Understand how to structure geological and geotechnical risk in construction using GBRs, variation bands, risk matrices, investigation data, and objective contractual criteria.
Check it out!
Geological risk in construction is the exposure to ground, rock, groundwater, and other subsurface conditions capable of changing the construction method, productivity, quantities, cost, schedule, or safety. Geotechnical risk is the portion of this exposure associated with the engineering behavior of the ground mass, foundations, excavations, retaining systems, fills, tunnels, and other systems that interact with the ground. In engineering contracts, the central issue is not simply deciding whether “soil conditions are the contractor’s risk,” but defining which conditions were investigated, what reference condition was made available to bidders, which variations are ordinary, and which exceed the contractual baseline.
The Geotechnical Baseline Report — GBR — is a way to make this boundary objective. It consolidates reference geological and geotechnical conditions for contractual purposes, establishes parameters and variation bands, and makes it possible to compare what was actually encountered during execution with what the market could reasonably have priced into the bid. The GBR does not replace borings, testing, the geological-geotechnical model, design, or the risk matrix; it connects these elements to contractual risk allocation.
What is geological risk and when does it become contractual risk?
Construction is always carried out in a physical medium that contains variability. Even an extensive investigation program represents only a sampling of subsurface conditions. Between two boreholes there may be changes in lithology, groundwater level, fracturing, strength, density, the presence of boulders, anthropogenic materials, or contamination that were not directly observed.
This physical variability, by itself, does not determine who bears the economic impact. Risk becomes contractual when an uncertain condition may change an obligation, cost, schedule, or construction method and the contract needs to define how that consequence will be addressed.
Brazilian Law No. 14,133/2021 requires risk analysis during the preparatory phase and allows the bidding documents to include a risk allocation matrix. For large-scale public works and integrated and semi-integrated contracting, the matrix is mandatory. The law also requires the allocation to seek efficiency and risks to be quantified so their effects can be reflected in the estimated value.
This prevents a simplistic solution: generically transferring “all geological risk” to the contractor without providing sufficient information for pricing. Economically efficient risk transfer needs to be supported by data, a defined baseline, and the party’s actual ability to manage the event.
Geological risk and geotechnical risk are not exactly the same
The terms are often used as equivalents, but they play different roles in engineering.
Geological risk is broader. It may involve the formation and distribution of geological materials, structures, faults, fractures, karst, boulders, cavities, mineralogy, abrasivity, natural gases, contamination, hydrogeological dynamics, and other phenomena of the physical environment.
Geotechnical risk is related to the engineering behavior of these materials and conditions: bearing capacity, settlement, stability, deformability, excavability, earth pressures, permeability, internal erosion, slope failure, soil-structure interaction, and the performance of foundations and underground works.
The distinction is useful because contractual documentation needs to translate ground characteristics into measurable consequences for the scope. Knowing that a geological unit exists is not enough; it is necessary to understand which properties and behaviors affect execution.
Geotechnical investigation produces data; it does not allocate risk
Borings, field tests, laboratory tests, geophysical surveys, geological mapping, instrumentation, and monitoring produce evidence about ground conditions. The greater the complexity and sensitivity of the project, the greater the need for an investigation program compatible with the decisions to be made.
The mistake is to confuse three different layers:
- factual data, such as borehole logs, samples, test results, and observed groundwater levels;
- technical interpretation, such as the geological-geotechnical model, behavior units, representative parameters, and likely mechanisms;
- contractual baseline, which defines the conditions to be used as the reference for price, schedule, and risk allocation.
Decision 609/2026 of the Second Chamber of Brazil’s Federal Court of Accounts (TCU) illustrates the consequence of skipping the first two layers. In the case analyzed, the absence of borings in the critical area and stability studies led to a generic solution that did not adequately represent local conditions. The issue was not merely documentary: insufficient investigation directly affected the design’s ability to predict embankment behavior.
What is a Geotechnical Baseline Report — GBR?
A GBR is a contractual geotechnical baseline report. Its purpose is to state, objectively and verifiably, the reference conditions that bidders are expected to consider when preparing their proposals and that will be used during execution to distinguish ordinary variation from exceptional conditions.
It is especially relevant in projects where subsurface conditions dominate cost and schedule, such as tunnels, deep excavations, special foundations, dams, port works, underground structures, complex retaining systems, and linear infrastructure in heterogeneous ground.
In Brazil, the GBR is not a universally mandatory document created by Law No. 14,133/2021. Its use derives from contractual practice and engineering risk management. Brazil’s Federal Court of Accounts (TCU) has expressly recognized the instrument’s usefulness. In Decision 4,036/2020-Plenary, the risk matrix for a procurement was revised using objective parameters to distinguish ordinary and extraordinary risks and expressly referred to the GBR as support for analyzing geological and geotechnical risk.
This interpretation is important: the value of the GBR lies in the objectivity it introduces into the risk boundary.
A GBR is not a boring report, geotechnical design, or risk matrix
Each document answers a different question.
| Document | Main question | Function |
| Investigation reports | What was observed and measured? | Record factual field and laboratory data |
| Geological-geotechnical model | How is the subsurface interpreted? | Integrate data and represent units and behaviors |
| Geotechnical design | What solution should be designed? | Define foundations, retaining systems, excavations, stabilization, and technical criteria |
| GBR | Which conditions should be treated as the contractual baseline? | Establish an objective reference for price, schedule, and risk |
| Risk matrix | Who bears the consequence of each event? | Allocate responsibilities and economic effects |
The GBR should be consistent with the other documents. A baseline that conflicts with boring data loses credibility; a matrix that transfers risk without reflecting the baseline loses objectivity; and a design that ignores identified variability creates false precision.
What a geotechnical baseline needs to make verifiable
The content depends on the project type, but the logic remains constant: transform variables that affect execution into observable and comparable criteria.
When relevant, it may include:
- relevant geological and geotechnical units;
- ranges of strength and deformability;
- frequency and orientation of fractures;
- rock mass quality;
- abrasivity and wear potential;
- distribution of blocks, boulders, or obstructions;
- groundwater levels and pressures;
- permeability and seepage flow rates;
- occurrence of artesian conditions;
- potential for piping, internal erosion, or hydraulic instability;
- presence of expansive, collapsible, or organic materials;
- contamination limits or presence of subsurface gases;
- excavatability classes;
- mixed-face and clogging conditions in mechanized excavation;
- parameters or criteria that change support, retaining systems, foundations, or construction methods.
The baseline should not become an endless list of properties. It should focus on variables capable of producing material contractual consequences.
Variation bands: the boundary between ordinary and extraordinary conditions
One of the most important contributions of a GBR is allowing risk allocation to move beyond a binary and generic approach. Instead of merely stating that “geological conditions are the contractor’s risk,” the contract can establish variation bands.
Within the band, the condition is considered compatible with the baseline and forms part of the ordinary risk priced by the contractor. Outside the band, the condition may trigger the treatment defined in the matrix: additional investigation, change of method, schedule review, change procedure, economic rebalancing analysis, or another mechanism established in the contract.
Conceptual example: a rock excavation may have a baseline for strength, abrasivity, fracturing, and water inflow. Small variations within the stated ranges are part of normal productivity. A materially different combination outside those ranges may characterize a condition different from the basis used for the bid.
The band needs to be technical and measurable. Terms such as “poor soil,” “very hard rock,” or “large amounts of water” do not create an auditable baseline.
The risk matrix should use the baseline, not replace it
The matrix answers who bears the impact. The GBR answers against which condition that impact will be compared.
Brazil’s Federal Court of Accounts (TCU) guidance on risk matrices itself notes that construction risks are commonly related, among other factors, to environmental and geotechnical conditions and that allocation should not be arbitrary. The guidance observes that a matrix may, for example, assign geological risks to the contractor, but it also warns that allocation needs to be technical, economically efficient, and compatible with the party’s ability to manage the risk.
In a robust structure, each relevant geotechnical risk has:
- an identifiable physical cause or condition;
- a baseline or reference range;
- a defined risk event;
- a technical consequence in cost, schedule, safety, or performance;
- the party responsible within the ordinary range;
- an objective trigger for exceptional conditions;
- a notification and substantiation procedure;
- a method for measuring impact and treating the occurrence.
This structure reduces subjective discussion after the excavation is already open and the impact has already occurred.
Integrated and semi-integrated contracting require special attention
Under integrated and semi-integrated contracting, the relationship between design freedom and risk needs to be explicit. Brazilian Law No. 14,133/2021 requires risks associated with the contractor’s choice of the basic-design solution to be allocated to the contractor in the legally applicable cases.
This does not turn nonexistent geotechnical information into risk that can automatically be transferred. The contractor may assume risks associated with the solution it selects, but it needs to receive the data and technical elements that the public authority is required to provide according to the stage of procurement.
The problem arises when the bidding documents simultaneously transfer:
- the obligation to discover existing conditions;
- the risk of any condition discovered;
- responsibility for the solution;
- and the economic impact of information that was unavailable for pricing.
This combination tends to produce risk premiums, non-comparable bids, or future claims.
Decision 4,036/2020 and the contractual function of the GBR
Decision 4,036/2020-Plenary is a relevant Brazilian reference because it addresses precisely the need to expand the information made available to the market and to make the division between ordinary and extraordinary risk objective.
The proceeding records the evolution from an older practice in which contracting authorities provided only factual boring records and avoided interpretive reports. The solution analyzed by the Court moved in the opposite direction: greater transparency, objective parameters in the risk matrix, and express reference to the GBR.
The lesson for contract structuring is direct: unavoidable uncertainty should not be confused with withheld information. The market prices risk more effectively when it knows the interpretive basis used by the public authority.
Decision 609/2026 shows why investigation comes before risk transfer
In the case reviewed by the TCU in 2026, the solution for a significant-height embankment was designed without stability studies and without borings in the landslide area. The Court noted that the design did not adequately consider local conditions, including groundwater levels and low-strength soils.
This type of case defines a fundamental boundary. An adverse condition may remain a residual risk even after adequate investigation. But a condition that was not investigated because of a design deficiency should not automatically be treated as unavoidable randomness.
The decision on how much to investigate should be proportional to the project. The same investigation program is not required for a small conventional building and a tunnel in a heterogeneous rock mass. The criterion is to produce sufficient information for the engineering and contracting decisions that will be made.
Five risk families that a GBR can structure
Soil and rock conditions
These include variations in strength, density, fracturing, abrasivity, excavability, occurrence of boulders, mixed-face conditions, and materials that change productivity, support requirements, or equipment wear.
Hydrogeology
Groundwater level, pressure, flow, confined aquifers, permeability, water chemistry, piping, and other conditions may change retaining systems, dewatering, waterproofing, safety, and construction methods.
Structures and discontinuities
Faults, fractures, shear zones, cavities, and geological contacts may change stability and support requirements.
Natural and anthropogenic obstructions
Boulders, old foundations, buried structures, abandoned utilities, and waste may cause significant impacts on excavation and foundation work.
Subsurface environmental risks
Contaminated soil, contaminated groundwater, gases, or materials requiring special handling affect excavation, disposal, occupational safety, permitting, logistics, and cost.
How to quantify risk without turning uncertainty into an arbitrary percentage
After the baseline is defined, risk can be quantified according to materiality and data availability. The analysis may combine the history of comparable projects, parameter distributions, scenarios, simulation, parametric modeling, and documented expert judgment.
The objective is not to choose a sophisticated method for appearance’s sake. It is to preserve the chain between physical condition, probability or frequency, technical consequence, and economic impact.
The TCU has reinforced this evidence standard in recent analyses of contingency and risk. When modeling depends on empirical parameters, isolated expert opinion should not be presented as a substitute for historical data or verifiable studies.
In geotechnical engineering, this means recording where the ranges came from, which investigations support the model, which data were extrapolated, and how residual uncertainty was treated.
Geotechnical risk, contingency, and estimating need to work together
The existence of a risk matrix does not eliminate the need to reflect risks in the estimate. Brazilian Law No. 14,133/2021 requires risk allocation to be quantified so its effects can be projected into the estimated value.
A risk allocated to the contractor may generate an appropriate risk premium or contingency. A risk retained by the public authority may require a provision or specific mechanism. A risk covered by insurance needs to be treated considering coverage, deductible, and residual exposure.
The critical control is avoiding double counting. The same risk should not appear in full in the base estimate, BDI markup, contingency, insurance, and still be treated as an extraordinary event without reconciliation.
Contingency reserves in engineering projects should derive from identified risks, rather than from a percentage selected before the exposure is understood.
The cost of investigation should be compared with the cost of carrying uncertainty
There are situations in which the public authority saves on investigation and later pays for the uncertainty through bid pricing. If an additional borehole, test, or geophysical program significantly narrows a range that dominates project cost, producing information may be economically better than transferring broad uncertainty to the market.
This reasoning does not mean investigating indefinitely. There is a point of diminishing returns. The question is whether the next piece of information has the potential to materially change the solution, quantities, method, risk, or price.
Design Review in Engineering Projects can apply this logic to identify which design gaps still justify investigation before bidding.
How to structure triggers for differing conditions
A good contractual trigger avoids open-ended expressions and describes what must be demonstrated.
It may include:
- baseline parameter or condition;
- measurement or testing method;
- ordinary range considered;
- threshold that characterizes deviation;
- minimum extent or persistence, when applicable;
- notification deadline;
- preservation of evidence;
- joint field verification;
- effect on the construction method;
- rule for measuring incremental impact.
The trigger does not automatically create entitlement to additional payment. It opens the contractual procedure established for analyzing the occurrence. Causation, responsibility, documentation, and the criteria in the risk matrix remain necessary.
Geotechnical inspection needs to record what was encountered, not only what was executed
When a contract uses a geotechnical baseline, inspection needs to produce contemporaneous evidence of the conditions encountered. Without it, later comparison between the baseline and reality becomes an indirect reconstruction.
Useful records may include:
- excavation-face mapping;
- classification of material encountered;
- georeferenced photographs;
- supplementary boring logs;
- test results;
- flow rates and groundwater levels;
- instrumentation readings;
- equipment cycles and productivity;
- tool consumption and wear;
- volumes classified by condition;
- daily construction reports and formal notices linked to the event.
The evidence needs to be sufficient to answer three questions: what was encountered, to what extent it differs from the baseline, and what technical consequence was caused by that divergence.
How to analyze a claim related to geotechnical conditions
The existence of a difficult condition is not, by itself, evidence of economic imbalance. The technical analysis needs to reconstruct the causal chain.
A robust sequence verifies:
- which condition was defined in the GBR and other bidding documents;
- which risk and which party were identified in the matrix;
- which condition was actually observed;
- whether the deviation exceeds the contractual band or trigger;
- whether notification was made and evidence preserved;
- whether the condition actually changed the method, productivity, quantity, or schedule;
- which portion of the impact is incremental relative to what was already priced;
- whether there was contribution from execution error, a contractor decision, or a scope change;
- which contractual mechanism applies.
This logic connects the GBR, risk matrix, inspection, and economic-financial rebalancing in engineering contracts.
Recurring errors in geological-risk allocation
The most common problems arise when the documentation creates the appearance of risk transfer without creating a measurable boundary.
- stating that “all subsurface risk belongs to the contractor” without a baseline;
- providing only raw boring data while withholding relevant interpretation;
- using too little investigation for a high-sensitivity decision;
- defining a baseline with non-verifiable qualitative terms;
- creating bands so broad that allocation becomes fictitious;
- creating bands so narrow that any ordinary variation becomes a claim;
- failing to connect the GBR to the risk matrix;
- failing to reflect transferred risk in the estimated value;
- failing to define notification and verification procedures;
- failing to record actual conditions during execution;
- paying for an impact that was already included in ordinary risk;
- automatically denying the impact of a condition outside the baseline merely because the matrix uses generic language.
Checklist for a contractually useful GBR
Before bidding, it is worth checking whether:
- the investigation program is proportional to the project type and risks;
- factual data are available to bidders;
- there is a geological-geotechnical model consistent with the data;
- the variables that dominate cost and schedule have been identified;
- each baseline has a unit and verification method;
- variation bands are objective;
- the risk matrix references the relevant conditions;
- the estimate consistently reflects transferred risk;
- triggers for differing conditions are defined;
- the notification procedure is established;
- inspection and instrumentation will be able to verify field conditions;
- insurance and contingency treatment do not duplicate the same risk;
- the change or rebalancing mechanism preserves the original allocation.
How Engineering Consulting turns subsurface risk into contractual criteria
The contribution of engineering consulting is not to eliminate geological uncertainty. That is not possible. The technical work is to reduce avoidable uncertainty, characterize residual uncertainty, and convert that exposure into documentation that the market can price and inspection teams can verify.
For A3A Engenharia, the structuring process can integrate surveys and data review, maturity assessment, design interfaces, risk registers, the baseline, verification criteria, the risk matrix, estimating, contracting conditions, and evidence procedures during execution.
The Engineering Risk Management service connects event identification to quantification, treatment, and contractual governance. Together with planning and Design Review, it prevents an unstudied uncertainty from simply being shifted into the bid.
Final considerations
Geological risk is not synonymous with surprise, and geotechnical risk should not be treated as a generic line in the matrix. The more subsurface conditions influence cost, schedule, and method, the more important it is to define the reference condition before economic competition begins.
The GBR is valuable because it turns geotechnical interpretation into a contractual baseline. Variation bands make it possible to separate ordinary variability, which can be priced by the contractor, from materially differing conditions that require the procedure established in the contract.
The quality of this allocation depends on a complete chain: sufficient investigation, interpretation, baseline, matrix, estimate, inspection, and evidence. If any link is missing, the contract loses the ability to distinguish assumed risk from an unrepresented condition.
The best matrix is not the one that transfers the most risk. It is the one that makes explicit which risk exists, what information supports it, who can manage it, and how its effect will be treated if reality exceeds the reference condition.
Technical references
[1] BRAZIL. Law No. 14,133, of April 1, 2021. Public Procurement and Administrative Contracts Law. Available at: [Planalto — Law No. 14,133/2021](https://www.planalto.gov.br/ccivil_03/_ato2019-2022/2021/lei/l14133.htm).
[2] BRAZILIAN FEDERAL COURT OF ACCOUNTS. Procurement and Contracts: TCU Guidance and Case Law. Item 4.5.5 — Risk Matrix. Updated Aug. 29, 2025. Available at: [TCU — Risk Matrix](https://licitacoesecontratos.tcu.gov.br/4-5-5-matriz-de-riscos/).
[3] BRAZILIAN FEDERAL COURT OF ACCOUNTS. Decision No. 4,036/2020-TCU-Plenary. Discussion of objective parameters for ordinary and extraordinary risks and reference to the Geotechnical Baseline Report. Available at: [TCU — Decision 4,036/2020](https://pesquisa.apps.tcu.gov.br/documento/acordao-completo/4.036%252F2020/%2520/DTRELEVANCIA%2520desc%252C%2520NUMACORDAOINT%2520desc/0).
[4] BRAZILIAN FEDERAL COURT OF ACCOUNTS. Decision No. 609/2026-TCU-Second Chamber. Geotechnical studies, stability, and suitability of the solution to local conditions. Available at: [TCU — Decision 609/2026](https://pesquisa.apps.tcu.gov.br/doc/acordao-completo/609/2026/Segunda%20C%C3%A2mara).
Frequently asked questions
It is exposure to ground, rock, groundwater, geological structures, obstructions, or other subsurface characteristics that may change the solution, construction method, productivity, quantities, cost, schedule, or safety.
Geological risk is broader and describes uncertainties in the physical environment. Geotechnical risk mainly addresses the engineering behavior of soil, rock, and water in interaction with foundations, excavations, fills, retaining systems, tunnels, and structures.
A GBR, or Geotechnical Baseline Report, establishes reference geological and geotechnical conditions for contractual purposes. It helps separate ordinary variations that can already be priced from conditions that differ from the baseline.
No. Law 14,133 does not create a universal document called a GBR. The instrument is an engineering and contracting technique that may be adopted when the materiality and complexity of geotechnical risk justify a more objective contractual baseline.
No. Borings produce factual data about investigated points. The GBR uses data and interpretations to declare reference conditions that will be used in the contract for pricing, risk, and treatment of differing conditions.
No. A condition outside the baseline may trigger the established contractual procedure, but the risk matrix, causation, evidence, incremental impact, and other requirements for change or rebalancing still need to be verified.
The bands should use measurable parameters that materially affect execution, with a verification method, ordinary range, divergence threshold, and recording procedure. Their width should reflect available data, variability, and investigation maturity.
There is no single answer. Allocation should consider control over the cause, access to information, ability to prevent or mitigate, ability to price and insure the risk, delivery model, and economic efficiency of the transfer.
Additional technical materials
Related solutions
Related services
- Engineering Risk Management: identification, analysis, mitigation, and contingency
- Design Review in Engineering Projects: technical review, interfaces, and project maturity
- Technical Planning for Engineering Procurement: strategy, requirements, risks, and documentation
Core content on this topic
- Risk Allocation Matrix in Engineering Contracts: how to allocate responsibilities under Law 14,133
- Risk Management in Public Works: integrating planning, contract, and execution
