What a QA/QC audit actually covers in utility vegetation work

A vegetation QA/QC audit checks finished work on identifiable spans against named standards, records who saw what and when, grades any defects, and routes them into corrective and preventive actions with due dates. The file is photos, measurements, location, auditor identity, a pass or fail, and a clock that can be verified.

What QA is, what QC is, and where each starts on one span

A vegetation QA/QC audit starts with two distinct jobs. Quality control is the inspection of finished work on a defined unit, recorded as a pass or a fail against a named standard. Quality assurance is the system that makes those inspections hold up, including the method for choosing what to check, the evidence each check must produce, how findings are graded, and how corrections are tracked to closure.

On a three pole span trimmed on July 18, 2026, QC is a second set of eyes on that span with a camera, a rangefinder, and the applicable rule or spec open. QA is the checklist the auditor is required to follow, the definition of a Critical versus a Major defect, the rule for when a sample triggers 100 percent review of a circuit, and the report that rolls those unit results up by district and contractor.

A utility quality standard is what the auditor inspects to. Pacific Gas and Electric publishes a formal standard for routine distribution vegetation quality control, SW-0003 Quality Control - Vegetation Management for Routine Electric Distribution with a stated publication date of May 31, 2024.

That sort of document sets the workmanlike details QC checks on a span. The legal clearance requirements that govern the span are a separate body of rules. What Rule 35 actually requires and how Table 1’s distances are read by voltage are the sources for the clearance numbers an audit applies when it asserts a code nonconformance.

The unit under audit, and the evidence file that proves it

Define the thing you are checking, or the file will not tie back to the work. On overhead distribution, the unit of vegetation work is commonly a span from pole P13421 to P13422 on Circuit 12JY, or a tree work order W-009334 linked to that span. Pick one and use it in every line of the evidence.

A defensible span audit file carries, at minimum:

• The unit identifier, for example Span P13421-P13422 on Circuit 12JY, or Work Order W-009334. The same ID appears on the photos, the notes, and the pass or fail.

• Location, as a GIS feature ID, coordinates, or a map tile reference. The point is to put the auditor and the work in the same place.

• Date and time of audit. If the span was trimmed on July 18, 2026 and audited on August 2, 2026, both dates belong in the file.

• Auditor identity. A name tied to an employer, with contact or badge ID. The signature authority that can accept work is covered in who signs off on a span, and what that signature means.

• Photos that show the condition claimed, labeled with the span ID and the finding ID. If the finding is a 9 foot radial to a 12 kV phase, the photo must show the measuring device and the object measured.

• Measurements, with method. If a clearance is cited, state whether it was measured with a laser rangefinder, a fiberglass stick, or derived from a photo with a scale in frame, and state the device model.

• The governing rule or spec section. If the alleged gap is a GO 95 Case 14 maintained clearance, cite it and link to how the rule binds this span. If a utility spec requires 12 inches of final trim on subtransmission guy strain insulators, cite that spec.

• The verdict, pass or fail, and the defect class if it failed.

• The corrective action, the owner, and the due date. If a CAPA is opened, the CAPA ID joins the span evidence.

The record obligations that come from the inspection order itself live in GO 165 inspection cycles, and the one thing they do not cover. Many utilities build their vegetation QA/QC records to sit cleanly beside those GO 165 inspection records so a reviewer can follow the chain from inspection to finding to completion without re-keying identifiers.

Defect classes and acceptance criteria, stated as pass or fail

An audit that cannot say what a fail is will not survive an outside review. Use three classes and state what each one means on a span.

• Critical defect. A condition that is an immediate safety hazard or an unambiguous code nonconformance. A measured maintained clearance below the required GO 95 distance on a 12 kV conductor is Critical. A live branch in contact where contact is not permitted is Critical. A Critical defect fails the unit and fails any lot it is sampled from.

• Major defect. A work-spec deviation that affects function, safety margin, or near-term compliance. A missed tree on a prescription. Failure to remove required slash on a PRC 4292 subject pole. A Major usually counts against acceptance limits, and beyond a stated count will fail a lot. The count and the rule that ties it to acceptance are defined in the program, not here, and the math lives in sampling versus a full census, and when each is defensible.

• Minor defect. A workmanship or documentation issue that does not affect function or compliance. A photo missing the span ID. A poorly trimmed stump outside a firebreak footprint. Minors are tracked and trended and can trigger retraining, but do not usually fail a lot unless they repeat or are widespread.

A Critical, Major, and Minor scheme aligns with cross-industry practice for acceptance sampling and audit grading. Generalized acceptance sampling rules, such as those built around ISO 2859-1 and ANSI/ASQ Z1.4, define how many Majors or Minors can be tolerated in a sample before the lot fails. The framework and worked examples for vegetation audits are in sampling versus a full census, and when each is defensible.

The standard methods those examples are based on are published by the bodies that issue them: ISO 2859-1 and ANSI/ASQ Z1.4. A common AQL decision scheme in practice sets Critical at 0, Major at 1.5, and Minor at 4.0.

On a real span, this section reads as three checkboxes beside each finding, and a decision at the end of the form. A 7 foot radial to a 12 kV bare phase where 8 feet is required inside the High Fire-Threat District is a Critical. A missed side-prune on a prescribed blue oak in the same span is a Major. A mislabeled photo is a Minor. The lot passes or fails based on the rule the program states.

How an auditor selects spans and records what they saw

Selection is the first place a QA/QC audit fails if it is casual. Define the lot, say how spans are drawn from it, and write down the randomization method used.

• One circuit, one crew, one week, is a defensible lot definition on routine distribution work. If Circuit 12JY had 480 spans worked by Crew 44 between July 15 and July 21, that is a lot size of 480 for selection.

• The draw is randomized, not haphazard. Document the method. For example, if you used a random number generator seeded by the circuit ID to pick 29 spans from a numbered list, include the seed, the list version, and the 29 IDs drawn.

• Risk-based targeting is acceptable, but it is not random. Label it that way and do not mix it into a claim of random sampling. If the HFTD Tier 3 reaches the west half of the circuit, you can stratify and oversample there, but you have to say that is what you did.

A full sampling plan, with AQLs, switching rules, and math, belongs to sampling versus a full census, and when each is defensible. This page only asks that the auditor be able to show how the 29 spans audited were picked from the 480 spans worked, on the day the question is asked.

Once in the field, record what you actually saw in a way that cannot be confused with the pre-inspection. PacifiCorp’s program illustrates why this matters. California Energy Safety’s Independent Evaluator reported that PacifiCorp has a forester sign off work as it is complete, and that the utility reviews pre-inspections when conducting audits of completed work (2022 IE report excerpt). The pre-inspection is not the audit. A QC photo and measurement taken on August 2, 2026 needs to stand on its own and be tagged as the audit, not the prescription.

Turning a finding into CAPA with a due date that can be audited

A checked span that failed is only half of the file. The other half is the corrective and preventive action that closes it, and the date on that action is what makes the program auditable.

Make the CAPA traceable:

• Assign a CAPA ID at the time of the audit. Do not leave a finding unnumbered.

• State the correction, not just the problem. For example, Side prune blue oak at midspan to restore maintained clearance to 12 kV phase to 8 feet plus one growing season, remove all cuttings from PRC 4292 footprint around P13421.

• Name an owner and a due date. If the owner is Crew 44 foreman Daniel Ruiz, and the due date is August 16, 2026, write those facts in the record. The due date is an internal program fact. It can be set by a utility rule or by business practice. What matters to the audit is that it can be checked.

• Link the CAPA record and the span record. The CAPA ID belongs on the span audit form, and the span ID belongs on the CAPA.

Preventive steps, like retraining or a spec change, also belong in the CAPA when defect patterns demand them. T&D World has argued that vegetation management needs stronger metrics to drive programs (Vegetation management metrics).

That is the same point here. A set of Majors on overhang clearance in riparian spans can be a reason for a targeted training block or a revision to the photo checklist. It is also a reason to raise the sample size on the next pass. The place where those decisions are recorded is the CAPA file, not the auditor’s notebook.

Where CAPA timelines collide with regulatory correction timelines, the regulatory clocks govern. The mechanics of inspection recordkeeping and reporting are in GO 165 inspection cycles, and the one thing they do not cover, and enforcement and plan verification roles are split as described in who enforces what: CPUC, CAL FIRE, NERC and the utility.

Calibrating auditors so two raters make the same call

Two people can look at the same span and score it differently. That is not a theory. A peer-reviewed study measured inter-rater reliability of vegetation metrics between two identically trained field crews and found variation in agreement for both categorical and continuous metrics (Forests, 2018, Evaluating Inter-Rater Reliability and Statistical Power of Vegetation Metrics). A QA program anticipates that drift and corrects for it.

Use three simple calibration tools:

• Side-by-side rides. Once a month, have two auditors ride the same ten spans independently. Score them blind, then compare. If they do not agree on the class of the same defects, pull both back for retraining with examples.

• Photo panels. Build a set of 30 photos across common defects. Require new auditors to score them before fielding, and require all auditors to re-score them quarterly. Track scores and look for drift.

• Measurement checks. If the program allows photo-derived measurements, test them against a known scale on a handful of spans each quarter. If the deviation exceeds the program’s tolerance, adjust the method or the tools.

Write the threshold for action down. For example, action if agreement on defect class across the panel falls below 85 percent, or if mean absolute error on photo-derived distances exceeds 6 inches on spans with unobstructed sight lines. Those numbers are examples. The bar your program sets is the one it will be measured against, so set it deliberately and put it in the QA plan.

What the utility receives: counts, rates, rework, and trends

A QA/QC program is not just a stack of span forms. It is a set of numbers that a manager and a regulator can read.

In a single reporting month on Circuit 12JY, the utility should expect:

• Units checked, by contractor, circuit, and crew. For example, 96 spans checked out of 480 worked on Circuit 12JY, Crew 44.

• Acceptance results at the lot level. For example, 3 lots accepted, 1 lot failed when 2 Majors were found in a 29-span sample under a 1.5 Major AQL rule. The sampling rules and math that make that decision defensible are in sampling versus a full census, and when each is defensible.

• Defect rates by class, with counts. For example, Critical 1, Major 7, Minor 18.

• Rework, measured as CAPAs opened and closed, average days to close, and rework recurrence. For example, 8 CAPAs opened, 8 closed in a mean of 9.4 days, with 1 repeat Major on overhang clearance in riparian zone.

• Photo completeness and measurement method use, as process indicators. For example, 100 percent of Criticals have a measurement photo, 92 percent of Majors do.

• Trends over time and place. For example, Major rates fell from 4.8 percent to 2.1 percent over two months after a training refresh, with no change in Critical rates.

Those numbers are what allow a manager to redirect crews, to target training, or to adjust cycle timing where growth rates outpace assumptions. They are also what allow an Independent Evaluator to recognize a functioning QA/QC loop.

How the QA/QC record ties back to GO 165 and WMP verification

Two points tie a vegetation QA/QC audit back to the governing structure.

First, inspection programs and their records run under the inspection order. California’s General Order 165 sets definitions for inspections, establishes record content, and requires annual reporting. Many vegetation findings are born in those inspections and turn into corrective actions in vegetation work management.

Where those worlds meet, the QA/QC record has to carry the identifiers that let a reviewer step from a GO 165 inspection to a vegetation CAPA and back. The order and its record obligations are explained in GO 165 inspection cycles, and the one thing they do not cover.

Second, wildfire mitigation plans commit the utility to vegetation practices and to QA/QC controls, and the State’s Independent Evaluator tests those commitments. California Energy Safety’s 2025 Independent Evaluator report for PacifiCorp states that its aim includes verification of WMP activities and control programs, which encompasses QA and QC (2025 IE report, PacifiCorp). The 2022 evaluation documented that PacifiCorp uses foresters to sign off completed work and reviews pre-inspections during audits of that work (2022 IE report excerpt).

Those facts illustrate what an evaluator expects to see in the field and in the file. If your QA/QC system is digital, the features matter. PacifiCorp has stated an initiative to implement new vegetation work management software (2026-2028 WMP responses). A system that ties span IDs, auditor identity, geotagged photos, and CAPA clocks in one place will make both internal and external reviews faster and less error-prone.

Finally, utility QC standards exist and are public in some jurisdictions. PG&E’s SW-0003 Quality Control - Vegetation Management for Routine Electric Distribution is one example. Your audit should cite the specific section of the standard that supports a finding. When the finding asserts a code nonconformance, cite the rule and measure to it using the methods stated in what Rule 35 actually requires. That is what lets a reviewer stand under the same conductor a month later and agree with the original call.

Further reading

• Sampling versus a full census, and when each is defensible: Choose a sample size and acceptance rules that an auditor can recalculate, and know when the answer has to be 100 percent inspection.

• Who signs off on a span, and what that signature means: Set signer authority, what a signature attests to, and how to make e-signature records defensible under ESIGN and UETA.

Frequently asked questions

What is the difference between QA and QC in utility vegetation management?

QC is the inspection of completed work on defined units like spans against specific standards, with a pass or fail recorded. QA is the system that makes QC reliable, including audit methods, rater training, defect grading rules, CAPA flow, and the reporting that proves the program worked.

What evidence should a span-level vegetation audit collect?

At minimum, the span identifier, GPS or mapped location, audit date and time, auditor identity, photos tied to the span and finding, measured distances where claimed, the governing standard cited, a pass or fail, and any corrective action with a due date and owner.

How are vegetation audit findings graded?

Most programs use three levels. Critical is an immediate hazard or code nonconformance. Major is a work-spec deviation that affects function or compliance soon. Minor is a workmanship or documentation issue. Acceptance rules decide how many Majors or Minors can pass. Critical defects fail the lot outright.

How do QA/QC audits connect to GO 165 and WMP verification?

GO 165 governs inspection programs and records, and many vegetation findings are born in those inspections. California Energy Safety’s Independent Evaluator reviews QA and QC as part of Wildfire Mitigation Plan verification, so audit methods, evidence, and results need to be ready to produce.

How do you keep different auditors from making different calls on the same span?

Run side-by-side calibration rides, score identical photos and spans blind, measure agreement, and retrain until raters converge. Published research shows inter-rater reliability can vary without calibration, so plan to measure and correct drift.