What to expect
| Typical duration | Trial pits 1 day · foundation works 2–6 weeks |
|---|---|
| Cost band | Figures live in the cost index |
| Permit required | Yes, as part of the structural design |
| Seasonal constraint | Year-round |
What is under a Santorini plot
Reading from the surface down, most of the island presents some version of: loose ash and pumice from the eruption sequence, then welded tuff, then older layered pyroclastics, then pre-eruption bedrock. The thicknesses vary enormously, layers pinch out, and old ravines and quarry workings interrupt the sequence entirely.
The practical result is the thing owners find hardest to accept: your neighbour's foundation depth is not evidence about your plot. Two houses fifteen metres apart can need completely different solutions.
Why loose pumice is the problem
It is light, free-draining and extremely easy to excavate, which makes it feel benign. It is not.
- Poor bearing capacity. A spread footing on loose pumice fill will settle, and unevenly. The symptom is diagonal cracking at door and window openings a year or two after completion.
- Differential settlement. Because the depth of loose material varies across a plot, one part of a building settles more than another. This is the single most common foundation failure here.
- Trench instability. An open face stands up long enough to feel safe and then fails suddenly and in bulk. Shoring is a safety requirement, not a precaution. See excavation.
- Existing fill. Old terracing, quarry backfill and levelling material are common and are not always distinguishable from natural deposit without investigation.
What a trial pit tells you, and what it costs to skip
A trial pit is a day's work. It establishes the sequence, the depth to competent material, how it varies across the plot, and whether groundwater is present.
Skipping it means the foundation design is an assumption. The assumption is discovered to be wrong at excavation, which is after the permit has been issued on the strength of that design. Changing foundation type at that point means revised structural design, a permit revision, and a stopped site.
The options, once you know
| Ground condition | Typical approach | Cost implication |
|---|---|---|
| Competent tuff at or near the surface | Conventional pads or strips directly on it | Lowest — this is the good outcome |
| Shallow loose material over competent tuff | Dig through to the tuff; deeper excavation, standard footings | Moderate; excavation and spoil volume increase |
| Deep loose material | Ground improvement, a raft to spread the load, or deeper foundations to reach competent ground | Highest; this is what the trial pit exists to find early |
| Highly variable across the plot | Design for the worst case, or design to accommodate differential movement deliberately | Depends; the risk is doing neither |
| Old fill, quarry backfill or made ground | Investigate the extent, then remove or found below it | Variable and occasionally severe |
Founding next to what is already there
In a dense settlement, new foundations are frequently within a metre or two of an existing building whose own foundations are shallow, undocumented and sitting in the same loose material.
Where a party wall or shared structure is involved, that gets established before excavation rather than discovered during it. See demolition.
Questions people actually ask
How deep do foundations need to be here?
There is no island-wide answer, and that is the honest response rather than an evasion. It depends on what depth competent material sits at on your specific plot, and that is what the trial pit establishes.
Can I use my neighbour's ground investigation?
As background, usefully. As a basis for design, no. The deposits vary over short distances and old ravines and fill are common, so borrowing a neighbour's findings is exactly the assumption that produces a mid-build redesign.
My house has cracks above the doors. Is it the foundations?
It can be — diagonal cracking at openings is a classic differential settlement signature. It can also be thermal, or the junction between an original and a later addition. It warrants an engineer's opinion. See structural cracks.
Is a raft foundation the safe default?
Not automatically. A raft spreads load and can be the right answer over variable ground, and it is more expensive and not always necessary. Choosing it as insurance instead of investigating is spending money to avoid spending less money.
Does the ground on Santorini move because it is volcanic?
The relevant issues for foundations are bearing capacity, variability and settlement rather than volcanic activity itself. Seismic design is a separate and equally real consideration — see seismic design.