A borehole log rarely fails because the ground is complicated. More often, it fails because the reader treats it as a neat vertical list of layers instead of a record of method, disturbance, uncertainty and local geology. If you want to know how to interpret borehole data well enough to support design decisions, you need to read beyond the stratigraphy column.
For geotechnical and tunnelling work, borehole data sits somewhere between observation and inference. It gives direct evidence from a very small footprint, then asks the engineer to extend that evidence laterally, sometimes across a slope, a station box or a tunnel alignment. That is where judgement matters. The practical task is not simply to read the log, but to decide what the log can support and what it cannot.
How to interpret borehole data in context
Start with the purpose of the investigation. A borehole drilled for shallow foundations, a cored hole for rock classification and a hole intended for groundwater observation will not deliver the same kind of confidence. The drilling method, sampling quality and logging standard all shape what the data means.
A common error is to treat all boreholes as equivalent because they share the same final depth and coordinate system. They are not. A hollow stem auger hole with disturbed samples tells a different story from rotary core drilling with good recovery. If the method has smeared soft clay, washed out fines or failed to recover weak rock, the apparent profile may be cleaner than the ground really is.
Before looking at engineering parameters, read the metadata. Check location, elevation datum, drilling date, method, casing, groundwater observations, sample types and any remarks on drilling progress. A sudden loss of flush, refusal, heave in the bore, or unstable walls can be as informative as the formal soil description.
Read the borehole log as a chain of evidence
The stratigraphy column is only one part of the record. A sound interpretation links several parallel observations – soil or rock description, sample depth, recovery, in situ test results, groundwater level and drilling comments.
If the text says “firm clay” but the vane results are low and the sample recovery is poor, pause. The descriptive term may reflect a brief tactile impression from a disturbed specimen rather than the in situ mass behaviour. Likewise, a sand layer shown as 0.3 m thick may represent a true lens, or it may reflect mixing during drilling at an interface.
Depth control deserves attention. Small errors accumulate quickly, especially where casing is advanced in stages or where sample intervals do not align cleanly with layer boundaries. When a key boundary controls excavation support or pile length, do not assume that the plotted depth is exact to the nearest centimetre.
Soil description and classification
Read the soil description with a practical eye. Terms such as clayey silt, silty clay, sandy till or gravelly sand are useful, but they are not design parameters by themselves. They help you infer likely drainage behaviour, compressibility and variability. They also tell you how much caution to apply when using laboratory data.
In glacial and post-glacial deposits, transitions can be abrupt or highly interbedded. A borehole may show a clean sequence on paper while the actual ground contains thin seams that matter for pore pressure response or face stability. That is one reason why experienced engineers look for consistency across neighbouring holes rather than relying on a single attractive log.
Samples, recovery and disturbance
Sample type often matters as much as sample value. Undisturbed tube samples in soft clay can support strength and compressibility assessment if quality is acceptable. Disturbed bag samples are useful for identification and index testing, but less so for direct strength interpretation.
Recovery figures should never be ignored. Poor recovery in very soft or very weak material may indicate washout, collapse or extreme disturbance. In rock, high total recovery with low RQD points to fractured ground. Low recovery with drilling difficulties may indicate weathered zones, crushed rock or open joints. The numbers are only meaningful when read together.
Interpreting in situ test data
Many boreholes include SPT, vane, CPT tie-in levels or packer information nearby. These results should be used to challenge the descriptive log, not merely decorate it.
SPT values need caution. They are influenced by equipment, energy transfer, borehole condition and operator practice. An N-value can still be useful for broad trends in density or consistency, but only if you know how the test was carried out and whether corrections are needed. Comparing raw values from different investigations without checking method is an easy way to create false precision.
Field vane results in clay are often more directly useful, but even here rate effects, rod friction and sample disturbance can mislead. The remoulded strength may be as valuable as the undrained peak result if sensitivity affects excavation or stability.
Where boreholes are interpreted alongside CPT, the combination is usually stronger than either method alone. CPT gives continuity, boreholes give material confirmation. If they disagree, do not force them into artificial alignment. Ask why. The answer may reveal interbedding, gravel content, fissuring or a logging issue.
Groundwater is not a side note
Groundwater entries on borehole logs are frequently over-trusted. A measured water level at the end of drilling may reflect drilling disturbance, incomplete equilibration or a temporary perched condition. In low-permeability soils, the true piezometric level may take time to establish. In fissured rock, one observed level may hide several hydraulic domains.
For excavation, tunnelling and grouting work, this distinction matters. The question is not just “where was water seen?” but “what hydraulic condition does this observation represent?” If groundwater level controls uplift, inflow or effective stress, use standpipes, piezometers or repeated monitoring where possible.
Signs such as artesian response, loss of water during drilling, or wet seams at specific depths can indicate far more than the final recorded groundwater level. Those details often carry the risk picture.
How to interpret borehole data for design
Interpretation becomes engineering when you convert observations into a ground model. That means defining units with expected behaviour, not merely copying the log descriptions into a section drawing.
A useful ground model groups materials by design relevance. Two clay layers with slightly different colours may behave similarly for settlement. A thin silt seam within clay may deserve separate treatment if it controls drainage or stability. In rock, the key division may be fresh versus weathered, or massive versus fractured, rather than lithology alone.
This is where “it depends” is not evasive but necessary. The level of detail should match the decision being made. For a preliminary alignment study, broad units may be enough. For a deep shaft, tunnel face or compensation claim, much tighter interpretation is needed, including uncertainty ranges.
Avoid extracting characteristic values directly from isolated test numbers. Borehole data supports parameter selection, but it does not remove engineering judgement. Outliers may represent error, or they may represent the critical weak zone that governs performance. The right response depends on geology, spacing and consequence of failure.
Correlation across multiple boreholes
No borehole should be interpreted alone if nearby data exists. Correlation is where local anomalies become either noise or a real feature. Look for repeatable boundaries, recurring weak zones, groundwater patterns and changes linked to topography or geomorphology.
Cross-sections are especially useful when kept honest. Do not draw smooth continuous horizons where the evidence is sparse. Sometimes the correct interpretation is a probable range or an uncertain lens. Straightforward presentation of uncertainty is better engineering than an elegant but unsupported profile.
For engineers working digitally across desktop and mobile workflows, the value is speed with traceability. A good interpretation process lets you move from field observations to sections, checks and revised assumptions without losing sight of the original log data. That practical clarity is one reason specialised tools matter more than generic drafting software.
Common mistakes when reading borehole data
The most common mistake is taking the wording of the log as final truth. The second is ignoring how the hole was drilled. The third is treating a single groundwater reading as a settled hydrogeological model.
Another frequent problem is over-correlation. Engineers naturally want continuity, but the ground does not owe us neat geometry. Thin granular seams pinch out. Weathering fronts undulate. Fracture zones split and rejoin. If the model looks too tidy, check whether interpretation has overtaken evidence.
A final point is scale. Borehole data is detailed vertically and sparse horizontally. That imbalance is easy to forget during design. Confidence near the hole may be high, while confidence between holes may be modest. Good interpretation keeps both truths visible.
Borehole data is at its most useful when read with discipline and a little scepticism. The log is not the ground. It is a carefully collected sample of evidence from which the engineer builds a defensible model. Read it that way, and the data becomes far more valuable than a stack of plotted layers.
