Vegetation-caused outage reporting: what actually gets counted

The only report whose subject is a vegetation-caused outage is the FAC-003-5 periodic data submittal, filed quarterly to the Regional Entity 20 days after the quarter ends, covering Sustained Outages on applicable lines. California's reliability reporting under D.16-01-008 has no vegetation cause category, and the CPUC dropped its vegetation-specific incident reporting requirement in 2006.

On November 20, 2024, a tree caused an outage in PacifiCorp's Mt. Shasta district that cost 637,202 customer minutes across 1,037 customers, roughly ten hours each. In the annual electric reliability report the company filed with the CPUC on July 15, 2025, that event appears exactly once.

It is one row in a ten-row table, described as "Tree - Non-Preventable," flagged as a Major Event, and therefore outside the SAIDI the same report gives as the utility's 2024 performance. The words "Tree - Non-Preventable" are PacifiCorp's. No decision, general order or statute defines them.

That is the ordinary condition of a vegetation-caused outage in California. Several bodies collect outage data from the same utility, none of them counts the same set of events, and only one of them is looking for vegetation in the first place.

The only report in the stack whose subject is a vegetation outage

FAC-003-5 carries it, and not as a requirement. It sits in Section C.1.4, Additional Compliance Information, under the heading Periodic Data Submittal:

The applicable Transmission Owner and applicable Generator Owner will submit a quarterly report to its Regional Entity, or the Regional Entity's designee, identifying all Sustained Outages of applicable lines operated within their Rating and all Rated Electrical Operating Conditions as determined by the applicable Transmission Owner or applicable Generator Owner to have been caused by vegetation, except as excluded in footnote 4, and including as a minimum the following: The name of the circuit(s), the date, time and duration of the outage; the voltage of the circuit; a description of the cause of the outage; the category associated with the Sustained Outage; other pertinent comments; and any countermeasures taken by the applicable Transmission Owner or applicable Generator Owner.

Six fields, then a category. Because it is in Section C rather than Section B, it has no Requirement number, no Measure, no Violation Risk Factor and no Violation Severity Level. It is a data submittal the Compliance Enforcement Authority requires under the NERC Rules of Procedure, Appendix 4C Section 4.6, not a performance obligation the way R1 or R6 is. Which lines are applicable lines decides the whole population it covers, and on most California systems that is a short list.

The standard sets no filing date. The date comes from the 2026 ERO Enterprise Periodic Data Submittals Schedule, which lists FAC-003-5 C.1.4 as a quarterly submittal to the Regional Entity due 20 days after the end of the quarter, made through the Align tool by entities that use it.

Twenty days is a short window for a determination that requires a field investigation, and the determination is the utility's own: the standard says the outages are those "as determined by the applicable Transmission Owner or applicable Generator Owner to have been caused by vegetation."

One detail in that schedule is worth checking against the standard before quoting either. The schedule reproduces the submittal text under the heading FAC-003-5 but with the version 4 wording, "except as excluded in footnote 2." In FAC-003-4 footnote 2 read, in its entirety, "Id." The exclusion has been in footnote 4 in both versions, and FAC-003-5 corrected the pointer. The document that tells a program when the report is due still quotes the sentence that pointed at the wrong footnote.

The seven categories a Sustained Outage is sorted into are 1A, 1B, 2A, 2B, 3, 4A and 4B. What the A and B letters test, and the gap version 5 opened in that test belongs with the reservation of R2, because the report kept a split the requirements gave up.

A Sustained Outage is defined by a relay, not by a clock

The NERC Glossary definition, which is the one FAC-003 uses, contains no duration at all:

The deenergized condition of a transmission line resulting from a fault or disturbance following an unsuccessful automatic reclosing sequence and/or unsuccessful manual reclosing procedure.

An outage is sustained when reclosing fails. Whether that takes 40 seconds or four hours is not part of the test.

NERC itself declined to use that definition in the other place it collects outage data, and said so in writing. The TADS System Definitions document defines a Sustained Outage as "An Automatic Outage with an Outage Duration of a minute or greater," and attaches a footnote explaining the departure:

The TADS definition of Sustained Outage is different than the NERC Glossary of Term Used in Reliability Standards definition of Sustained Outage which is presently only used in FAC-003-1. ... The definition is inadequate for TADS reporting for two reasons. First, it has no time limit that would distinguish a Sustained Outage from a Momentary Outage. Second, for a circuit with no automatic reclosing, the outage would not be "counted" if the TO has a successful manual reclosing under the glossary definition.

That second reason has aged into a California problem. Circuits with automatic reclosing cut out are now routine here for months at a time. PG&E's own utility procedure TD-2700P-26, revision 3, effective September 10, 2024, sets out the enablement steps for Enhanced Powerline Safety Settings, and step 2.2 is a three-item list that begins: navigate to the device, "CUT OUT the reclosing relay," then enable EPSS. Field personnel are dispatched to any device without SCADA to do the same by hand.

EPSS runs on distribution and on prescribed high-side devices, not on the 200 kV and above population FAC-003 reaches, so the two do not collide directly.

The point is about the definition rather than the program: the boundary between a momentary and a Sustained Outage under FAC-003 is the state of a relay, and the state of that relay is now a seasonal operating decision made for fire risk. A definition written when automatic reclosing could be assumed is being applied on a system where it is switched off on purpose.

Footnote 5 closes the gap from the other side, and it is the footnote most often missed:

If a later confirmation of a Fault by the applicable Transmission Owner or applicable Generator Owner shows that a vegetation encroachment within the MVCD has occurred from vegetation within the ROW, this shall be considered the equivalent of a Real-time observation.

So a fault that clears on reclose, investigated a week later and confirmed as vegetation inside the right of way, produces an R1 part 1.1 finding without ever producing a Sustained Outage. It is a violation that never enters the quarterly report. The encroachment count and the outage count are fed by different events, and a program that reconciles one against the other will not get them to agree.

What the count leaves out, and what it merges

Footnote 4 is the exclusion, and it is long:

This requirement does not apply to circumstances that are beyond the control of an applicable Transmission Owner or applicable Generator Owner subject to this Reliability Standard, including natural disasters such as earthquakes, fires, tornados, hurricanes, landslides, wind shear, fresh gale, major storms as defined either by the applicable Transmission Owner or applicable Generator Owner or an applicable regulatory body, ice storms, and floods; human or animal activity such as logging, animal severing tree, vehicle contact with tree, or installation, removal, or digging of vegetation.

Read the phrase "major storms as defined either by the applicable Transmission Owner or applicable Generator Owner." The exclusion that removes the largest share of vegetation outages from the count turns on a definition the reporting entity writes for itself. Whether a November wind event on the north coast was a major storm is answered by an internal document, and that document is the first thing an auditor should be given rather than the last.

A second carve-out sits in the supplemental material and never appears in the requirement text. Where a conductor is operated beyond its Rating and Rated Electrical Operating Conditions, whether through emergency action to protect an Interconnection or through ice loading, the standard says that "Such vegetation-related encroachments and outages are not violations of this standard." An outage on an over-Rating line is a fact about the loading, not about the trees.

Then the merge rule. Footnote 6, repeated by reference at footnotes 7 and 8:

Multiple Sustained Outages on an individual line, if caused by the same vegetation, will be reported as one outage regardless of the actual number of outages within a 24-hour period.

The supplemental material explains what it is for: an initial investigation may fail to find the cause, the line is re-energized, conductor temperatures climb back, and it goes out again. Three trips become one entry. Two facts have to be in the file for that entry to survive review, and neither is supplied by the rule.

The first is the finding that it was the same vegetation, which is a physical determination somebody made on the ground. The second is the start of the 24 hours, which the standard does not fix. A program that merges outages needs its own written convention for which trip starts the clock, applied the same way every quarter.

The category that gets published is the one that is not a violation

Category 3 is Sustained Outages "caused by vegetation falling into applicable lines from outside the ROW." R1 has four parts, and none of them corresponds to it: parts 1.2 and 1.3 place the fall-in and the blowing together inside the right of way, and part 1.4 covers growth with no location qualifier. Where the corridor boundary is drawn, and out of which documents therefore decides not just the severity of a finding but whether there is one, and the recorded easement is frequently not the answer.

Now look at what NERC publishes out of the submittal. Vegetation-Related Transmission Outages, 2024 Annual Report, dated May 6, 2025, is twelve pages, and its entire data section is Category 3:

• 35 Sustained Outages from outside the right of way in 2024, up from 18 in 2023, against a five-year average of about 30

• 30 of the 35, or 86 percent, attributed to weather-related events

• most of them on 230 kV lines

• nearly 69 percent in the Eastern Interconnection

There is no published count of Categories 1A, 1B, 2A, 2B, 4A or 4B. The report handles the violation side in one sentence near the end: "In the past 5 years, 7 violations have been filed that involved vegetation contact or encroachment into the line," and points the reader at the enforcement spreadsheet on NERC's website.

So the only public annual accounting of the quarterly vegetation outage report is an accounting of the one category nobody can be cited for. That has a practical consequence for anyone benchmarking a program against industry numbers: the published figure is not a floor on grow-ins and fall-ins from inside the corridor, it is silent about them.

TADS counts the same tree by a different rule

The Transmission Availability Data System is the second federal collection, run by the same organization, on an overlapping population, using incompatible definitions. Reporting is mandatory under Section 1600 of the NERC Rules of Procedure for Transmission Owners with elements at 200 kV and above.

Element	Threshold that brings it in
Overhead and underground ac circuits	At or above 200 kV
Transformers	Low side at or above 200 kV
Back-to-back converters	At or above 200 kV on both sides
dc circuits	At or above plus or minus 200 kV

Vegetation is Cause Code 13, and the differences from FAC-003 are systematic rather than incidental.

It counts momentaries. Cause Code 13 covers "Automatic Outages (both Momentary and Sustained) caused by vegetation." FAC-003's submittal covers Sustained Outages only. A limb that faults a 230 kV line and clears on reclose is a TADS entry and not a quarterly report entry.

A Sustained Outage is one minute. Not a failed reclose. The two systems will disagree about the same event whenever reclosing succeeds after more than a minute, or fails in less.

The exclusions do not delete the outage, they move it. TADS borrows FAC-003's exclusions almost verbatim, then adds a sentence FAC-003 has no equivalent of: "Outages that fall under the exclusions should be reported under another Cause Code and not the Vegetation Cause Code." Under FAC-003, an excluded vegetation outage is simply absent from the report. Under TADS it is present under a different cause, which means the total is preserved and only the attribution changes.

Its outside-the-ROW exclusion is narrower. TADS excludes outside-ROW fall-ins only where they "result from natural disasters." FAC-003 reports all of them as Category 3. The same November windstorm can produce a Category 3 line in the quarterly report and a non-vegetation cause code in TADS.

The consequence for a file is that a single 230 kV vegetation outage generates two federal entries that are supposed to describe the same event and will not reconcile on cause, on duration or on inclusion. Recording which system each entry was written for, and under which definition, is the only thing that makes the difference explainable later.

California's annual reliability report has no vegetation line in it

The state's reliability reporting comes from Decision 16-01-008, issued January 14, 2016 in Rulemaking 14-12-014. Ordering Paragraph 1 requires the investor-owned utilities to submit system level and district or division level reliability information on July 15 of each year, and Ordering Paragraph 5 requires them to use the template at Appendix B of the order.

The indices are the IEEE 1366 set, SAIDI, SAIFI, CAIDI and MAIFI, reported both including and excluding Major Event Days, with the threshold set by the 2.5 beta method at the state level and applied to each reporting region.

Nowhere in that structure is there a cause-code table. PacifiCorp's CY 2024 report runs to fifty-eight pages:

• indices by district

• ten-year histories with and without major events

• planned outage tables

• a major event summary

• worst performing circuit narratives

The word vegetation appears in exactly two places: the top ten unplanned events table, and the discussion of two circuits.

The cause vocabulary in that table is each utility's own, and two reports filed under the same ordering paragraph do not share it. PacifiCorp's 2024 top ten uses Wind, Loss of Transmission, Equipment Failure and "Tree - Non-Preventable." SDG&E's 2023 top ten uses Severe Weather, Faulted Cable, Vehicle Contact, Undetermined and Crew Error, with no vegetation term in it at all. Neither list is wrong. There is no prescribed list.

Two further features of that report decide how much of a vegetation outage survives into the reported number.

The Major Event Day removes the day, not the outage. All ten of PacifiCorp's largest 2024 unplanned events, including both tree outages, are flagged as major events, so they sit in the 312.9-minute distribution SAIDI that includes major events and not in the 123.6-minute figure that excludes them.

On the same system in 2023, SDG&E recorded no Major Event Day at all, which means every vegetation outage it had that year is inside its reported indices. The same tree on the same day is in or out depending on how bad the rest of the day was.

Fire-season settings are tracked in a separate column. PacifiCorp's note to the index tables excludes outages "extended as a result of enhanced safety settings (established 2021), emergency de-energization (established 2024), public safety power shutoff (established 2021)."

The report gives those their own figures: Enhanced Safety Settings, SAIDI 370.8 and SAIFI 2.447; emergency de-energization, SAIDI 18.9 and SAIFI 0.080. The ESS number alone is larger than the entire distribution SAIDI the utility reports for 2024 on either basis. A branch contact on an EPSS-style circuit produces a lockout that is counted in one place and an index that is reported in another.

Where vegetation does show up by name is the worst performing circuit section, and it shows up at a scale the system indices cannot see. Patrick's Creek, circuit 6R3, serves about 14 customers over seven circuit-miles in Del Norte County, inside Tier 2 of the High Fire-Threat District.

Since 2020 it has had ten vegetation-related outages, 54 percent of its total customer minutes interrupted, and it has been on the deficiency list five years running. Fourteen customers cannot move a state SAIDI. They can put a circuit in a public filing with a five-year record attached to it.

Where a California vegetation outage does get counted by name

Two places, and the more surprising fact is about the third.

The Quarterly Data Report to Energy Safety. A vegetation-caused outage is a risk event in the wildfire mitigation framework, reported in the QDR's risk event driver table, which is kept separate from the ignition driver table. Where the numbers come from is stated in the filings themselves.

SCE's Q4 2021 QDR says that to populate outage data for each driver, it "used ODRM outage cause codes," from a database that "records and catalogs outage impacts and causes, determined by the cooperation of field, operations, and engineering employees." The same filing defines its own transmission and distribution split for reporting purposes:

• at or above 65 kV is transmission

• below 65 kV is distribution

• "from an operational perspective, SCE also treats its 55 kV lines as transmission."

So the driver taxonomy in the state's wildfire outage data is a utility cause-code list assigned by people in the field, and the voltage boundary in it is an operating convention rather than a regulatory threshold. Both are worth knowing before comparing one utility's vegetation outage rate to another's. It also matters more than it looks: Energy Safety selects its field inspection locations out of the QDR, so what a program reports is what gets visited.

The utility's own program. A vegetation outage on a span is evidence about a clearance that was set on a cycle, and nothing in any of the reports above sends it back to the people who set the cycle. The finding, the clearance achieved and the date it was achieved live in a different file from the outage record, usually in a different system, and the join between them is the field no regulator requires.

And the report that used to name vegetation and stopped. Until 2006, California's electric incident reporting rules had a vegetation category. Appendix B required a utility to report incidents that "involve or allegedly involve trees or other vegetation in the vicinity of power lines and result in fire and/or personal injury whether or not inpatient hospitalization is required." That captured tree-related fires of any size, including ones under a square foot with no property damage.

PG&E, SCE and SDG&E filed Application 06-01-008 on January 10, 2006 to amend it, and in Decision 06-04-055, dated April 27, 2006, the Commission granted it and eliminated the vegetation-specific category. Incidents involving trees have since been reportable only if they meet the general criteria, which turn on injury, property damage and public attention rather than on cause.

What the incident and ignition reporting rules require now is a separate subject with its own thresholds. The boundary between the two is the trigger: an ignition report is about a fire, an outage report is about an interruption, and one tree can produce either, both or neither.

The biggest one leaves no federal record of the cause

The last count is the emergency one. Form DOE-417, the Electric Emergency Incident and Disturbance Report, is mandatory, and criterion 12 is "Loss of electric service to more than 50,000 customers for 1 hour or more," filed within six hours of the incident. A vegetation-caused transmission outage that sheds enough load meets it the same as any other.

Line K of the form is Type of Emergency: Cause, "check all that apply." The list runs:

• Unknown

• Physical Attack

• Threat of Physical Attack

• Vandalism

• Theft

• Suspicious Activity

• Cyber Event with IT and OT sub-options

• fuel supply emergencies

• generator loss or failure

• transmission equipment failure

• failure at a high voltage substation or switchyard

• Weather or Natural Disaster

• Operator Actions

• Other

There is no vegetation box. The largest vegetation-caused outage a California utility is capable of having gets filed federally as weather, as transmission equipment failure, or as "Other" with a free-text comment. The form's own instruction for Other is to "Include any description of the cause which can be publicly released," so the word tree, if it appears at all, appears in a narrative field rather than in a coded one, and it will not aggregate.

What the record has to name

An outage entry becomes useful to more than one reader only if it carries the facts each of them sorts on. Most of these are not required by any single report, and all of them are recoverable on the day and unrecoverable later.

The line, the structures and the voltage, in the identifier the outage system uses. The circuit name is what the FAC-003 submittal asks for. It is not enough to locate the vegetation. A span written as an ordered pair of structures is what lets an outage be joined to a clearance record.

The interrupting device and what it did. Whether reclosing was in service, how many attempts were made, whether it locked out, and whether it was restored automatically, by SCADA or by hand. That sequence is the whole of the Sustained Outage test under FAC-003 and it is the whole of the one-minute test under TADS, and both are unreconstructable from a duration alone.

Two durations. Time to reclose or lock out, and time to restoration. One decides the federal classification, the other drives the customer minutes in the state report.

The vegetation, as a physical finding. Species, approximate height and diameter, where it stood, and whether it was inside or outside the right of way as that width is documented. That last determination is the difference between an R1 violation and a Category 3 line, so the record should also name the source the width came from.

Grow-in, fall-in or blowing together. The three mechanisms are three different categories in the report and three different parts of R1, and they cannot be inferred from a photograph of the aftermath.

The exclusion claimed, if one is. Which footnote 4 circumstance applied, and the document that establishes it. Where the claim is a major storm, the entity's own storm definition and the data showing the event met it.

The merge, if there was one. Which trips were combined under footnote 6, the finding that the same vegetation caused them, and the convention used to start the 24 hours.

The cause code, and which taxonomy it came from. A vegetation outage carries at least three cause labels in a large California utility: a TADS cause code, an internal outage management code that feeds the reliability report and the QDR, and the FAC-003 category letter. Storing one and deriving the others later does not work, because the definitions do not map.

Every filing the event went into, with its date. Quarterly submittal, TADS, the July 15 reliability report, the QDR, a DOE-417 if it qualified, and any ignition report. The list is short and nobody keeps it, which is why the same event so often gets described five different ways by five different departments of the same company.

Assembled, one outage reads as a line another person can test:

Line 5R2, structures 41176 to 41177, 230 kV, applicable under FAC-003-5 part 4.2.1. Fault 14:07 PDT August 22, 2026; two automatic reclose attempts, lockout; restored 17:41. Douglas fir, approximately 90 feet, failed at 22 feet, base 31 feet from centerline, inside a documented 100-foot ROW per 1961 construction drawing. Mechanism: fall-in from inside the ROW. FAC-003-5 Category 2B, R1 part 1.2, Severe VSL, filed in the Q3 2026 submittal on October 20, 2026. TADS Cause Code 13, Sustained. No footnote 4 exclusion claimed. Not merged.

Nothing in that line required a decision that could only be made in an office, and every fact in it is checkable against a published definition or another entry in the same file. The alternative is the row this page opened with: a date, a district, and two words nobody else defines.

Further reading

• NERC FAC-003-5: transmission vegetation management after April 2024: The standard the quarterly report sits inside, and the requirements it does not sit under.

• FAC-003 Requirement R1 after R2 was reserved: The A and B category letters, and the gap the version 5 rewrite opened in them.

• Fire incident reporting: The other event report, on the other trigger: ignition rather than interruption.

Frequently asked questions

When is the FAC-003 quarterly vegetation outage report due?

Twenty days after the end of the quarter, per the 2026 ERO Enterprise Periodic Data Submittals Schedule, which lists FAC-003-5 C.1.4 as a quarterly submittal to the Regional Entity. The standard itself sets no date. It says only that the applicable Transmission Owner and applicable Generator Owner will submit a quarterly report to its Regional Entity or the Regional Entity's designee, and that the Regional Entity will pass the information to NERC quarterly.

What counts as a Sustained Outage under FAC-003?

The NERC Glossary definition: the deenergized condition of a transmission line resulting from a fault or disturbance following an unsuccessful automatic reclosing sequence and/or unsuccessful manual reclosing procedure. It has no time threshold. That is a different definition from the one NERC uses in the Transmission Availability Data System, where a Sustained Outage is any Automatic Outage lasting one minute or more, and the TADS definitions document says in a footnote why it declined to use the glossary version.

Does a vegetation outage caused by a storm have to be reported under FAC-003?

Not where footnote 4 applies. It removes circumstances beyond the entity's control, including major storms as defined either by the Transmission Owner, the Generator Owner or an applicable regulatory body, along with earthquakes, fires, landslides, wind shear, fresh gale, ice storms, floods and human or animal activity such as logging or vehicle contact with a tree. The storm definition that decides it is normally the utility's own, which makes it a document an auditor can ask to see.

Does California require utilities to report vegetation-caused outages to the CPUC?

Not as a vegetation category. The annual electric reliability report required by D.16-01-008 is filed each July 15 and reports IEEE 1366 indices with and without Major Event Days; cause words appear only in the top ten unplanned events table and the worst performing circuit narratives, in each utility's own vocabulary. The Commission eliminated its vegetation-specific incident reporting requirement in D.06-04-055 on April 27, 2006.

Is a tree that falls into a line from outside the right of way reportable?

Yes, as Category 3, and it is not a violation. R1 has no part matching it, so the category exists in the report and nowhere in the requirement. NERC's public annual report on the submittal covers Category 3 only: 35 such outages in 2024, up from 18 in 2023, against a five-year average of about 30.