01 · Review scope
Define what is being reviewed, and why.
The starting point is the decision the evidence will support: an operating change, asset intervention, design acceptance or investment pathway. Record the review boundary, available documents, applicable criteria and the version of each model or calculation.
Independence should be described in practical terms, including the reviewer’s relationship to the original work and any conflicts. A review of a report is not equivalent to reproducing the model, and a targeted technical check is not a complete design certification.
02 · Technical challenge
Look beyond whether the calculation runs.
Review the chain from input data to decision. For hydraulic network modelling, this includes topology, demand allocation, pump and valve behaviour, calibration evidence and outage assumptions. For CFD, it includes geometry, boundary conditions, mesh and time-step sensitivity, conservation checks and the suitability of the physical model.
| Review area | Question to resolve | Evidence to retain |
|---|---|---|
| Inputs | Are the sources, units and assumptions traceable? | Data register; calculation basis; model version |
| Behaviour | Are the results physically plausible? | Balance checks; benchmark or field comparisons |
| Uncertainty | Could an uncertain input change the decision? | Sensitivity cases; limitations; further evidence needed |
| Interpretation | Does the conclusion stay within the evidence? | Decision criteria; option comparison; unresolved risks |
Scroll the table sideways to read all 3 columns.
Residuals and convergence plots are part of the evidence, not a substitute for engineering checks. Where full reproduction is outside the agreed scope, identify the checks performed and the elements that remain unverified.
03 · Findings
Make each finding actionable.
A finding should identify the issue, the supporting evidence, its possible consequence and the action needed. Distinguish a conclusion-changing uncertainty from a documentation improvement; a long list without priorities makes resolution harder.
- Decision-critical: a concern that could change acceptability, option selection or operational safety. Resolve it, or explicitly withhold reliance on the affected conclusion.
- Material clarification: an assumption or evidence gap that needs a documented response before the next decision stage.
- Improvement: a correction to clarity, traceability or presentation that does not currently alter the main conclusion.
These are an illustrative classification, not a universal standard. Use the project’s agreed risk and assurance framework where one applies. Assign an owner and retain the response, revised evidence and reviewer’s closure status for each material finding.
04 · Decision readiness
Close the reasoning, not just the comments.
For an investment case, check that the need, option comparison, phasing and risk treatment are supported by a consistent evidence base. A technically attractive option may still depend on a delivery interface, operating change or demand assumption that has not been demonstrated.
The review record should say what can be relied on, for which purpose, under which assumptions, and what remains open. An explicit limitation can be more useful than an unconditional-looking conclusion that exceeds the review scope.
The water-resource appraisal example shows a generalised approach to scenario-led decisions. The CFD contact-tank example makes numerical limitations visible alongside the results; it is not evidence of validated treatment performance.
Common questions
What a review does — and does not — establish.
Does peer review transfer responsibility for the design?
A technical review does not, by itself, redefine the responsibilities of the designer, operator or decision-maker. Its scope, findings and limitations should be documented alongside the project’s existing responsibilities.
Can a review be useful before the model is complete?
Yes. Early review can test the problem definition, data plan and modelling approach before effort is committed. Later review examines the resulting evidence. These are distinct stages, and an early review should not be presented as approval of later results.
Technical references
These references support the methods discussed; they do not imply endorsement of KOMEC GLOBAL or certification of a particular model.
- US EPA · Guidance on the Development, Evaluation, and Application of Environmental Models
Background on model evaluation, intended application, sensitivity and uncertainty. General guidance, not a project-specific approval standard.