture barrier. The first backfill area between the tire wall and insulation/moisture barrier is a space that can and I believe should be completely covered by a moisture barrier. This means that this soil humidity is going to be relatively stable. If it is filled with mostly dry soil then it will also not change much, if it is filled with moist soil – then it may shrink as the envelope of the house dries over the first years of operation. Either way that part of the backfill is relatively stable and becomes and extends the fabric of the house. To my understanding it should absorb most of the additional pressures that come from the surrounding soil BECAUSE it isn’t structurally packed liked the tires – it is a more dynamic wall up against the more static tire-wall. Then there is the second – outer backfill – the one that is outside the moisture barrier. For this backfill I would prefer to use a good draining soil. It is a smaller volume of backfill and therefore less expensive to do so. It would serve two purposes. One is faster draining of any moisture that comes near the fabric of the house. The other is an additional pillow against the pressures of the surrounding earth. So already there is plenty of support against the potential pressure of the surrounding expansive clay soil.


I am thinking of starting the build by placing (on the excavated undisturbed soil) ~30cm of gravel (with built in drainage – see below). My thoughts are to excavate in such a way that the resulting surface will be slightly downhill (more elevated towards the back of the house. A level layer of gravel would then be placed on it. The gravel would cover the entire construction area up to and including the perimeter insulation wall and future tire walls (which will be built on the layer of gravel).

Common sense tells me that the gravel layer may also act as a flexible absorption layer should their be any excess pressure due to expansive soil from below.


On the gravel I would place a french-drain system made up of:

  1. A perimeter drain pipe.
  2. An inverted U drain pipe in every U module – with a T joint which leads out of the U and into the greenhouse/corridor (I feel it is better to avoid running any pipes under the tire walls). The  corridor connections would need to eventually pass through the stem wall of the inner corridor wall.
  3. Two main main drain pipes which collect flow from the U-drains and lead out to the two sides of the house and connect to the perimeter drain.

This would both remove excess moisture should it ever accumulate and create (as our architect suggested) a pleasant and mud-free work zone.

Living Roof

Our intentions are to install a living roof instead of a rain-collecting roof. The weight of the living roof as carried by the all the (inner and outer) structural tire walls is an additional counterweight to pressures from the surround soils. Most of the weight of the roof will be transferred down into the ground below the house. Some of the weight will be transferred to the side walls to do their outward leaning angle. This is unexpected and welcome benefit of the living roof.

Moisture Barriers

Moisture barriers are, I believe, a given in Earthship design (and any other well designed overground/underground house). Though in an Earthship I believe there are two aspects to this challenge. One is during construction (which in a self-build can take years) and the other is the typical finished house. My thinking is to start with the moisture barrier from within the tire walls – so it would be placed on top of the gravel. I think that a 4 meter wide would sheet would be enough to go from within the wall, underneath the tires & the the inner backfill and over the top of the first layer of insulation panels. Then each course of insulation would be covered by another overlapping sheet. In the end a top sheet of moisture barrier will extend from the roof and will overlap the top course of insulation sheathing. During construction a temporary cover will be needed to cover the breadth of the tire-walls + infill area + insulation panels. The floor area can remain uncovered as rain water will be diverted by the drains.


Though insulation is not directly related to the structural aspects of expansive soils it can effect the thermal performance of the house within these soils. Expansive soils hold a lot of moisture content and wet-earth can suck warmth from a house much more then dry earth.

Thorough insulation (as would be required in the Romanian climate), in my opinion, has not yet been achieved in Earthship designs (based on freely available information online). This is a testament to the fact that Earthships do not originate in cold and soggy climates and soils. Insulation was added in later Earthship designs and is now standard in the Global Model, but I believe it is still not up to the task of dealing efficiently with the Romanian climate. First, as designed in the Global Model, the insulation panels are better off protected from moisture – so it is sensible to install them within the moisture barrier sheath. In addition to that I would like to extend the insulation to close off additional energy bleeds from the house:

  • Floor insulation will be added throughout the house – above the gravel drainage layer and beneath the earth floor.
  • Floor insulation will also extend beneath the tire walls – it will be laid out around the perimeter and beneath the inside walls before tires are put in place and filled with dirt.
  • Floor insulation will also extend beneath the inner backfill area and through to the perimeter insulation panels.
  • The stem-wall for the inner corridor wall will also be insulated beneath ground level with R5 insulation panels to prevent energy bleed through the concrete.
  • Similarly the concrete footers for the front wall (living roof load bearing) posts will be insulated below ground.
  • Roof insulation will continue and meet the perimeter insulation panels using R10 panels (this insulation is closer to ground-level and therefore exposed to more sever ground-frost).
  • On the front face wall frost-blocking (45 degrees) insulation panels will be installed.

I am still debating what to do with the planters. I believe that the presence of composting soil and living plants and solar gain makes adding ground insulation in the planters redundant … we’ll see. Together with the bermed earth and living roof this should provide an effective shell of insulation that should prevent energy bleeds from the core of the house to the surrounding earth.


It seems to me that if properly dried soil (still no clear idea on how to achieve this) can be created and maintained on the work site, together with uncompromising moisture barriers and insulation should make it OK to build an Earthship in clay soils. I would be grateful to hear other opinions and other experience on this issue.


  1. Advice
    Posted July 17, 2012 at 4:36 am | Permalink

    I have 2 major issue with your design. First, all earthships need to be built on undisturbed ground and not on gravel. Second is the load on the insulation below the tire wall. Determine the weight of the wall and the crush or deformity rating on the insulation. Another type of thermal break might be better than the insulation at bearing the weight of the tire wall.

    • iamronen
      Posted July 22, 2012 at 4:58 pm | Permalink

      those questions did come up and I don't have clear answers …

      if I understand correctly then undisturbed soil mean soil without plant life residue … which does not necessarily exclude gravel.

      insulation would have to be special load-bearing insulation – supposedly such stuff exists to be used under foundation-footings so …

      BUT … we are now looking at a completely different direction that does away with these questions altogether:

  2. Posted March 16, 2014 at 11:49 am | Permalink

    I like your design.

    I will add a 6mm moisture barrier on the 45 degrees insulation panels in the front of the ES, a good compromise with the PAHS. Moreover, I read that PAHS is too expensive compared to the expected results.

    Insulation under foundation-footing is possible, I saw it in the ES located in the south France :

  3. oscar
    Posted March 16, 2014 at 6:07 pm | Permalink

    I do not know if I will understand, but what I build an earthship in sandy soils? have a problem?

Leave a Reply