Geotechnical Engineering: What It Is and What It Covers

Geotechnical engineering is the branch of civil engineering that deals with the ground, the soil and rock that structures stand on or hold back. It investigates the ground under a site, predicts how it will respond to a project, and designs what depends on it: foundations, retaining walls, slopes, excavations and earthworks.

Steel and concrete are made to a specification, while the ground is found as it is, and two sites a few hundred metres apart can behave in opposite ways. The section below puts the field’s main problems on one site, over the ground that the effective stress example works through further down.

One site and the problems geotechnical engineering solves on it. The ground is the worked example's: 3 m of sand, 6 m of soft clay and dense sand below, with the water table 2 m down.

Geotechnical engineering, geology and soil mechanics

The three study the same material and are often confused. What separates them is the question each one asks:

Discipline	The question it answers	What it produces
Geology	What is this ground made of, how did it form and how is it arranged?	Geological maps and sections, and the history of the site: faults, old landslides, cavities
Soil mechanics (and rock mechanics)	How do soil and rock deform, fail and let water through?	The theory and the parameters: effective stress, shear strength, compressibility, permeability
Geotechnical engineering	What can be built on, in or with this ground, and how?	Investigations, designs and site control of foundations, walls, slopes, excavations and earthworks

Soil mechanics is the science, and geotechnical engineering applies it to design, together with rock mechanics and geology. Soil mechanics became a discipline in the 1920s with Karl Terzaghi, who formulated the principle of effective stress and the theory of consolidation and is often called its father. Engineering geology sits between the two: its geologists map and interpret the ground for engineering works such as dams and tunnels.

What a geotechnical engineer does

A geotechnical engineer turns an unknown piece of ground into numbers a designer can trust, and designs with them:

plans the site investigation: where to drill, how deep, and which samples and tests to take;
builds the ground model, with the strata, the water table and each layer’s design parameters;
writes the geotechnical report: foundation type, allowable bearing pressure or pile capacities, earth pressures, slope angles, compaction and groundwater control;
designs foundations, retaining walls, excavation support, slopes, embankments and ground improvement;
checks on site that the ground matches the design, from footing excavations and pile driving records to compaction tests and instruments such as piezometers and inclinometers.

Geotechnical engineers work for consultancies, contractors, testing laboratories and public agencies, and some specialize in dams, tunnels, offshore foundations, mining or earthquakes.

Soil as a material

Soil is mineral grains with water and air in the spaces between them. Its behavior depends on the size and minerals of the grains, on how tightly they are packed and, above all, on the water in the pores. Each property is measured in the laboratory, and each test has its lesson in the Academy and its free calculator.

Three phases: solids, water and air

A soil sample is solids plus voids, and the voids hold water, air or both. The phase relations are the ratios between those volumes and weights, and nearly every calculation starts from them: the void ratio $e = V_v/V_s$ , the porosity $n = V_v/V$ , the water content $w = W_w/W_s$ and the degree of saturation $S = V_w/V_v$ , which goes from 0 in a dry soil to 1 in a saturated one. They are tied together by

$S e = w G_s$

where $G_s$ is the specific gravity of the solids, about 2.65 for a quartz sand. The phase relations lesson derives these and the unit weights that follow from them, the phase relations calculator works them out from laboratory weights, and the specific gravity test gives $G_s$ .

Grain size and plasticity

Soils are classified first by grain size. Gravel and sand are coarse grained, and a sieve analysis gives their gradation curve; silt and clay are too fine for sieves, so the hydrometer analysis sizes them by how fast they settle in water. The boundary is the No. 200 sieve, 0.075 mm.

A fine soil’s behavior depends on its minerals and its water, which the Atterberg limits capture. The liquid limit $LL$ and the plastic limit $PL$ are the water contents at which the soil passes from liquid to plastic and from plastic to semi-solid, and the plasticity index $PI = LL - PL$ is the range over which it stays plastic. A high $PI$ usually means a clay whose volume changes with its water content and which compresses more under load.

The Unified Soil Classification System (USCS, ASTM D2487) turns both into a two-letter symbol, such as SP for a poorly graded sand or CL for a lean clay. When half or more of the soil by weight passes the No. 200 sieve, its letters come from Casagrande’s plasticity chart. These have free calculators too: sieve analysis, hydrometer analysis, Atterberg limits and USCS classification.

Effective stress

The principle of effective stress is the base of soil mechanics. Terzaghi stated it in 1936: every measurable effect of a change of stress on a saturated soil, whether compression, distortion or a change in its shear strength, is due to a change in the effective stress,

$\sigma' = \sigma - u$

The total stress $\sigma$ is the weight of everything above a point, and the pore pressure $u$ is the part the water carries. The grains carry what is left, $\sigma'$ , from contact to contact, and it is $\sigma'$ that gives soil its strength and makes it compress. With the water at rest, $u = \gamma_w (z - z_w)$ below a water table at depth $z_w$ .

Take the ground of the section above and point A, 11 m below the surface, where the piles end:

Depth	Stratum	Unit weight
0 to 2 m	Sand, above the water table	$\gamma$ = 17 kN/m³
2 to 3 m	Sand, below it	$\gamma_{sat}$ = 19 kN/m³
3 to 9 m	Soft clay	$\gamma_{sat}$ = 17 kN/m³
9 to 11 m	Dense sand	$\gamma_{sat}$ = 20 kN/m³

The total stress at A adds the weight of each stratum above it, and the pore pressure is that of 9 m of water:

$\sigma_v = 2(17) + 1(19) + 6(17) + 2(20) = 195 \text{ kPa}$

$u = 9.81 \times (11 - 2) = 88.29 \text{ kPa}$

$\sigma'_v = 195 - 88.29 = 106.71 \text{ kPa}$

Rischio’s engine gives the same 106.71 kPa: a driven pile 0.4 m square and 11 m long in this ground has $q' = \sigma'_v(L) = 106.71$ kPa at its toe, the stress its toe resistance is computed from. It also warns that the toe is only 2 m below the soft clay, closer than ten pile widths (4 m), so the clay may lower that resistance.

If a flood raises the water table to the surface, the top 2 m of sand saturate, the pore pressure becomes that of 11 m of water, and $\sigma'_v$ at A falls to 91.09 kPa, about 15% less. The strength of the sand falls with it, which is why so many slopes fail after heavy rain. Pumping does the opposite: the effective stress rises and the clay consolidates, and that is how pumping from the aquifer under Mexico City’s lake clays has made large parts of the city sink.

Shear strength

Soil fails in shear, with the grains sliding and rolling over each other along a surface. The strength along that surface follows the Mohr-Coulomb criterion,

$\tau_f = c' + \sigma' \tan\phi'$

with the friction angle $\phi'$ and the cohesion intercept $c'$ in effective stress. Since the strength grows with $\sigma'$ , water weakens soil: at the same depth, a sand below the water table is weaker than the same sand above it.

A saturated clay loaded quickly cannot drain, so the load raises its pore pressure and its undrained shear strength $c_u$ governs. Over months or years the water drains and $c'$ and $\phi'$ take over. A footing or an embankment on soft clay is usually most critical in the short term, and a cut slope in clay in the long term.

The direct shear test is the simplest way to measure $\phi'$ : specimens are sheared along a fixed plane under several normal stresses, and the direct shear calculator fits the Mohr-Coulomb line to their peaks. The triaxial test adds control over drainage and pore pressure, and in the field the vane test measures the $c_u$ of a soft clay in place.

Compaction

Fill for an embankment, a wall backfill or a slab subgrade is rolled in layers to drive the air out, so it ends up denser and stiffer and settles less. The Proctor test finds the optimum water content, at which a given compaction effort gives the highest dry density. Drier, the grains resist rearranging; wetter, water fills voids the grains could have filled. The modified test (ASTM D1557) applies about 4.5 times the energy of the standard one (ASTM D698), closer to modern rollers.

On site, the specification sets a minimum relative compaction, the field dry density as a percentage of the Proctor maximum. The soil compaction guide works an example, and the Proctor calculator draws the compaction curve from your trials.

Consolidation and settlement

Load a saturated clay and the water in its pores takes the load first, as excess pore pressure. The clay compresses only as that water drains out, which in a clay of low permeability takes months or decades: consolidation is that slow transfer of the load from the water to the grains. A sand drains as fast as it is loaded and settles at once.

The oedometer test (ASTM D2435) gives the compression index $C_c$ , the preconsolidation pressure $\sigma'_p$ , which is the highest effective stress the clay has carried and past which it is much softer, and the coefficient of consolidation $c_v$ , which sets the time. A foundation’s settlement adds an immediate part, primary consolidation and secondary compression, and our guide to foundation settlement works out each one.

Permeability

Water flows through soil according to Darcy’s law, $v = k\,i$ : the flow velocity is the coefficient of permeability $k$ times the hydraulic gradient $i$ . A clean gravel can be more than a million times more permeable than a clay, and that range decides how fast an excavation floods, how long a clay takes to consolidate, how much water seeps under a dam, and whether a soil behaves drained or undrained under a given load.

Site investigation

A design is only as good as the ground model behind it, and that model comes from the site investigation:

A desk study and a walkover: geological maps, old aerial photographs, nearby investigations, and signs of fill, springs or past movement.
Boreholes, drilled with augers or a rotary rig and logged on the way down, with disturbed samples for classification and undisturbed ones, from a thin-walled tube pushed into clay (ASTM D1587), for the strength and consolidation tests.
The standard penetration test (SPT, ASTM D1586): a 63.5 kg hammer falling 760 mm drives a split-spoon sampler 450 mm into the bottom of the borehole, and the blows for the last 300 mm are the $N$ value, corrected to 60% of the hammer’s energy as $N_{60}$ .
The cone penetration test (CPT, ASTM D5778): a cone pushed at a steady 20 mm/s records the tip resistance and the sleeve friction continuously, and the pore pressure with a piezocone (CPTu). It takes no sample, so it is paired with boreholes.
Test pits, piezometers for the groundwater, geophysics between the boreholes, and laboratory tests on the samples.

All of it ends up in the geotechnical report, and our guide on how to read a geotechnical report goes through one, from the boring log to an allowable bearing pressure.

The problems geotechnical engineering solves

Shallow foundations

A footing spreads the load of a column or a wall over enough ground to carry it. Its design checks that the soil under it will not fail in shear, its bearing capacity, from Terzaghi’s equation or the general bearing capacity equation, and that it will not settle more than the structure tolerates. Isolated, strip and combined footings suit most buildings, and a raft foundation carries the whole building when footings would cover most of the plan. The bearing capacity calculator solves a single footing, and the shallow foundation tool checks bearing, eccentricity and immediate settlement together. Whether a report’s allowable pressure is gross or net changes the size of the footing.

Deep foundations

When the ground near the surface is too weak or too compressible, or the loads too large, piles and drilled shafts carry the load down to firmer strata, by friction along the shaft and bearing at the toe. The alpha method gives the shaft friction in clay from $c_u$ and the beta method gives it in sand from the effective stress. Groups are checked pile by pile and as a block, and on real projects a load test confirms the calculation. The pile foundation design guide works a full example, the pile capacity calculator and the deep foundation tool run Rischio’s engine, and under a building already standing, the foundation is taken deeper by underpinning.

Retaining structures and earth pressure

A retaining wall holds ground at a slope it would not keep on its own. The pressure on the wall depends on how much it moves: at rest ( $K_0$ ) if it does not move, active ( $K_a$ ) once it yields away from the soil, and passive ( $K_p$ ) where it pushes into the soil. For a level backfill against a smooth vertical back, Rankine gives

$K_a = \tan^2\left(45° - \frac{\phi'}{2}\right)$

which is one third for $\phi' = 30°$ . Gravity, cantilever, gabion and MSE walls are then checked against sliding, overturning and bearing failure. Full examples are in the step-by-step retaining wall design and MSE wall design guides, and the lateral earth pressure and retaining wall calculators and the retaining wall tool run the checks.

Excavations and shoring

A basement or a deep trench is dug behind sheet piles, soldier piles and lagging, or secant pile or diaphragm walls, braced with struts or tied back with anchors. The supports carry an apparent earth pressure envelope, and the bottom of the cut is checked for heave in soft clay and, in sand below the water table, for upward seepage that can make it boil unless the water is lowered first. The shoring design guide works through a 6 m braced excavation.

Slopes

A slope fails when the shear stress along some surface inside it reaches the strength of the soil. Limit equilibrium methods (Bishop, Janbu, Spencer, Morgenstern-Price) try many trial slip surfaces, like the arc in the section, divide the mass above each into slices and compare the strength available with the strength needed for equilibrium. The lowest ratio over all the surfaces is the slope’s factor of safety, and for long-term steady seepage the US Army Corps of Engineers asks for 1.5 (EM 1110-2-1902). Water is the most common trigger: rain or a rising water table raises the pore pressure, and the effective stress and the strength drop with it.

Ground improvement and earthworks

Ground that is too weak for the project, and too costly to replace, is improved in place: densified by vibration or dynamic compaction, reinforced with stone columns, drained under a surcharge with vertical drains, mixed with cement or lime, or bridged with rigid inclusions. Our guides compare the ground improvement techniques and the soil stabilization methods. Earthworks, the cuts and fills that shape a site, depend on the compaction specification above.

Water: seepage, uplift and drainage

Seepage under a dam or into an excavation is drawn with flow nets or computed numerically. Where the upward gradient reaches the critical value $i_c = \gamma'/\gamma_w$ , the effective stress falls to zero and the soil boils; for the dense sand of the example, $i_c = (20 - 9.81)/9.81 = 1.04$ . A basement or a tank below the water table must weigh enough, or be anchored, to resist the uplift, and a drain behind a retaining wall keeps the water pressure off its back. At the surface, ditches and channels carry the runoff away, sized with Manning’s equation, which the Manning calculator and the open channel tool solve.

How designs are checked

Every check compares what the ground can give with what the project asks of it. Allowable stress design (ASD) divides the ultimate resistance by a global factor of safety and compares it with the service loads:

$\frac{R_u}{FS} \geq Q$

A footing on ground that would fail at 750 kPa may carry 250 kPa with $FS = 3$ . Common factors are 3 for the bearing capacity of footings, 1.5 against sliding and 2 against overturning of walls, 2 to 3 for piles depending on load testing, and 1.5 for long-term slopes.

Load and resistance factor design (LRFD) multiplies the loads by load factors above 1 and the nominal resistance by a resistance factor $\phi$ below 1, each set by how uncertain its load or its method is:

$\sum \gamma_i Q_i \leq \phi R_n$

In the United States, bridge foundations follow the AASHTO LRFD Bridge Design Specifications, while building foundations are still often sized with the allowable bearing pressure in the geotechnical report. Eurocode 7 applies partial factors to the actions, the soil parameters or the resistances. Either way, the design checks ultimate limit states, where something collapses, and serviceability limit states, where the structure settles, tilts or cracks more than its use allows.

Rischio’s tools design by factors of safety, with default criteria taken from FHWA guidance: 1.5 against sliding and 2.0 against overturning for walls, 3.0 against bearing failure for walls and footings, 3.0 for piles, and 25 mm of settlement for a footing. Below is the studio’s editor with a footing in it: change its size, its load or the soil and calculate, and the engine checks its bearing capacity, its eccentricity and its immediate settlement against those limits.

Try it: change a number, calculate

The tools of the trade

Most geotechnical design still starts with a hand calculation or a spreadsheet. Software takes over as the geometry, the strata or the load cases multiply: limit equilibrium programs for slopes, finite element programs for deformation and soil-structure interaction, and design programs for walls, foundations and piles. Our comparison of geotechnical software sets the main packages side by side.

Rischio is a design program of that last kind, in the browser, with four tools: retaining walls, shallow foundations, deep foundations and open channels. Each result shows the formula, the data, the assumptions and the references behind it, and the engine is checked against worked examples from recognized references, published as verified examples. Sign up free to model your own, or start with the free calculators, which need no account.

Frequently asked questions

How long does it take to become a geotechnical engineer?

In the United States the usual path is a four-year bachelor’s degree in civil engineering, often followed by a master’s in geotechnical engineering. Signing designs takes a Professional Engineer (PE) license: the Fundamentals of Engineering (FE) exam, usually four years of supervised experience, and the PE Civil exam, which has a geotechnical module. California also licenses geotechnical engineers separately, with an exam of its own. Counting the degree and the experience, a license comes eight years or more after starting university.

Is geotechnical engineering a good career?

The work follows construction and infrastructure everywhere, since every building, road, dam and tunnel rests on or in the ground. It mixes field work, laboratory work and design, and because no two sites are alike, judgment counts as much as calculation.

What is the difference between geology and geotechnical engineering?

Geology is a science that explains what the ground is made of and how it formed. Geotechnical engineering uses that knowledge, together with soil and rock mechanics, to decide what can be built there and how.

Who is the father of soil mechanics?

Karl Terzaghi, the Austrian engineer who formulated the principle of effective stress and the theory of consolidation in the 1920s. Soil mechanics as a discipline starts with his work.