A site investigation reports competent ground beneath most of a proposed bridge. Does that mean the foundation is low risk? Not necessarily: untested ground, changes in groundwater, construction staging and neighbouring structures can create consequences the available boreholes have not resolved. Geotechnical risk assessment asks what we know, what remains uncertain and which decisions need further evidence.
The gold standard of geotechnical risk assessment connects geological models, failure mechanisms, investigation, quantified and qualitative uncertainty, engineering controls and accountable monitoring. This guide covers twenty-four chapters, five original worked scenarios, FAQs and traceable engineering references.
Part of The Gold Standard. This article is general educational material, not professional design approval, a site risk certificate, engineering instructions or advice to proceed with unsafe ground works.
Quick information: six questions a sound ground-risk review answers
- Which ground failure or movement mechanisms could affect people or assets?
- Which geological and groundwater facts are observed and which are inferred?
- Which uncertainties would change design, cost or safety decisions?
- How will the project prevent, reduce or monitor relevant risk?
- Who has authority to review new evidence and act?
- What residual uncertainty remains after controls are implemented?
Chapter index
- 1. Begin with a consequential ground decision
- 2. Treat risk and uncertainty as different but connected
- 3. Construct the geological and hydrogeological model
- 4. Describe failure modes before calculating scores
- 5. Use desk study and historical evidence responsibly
- 6. Design investigations around decisions, not borehole counts
- 7. Distinguish representative parameters from single tests
- 8. Include water as a dynamic design condition
- 9. Separate ultimate limit states from serviceability
- 10. Consider slopes, retaining structures and excavations independently
- 11. Assess ground improvement claims with verification
- 12. Identify seismic and liquefaction conditions
- 13. Use risk registers that retain traceability
- 14. Avoid pretending a likelihood category is a measured probability
- 15. Compare mitigation by its effect on consequences
- 16. Use staged and observational approaches carefully
- 17. Define alert thresholds with actual engineering authority
- 18. Investigate adjacent assets and cumulative effects
- 19. Manage procurement and differing site conditions
- 20. Measure risk reduction rather than activity volume
- 21. Report uncertainty in model outputs and forecasts
- 22. Govern construction changes and handovers
- 23. Transfer the method to the next ground problem
1. Begin with a consequential ground decision
Geotechnical risk assessment investigates uncertain ground conditions and how they could affect safety, serviceability, schedule, cost and neighbouring assets. It is not simply a catalogue of dangerous soil types. A retaining wall might have strong laboratory material data yet face uncertain groundwater behaviour; a highway might cross mostly stable ground with one poorly characterised landslide.
The FHWA Geotechnical Technical Guidance Manual addresses risk within investigation, design, construction and monitoring. Begin with what can go wrong, who would be harmed and which decision needs evidence. Distinguish physical failure mechanisms from contractual risks caused by incomplete information. The gold standard identifies what is known, what is uncertain and who has authority to accept or reduce each consequential risk.
2. Treat risk and uncertainty as different but connected
Uncertainty concerns incomplete knowledge about strata, loading, water, construction and system response. Risk concerns potential consequences of uncertain outcomes under a defined objective. A rare failure with severe human consequences may demand attention even when its probability is uncertain; a frequent nuisance may be manageable without extraordinary intervention.
ISO 31000:2018, confirmed current by ISO, provides broad guidelines for identifying, analysing, evaluating, treating, monitoring and communicating risk. It is not a certifiable declaration of safety. In geotechnical work, first identify plausible mechanisms, exposures and consequences, then evaluate the credibility of available evidence. A number in a risk matrix is not proof the underlying probability was measured accurately.
3. Construct the geological and hydrogeological model
A site model describes rock and soil units, faults, weathering, groundwater, relevant structures and historical changes. It should include alternative interpretations where boreholes or exposures are sparse. A model that assumes one continuous layer between two boreholes may be contradicted by an old river channel or buried fill between them.
Groundwater can alter pore pressure, effective stress and excavation behaviour, while soil and rock heterogeneity affect foundations and slopes. Document direct observations separately from extrapolation. A strong risk assessment identifies which uncertainties change the engineering decision and where more investigation could reasonably reduce them. Do not present a smooth cross-section as a proven underground photograph.
4. Describe failure modes before calculating scores
Different hazards require different models: slope instability, excessive settlement, bearing-capacity failure, ground heave, erosion, liquefaction, subsidence and unforeseen obstruction are not interchangeable. Each has potential triggers, exposed assets and warning signs. An excavation near an existing building can create serviceability concerns even without ultimate collapse.
Write each credible failure mode as a chain from ground condition to possible consequence. Avoid vague entries such as ‘bad soil risk’ that cannot guide investigation or mitigation. The FHWA guidance enumerates multiple failure mechanisms and emphasises designing for relevant consequences. A useful risk register should tell a qualified team what to investigate, not merely which colour to assign.
5. Use desk study and historical evidence responsibly
Previous geological maps, site records, aerial imagery, land-use history, old boreholes and utility plans can reveal information inaccessible in a single new test. But old plans may have different reference systems, uncertain dates or coverage limited to another project. Reusing a historical ground report without reviewing its purpose can create false reassurance.
BS 5930:2015+A1:2020 addresses ground investigation as a staged practice in the UK context. Check original reports against present use, current conditions and applicable local standards. A former watercourse or quarry may matter to a new foundation design even when its surface appearance has been erased by construction. Preserve the provenance and limits of the desk-study evidence.
6. Design investigations around decisions, not borehole counts
An investigation aims to reduce specific uncertainties. A large number of shallow boreholes in an accessible area may still miss a deeper weak stratum beneath a critical support or a fracture connected to groundwater. The right scope depends on design consequences, variability, existing data and the physical model.
Define what each test is expected to reveal and how the result could affect design or construction. Sample placement, depth, quality, representativeness and lab methods matter as much as quantity. Do not assert that a standard minimum number of boreholes guarantees safety. Qualified engineers must apply the relevant codes and local requirements for actual investigations.
7. Distinguish representative parameters from single tests
Material properties vary spatially and depend on testing, moisture, stress history and disturbance. Laboratory strength tests, in-situ penetration resistance and geophysical measurements describe different quantities. Selecting the most favourable result can conceal a critical weak layer; selecting the worst isolated result without context can create an uneconomic and still poorly understood design.
Use defensible characteristic or representative values under the applicable design framework, with uncertainty and sensitivity. A sample’s reported decimal precision is not equal to confidence in the entire ground volume. State whether a parameter represents a layer, location, orientation or specific operating state. Keep the original measured data available for review.
8. Include water as a dynamic design condition
Groundwater heads may respond to season, rain, nearby pumping, tidal influence, dewatering and construction. A water level observed once in a borehole cannot automatically establish the design head across decades. Pore pressure can influence effective stress and stability, while water flow may create erosion or uplift hazards.
Monitor where relevant and keep reference elevations and screen intervals consistent. Explain planned excavation, drainage and pumping assumptions. A change in groundwater conditions can create new risk to neighbouring foundations or ecosystems even if the primary structure remains stable. Groundwater consequences require qualified hydrogeological assessment and appropriate permissions.
9. Separate ultimate limit states from serviceability
A geotechnical design can satisfy an ultimate resistance criterion and still produce movements that affect a sensitive adjacent building, buried service or railway. Excessive settlement, differential movement and tilt need their own measurable performance objectives. Conversely, small visible cracking does not by itself prove imminent ultimate failure without context.
Define the structures and tolerable performance under the relevant professional standards. Distinguish imposed loading, soil–structure interaction and uncertainties about deformation over time. A design review should not replace a serviceability question with a single factor-of-safety claim. Monitoring must track the quantities relevant to each consequence.
10. Consider slopes, retaining structures and excavations independently
Slope failure may involve rock discontinuities, weak soil layers, rainfall infiltration or erosion. Retaining systems add structural components, staged excavation and interaction with neighbouring ground. A common slope-stability diagram does not address every relevant wall, anchor or adjacent asset failure mechanism.
Identify construction stages, potential movement paths, water conditions and the people or assets exposed. Analyse credible alternatives rather than choosing a method merely because software provides it. A numerical result should be checked against the conceptual model and field evidence. Safety design and temporary works must remain under competent professional control, never dictated by a general educational article.
11. Assess ground improvement claims with verification
Ground improvement may seek to change density, strength, permeability or settlement behaviour in certain ground. Its effectiveness is site- and method-specific. A subcontractor’s completion certificate or number of treatment points does not automatically prove that the intended ground response was achieved.
Define the required properties and suitable verification tests in the professionally approved specification. Compare treated and untreated evidence appropriately and inspect areas where delivery could have been incomplete. Consider environmental or adjacent-structure effects. The risk assessment should state what remains uncertain after treatment and what observations could reveal an unsuccessful zone. Counting installed improvements is a process metric, not a performance guarantee.
12. Identify seismic and liquefaction conditions
Earthquake-related ground risks can include shaking, settlement, lateral spreading and liquefaction under suitable soil, groundwater and seismic conditions. The event magnitude alone does not determine site response. Soil state, depth, cyclic demand and structural vulnerability matter.
Use the relevant jurisdiction’s hazard and design framework and qualified geotechnical earthquake expertise. A screening map can indicate need for investigation but cannot substitute for site-specific design. Distinguish seismic hazard, susceptibility and the actual consequence to a foundation or lifeline. Avoid claiming a site is safe because its historical record lacks a damaging earthquake.
13. Use risk registers that retain traceability
A risk register should name the failure mechanism, relevant locations, cause, evidence, consequence, owner, proposed control, review date and residual uncertainty. Numbers for probability and severity can help prioritise discussion, but their definitions must be explicit. Identical nominal risk scores may conceal vastly different consequences.
The National Academies’ NCHRP Research Report 884 discusses geotechnical risk-management strategies and tools for design–build projects, including reducing uncertainty before contracts and allocating remaining risks appropriately. The register should support decisions and accountability, not become a spreadsheet whose scores never change when new ground evidence arrives.
14. Avoid pretending a likelihood category is a measured probability
A team may label slope failure ‘unlikely’ or ‘possible’ based on qualitative judgement. Those terms do not automatically correspond to a defensible annual event probability. Converting ordinal category numbers into exact expected losses can create numerical confidence unsupported by evidence.
State whether probabilities come from statistical records, mechanistic modelling, expert elicitation or simple relative ranking. Document the time horizon and scenario. Where the uncertainty is too deep for a meaningful probability, use scenario-based consequence analysis and decisions robust across plausible conditions. Severe safety consequences require professional and regulatory attention independent of a low-colour matrix cell.
15. Compare mitigation by its effect on consequences
Risk treatment can involve avoiding the location, modifying design, collecting better information, staged works, monitoring or other qualified controls. A control may reduce event probability, limit consequences or improve detection. Merely transferring contractual liability does not reduce physical danger to workers or neighbours.
Compare options against their actual objective and residual uncertainty. A new monitoring sensor may improve awareness but not strengthen an inadequate foundation. A design change might reduce the need for monitoring or introduce a new exposure elsewhere. The register should record which mechanism each control affects and what evidence will verify its performance.
16. Use staged and observational approaches carefully
Some projects can manage uncertainty through staged construction combined with predictions, measurements and preplanned responses. This is not permission to discover unsafe limits by trial and error. The method requires an adequate safe design basis, credible prediction ranges, qualified monitoring, action authority and contingency plans.
Link each stage to relevant measurements and decisions. Establish what would trigger review, change or work stoppage under the project’s approved safety system. A numerical alert without staffed interpretation and actionable contingency is not a sufficient risk control. The observational approach must be compatible with the legal and engineering requirements of the specific project.
17. Define alert thresholds with actual engineering authority
Generic movement or pore-pressure thresholds cannot be copied safely between walls, tunnels, slopes and neighbouring buildings. The significance of displacement depends on location, rate, direction, material, loading and exposed structures. An apparently small movement can be concerning in one context while a larger movement is anticipated in another.
Qualified professionals establish site-specific alert levels and the required response. Monitoring reports should show baseline, dates, quality flags, comparison to predictions and review decisions. An automated notification can support attention but cannot replace competent interpretation. Avoid describing an educational worked number as a real stop-work criterion.
18. Investigate adjacent assets and cumulative effects
An excavation may affect neighbouring foundations, utilities, roads, trees, groundwater and other works. Several separate projects can interact through groundwater drawdown or changed loading. A narrow analysis limited to the owner’s foundation may omit serious off-site consequences.
Identify affected parties and relevant historic or ongoing activities. Compare alternative stages and construction methods within professional and legal frameworks. Keep communications and monitoring responsibilities clear across owners and contractors. Risk allocation in a contract should not be confused with scientific evidence that the hazard has been controlled. A robust assessment includes consequences beyond the immediate worksite boundary.
19. Manage procurement and differing site conditions
A tender can be affected by incomplete ground information, ambiguous assumptions and different interpretations of subsurface uncertainty. Design–build contracts may allocate risk through defined provisions, but no contract can make the physical ground more predictable. Suppressing known uncertainties may increase disputes and unsafe pressure to proceed.
The NCHRP guidance encourages early reduction of geotechnical uncertainty and clearer allocation of remaining risks. Use a transparent ground baseline and reporting process appropriate to the procurement model. Distinguish contractual compensation triggers from safety action levels. Legally binding terms require qualified contractual and legal review, not generic guidance from a subject primer.
20. Measure risk reduction rather than activity volume
More boreholes, meetings, dashboard charts and installed sensors may indicate effort without establishing lower risk. A successful investigation can reveal a previously unknown hazard and increase its recorded severity while improving the ability to manage it. Counting fewer risk entries after a review can likewise reflect omission rather than safer ground.
Choose measures linked to the decision: uncertainty reduced, failure modes addressed, safe operating conditions verified and contingency available. Preserve original and revised assessments so changes are explainable. Audit how information affected design choices. A risk programme should reward the discovery of a serious unknown before construction, not penalise teams for revealing inconvenient evidence.
21. Report uncertainty in model outputs and forecasts
Slope, settlement and groundwater models depend on uncertain inputs and sometimes alternative conceptual structures. Solver convergence or a low calibration error does not establish that the chosen model is physically correct. Several plausible parameter sets can fit limited measurements and produce different future outcomes.
Test sensitivity to material properties, water levels, loading and model boundaries. Compare predictions with independent observations and explain where results diverge. A figure rounded to two decimals may not deserve that precision in the site context. If different credible assumptions change a consequential decision, the uncertainty must be escalated rather than hidden in a technical appendix.
22. Govern construction changes and handovers
Ground evidence can change as excavations expose layers, instrumentation records movements or unexpected materials are found. Design approvals, site instructions and risk registers must remain aligned. A new contractor may lack the assumptions behind an inherited ground report unless the handover is structured.
Record the authorisation for changes, observations, interpreted risks and remaining controls. Preserve raw data and dates instead of overwriting previous reports. Assign the current decision owner for each unresolved hazard. A successful transfer is one where a qualified successor understands why the design is safe under its specified conditions, not merely where all files have been emailed.
23. Transfer the method to the next ground problem
A bridge foundation, railway tunnel, port reclamation and landslide repair require different geological models, measurements and consequences. General principles transfer—define the decision, recognise uncertainty, investigate, evaluate, control and monitor—but the numerical limits and specific design rules do not.
A gold-standard geotechnical risk report enables a second competent team to trace each important hazard from geological evidence to decision and review. It identifies which risks remain and whose authority governs them. It is valuable because the team can act on uncertainty responsibly, not because a risk matrix is mostly green.
Five fictional decision cases
Fictional case A: a risk matrix cannot predict collapse
An invented project ranks five ground uncertainties on categories from one to five for likelihood and consequence. One slope concern receives category 2 for likelihood and 5 for consequence, yielding an ordinal product of 10. A settlement concern receives 5 and 2, also yielding 10. Identical numbers do not mean identical physical risk, expected losses or safety requirements. The score depends on categories defined by the team and is not a probability.
The responsible review describes plausible outcomes, the people or assets affected, current evidence and relevant controls separately. Safety-critical scenarios require competent professional attention regardless of the similarity of risk-matrix products. These categories are fictional and not approved design thresholds.
Fictional case B: investigation changes the estimated ground model
An imaginary building site has six boreholes, all outside an inaccessible central portion. The initial cross-section assumes a uniform competent layer. A later professionally planned investigation finds a deeper weak unit in the central area. The newly recorded ground-risk count rises, but the project is better informed than before.
A revised design and construction plan may then address the discovered condition, with new monitoring or approvals as appropriate. The change from ‘low’ to ‘high’ recorded risk should be documented as an information gain rather than interpreted automatically as worsening physical ground. No real drilling method or location is prescribed here.
Fictional case C: a control reduces uncertainty but not a physical hazard
A team installs monitoring beside a hypothetical retaining structure. The quality of movement observations improves, yet no physical reinforcement or change to loading has occurred. The new monitoring may reduce uncertainty and support timelier intervention; it does not by itself increase the structure’s resistance.
A risk report should distinguish detection effectiveness, actionable response capacity and hazard reduction. A dashboard that records small movements is useful only within a professionally approved engineering plan and operational review. These are invented conditions, not a claim that a particular structure is safe.
Fictional case D: baseline comparison with changed references
A surveying team records a monitoring point as 10 millimetres lower at the end of construction, but the benchmark used for the second reading may itself have shifted by 4 millimetres. A naive settlement calculation of exactly 10 millimetres would conceal reference uncertainty. The true differential under these simplifying facts may need remeasurement or correction, not an arbitrary subtraction without evidence.
A qualified team would investigate stable benchmarks, instrument precision, observation methods and the design’s relevant movement patterns before evaluating consequences. These numbers are fictional and are not site alert values.
Fictional case E: a contract baseline is not a geotechnical certainty
An invented design–build contract allocates unforeseen subsurface variations according to a baseline report. During construction, new ground conditions differ from the stated baseline. The parties may need the contract’s dispute procedures, but the physical engineering response must prioritise applicable safety and design duties.
A rigorous report preserves observations, comparisons, effects on work and the responsible technical decisions. Allocation of commercial consequences is distinct from identifying ground failure mechanisms. No contractual recommendation or legal interpretation is supplied by this teaching case.
Geotechnical risk assessment evidence audit
- Define project purpose, exposure and engineering decision authority.
- Document geological, groundwater and historical ground evidence.
- Record plausible failure modes and their uncertainty separately.
- Verify investigations match consequential unknowns.
- Distinguish risk categories from justified probabilities.
- Evaluate serviceability and adjacent assets as well as ultimate safety.
- Link controls and monitoring to mechanisms and preapproved responses.
- Preserve contractual and engineering responsibilities separately.
- Update and archive risk registers after new observations.
Applied geotechnical risk reviews: ten field and governance tests
Test 1: the investigation that measured the wrong ground volume
An imaginary project commissions twenty shallow tests distributed around a proposed deep foundation. The test count sounds comprehensive, but none of the measurements reach a layer that may influence deep settlement. The inadequacy is not a lack of diligence measured by staff hours; it is a mismatch between the investigation and the physical mechanism. A competent team should use the conceptual site model to identify influential depths, uncertain strata and the limits of available investigations. Additional work may be justified, but this educational example does not prescribe a number or placement of boreholes. The risk register must state why particular unknowns affect the proposed foundation and how the new evidence would change the authorised design choice.
Test 2: a small measured movement with a significant consequence
An invented monitoring scheme observes a modest movement in soil beside an existing service connection. Another location shows a larger movement well away from sensitive assets. The larger number is not automatically the more important risk, because consequence depends on what could be affected, the movement pattern and relevant structural tolerance. Report reference positions, instrument quality, observation timing and changes in movement rate. The engineer must relate the observations to the approved design and affected structures. Never use a generic millimetre threshold from another site to authorise work. The illustrative lesson is that a measurement becomes useful only when it answers an actual safety or serviceability question.
Test 3: groundwater change outside the project boundary
A hypothetical excavation includes groundwater control, and nearby well observations change during works. Some adjacent buildings and vegetation may respond to altered subsurface water conditions even if they lie outside the property. A risk register focused exclusively on the new building’s foundations would omit an important group of potential consequences. A hydrogeological assessment should examine connectivity, reference elevations, baseline observations and potential alternative causes. The project needs named responsibility and appropriate professional and regulatory review for any relevant effects. The exact water level and mitigation method cannot be inferred from the fact that groundwater changed. It is the causal and spatial reach of the change that requires investigation.
Test 4: a report reused for a different purpose
An older geotechnical report was prepared for a low-rise industrial site. The land is later considered for a different structure and changed loading conditions. Reusing the original parameter table without reviewing scope, records and site changes can transform evidence suitable for one design into a misleading premise for another. Review original investigation depths, sampling purpose, groundwater period and design assumptions. Distinguish data that remain relevant from conclusions that require new assessment. A report’s age is not the only issue: its original question and the quality of its source observations matter. The professional team should document whether and how the inherited evidence is suitable for the present decision.
Test 5: a risk score falls while uncertainty rises
A fictional register initially lists a high-priority slope concern. After a meeting, the entry receives a lower likelihood category because no recent movement was noticed, but field access was actually restricted and monitoring stopped. The score now looks more favourable while the evidence has weakened. That is an information-quality failure. A robust review would record the unavailable measurements and distinguish lack of observed movement from lack of observation. It might retain or increase uncertainty rather than mechanically lower risk. Risk matrices are useful only when their inputs, assumptions and reasons for change remain visible. A programme that rewards green cells can discourage disclosure of exactly the gaps it needs to discover.
Test 6: a temporary design stage has different ground behaviour
A hypothetical retaining system is designed for its completed configuration, but excavation stages change loads and groundwater before permanent supports are in their final positions. A calculation for the finished structure does not automatically prove that intermediate stages satisfy relevant requirements. The construction sequence itself may influence risk. The relevant engineering team must evaluate stages and temporary works through the applicable professional framework. This article does not provide construction procedures or staging limits. The risk assessment should identify when the system changes, which observations matter at each stage and who may approve progression. A valid final design and a safe construction plan are related but distinct deliverables.
Test 7: risk transfer leaves the physical hazard unchanged
A contract assigns the financial consequences of certain unforeseen ground conditions to a specific party. That allocation may change incentives, pricing and dispute procedures, but the geological uncertainty and potential physical dangers remain. An excavation cannot be made safe through a contractual clause alone. Separate commercial responsibility from the obligation to investigate, design, monitor and respond competently. The National Academies’ NCHRP report addresses both reducing uncertainty and allocating remaining risk in design–build projects. A clear contractual baseline may improve coordination but should never be described as a substitute for controlling threats to workers, neighbouring people or public infrastructure. Legal interpretation belongs to qualified advisers.
Test 8: two models agree because they share one assumption
A team runs two numerical methods and obtains similar predicted settlement. Both models, however, use the same assumed weak-layer depth and water level, derived from one incomplete investigation. Agreement could reflect shared input rather than independent validation. A good technical review identifies common assumptions and considers plausible alternative ground profiles or groundwater conditions. Field evidence and independent monitoring can help resolve what the models cannot. A small difference between software outputs is not proof the site is well characterised. The review should focus on whether changing credible assumptions changes the safety or serviceability decision.
Test 9: a monitoring alert needs a responsible recipient
An automated system sends a warning message when an invented instrument crosses an alert level. No qualified reviewer is available to investigate it, and the system has no documented route for acting if readings fail. The technology has improved detection but not completed the management process. A site-specific professional plan needs data quality checks, assigned review, escalation and a basis for deciding protective actions. A sensor failure or transmission gap should have an appropriate exception process. This example deliberately gives no numerical threshold or operational construction advice. A dependable system links the observation to accountable decisions, not just the arrival of an email.
Test 10: a contractor handover erases the basis of an old decision
The original project engineer documented an uncertain fault crossing and obtained field evidence supporting a chosen design condition. Years later, the asset is transferred to a new owner, but the handover contains only a drawing and not the investigation or interpretation history. The next team may treat inferred geology as verified and misunderstand why monitoring was required. Preserve source records, approvals, design revisions, assumptions and monitoring outcomes through the asset’s life. Note which matters remain unresolved and who must review them before alterations. Geotechnical risk assessment is not complete when construction ends; its evidence supports safe maintenance and future decisions. A record that future competent professionals can challenge is more valuable than one polished, unqualified final number.
Frequently asked questions
What is geotechnical risk assessment?
Evaluating uncertainties in soil, rock, groundwater and ground–structure behaviour and their possible safety, service or cost consequences.
How is risk different from geological hazard?
A hazard is a potentially damaging process; risk includes consequences and relevant exposure under defined conditions.
Does a green risk matrix certify a safe site?
No. Its categories and underlying evidence must be assessed; actual site safety requires competent engineering.
What is ISO 31000?
A current guideline standard for organisational risk management, not a certifiable safety approval.
What is a ground conceptual model?
An evidence-based interpretation of geology, groundwater and relevant structures guiding investigation and design.
Why investigate the ground before construction?
To reduce uncertainties that could affect safety, serviceability, costs and planned methods.
What are differing site conditions?
Subsurface conditions that differ from assumptions or contract descriptions under relevant project terms.
Can monitoring alone stabilise a slope?
No. It measures or detects changes; hazard treatment and response are separate.
Who sets safe movement thresholds?
Competent engineering professionals and applicable authorities for the actual site and asset.
Is a high numerical factor of safety enough?
Not necessarily. Design assumptions, modes of failure, serviceability and adjacent consequences also matter.
Why keep original risk assessments?
To trace changes in evidence, decisions and residual uncertainty as a project progresses.
Can I apply this guide to approve earthworks?
No. Actual work requires locally applicable codes, professional review and permissions.
Related Gold Standard engineering and Earth-science library
Geological Mapping establishes evidence about the site, Hydrogeology examines water conditions, Geotechnical Monitoring measures change and Risk Management gives a wider governance framework. Use the related subject roots for their distinct purposes and retain professional accountability for actual construction decisions.
- The Gold Standard of Geotechnical Engineering
- The Gold Standard of Geotechnical Monitoring
- The Gold Standard of Geological Hazards
- The Gold Standard of Geological Mapping
- The Gold Standard of Geology
- The Gold Standard of Geomatics
- The Gold Standard of Geophysics
- The Gold Standard of Hydrogeology
- The Gold Standard of Groundwater Modelling
- The Gold Standard of Risk Management
- The Gold Standard of Seismology
- The Gold Standard of Civil Engineering
Official standards and risk-management sources
- FHWA: Geotechnical Project Development Manual, Chapter 6
- FHWA: Geotechnical Technical Guidance Manual
- National Academies: Managing Geotechnical Risks in Design–Build Projects, NCHRP 884
- ISO: Risk Management Guidelines ISO 31000:2018
- BSI: BS 5930 Ground Investigations, Current Edition
- FHWA: Geotechnical Risk Management, NHI Manual
FHWA, ISO, BSI and National Academies sources reviewed on 5 November 2026. ISO 31000:2018 remains a current risk-management guideline and cannot be used for certification; BS 5930:2015+A1:2020 is the UK current ground-investigation code. These standards have specific jurisdictions and scopes. All numerical and construction examples here are fictional and do not constitute safe operating limits.
Return to chapter index · The Gold Standard hub · How X Works hub