Saturday 19 January 2013

Earth Sheltered Insulation - Southern Europe

From my original research it emerged that the insulation strategy for earth sheltered housing needs to be adjusted as one moved further south from the well documented examples I quoted in my last post.

At the depths earth sheltered houses are set the earth temperatures in southern regions will be warmer in both summer and winter. With the higher earth temperatures dew point in the form of sweating is less likely. Also the further south we go the more important cooling the house becomes in terms of both comfort and energy cost. By lowering the amount of insulation the moderating effect of the earth's temperature performs more effectively as a means of natural cooling.

In his classic book "Earth Sheltered Homes" Rob Roy states that:

"In areas with climates of 4000 - 7000 heating degree days insulate with 2" (R10) of extruded polystyrene down to footings, 1' adjacent to and under the footings and 1" under the floor for a distance of 4 feet in from the footings. Leave insulation out from under the rest of the floor to promote natural cooling.

In areas of less than 4000 heating degree days insulate with an inch of extruded polystyrene down to the footings and leave it out around the footings and floor."

As a reference point the PVGIS Estimates of long-term monthly averages for my location suggested that I should expect an average of 1600 heating degree days.

This variation in insulation requirements necessitated  further consideration.

Researching further I found a recent paper (18th September 2012) by Christos Viskaourakis (1) very enlightening.

He starts by considering soil and notes that:

"According to Golany (2), the thermal conductivity and retention of soil is affected mostly by its density, the size and shape of its particles and by its moisture content along with the site's climatic conditions.

Soil temperature itself is related mostly to two parameters, its thermal conductivity and the depth. Heat moves slowly through soil and as a consequence, temperature swings (daily or annually are limited as we go deeper within it (2,3). As Hait (3) argued, the soil temperature remains stable throughout the year at a depth of 6 metres and this value represents the average annual soil temperature. In addition, Givoni (4) claimed that the values on average annual air temperature and soil temperature could be considered as the same. These arguments are correlated, as heat requires an estimated period of six months to be transmitted from air to 6 metres deep into the soil.

Soil temperature is also affected by the nature of its surface. As Jacovides et al (5) proved, there can be differences on the annual pattern of soil temperatures between bare and short grass covered soil surfaces. According to Mihalakakou et al. (6), the average annual soil temperature of bare and short grass covered soil in Athens is 20.9 degrees celcius and 17 degrees celcius respectively."

He further goes on to note that Givoni (4) created an equation based on several experimental studies to predict the soil temperature at any depth and at any day of the year (see his paper for mathematical details) and by using Mihalakakou's et al. findings through Givoni's equation he produced the pattern of annual soil temperature at 3 metres depth.

The depth of 3 metres was chosen as a representative depth into the soil, as a practical depth for building an earth sheltered house and where the temperature there would influence the structure. Obviously the grass covered soil creates lower temperatures throughout the year but more interesting is that, when compared with climate data the average soil temperatures at 3 metres lag behind the average air temperatures by approximately 3 months."

Viskakourakis (1) went on to note that:

"Golany (2) stated that soil is considered, incorrectly, as an effective insulator for earth sheltered structures. The thermal conductivity of soil is approximately 25 times greater than the recent insulation materials on the market. In other words, soil should be considered as an effective thermal moderator and not insulator (3)

From further reading I found that another classic earth-sheltered book ("Earth Sheltered Housing Design: Guidelines, Examples and References," by the Underground Space Centre University of Minnesota 1979) (7) stated that the question of how much insulation and where it should be placed is an important consideration for earth sheltered houses. It seems that the position of the insulation in relation to the structure as well as the sizing of the insulation itself have some impacts on the inside air temperature of the structure.

Basically the authors used some computer models to predict energy flow and usage of several earth-sheltered and underground designs. In their discussion of walls and floors, the authors argue that insulating the entire wall/floor structure defeats a main advantage of earth-sheltered housing: the great thermal mass of the surrounding soil which moderates both cold winter temperatures and warm summer temperatures. After some site experimental studies and computer simulations that took place in 1987 they claimed that it would be more energy efficient to leave the lower part of the walls uninsulated. In that way a thermal balance between the soil and the air temperature of the space will be achieved and as a result heat losses through the walls are limited. Soil stores heat proportionately to its depth, consequently the temperature range at the depth of the uninsulated wall is narrow daily or annually. As a result the structure takes advantage of the moderate soil temperature at the depth of the uninsulated walls.

The conductive path between the surface and the roof or the upper part of the walls is much shorter compared with the lower part of the walls or the floor. Therefore, these are much more sensitive to variations of climate conditions, hence they should be insulated.

As for floors their model predicted that 2.5 cm of insulation under the flooring would reduce winter heat loss through the floor only 5% (i.e. a 1% decrease in the total building heat loss). Adding the same 2.5 cm in the roof, however, resulted in a 20% decrease in ceiling heat loss (i.e. a 11% decrease in the total building heat loss). Thus it made more sense to insulate the roof rather than the floor. This conclusion was reinforced by their model for summer use, where an insulated floor would significantly reduce the effective mass of the floor as an aid to cooling.

Specifically they compared three insulation schemes: one where insulation is run down the entire length of the 3 metre concrete wall (on the outside next to the soil); one where insulation was only run on the 1.5 metre of wall closest to the surface and the last where there was no wall insulation but the roof insulation was extended, say, 1.5 metres all round. This research was based on the fundamental principle of keeping the amount of insulation stable throughout these options.

Using their models they found that although the full insulation method saved 5% in winter heat losses, that came at the sacrifice of a 10% improvement in summer cooling for the partial insulation method. However it has been proved that the option that extends the roof insulation has the potential to win this unofficial competition, although it was not clearly stated. (7) Furthermore this option would provide efficient shedding of rainwater from the earth immediately adjacent to the house walls, hence it would increase the thermal performance of the house, since the moisture content of soil can significantly reduce its thermal storage capacity and its temperature.

This option of extending the roof insulation some distance from the walls can be characterised as the forerunner of the Passive Annual Heat Storage (PAHS) technique developed by John Hait (3). According to Hait, PAHS is based on the principle that thermal properties of the earth are cumulative. In other words, the longer the conductive path or depth, the greater the temperature moderation inside the soil. Since it is not practicable to create a house 6 metres underground where the temperature is always stable and close to the comfort band. As Hait claims the PEHS technique artificially arranges a 6 metre conductive path.

The installation of an insulator inside a conductor like the earth will force heat to flow around the insulation. In addition the small cumulative resistance of the earth will cause thermal lag. Six months of thermal lag between the soil surface and the walls of the structure can, therefore, be created so that the structure takes advantage of the summer "heat" during the winter period and winter "coolth" during the summer period.

Whilst the PAHS approach still falls under the general passive solar earth sheltering concepts of energy efficient construction it specifically uses the earth surrounding the house to store solar heat by treating this dirt as part of the dwelling's thermal mass by insulating it from the elements but NOT from the walls.

The technique calls for a specially designed cap, called an insulation/watershed umbrella, that is placed a few feet ABOVE an underground building's roof (not against it), extending outward to isolate the earth around the structure from the temperature fluctuations of surface layers.

As with other passive solar earth sheltering, windows on the south side of the dwelling let sunshine in to heat all the mass within the insulating umbrella. Slowly, ever so slowly over the whole year, a balance is achieved between the warmth of the summer sun and winter heat loss Thus, an artificial average annual air temperature is established at the junction of the house's walls and the earth. Prevailing temperatures inside the building will be transmitted through the walls and into the earth, extending to a radius of at least 6m from the structure. By controlling the amount of sunshine let into the house and the amount of heat rejected, by shading and ventilation, it is possible to adjust the temperature of the surrounding soil.

Because of the tremendous mass of the building and surrounding soil for the 6 metres around and below the building the interior temperature should only vary a few degrees throughout the year.

Earth sheltered homes have long been known to have slowly changing temperatures that are largely controlled by the earth around them. The average of this flux is often referred to as the "floating" temperature. Many designers at first assumed that an earth sheltered house would take on the natural soil temperature, but experience has shown that this is not the case. Even a conventional underground house modifies the temperature of the earth around its walls because occupiers add heat to the building (and therefore to the earth around it) for comfort. The result is an adjusted floating temperature and PAHS's aim is to get that temperature into the comfort zone.

An entire year's worth of heating cooling (say, 3 - 5 million BTU) can be contained in an area that extends about 6 metres from the walls of the house. Furthermore, over this distance the accumulated resistance to heat flow (R-factor) is sufficient to block 90% of the loss.

The insulation/watershed umbrella extending 6 metres from the walls is only sufficient, however, if the earth under the umbrella is DRY. Although damp earth has greater heat capacity than dry earth, the greater conductivity of water allows too much heat to escape the confines of the insulated cap. Since it is inadvisable to build any earth sheltered home where there is a high water table that same restriction applies to a PAHS dwelling. It is however also important to protect the earth within the insulating umbrella from surface water hence layers of insulation in the cap should be interspersed with water barriers to shunt liquid down the upper surface if the umbrella to a suitable drainage system.

According to Hait, the PAHS technique described separates the surrounding earth into thee zones:

- the moderation zone,
- the isolation zone,
- the storage zone.

The moderation zone is only a sufficient layer of soil above the insulation umbrella which isolates the storage mass below and protects this insulation layer. He further suggests that its depth be about 0.5 metres otherwise construction costs become excessive. The isolation zone replacing the normal insulation layer which would wrap the entirely around the thermal mass of the walls. Heat therefore conducts through the isolation zone before it reaches the storage zone and finally the walls. Conventional earth sheltered houses lack the storage zone as they are built with only a moderation zone and an isolation zone and rely on the high thermal mass of the structure alone to store the seasonal "heat" and "coolth".

Viskadourakis(1) uses ECOTECT Analysis to compare a number of earth sheltered insulation techniques for a 3 dimensional chamber atrium model under the same structure type and climate. These insulation techniques were tested under bare and short grass covered soil and soil temperatures for both surfaces were calculated at a depth of 3 metres. Typically polystyrene was used as the insulation material in all cases. He noted that, although Hait specified some particular requirements and optimised the shape of the insulation in PAHS, he did it mostly for low cost and waterproofing reasons and as a result his technique was modelled in a simplified way. i.e. with a 20cm earth layer between the concrete roof and the insulation layer with the roof insulation layer projecting at least 4 metres beyond the non insulated walls.

The effectiveness of each technique was measured according to the simulated air temperature of the "house". Since the research was based on free running structures Humphrey's equation for adaptive comfort temperatures was used. i.e.

Th = 0.534* T + 11.9         where:     Th = the monthly comfort temperature
                                                               T = the average monthly outside air temperature.

Each monthly average inside air temperature, from the simulation, was compared with the corresponding monthly comfort temperature and the difference calculated. A percentage monthly failure rate compared with the monthly comfort temperature and the modulus of the results used to calculate an annual average monthly failure rate for each technique.

The results of Viskadourakis' analysis shows, as expected, the temperatures for the grass covered soil are much lower than those for the bare earth. See charts below.

He states that although each technique has its benefits and drawbacks, it appears that the most successful technique on bare soil (which would best approximate to the earth sheltering I envisage for my house - with a few mountain herbs and shrubs to blend with the natural vegetation) is the one with insulation only on the top half of the walls with an annual average failure rate of 5.84%.

 In contrast he found that PAHS was the most unsuccessful and would produce the most uncomfortable conditions in the house. However, my first reaction (after many years of examining graphs and data sets) on seeing his charts and making a few quick calculations was that a shift by one month (ie. increase the thermal lag by one month) in the inside air temperature would produce an excellent result for the PAHS technique with an annual average failure rate of only 2.86%. Viskadourakis (1) bears this out and states that "the PAHS technique would have the potential to be the most successful for the baresoil of Athens, if another forced time lag was arranged properly." He later states that in general it appears that the bare soil options produce the better conditions in an earth sheltered house in Athens throughout the year.


My conclusion from the above is that I need to adopt a different insulation technique to the wrap everything in insulation approach  I had in my mind( based upon my reading that had centred around examples in more northern climes both in the UK and in the USA).

My instinct is that a PAHS approach offers the most exciting possibilities, especially if I am able to extend the waterproof/insulation "umbrella" by a metre or so as suggested above. However there are serious physical limitations with achieving the ideal "umbrella" on my particular site as I am cutting into the hillside to "bury" the house in order to not only create the earth sheltering but also to blend it, unobtrusively, into the natural habitat of the site.

Over the next couple of months I will redo all the calculations using as much local data as I can, and making reasoned estimates where none is available and investigate further how the PAHS approach could best be applied in my situation.


  (1) Christos Viskaourakis - "The Effect of Six Distinct Insulation Techniques on Earth Sheltered Houses: a Sensitivity Analysis" (2012)

(2) Golony S. Gideon (1983) - "Earth-Sheltered Habitat"

(3) Hait John (1983) "Passive Annual Heat Storage: Improving the Design of Earth Shelters"

(4) Givoni Baruch (1994) "Passive and Low Energy Cooling of Buildings".

(5) Jacovides C.P. et al (1996) "On The Ground Temperature Profile For Passive Cooling          Applications In Buildings".

(6) Mihalakakou G. et al (1997) "On The Applicaton of The Energy Balance Equation To     Predict Ground Temperature Profiles".

(7) The underground space centre (University of Minnesota) (1979). "Earth sheltered housing design".