QA/QC for Environmental Data on Korean Projects: What Makes Results Defensible

Seventh in the methodology series. Every environmental decision downstream — the deal price, the remediation budget, the statutory position — stands on laboratory numbers. A number without its quality record is an anecdote with decimal places. Here is the QA/QC architecture that turns results into evidence, and where Korean projects need it most.

What the discipline actually asks for

• QC samples that interrogate the process. Field duplicates test reproducibility — two samples from the same point that should agree, and when they don't, the variability is itself a finding about the medium or the method. Blanks test cleanliness: equipment blanks prove decontamination worked, trip blanks catch contamination acquired in transit. Each QC sample answers a specific question about whether the process, not the site, produced the number.
• Custody without gaps. Every transfer of every sample documented — who, when, condition, temperature. The chain of custody is boring precisely until a result is disputed in a negotiation or a proceeding, at which point it is the only thing standing between "our data shows" and "so you claim."
• Detection limits chosen against the decision. A reporting limit above the applicable standard makes the analysis ornamental — the result cannot answer the question it was ordered for. Limits are checked against the relevant Korean tables at scoping, per the design discipline, not discovered at data review.
• Review with authority. Someone qualified examines the package — holding times, blank hits, duplicate agreement, calibration flags — and has the standing to qualify or reject data. Review that cannot reject is decoration.

Where it meets the Korean project

• The two-question laboratory test. Is the laboratory analytically competent for these parameters at these levels — and is it positioned under the Korean framework (official methods, recognized standing) for results that must land in statutory comparisons? On dual-track programs the answer may be two laboratories, or one that satisfies both — but the question is asked per data use, before samples ship, as part of the field program architecture.
• Trace work raises the stakes. The parameters increasingly in play on Korean sites — PFAS at nanogram levels, low-level volatiles — are exactly where field technique and laboratory background dominate results. At trace levels, the QC samples aren't overhead; they are most of the interpretation.
• Two files, one truth. The international deal file wants the QC narrative in E-series terms; the Korean file wants conformity with official methods. The data quality section is where the two-frame discipline becomes concrete: one dataset, its quality demonstrated in both languages, no result claimed in one file that the other file's rules would reject.

The mistakes that repeat

1. QC as budget cut. Blanks and duplicates are the first line deleted when the quote needs trimming — removing precisely the samples that make the rest defensible. A program that cannot afford QC cannot afford to be challenged.
2. The unexamined blank hit. The equipment blank shows the solvent; the report reports the site detections anyway, unqualified. One contaminated blank can hollow out a whole round's volatiles — acknowledging it is painful and mandatory.
3. Custody by courier receipt. Samples handed to a shipping service with no custody documentation, temperatures unrecorded. The results may be perfectly accurate; they are also perfectly deniable.
4. Suspiciously smooth data. Field duplicates in implausibly perfect agreement, no qualified results across hundreds of analyses — the same instinct that applies to too-clean self-measurement records applies to investigation data. Real programs have friction; friction documented is credibility.
5. Quality bolted on at reporting. The QA/QC section written last, describing procedures that were never actually specified. Reviewers — and opposing consultants — can tell. Quality is either in the scope or in the imagination.

Last reviewed: 31 August 2026 · Dime Works