The Roadmap We Already Had

There is a version of the AISCT® story that goes: here is a new technology, and here is why you should consider adopting it[cite: 4]. It is not the most useful version, because it puts the technology first and leaves the harder question untouched — why has the industry spent three decades agreeing with guidance it does not follow?[cite: 4]

Because that is what happened[cite: 4]. The practice that high-density, real-time, adaptive field data supports is not a novel proposal[cite: 4]. It has been the published recommendation of the U.S. EPA, CCME and CSA Group for a very long time[cite: 4]. What the industry lacked was never the roadmap[cite: 4].

The EPA Triad: written for exactly this

The EPA's Triad approach is one of the most definitive statements ever published on how environmental site work should be conducted[cite: 4]. It rests on three integrated practices, and it is worth reading them with a field programme in mind[cite: 4]:

• Systematic project planning — Planning built around data quality objectives, decision logic and adaptive strategy — rather than a fixed, predetermined sampling grid drawn before anyone has seen the site[cite: 4].
• Dynamic work strategies — A sequence of data collection activities implemented in real time, using field data to guide what happens next rather than executing a static plan regardless of what is found[cite: 4].
• Real-time measurement technologies — Data generation that enables reliable measurement and analysis in the field, in time frames that let a dynamic work strategy actually run[cite: 4]. EPA notes these "typically result in a much greater density of information."[cite: 4]

The EPA's own assessment of what this delivers is not hedged: the Triad approach "can be used to significantly reduce data collection costs, expedite project schedules, enhance stakeholder communication, and improve the quality of project and site decisions."[cite: 4]

Read the third pillar again[cite: 4]. Reliable measurement, in the field, fast enough to change what happens next, at much greater density[cite: 4]. That is a specification for a field instrument, published in 2010, describing something the field did not have[cite: 4].

The DQO process: decisions first, samples second

The EPA's guidance on systematic planning using the Data Quality Objectives process makes the same argument from a different angle[cite: 4]. It establishes that environmental data collection should be planned around specific decisions rather than arbitrary sampling protocols, and it asks the practitioner to define three things up front[cite: 4]:

• The specific decision the data is intended to support[cite: 4].
• The tolerable level of decision error — how often you can accept being wrong, and in which direction[cite: 4].
• The quantity and quality of data needed to make that decision at acceptable confidence[cite: 4].

That second bullet is the interesting one, because it is the false-negative and false-positive framework in everything but name[cite: 4]. A tolerable decision error is a statement about how often a screen may miss an exceedance and how often it may over-call a clean sample[cite: 4]. In Triad practice those bars are conventionally set at fewer than 5% false negatives and fewer than 20% false positives — the asymmetry reflecting that an over-call costs a confirmatory sample while a miss can cost a site[cite: 4].

Here is the difficulty this creates for conventional practice[cite: 4]. You cannot state a tolerable decision error for an instrument whose readings are not comparable to one another[cite: 4]. A bag-and-PID programme has no error rate, because it has no reproducible relationship to anything[cite: 4]. Which means a programme built on it cannot, strictly speaking, satisfy the DQO process it claims to follow[cite: 4].

CCME: the Canadian expectation is not softer

CCME's Guidance Manual for Environmental Site Characterization echoes and extends the same principles in the Canadian context[cite: 4]. It recognizes headspace vapour and solvent-extraction testing as legitimate field analytical methods, provided they are performed under controlled, repeatable conditions[cite: 4]. It emphasizes high-density sampling to capture heterogeneity and reduce conceptual site model uncertainty[cite: 4]. It describes site characterization as iterative and adaptive — a model that refines as data accumulates, not a plan executed to completion[cite: 4].

And it sets an expectation that quietly indicts a great deal of ordinary practice: field precision assessment, including duplicates and triplicates, is standard professional practice rather than an optional enhancement[cite: 4].

CSA Z769 goes further still on responsibility[cite: 4]. The professional overseeing field activities bears responsibility for the adequacy of the sampling programme — and that responsibility cannot be discharged by adhering to a fixed protocol that does not adapt to site conditions[cite: 4]. In the context of professional standards, adaptive, high-density, real-time-informed sampling is not merely better practice[cite: 4]. It is the expected practice[cite: 4].

The guidance has been in place for roughly a decade in its current form, and far longer in substance[cite: 4]. Long enough for a recommended practice to have become standard practice[cite: 4]. In field vapour screening, it has not[cite: 4].

So why didn't it happen?

The honest answer is that the instruments could not do it[cite: 4]. A dynamic work strategy needs reliable data faster than the decision[cite: 4]. A DQO needs a stated error rate[cite: 4]. High-density sampling needs throughput and per-sample economics that a laboratory-only programme cannot provide[cite: 4]. Every pillar of the guidance depended on a class of field measurement that did not exist at the required quality — so practitioners did the reasonable thing and executed the parts they could: a fixed grid, sparse sampling, laboratory confirmation, professional judgment filling the gaps[cite: 4].

That constraint has now lifted, which is why the second half of the answer matters more than the first[cite: 4]. Once the tool exists, what remains between the guidance and the practice is organizational, not technical — and it is worth naming, because these patterns are recognizable in every firm[cite: 4]:

• Ontological friction — A conventional field reading is a single number[cite: 4]. A modern one is a prediction, a precision metric and a probability[cite: 4]. That is not more complicated — it is more honest — but it asks for a different kind of interpretation, and unfamiliarity reads as complexity[cite: 4].
• The competency trap — Expertise in an established method can quietly become a barrier to adopting a better one[cite: 4]. It shouldn't: the professional who understands conceptual site models, regulatory requirements and field logistics is exactly the person best positioned to use better data[cite: 4].
• The hourly value paradox — In a time-and-materials world, a machine that isn't moving looks like a cost[cite: 4]. So a three-minute pause to run a duplicate reads as lost time — even when it avoids hauling fifty tonnes of clean soil[cite: 4]. The relevant unit is lifecycle cost, not hourly utilization[cite: 4].
• Unconscious resistance — Nominally adopting the tool while continuing to operate on the old mental model — screening every sample but never expanding the programme when the data clearly warrants it[cite: 4]. Only structured, organizationally endorsed decision triggers fix this[cite: 4].
• Fear — That real-time data will surface problems, or that field-confident decisions will be questioned[cite: 4]. The regulatory record says the opposite: programmes demonstrating thoroughness and decision logic aligned with Triad and CCME are more defensible, not less[cite: 4].

The uncomfortable conclusion

If the guidance describes adaptive, high-density, real-time practice as the expectation, and the instrumentation to execute it now exists, then continuing with sparse, static, qualitative screening is a choice — and it is the choice that carries more professional risk, not less[cite: 4].

That is the part worth sitting with[cite: 4]. The greater exposure has never been in over-characterizing a site[cite: 4]. It has always been in the ten-sample programme that asked a reviewer to accept a conceptual model on faith, and in the boundary that was assumed rather than measured[cite: 4]. The roadmap was right the whole time[cite: 4]. What changed is that following it is now possible[cite: 4].

The One Thing to Remember

AISCT® does not ask a regulator to accept a new idea[cite: 4]. It executes an old one — the EPA Triad, the DQO process, CCME Volume 1 and CSA Z769 — at a density and reliability that the field has never previously been able to deliver[cite: 4]. The industry has had the roadmap for thirty years[cite: 4]. It now has the vehicle[cite: 4].