Field data is never perfectly complete. Soil borings cover selected points, surveys contain measurement limits, material tests represent sampled specimens, and monitoring equipment has its own accuracy range. Engineering decisions therefore depend not only on calculations, but also on how well the available information represents the real system. Understanding that uncertainty is an important part of PE continuing education, especially for engineers who make decisions from incomplete or variable field information.
The Number Is Not the Whole Story
An engineer may receive a soil strength value, groundwater elevation, structural measurement, flow rate, or material test result and place it directly into a calculation. The value looks objective, but every measurement has some level of uncertainty.
NIST describes measurement uncertainty as an indication of the doubt associated with a measurement result. It can arise from factors such as equipment, sampling, environmental conditions, and the measurement process itself.
The practical issue is simple: a number can be useful without being exact.
Sampling Creates a Picture, Not the Entire Site
Geotechnical work shows this clearly. A few borings provide information about selected locations, not every inch of the subsurface.
One boring may encounter dense soil while another reaches a weaker layer nearby. Rock depth can change. Groundwater can vary across a site or over time. A foundation design based on limited investigation data therefore needs an understanding of what the investigation did not capture.
That does not make the investigation inadequate. It means the engineer must consider how much variation could exist between known points.
Not Every Unknown Deserves the Same Attention
Engineers do not need perfect information about every variable. They need to know which variables can materially change the result.
Consider a hydraulic model containing several inputs. One parameter may change the predicted flow by only a small amount when adjusted. Another may cause a large change. The second variable deserves closer investigation because uncertainty in that input has greater influence on the decision.
This is the basic idea behind sensitivity analysis.
A useful review asks:
- Which input has the largest effect on the result?
- How much can that input realistically vary?
- Does the design remain acceptable across that range?
- Would additional field data meaningfully reduce the uncertainty?
Uncertainty Can Change the Engineering Approach
Suppose an engineer is evaluating foundation settlement and the available subsurface information indicates a wide range of soil properties. Selecting one average value may produce a neat calculation, but it can hide the variation. A better approach may involve evaluating several reasonable conditions. The engineer can then see how the predicted settlement changes as soil stiffness, groundwater conditions, or layer thickness changes.
This does not mean designing for every imaginable condition. It means understanding the range of plausible behavior before making the final decision.
Field Conditions Can Change After the Investigation
A field investigation represents conditions observed at a particular time. Those conditions can change.
Groundwater levels may rise after heavy rainfall. Excavation can alter drainage paths. Construction can disturb soil. Nearby development can change surface runoff. Existing structures can deteriorate after the original investigation.
Engineers working on existing infrastructure therefore need to ask if old information still represents current conditions. A report from ten years ago may still provide useful background, but it should not automatically be treated as a perfect description of today’s site.
When More Data Is Worth the Cost
Additional testing is not always the answer. Field investigation costs money and takes time. The engineer has to decide if new information could change the design enough to justify collecting it.
Imagine a retaining wall analysis that remains stable across a broad range of reasonable soil parameters. More testing may provide little practical benefit. Another project may sit close to a critical stability limit, making better soil data much more valuable.
The right question is not simply, “Can we collect more data?” It is, “Can better data change the engineering decision?”
Uncertainty and Safety Margins
Engineering design already includes factors that account for variability and uncertainty. Load factors, resistance factors, material design values, and other provisions help engineers manage known sources of variation.
Those tools do not eliminate uncertainty in field information. A safety factor cannot correct an incorrect site model or a missing failure mechanism.
Engineers still need to understand the quality of the information behind the calculation. A conservative number based on the wrong physical condition can remain misleading.
Models Can Hide Uncertain Inputs
Modern software can make uncertain information look remarkably clean. A model may display a single groundwater level, soil parameter, flow rate, or material property even when the field data supports a range of values.
The output may then appear more certain than the input deserves.
NIST’s guidance on uncertainty emphasizes identifying sources of uncertainty and evaluating how they affect measurement or modeling results.
Good engineering practice keeps that distinction visible. A calculated value is an estimate based on assumptions and available evidence.
When Field Data Conflicts With the Model
Sometimes an engineer sees something that the calculation did not predict. A structure settles more than expected. A drainage area produces higher flows. A measured temperature differs from the thermal model. A pavement section deteriorates sooner than predicted.
The first response should not be to force the field observation to match the model. The difference is information.
Engineers can review the inputs, assumptions, boundary conditions, construction records, and measurement methods. The mismatch may reveal an incorrect assumption or a condition that was not included in the original analysis.
Professional Judgment Fills the Gap
Engineering decisions rarely come from equations alone. An engineer has to decide if the available information is adequate, if an assumption is reasonable, and if additional investigation is needed.
That judgment becomes especially important when the consequences of an incorrect decision are significant.
A small uncertainty may be acceptable for one application and unacceptable for another.
Strong PE continuing education courses can help engineers revisit these decisions through technical examples, modeling methods, risk analysis, and lessons from actual engineering practice.
Learning to Work With Imperfect Information
Uncertainty is not a defect in engineering. It is part of working with physical systems that cannot be measured completely.
The goal is to recognize uncertainty, estimate its effect, and make decisions that remain defensible when conditions differ from the assumed case. That skill becomes more valuable as projects become more complex and engineers rely on larger amounts of field and model data.
Keeping Engineering Judgment Current
Engineering rarely gives professionals perfect information. The real skill is knowing what the available data can support, where the gaps are, and when those gaps could affect the outcome.
PE continuing education can help engineers sharpen that judgment through practical technical problems, updated methods, and real-world examples. The goal is not to remove every unknown, but to recognize the important ones before they become costly engineering problems.
