Wildfire Black Box · California, United States · 2018
The Mendocino Complex comprised Ranch and River, two fires with separate identities and a shared operational history. Its record connects a 459,123-acre reported total, satellite observations, two different firefighter accidents and conflicting administrative date fields. Understanding those connections requires keeping each source’s scope visible.
Was the Mendocino Complex one fire?
Ranch and River were separately reported on 27 July 2018. CAL FIRE’s current archive lists Ranch near Potter Valley, northeast of Ukiah, and River near Hopland. The displayed start times are 12:05 p.m. and 1:01 p.m., respectively; those are archive timestamps rather than independently verified ignition times. [1] [2]
The later facilitated learning analysis describes how management of the two incidents evolved, including coordination between CAL FIRE and federal teams. A shared command history provides an organisational connection. It does not establish that two origins were identical or that every observation elsewhere in the complex applies to both fires. [8]
The cause module retains that separation. Ranch’s current CAL FIRE page uses the broad category Human. The 2018 annual table classifies River as Undetermined; the retrieved current River page supplies no cause field. These sources support those labelled states. A detailed ignition mechanism or a later River finding would require additional original investigation records. [1] [2] [3]
Record identity. This study has one complex-level permanent record and two component identities. Comparisons below select either the component level or the aggregate level explicitly. This makes the arithmetic auditable and prevents the same reported area entering a total twice.
How big was it, and how many structures were destroyed?
| Reporting level | Area, acres | Structures destroyed | Provenance |
|---|---|---|---|
| Ranch component | 410,203 | 246 | CAL FIRE archive; annual Table 5 agrees [1] [3] |
| River component | 48,920 | 35 | CAL FIRE archive; annual Table 5 agrees [2] [3] |
| Mendocino Complex | 459,123 | 281 | Reported aggregate in annual Table 2 [3] |
Our reconciliation: 410,203 + 48,920 = 459,123 acres, and 246 + 35 = 281 destroyed structures. Both sums match the annual aggregate. They establish numerical agreement between the selected component values and that report; we have not independently measured a spatial union of the fire boundaries.
A spreadsheet containing all three reporting rows needs a hierarchy rule. Summing the complex alongside Ranch and River would produce 918,246 acres: exactly twice the reported aggregate. The error would arise from mixing an aggregate with its components, even though each input number was copied correctly.
Ranch contributes about 89.3% of the reported area and 87.5% of the reported destroyed structures. These percentages use different denominators. To estimate a building-loss rate, an analyst would need an appropriate count of exposed buildings for each component, compatible damage definitions and a geographic matching method. Acreage supplies none of those automatically.
The structure category also matters. Both portal pages describe counts covering residential, commercial and other structures. The 281 total should therefore retain the word structures. Turning it into a count of homes, households or displaced residents would introduce a new measurement without its own evidence. [1] [2]
Known At: observations, reporting and later learning
| Event or observation date | Document clock | Evidence contribution | Reading boundary |
|---|---|---|---|
| 6 August 2018 | NASA article published 9 August | Landsat 8 OLI smoke-column observation over Ranch [4] | Image acquisition precedes publication. |
| 13 August 2018 | NASA article published 14 August | Aqua image; caption reports River contained and Ranch still growing [5] | Caption status and image time are separate fields. |
| 13 August 2018 accident | Blue Sheet event cover date; release date not established here | Early tree-strike account [6] | Preliminary account explicitly open to revision. |
| Same 13 August accident | Green Sheet event cover date; release date not established here | Expanded accident sequence and injury descriptions [7] | Informational summary; cover says Board of Review approval had not occurred. |
| 19 August 2018 entrapment | Retrospective facilitated learning analysis | Ground, aviation and management perspectives [8] | Later reconstruction is distinct from incident-time knowledge. |
NASA’s 14 August caption reports a complex area of 354,410 acres and 68% containment. We preserve that exact caption value as an intermediate report rather than silently replacing it with the eventual annual total. The caption describes the River component as contained while Ranch continued expanding northward. [5]
The evidence timeline has several clocks: acquisition, described event, document release, website update and our review. A timestamp on a PDF cover can identify the accident date without establishing when the report became publicly available. Likewise, a later website modification date cannot establish what a firefighter or resident knew in August 2018.
For any future chronology, the useful evidence unit is a claim with its source, scope and time field. Retrospective accounts can explain decision context, but placing their complete contents into an earlier knowledge state would give participants information they did not necessarily possess. This record keeps that distinction explicit.
Why the containment dates need an unresolved flag
| Component | 2018 annual Table 5: “Cont.” | Current archive: “Date Contained” |
|---|---|---|
| Ranch | 17 August 2018 [3] | 4 January 2019, 9:33 a.m. [1] |
| River | 27 September 2018 [3] | 4 January 2019, 9:18 a.m. [2] |
The January archive dates conflict with the annual component dates, while NASA’s mid-August caption already calls River contained. The reviewed sources do not reconcile those fields. We retain them as a date conflict rather than selecting one as a verified operational milestone.
CAL FIRE’s glossary distinguishes containment from control: containment describes a surrounding boundary while fire can continue burning; control describes a further status concerning escape from the line. That distinction helps interpret status reports. It does not authorize relabelling a portal field to explain this particular discrepancy. [9]
Resolving the conflict would require the original operational updates and an explanation of the archive’s field history. The currently retrieved Ranch updates page lists no available updates. A historical progression-map URL also could not be recovered during this review. Search snippets were not used to fill those gaps. [11]
13 August and 19 August: preserve the separate accident records
The Blue Sheet describes a 13 August aerial retardant drop followed by falling-tree injuries: one firefighter fatally injured and three others initially described as having minor injuries. The Green Sheet provides expanded descriptions, including major and moderate injury classifications. It remains an informational, revisable summary according to its own cover. [6] [7]
The facilitated learning analysis concerns a different event on 19 August. Its executive summary describes a wind shift and rapid growth during a firing operation that compromised escape routes; six firefighters were injured. Later sections examine communication, travel and organisational difficulties, and frame their account as learning rather than judgment. [8]
| Event | Mechanism described by the source | How this record uses it |
|---|---|---|
| 13 August | Tree debris associated with an aerial drop | Keep early and expanded injury descriptions together with their report status. |
| 19 August | Entrapment during a firing operation | Separate incident-within-incident identity and retrospective learning account. |
A source-label conflict also matters: Ranch’s current portal displays one fatality under a civilian heading, whereas the annual table places one in the firefighter column and leaves the civilian cell blank. The accident reports document the firefighter fatality. We flag the portal classification and do not infer an additional death or turn an empty table cell into a confirmed zero. [1] [3] [6] [7]
For a permanent record, each accident needs its own identifier, date, mechanism, affected-person category and source version. Combining injuries from successive accounts of the same accident would count reports instead of people. Combining different accidents without an explicit scope would also conceal which mechanism produced each outcome.
Worked calculation: coverage level 6 in litres per square metre
The Green Sheet defines coverage level using gallons applied per 100 square feet and records a requested level of six for the drop described in its sequence. We convert that nominal quantity using the U.S. gallon convention and NIST’s listed factors: 3.785412 L per U.S. gallon and 0.3048 m per international foot. [7] [10]
Our unit conversion: 100 ft² = 100 × 0.3048² = 9.290304 m². Level 6 therefore corresponds nominally to (6 × 3.785412) / 9.290304 ≈ 2.44 L/m². The gallon factor is used at the precision listed by NIST; the displayed result is rounded.
| Nominal level | U.S. gallons per 100 ft² | Calculated L/m², rounded |
|---|---|---|
| 2 | 2 | 0.81 |
| 4 | 4 | 1.63 |
| 6 | 6 | 2.44 |
| 8 | 8 | 3.26 |
This calculation gives a common unit for reading the report. A reconstruction of ground deposition would additionally require the drop distribution, canopy interception and observations on the surface. Estimating impact forces would require a separate physical model and its inputs. The converted volume alone supplies neither calculation.
Likewise, a product comparison would need formulation, concentration, fuel, application method, drying conditions and a measured response. This case study makes no effectiveness comparison and contains no shopping placements. A dimensional conversion remains valid within its stated convention without establishing equivalent performance between treatments.
Fire DNA and the limits of reconstruction
The technical record has three complementary strands. The component statistics establish the accounting hierarchy. NASA identifies a Ranch smoke column in a 6 August Landsat acquisition and separately presents modelled black-carbon transport from multiple California fires. The facilitated learning analysis supplies local decision and organisational context for 19 August. [4] [8]
Each strand has a different spatial support. A plume image depicts an atmospheric feature at an acquisition time. A regional transport model describes its modelled quantity. An accident account focuses on a particular operation. Their combination helps build questions for further analysis; it does not produce a continuous, validated reconstruction of the entire complex.
Work completed here: primary-source review; visual verification of annual Tables 2 and 5 and the Green Sheet cover; transcription of component statistics; sum and hierarchy checks; percentage calculations; a nominal coverage conversion; and two original charts. The formulas and reported inputs above make the arithmetic reproducible.
Remaining evidence gaps: original containment milestones, archive-field reconciliation, complete incident-update coverage, native perimeter time series, address-level exposure, measured drop deposition and a deduplicated incident-wide injury register. We reviewed selected sections of the 43-page facilitated learning analysis, not every witness statement or an independent legal investigation.
Future revisions should add a new evidence entry for each resolved gap and explain what changed. The older, correctly attributed reporting state remains part of the record’s history. The value of the Mendocino case lies in connecting its components, operations and observations without letting one summary number replace that structure.
Primary sources and review scope
- CAL FIRE: Ranch Fire archive. Current retrieved fields; displayed update 24 October 2022. Cause, component figures and date/category conflicts.
- CAL FIRE: River Fire archive. Current retrieved fields; displayed update 24 October 2022. Separate component identity and figures.
- CAL FIRE: 2018 Wildfire Activity Statistics. Tables 2 and 5, printed pages 1 and 9, visually verified; selected tables, not a full-report audit.
- NASA Earth Observatory: Smoke Plumes Tower Over California. Published 9 August 2018; satellite acquisition and model-product descriptions.
- NASA: Mendocino Complex Still Grows Larger. Published 14 August 2018; image acquired 13 August and intermediate caption status.
- CAL FIRE Blue Sheet: firefighter injuries and fatality. One-page preliminary report concerning 13 August 2018; release date not established here.
- CAL FIRE Green Sheet: Ranch accident. Eight-page informational report; text reviewed through safety issues and lessons on page 5; cover visually checked.
- Mendocino Complex facilitated learning analysis. Selected executive, chronology, command and learning sections concerning 19 August 2018; retrospective perspectives with pseudonyms.
- CAL FIRE glossary. Containment and control definitions; not an explanation of specific archive dates.
- NIST SP 811, Appendix B.8. Listed U.S. gallon, foot and square-foot conversion factors.
- CAL FIRE Ranch update catalogue. Retrieved 13 October 2026; no updates listed in the returned page. Records an availability gap, not a new incident fact.