{"run_id":"2026-wildfire-southern-europe","categories":[{"id":"cat_h_aerial_suppression_surge_capacity","name":"Aerial Suppression and Cross-Border Surge Capacity","channel":"H","channel_label":"Hazard","tier":"confirmed","tier_rationale":"TIER MET (issue #52, 2026-09-23): three source-type axes now meet §4.1's bar for `confirmed` — regulator_filings, academic_studies and industry_reports. academic_studies: Brotons et al. 2013 (PLOS ONE) run suppression intensity as an isolated scenario variable for Catalonia under future climate and find suppression can compensate for climate-driven fire increases; unlike Turco 2016 it is not a joint historical attribution, though it covers all suppression means, not the aerial limb alone. industry_reports: Moody's July 2026 insight names suppression capacity as the assumption that has historically capped European fire losses, and its correlated strain across countries as how tail events are made — the cross-border surge problem, not any one operator. Neither gives an effect size for the forward aerial increment. PHASE X CONSEQUENCE, RETAINED PENDING LEAD REVIEW: the §5 `emerging` confidence cap (confidence: L on every active cell), applied while this row carried a declared tier deficit, is kept; lifting it is a Phase X judgement for the leads, not a consequence of meeting the axis bar. THE ORIGINAL AXIS: regulator_filings, which is genuinely strong and on-mechanism — the Commission's rescEU page (12 permanent firefighting planes for Portugal, Spain, France, Italy, Croatia and Greece, first planes expected early 2028), the UCPM/ERCC page (880+ activations since 2002, 64 in 2025), and the Italian Protezione Civile AIB page documenting COAU coordination and its deployment criteria. WITHDRAWN AT ITERATION 2 (critic M7 and M1): industry_reports is withdrawn because all three citations — h_aerial_suppression_operators, h_avincis_spain_contracts and h_avincis_ancala_ownership — are one constituent's (I009 Avincis) own disclosures of its asset base, and §4.1 requires sources discussing the category mechanism, \"not just constituent ventures\". Fleet counts and contract wins establish that Avincis exists and is large, not that aerial suppression reduces insured loss. academic_studies is withdrawn because Turco et al. 2016 is a joint attribution across suppression AND prevention over 1985–2011 — mitigation that is inside L_0 by construction — and therefore cannot evidence the forward increment this category is now scored on (see effectiveness.definition). Turco is retained in effectiveness.sources as a ceiling on the historical stock, not as an axis. Investment EUR 0: availability contracts and state fleet budgets are not disclosed per-venture anywhere in this pilot's evidence base; the absence was checked against all five constituent records and is a disclosure property, not a small-scale signal.","definition":"Physical suppression delivery from the air, and the institutional arrangements that make surge capacity available when a member state's own fleet is overwhelmed — contracted private operators flying fixed-wing and rotary assets on availability terms, national state fleets under central coordination, the EU strategic aircraft reserve, and the cross-border mutual-aid mechanism that mobilises another member state's idle capacity. The mechanism is HAZARD: water and retardant delivered onto the fire front arrest spread before the fire reaches built assets, and surge capacity acts precisely in the tail events that generate the largest insured wildfire losses.","cluster_rationale":"All five ventures deliver or coordinate the same physical act: putting suppression resource onto a fire that is already burning, at a scale the affected authority could not otherwise reach. I009 and I010 are the private operators that fly the aircraft under availability contracts; I031 is the Italian state fleet under COAU central coordination; I022 is the EU-funded strategic aircraft reserve; I023 is the mechanism that mobilises another member state's idle capacity into the burning one. Actor type varies (private operator, national fleet, supranational reserve, coordination platform) and is deliberately not the clustering variable — capacity delivered onto the fire front is. Boundary vs cat_h_detection_forecasting_decision_support: that category decides where and when suppression is applied; this one applies it. Boundary vs cat_h_fire_management_programmes_and_statutes: the programmes category is the budget line and the legal instrument; the aircraft, crews and mutual-aid activations are here. That boundary is a DOUBLE-COUNT RISK and is handled by compressing the programmes category's addressable share (see its cluster_rationale) rather than by moving ventures. Boundary vs cat_h_landscape_fuel_management: fuel treatment changes how the fire behaves before it arrives; suppression acts on the fire that is already running.","members":["I009","I010","I022","I023","I031"],"member_names":["Avincis","Eliance","rescEU wildfire fleet","EU Civil Protection Mechanism (UCPM / ERCC)","Italian state aerial fleet and COAU coordination"],"effectiveness":{"value":0.06,"lower":0.0,"upper":0.18,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. BASELINE-NETTED, and cut hard at iteration 2 from 0.25 for that reason (critic M1). phase_00_input.yaml#fit_check.rationale binds this pilot: \"the Phase X mitigation term must count only *new* adaptation above the already-embedded historical suppression baseline\". L_0 is anchored at base_year 2026, so the ENTIRE 2026 aerial fleet — the Avincis and Eliance contracted aircraft, the Italian state fleet under COAU, and the UCPM mutual-aid arrangement as it operates today — is inside L_0. THE COUNTERFACTUAL IS THE 2026 SYSTEM AS IT STANDS, not the thinner system of earlier decades. The iteration-1 definition named its counterfactual as \"materially thinner aerial and surge capacity\", which is a sub-baseline counterfactual and is exactly what Phase 0 forbids; it is withdrawn. What remains creditable is only the FORWARD INCREMENT over 2026 capacity across the ten-year horizon: the 12 permanent rescEU aircraft (first fixed-wing expected early 2028) and the AEGIS aerial deliveries, added on top of an already-large national fleet base. Turco et al. 2016's 66% burned-area decline over 1985–2011 is NOT the anchor for this value; it is retained in sources only as a CEILING ON THE WHOLE HISTORICAL MITIGATION STOCK, of which the forward ten-year increment is a small and unmeasured residual — and its own source page states the period ends in 2011 and establishes that decline is possible while warming, not that decline is continuing. No source anywhere in this pilot quantifies a forward suppression increment. Upper 0.18 assumes the rescEU and AEGIS additions land on schedule and bite in the tail; lower 0.00 reflects delivery slippage and the correlation problem — simultaneous Southern-European seasons are exactly when surge capacity is least available. PHASE V INSTRUCTION: because effectiveness is stated as a forward increment, adoption anchors must be set on that increment (y0 near zero, step at 2028), NOT on the installed fleet, or the 2026 baseline is counted twice.","effectiveness_sources":[{"slug":"h_eu_civil_protection","title":"rescEU - European Commission, Civil Protection and Humanitarian Aid Operations","tier":"T1","publisher":"european_commission","url":"https://civil-protection-humanitarian-aid.ec.europa.eu/what/civil-protection/resceu_en","wayback_url":"https://web.archive.org/web/20260725194433/https://civil-protection-humanitarian-aid.ec.europa.eu/what/civil-protection/resceu_en","snapshot_hash":"9f3ab8d7f1dcf078c566c0c4b84d680da0e2cd69641a9888ea5a8e39c37e10c3","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/9f3ab8d7f1dcf078c566c0c4b84d680da0e2cd69641a9888ea5a8e39c37e10c3.html","publication_date":null,"fetched_at":"2026-07-31T13:01:10Z","page_exists":true},{"slug":"h_ucpm_activations","title":"EU civil protection mechanism - European Commission","tier":"T1","publisher":"european_commission","url":"https://civil-protection-humanitarian-aid.ec.europa.eu/what/civil-protection/eu-civil-protection-mechanism_en","wayback_url":"https://web.archive.org/web/20260725120352/https://civil-protection-humanitarian-aid.ec.europa.eu/what/civil-protection/eu-civil-protection-mechanism_en","snapshot_hash":"c4bd8edb6a81e03a06796714435d305263176a20252dd2793cbcc7c8eff47afe","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/c4bd8edb6a81e03a06796714435d305263176a20252dd2793cbcc7c8eff47afe.html","publication_date":null,"fetched_at":"2026-07-31T13:21:12Z","page_exists":true},{"slug":"h_italy_aib_campaign","title":"Campagna antincendio boschivo 2025 - Dipartimento della Protezione Civile","tier":"T1","publisher":"protezione_civile_italia","url":"https://www.protezionecivile.gov.it/it/approfondimento/campagna-antincendio-boschivo-2025/","wayback_url":"https://web.archive.org/web/20260414154032/https://www.protezionecivile.gov.it/it/approfondimento/campagna-antincendio-boschivo-2025/","snapshot_hash":"a0e25aa8c03d8b0909a66362638de99790ad465dac7c711d3a260264f397e051","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/a0e25aa8c03d8b0909a66362638de99790ad465dac7c711d3a260264f397e051.html","publication_date":null,"fetched_at":"2026-07-31T13:49:09Z","page_exists":true},{"slug":"turco_2016_decreasing_fires_mediterranean","title":"Decreasing Fires in Mediterranean Europe","tier":"T1","publisher":"plos_one","url":"https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0150663","wayback_url":"https://web.archive.org/web/https://journals.plos.org/plosone/article?id=10.1371%252Fjournal.pone.0150663","snapshot_hash":"e6e474d000164c94776788292e661e41ca260802d2602795974cafdaa783be2b","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/e6e474d000164c94776788292e661e41ca260802d2602795974cafdaa783be2b.html","publication_date":"2016-03-16","fetched_at":"2026-07-31T12:17:17Z","page_exists":true}],"addressable_share":0.8,"axes_hit":["regulator_filings","academic_studies","industry_reports"],"citations":[{"axis":"regulator_filings","wiki_source_id":"h_eu_civil_protection","span":"rescEU: the Commission has allocated resources to acquire 12 new firefighting planes to be based in Portugal, Spain, France, Italy, Croatia and Greece; the first planes are expected to be delivered early 2028."},{"axis":"regulator_filings","wiki_source_id":"h_ucpm_activations","span":"The EU civil protection mechanism has been activated over 880 times since 2002, 64 times in 2025; the ERCC mobilises assistance from participating states and from the rescEU strategic reserve."},{"axis":"regulator_filings","wiki_source_id":"h_italy_aib_campaign","span":"The Dipartimento della Protezione Civile coordinates the national aerial fleet through COAU; deployment is set from climatological and meteorological forecasts, the national fire-danger bulletin, historical statistics, regional fleet strength and water-source availability."},{"axis":"academic_studies","wiki_source_id":"brotons_2013_fire_suppression_climate_mediterranean","span":"Fire suppression strategies had a strong capacity to compensate for climate change effects."},{"axis":"industry_reports","wiki_source_id":"moodys_europe_record_wildfire_season_2026","span":"Suppression capacity, the quiet assumption that has historically capped European fire losses, is being stretched thin at precisely the moment fire weather is intensifying."}],"total_disclosed":0.0,"unknown_count":5,"adoption":{"y0":{"value":0.02,"lower":0.0,"upper":0.06,"unit":null},"y3":{"value":0.3,"lower":0.1,"upper":0.5,"unit":null},"y5":{"value":0.55,"lower":0.3,"upper":0.75,"unit":null},"y10":{"value":0.8,"lower":0.5,"upper":0.92,"unit":null}},"saturation_cap":0.9,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced. Aerial suppression and cross-border surge capacity are state and contracted-operator capabilities with no contractual relationship to any insurance policy in the four countries. No source ties fleet strength, surge availability or rescEU pre-positioning to a premium rate, an underwriting appetite or a capacity decision. The near-miss is worth naming so it is not mistaken for a channel: Moody's observes that suppression capacity is 'the quiet assumption that has historically capped European fire losses' - which is a statement about LOSS, i.e. I3, not about premium. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No insurer statement that suppression capacity is priced, no regional rate differential attributed to it, no reinsurance term conditioned on it. Searched: h_eu_civil_protection, h_ucpm_activations, h_italy_aib_campaign, h_aerial_suppression_operators, h_avincis_spain_contracts, h_eliance_operator, h_infoca_2026_budget, insurance_journal_2026_marsh_europe_wildfire."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Firefighting aircraft fleets, availability contracts and the EU strategic reserve do not touch insurer asset allocation, duration or investment return. The one asset-ownership document in the set - Avincis' infrastructure-fund ownership - concerns the OPERATOR's capital structure, not an insurer's portfolio, and is not a channel. Searched: h_eu_civil_protection, h_ucpm_activations, h_italy_aib_campaign, h_avincis_ancala_ownership. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.952,"confidence":"L","sources":["wiki/sources/h_eu_civil_protection.md","wiki/sources/h_ucpm_activations.md","wiki/sources/h_italy_aib_campaign.md","wiki/sources/turco_2016_decreasing_fires_mediterranean.md","yaml-ref:phase_04_5_categories.yaml#cat_h_aerial_suppression_surge_capacity.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_h_aerial_suppression_surge_capacity.addressable_share"],"rationale":"[READ THIS WITH THE NUMBER - AGGREGATE CONTEXT, repeated on each of the five rows that dominate the product so it survives being read in isolation.] The matrix total is a 22.9% claims reduction at t=10 (total_factor_I3 = 0.770928, rebased on the observed L_0 per issue #53). NOT ONE VENTURE, PROGRAMME OR STATUTE IN THIS PILOT'S POOL PUBLISHES AN AUDITED INSURED-LOSS-REDUCTION FIGURE ANYWHERE IN SPAIN, PORTUGAL, ITALY OR GREECE, so that 22.9% is a structured analyst inference over sourced mechanisms, not a measured effect. TWO BASES ARE QUOTED BELOW AND THEY ARE NOT INTERCHANGEABLE: 'marginal' is the share of the 22.9% that disappears if the named rows are removed, and marginal shares do NOT sum to 100% across rows; 'log' is each row's share of the log-reduction, which does sum. Cells at confidence: L carry 71.4% marginal / 73.9% log. The two tier: emerging rows carry 25.7% marginal / 28.3% log, and removing both moves the headline from 22.9% to 17.0%. THIS ROW CONTRIBUTES 13.1% MARGINAL / 14.7% LOG. Full decomposition in loop_metadata.aggregation_disclosure, which does not render - see loop_metadata.disclosure_rendering_note. || value 0.952 = 1 - (effectiveness 0.06 x addressable_share 0.80), Phase IV.5 parameters unchanged. MECHANISM WELL DOCUMENTED, FORWARD INCREMENT ENTIRELY UNMEASURED. The one honest source-type axis is regulator_filings and it is strong and on-mechanism: the Commission's rescEU allocation of 12 new firefighting planes for Portugal, Spain, France, Italy, Croatia and Greece with first deliveries expected early 2028; the UCPM/ERCC record of 880+ activations since 2002 and 64 in 2025; and the Italian Protezione Civile AIB page documenting COAU coordination and its deployment criteria. BASELINE NETTING, BINDING AND SEVERE: L_0 is anchored at 2026, so the ENTIRE 2026 aerial fleet - the Avincis and Eliance contracted aircraft, the Italian state fleet under COAU, and the UCPM mutual-aid arrangement as it operates today - is inside L_0. The counterfactual is the 2026 system as it stands, not the thinner system of earlier decades; the iteration-1 counterfactual 'materially thinner aerial and surge capacity' was withdrawn as a sub-baseline counterfactual that phase_00_input.yaml forbids, and this cell does not reinstate it. Only the FORWARD INCREMENT is creditable: 12 rescEU aircraft from early 2028 plus the AEGIS aerial deliveries, added on top of an already-large national fleet base. Effectiveness was cut 0.25 -> 0.06 with a lower bound of 0.00 on precisely this ground. TURCO 2016 IS NOT AN ANCHOR: its 66% burned-area decline over 1985-2011 is a joint attribution across suppression AND prevention over a period that is inside L_0 by construction; it is cited here only as a CEILING on the whole historical mitigation stock, of which the forward ten-year increment is a small and unmeasured residual, and its own source page records that the series ends in 2011 and shows that decline is possible while warming, not that decline is continuing. CONFIDENCE L DESPITE THE `confirmed` TIER LABEL: Phase IV.5 records a DECLARED TIER DEFICIT here - one honest source-type axis and EUR 0 disclosed investment - and directs Phase X to apply the section 5 `emerging` cap rather than read the tier column as rigor. The industry_reports axis was withdrawn at Phase IV.5 iteration 2 because all three of its sources are one constituent's own disclosures of its asset base: fleet counts and contract wins establish that Avincis exists and is large, not that aerial suppression reduces insured loss. THE CORRELATION PROBLEM the 0.00 lower bound encodes: simultaneous Southern-European seasons are exactly when surge capacity is least available, which Moody's independently describes as correlated hazard plus correlated response strain."},"I4":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no independent operational channel. Nothing in the corpus attributes any insurer handling cost, adjuster deployment or expense movement to aerial suppression or surge capacity. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No claims-operations consequence of suppression capability exists in the corpus. Searched: h_eu_civil_protection, h_ucpm_activations, h_italy_aib_campaign, h_aerial_suppression_operators, r_claims_visual_intelligence, moodys_europe_record_wildfire_season_2026."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","regulator_filings"],"distinct_sources":5,"total_disclosed":0.0,"unknown_count":5,"reason":"3 evidence axes across 5 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_h_aerial_suppression_surge_capacity.md"},{"id":"cat_h_landscape_fuel_management","name":"Landscape Fuel Management and Grazed Fuel-Break Networks","channel":"H","channel_label":"Hazard","tier":"confirmed","tier_rationale":"Three source-type axes documented, and this is the best-evidenced Hazard category in the pilot. regulator_filings: the SGIFR fuel-management portal (the graded clearance geometry and its vertical- and horizontal-continuity rationale), the Junta de Andalucia's RAPCA page (annual grazier contracts, field-verified shrub- and pasture-reduction objectives, payment determined by measured result), and the ANEPC Aldeia Segura portal with live settlement-level counters. ngo_reports: the REMP good- practice record of RAPCA's scaling (one shepherd to 451 graziers and 15,371 ha), the Pau Costa Foundation's project and training portfolio, and the Ramats de Foc certification scheme. academic_studies: Turco et al. 2016, on the joint suppression-and-prevention attribution basis. Investment intensity EUR 0 disclosed — these are public contracts, statutory duties and a consumer-funded certification label, none of which discloses a venture-level figure; confirmed tier does not require investment evidence. NOTE: the one peer-reviewed evaluation that exists for this mechanism (Ruiz-Mirazo, Robles and Gonzalez-Rebollar 2011, Agric. Ecosyst. Environ. 141(1-2):13-22, 86%/82% classification accuracy on grazing-objective attainment across >2,000 ha of RAPCA fuelbreaks) was NOT ingested by this pilot and is recorded only as a reference on h_grazing_fuel_reduction; it measures programme integrity, not burned area, and is not cited as an axis here.","definition":"Physical reduction and re-structuring of vegetation fuel at landscape and settlement scale so that fire arriving from the wildland meets discontinuity instead of continuous fuel — statutory landscape fuel-management networks and the managed band around settlements, engineered protection zones at the wildland-urban interface, grazed fuel-break networks maintained by contracted livestock, and the technical intermediaries that specify the work and verify it was done to standard. The mechanism is HAZARD: breaking vertical and horizontal fuel continuity lowers fireline intensity and blocks propagation before the fire reaches the built edge. The category's intended geometry stops at the settlement boundary and the landscape network; the graded band dimensioned to the individual building envelope is a separate category in the V channel. SEE cluster_rationale FOR A LIVE EVIDENCE GAP IN THAT SPLIT — I015's own sources evidence only the building band, not the landscape limb assigned to it here.","cluster_rationale":"All five ventures change the physical fuel structure between the wildland and the built edge. I015 is the statutory Portuguese duty in its landscape and settlement-band limb; I018 (Aldeia Segura) engineers the managed protection zone around rural settlements and appoints a local safety officer; I019 (RAPCA) and I020 (Ramats de Foc) maintain fuel breaks by contracted or certification-funded grazing rather than mechanical clearing — different funding mechanism, identical physical effect; I024 (Fundacio Pau Costa) is the technical intermediary that specifies and verifies the work, and is clustered here because the capability it sells is inseparable from the treatment itself. THE DISJOINTNESS CONTROL — the single largest double-counting risk in the pilot, and MATERIALLY WEAKENED AND RE-STATED AT ITERATION 2 (critic M2). Portugal's fuel-management duty is deliberately entered twice in the inventory, split by mechanism: I015 is assigned here in H the landscape fuel-management network and the managed band around settlements; I043 sits in cat_v_defensible_space_clearance carrying the graded band dimensioned to the building envelope (all surface fuels eliminated 0–2 m, shrubs and trees removed or thinned 2–10 m, canopies held at least 5 m from the structure). THAT ASSIGNMENT IS NOT EVIDENCED BY I015'S OWN SOURCES, AND THE COMPOSER STATES SO RATHER THAN ASSERTING OTHERWISE. Three records in this repo describe I015 as a BUILDING-PERIMETER measure: its inventory record is named \"Portuguese fuel-management obligation around buildings\" with tags [defensible_space, graded_clearance, statutory_obligation, wui_interface, fireline_intensity] and a summary describing graded clearance around every building; its dossier wiki/ventures/pt_fuel_management_obligation.md is the 0–2 m / 2–10 m table and states in terms that \"the 50 m / 100 m outer-strip distances … are NOT stated on this page\"; and its primary source h_sgifr_fuel_management repeats that caveat, with every key_claim tagged to I015 being a building-band prescription. The 100 m settlement band exists in this pilot's evidence ONLY as a second-hand, un-snapshotted bundle claim on v_defensible_space_portugal — which is I043's source, in the V channel. So the split is defensible as a reading of DL 82/2021 the statute, and it is NOT defensible as a reading of I015's record: the evidence for the H limb sits on the V category's source page, and the H category's own source page disclaims it. CONSEQUENCE FOR PHASE X: the H limb of the Portuguese duty carries an EVIDENCE GAP that must be surfaced, not filled — a composer scoring this category must not assume the landscape/settlement geometry is independently evidenced, and must treat the risk that I015 and I043 describe the SAME building-perimeter act as live. ESCALATED TO THE LEADS as a Phase IV correction that Phase IV.5 has no authority to make: I015's mitigation_mechanism.summary, tags and sources should be rewritten to the landscape-network / settlement-band limb, citing the DL 82/2021 bundle claim, before Phase IV.5 signs off; if Phase IV is not reopened, this paragraph is the disclosure. THE SHARE PARTITION, relabelled (critic S1): this category's addressable_share 0.55 and cat_v_defensible_space_clearance's 0.45 sum to exactly 1.00. That is NOT a derived geometric partition — addressable_share is a share of LOSS, not of treated area, and no other pair in this lattice partitions (the H channel's five shares sum to 2.55, which matrix_schema.md's multiplicative combination permits). It is a DELIBERATE CONSERVATIVE DEVICE adopted because the two limbs may in fact overlap (see the evidence gap above), and forcing them to partition guarantees the same clearance cannot be credited twice. It carries a known DOWNWARD BIAS on both categories and should be revisited if Phase IV re-cuts I015. Separately, and regardless of the partition, the two limbs share ONE compliance rate (approximately 40% non-compliance in 2024, second-hand and unverified in this pilot) and ONE enforcement machinery (municipal substitution with cost recovery), so Phase V MUST drive both categories' adoption from a single Portuguese compliance series and must not treat their adoption as independent. Boundary vs cat_h_fire_management_programmes_and_statutes: that category is the SGIFR planning cascade and the budget; this one is the executed treatment. Boundary vs cat_h_utility_corridor_ignition_prevention: that category's linear strips exist to remove an ignition source from a powerline corridor, not to create landscape discontinuity ahead of an approaching front, and its ventures are scored there.","members":["I015","I018","I019","I020","I024"],"member_names":["Portuguese fuel-management obligation around buildings (Despacho 675/2026 regime)","Aldeia Segura, Pessoas Seguras","RAPCA - Red de Areas Pasto-Cortafuegos de Andalucia","Ramats de Foc","Fundacio Pau Costa"],"effectiveness":{"value":0.08,"lower":0.0,"upper":0.25,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. BASELINE-NETTED, and cut at iteration 2 from 0.20 for that reason (critic M1). phase_00_input.yaml#fit_check.rationale binds this pilot to count \"only *new* adaptation above the already-embedded historical suppression baseline\", and L_0 is anchored at base_year 2026 — so the RAPCA network as it stands (451 graziers, 15,371 ha), the executed SGIFR fuel-management network, the 2,435 Aldeia Segura settlements and the Ramats de Foc blocks are ALL inside L_0. THE COUNTERFACTUAL IS THE 2026 TREATED NETWORK AS IT STANDS. What is creditable is only the forward increment over ten years: closing the compliance gap (Portugal's reported ~40% non-compliance, itself unverified in this pilot, is the largest single piece of headroom) and expanding the treated network beyond its 2026 extent. Turco et al. 2016's 66% burned-area decline over 1985–2011 is NOT the anchor for this value; it is retained in sources as a CEILING ON THE WHOLE HISTORICAL MITIGATION STOCK, jointly attributable to suppression and prevention and shared with cat_h_aerial_suppression_surge_capacity, of which the forward increment is a small and unmeasured residual — and its own source page states the series ends in 2011 and shows that decline is possible while warming, not that it is continuing. The remaining inference legs are (a) the physical mechanism as stated by the Portuguese regulator — eliminating surface fuels and thinning the shrub and tree layers impedes propagation vertically and horizontally — which establishes direction but not magnitude, and (b) the RAPCA scaling record, which is an ADOPTION curve, not an effectiveness measurement. Upper 0.25 assumes the compliance gap is substantially closed and the network materially extended; lower 0.00 reflects that treated networks are a small fraction of the landscape, that extreme fire weather overruns fuel breaks, and that a forward increment over an already-executed 2026 network may be nil. PHASE V INSTRUCTION: anchor adoption on the increment (compliance improvement and network extension), NOT on the existing treated extent, which is already in L_0.","effectiveness_sources":[{"slug":"h_sgifr_fuel_management","title":"Gestao de Combustivel - Portal Publico do Sistema de Gestao Integrada de Fogos Rurais (SGIFR)","tier":"T1","publisher":"sgifr","url":"https://www.sgifr.gov.pt/gestao-de-combustivel","wayback_url":"https://web.archive.org/web/20260520011327/https://www.sgifr.gov.pt/gestao-de-combustivel","snapshot_hash":"b885ec69452410c500ce2358981afd70959be9e9105a69bfe4aa7aba42572ba1","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/b885ec69452410c500ce2358981afd70959be9e9105a69bfe4aa7aba42572ba1.html","publication_date":null,"fetched_at":"2026-07-31T13:12:59Z","page_exists":true},{"slug":"h_grazing_fuel_reduction","title":"Red de Areas Pasto-Cortafuegos de Andalucia (RAPCA)","tier":"T1","publisher":"junta_de_andalucia","url":"https://www.juntadeandalucia.es/medioambiente/portal/areas-tematicas/medio-forestal/incendios-forestales/prevencion-y-participacion-frente-incendios-forestales/red-de-areas-pasto-cortafuegos-de-andalucia","wayback_url":"https://web.archive.org/web/20231113130242/https://www.juntadeandalucia.es/medioambiente/portal/areas-tematicas/medio-forestal/incendios-forestales/prevencion-y-participacion-frente-incendios-forestales/red-de-areas-pasto-cortafuegos-de-andalucia","snapshot_hash":"839734112e725b087e2f6475c9ff9fa3f58cd377c2808636345c08cc22a9526f","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/839734112e725b087e2f6475c9ff9fa3f58cd377c2808636345c08cc22a9526f.html","publication_date":null,"fetched_at":"2026-07-31T13:01:06Z","page_exists":true},{"slug":"h_rapca_remp_corroboration","title":"Red de Areas Pasto-Cortafuegos. Andalucia (REMP buena practica)","tier":"T2","publisher":"remp","url":"https://www.remp.es/buenas-practicas/listado/red-de-areas-pasto-cortafuegos-andalucia","wayback_url":"https://web.archive.org/web/https://www.remp.es/buenas-practicas/listado/red-de-areas-pasto-cortafuegos-andalucia","snapshot_hash":"8b3f6ba0eb6e7220728ef3d49110fdd434680e077a0d6e4f98501b29b512b3a8","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/8b3f6ba0eb6e7220728ef3d49110fdd434680e077a0d6e4f98501b29b512b3a8.html","publication_date":null,"fetched_at":"2026-07-31T13:44:09Z","page_exists":true},{"slug":"turco_2016_decreasing_fires_mediterranean","title":"Decreasing Fires in Mediterranean Europe","tier":"T1","publisher":"plos_one","url":"https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0150663","wayback_url":"https://web.archive.org/web/https://journals.plos.org/plosone/article?id=10.1371%252Fjournal.pone.0150663","snapshot_hash":"e6e474d000164c94776788292e661e41ca260802d2602795974cafdaa783be2b","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/e6e474d000164c94776788292e661e41ca260802d2602795974cafdaa783be2b.html","publication_date":"2016-03-16","fetched_at":"2026-07-31T12:17:17Z","page_exists":true}],"addressable_share":0.55,"axes_hit":["academic_studies","ngo_reports","regulator_filings"],"citations":[{"axis":"regulator_filings","wiki_source_id":"h_sgifr_fuel_management","span":"Fuel management is defined as the reduction of vegetal and woody material so as to impede and reduce fire propagation both vertically, step by step from the herbaceous stratum to the canopies, and horizontally along the different strata."},{"axis":"regulator_filings","wiki_source_id":"h_grazing_fuel_reduction","span":"RAPCA contracts graziers annually to control vegetation on the forest's fire infrastructure; field staff of the Agencia de Medio Ambiente y Agua test attainment of the specified shrub- and pasture-reduction objectives and payment and its amount are determined by the measured result."},{"axis":"regulator_filings","wiki_source_id":"h_aldeia_segura","span":"Live programme counters at fetch: 2,435 settlements involved, 2,280 with a local safety officer, 1,057 evacuation plans, 735 drills, 1,582 shelter locations and 1,576 refuge locations, collected from the respective municipalities."},{"axis":"ngo_reports","wiki_source_id":"h_rapca_remp_corroboration","span":"RAPCA integrates extensive livestock as a remunerated service for preventive silviculture; controlled grazing in strategic areas creates discontinuity in the forest mass which impedes fire propagation. The network went from one shepherd to nearly 70 contracted in 2009 and 15,371 ha with 451 graziers by 2012."},{"axis":"ngo_reports","wiki_source_id":"h_pau_costa_foundation","span":"The foundation designs and coordinates integrated fire-management interventions and publishes an annual Memoria; Ramats de Foc is listed among its named projects as a scheme that manages wildfire risk through grazing."},{"axis":"ngo_reports","wiki_source_id":"h_ramats_de_foc","span":"The scheme sustains grazing in strategic forest zones through a consumer-facing certification label, with named participating graziers and municipalities published and a retail channel for labelled products."},{"axis":"academic_studies","wiki_source_id":"turco_2016_decreasing_fires_mediterranean","span":"Burned area fell 66% over 1985–2011 in Mediterranean Europe, attributed to increased fire-management and prevention effort — the peer-reviewed basis, jointly with suppression, for crediting fuel management with any effect at all."}],"total_disclosed":0.0,"unknown_count":5,"adoption":{"y0":{"value":0.03,"lower":0.0,"upper":0.08,"unit":null},"y3":{"value":0.15,"lower":0.05,"upper":0.28,"unit":null},"y5":{"value":0.25,"lower":0.1,"upper":0.4,"unit":null},"y10":{"value":0.42,"lower":0.15,"upper":0.58,"unit":null}},"saturation_cap":0.6,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced. Landscape fuel management and grazed fuel-break networks are public contracts, statutory duties and a consumer-funded certification label; none is priced into an insurance policy anywhere in the four countries. RAPCA's contracts pay GRAZIERS on field-verified fuel reduction - the only outcome-linked payment mechanism in the pilot - and the payer is the Junta de Andalucia, not an insurer. Ramats de Foc's price signal reaches a CONSUMER through a food label, not a policyholder through a premium. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No premium credit, rate relativity or underwriting-appetite statement attached to treated-landscape proximity in ES/PT/IT/GR. Searched: h_sgifr_fuel_management, h_grazing_fuel_reduction, h_rapca_remp_corroboration, h_aldeia_segura, h_aldeia_segura_sgifr, h_ramats_de_foc, h_pau_costa_foundation, v_insurer_wildfire_initiatives, insurance_journal_2026_marsh_europe_wildfire."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Fuel-break networks, grazing contracts and settlement protection zones do not touch insurer asset allocation, duration or investment return. Searched: h_sgifr_fuel_management, h_grazing_fuel_reduction, h_rapca_remp_corroboration, h_aldeia_segura, h_ramats_de_foc. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.956,"confidence":"M","sources":["wiki/sources/h_sgifr_fuel_management.md","wiki/sources/h_grazing_fuel_reduction.md","wiki/sources/h_rapca_remp_corroboration.md","wiki/sources/h_aldeia_segura.md","wiki/sources/h_pau_costa_foundation.md","wiki/sources/h_sgifr_activity_report_2024.md","wiki/sources/turco_2016_decreasing_fires_mediterranean.md","yaml-ref:phase_04_5_categories.yaml#cat_h_landscape_fuel_management.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_h_landscape_fuel_management.addressable_share"],"rationale":"value 0.956 = 1 - (effectiveness 0.08 x addressable_share 0.55), Phase IV.5 parameters unchanged. THE BEST-EVIDENCED HAZARD CATEGORY IN THE PILOT ON MECHANISM AND THE MOST SEVERELY BASELINE-NETTED ON MAGNITUDE. Three genuine on-mechanism axes. regulator_filings: the Portuguese SGIFR portal states the physical mechanism directly - fuel management is the reduction of vegetal and woody material so as to impede and reduce fire propagation both VERTICALLY, step by step from the herbaceous stratum to the canopies, and HORIZONTALLY along the different strata; Andalusia's RAPCA page documents annual grazier contracts in which field staff test attainment of specified shrub- and pasture-reduction objectives, with PAYMENT AND ITS AMOUNT DETERMINED BY THE MEASURED RESULT - the only outcome-linked mitigation contract anywhere in this pilot; the ANEPC Aldeia Segura portal carries live counters of 2,435 settlements involved, 2,280 with a local safety officer, 1,057 evacuation plans and 735 drills. ngo_reports: the REMP good-practice record of RAPCA's scaling from one shepherd to 451 graziers and 15,371 ha by 2012; Pau Costa Foundation's intervention and training portfolio; the Ramats de Foc certification scheme. academic_studies: Turco et al. 2016, on the joint suppression-and-prevention attribution basis. WHAT IS STILL MISSING, AND IT IS THE THING THAT MATTERS: not one of these measures avoided insured loss. The RAPCA scaling record is an ADOPTION curve; the SGIFR statement establishes DIRECTION but not magnitude; Aldeia Segura's counters are programme throughput. The one peer-reviewed evaluation that exists for this mechanism (Ruiz-Mirazo, Robles and Gonzalez-Rebollar 2011, 86%/82% classification accuracy on grazing-objective attainment across >2,000 ha of RAPCA fuelbreaks) was NOT ingested by this pilot, is recorded only as a reference on h_grazing_fuel_reduction, and measures PROGRAMME INTEGRITY, not burned area - it is not an effectiveness source and is not used as one here. BASELINE NETTING, BINDING: the RAPCA network as it stands, the executed SGIFR fuel-management network, the 2,435 Aldeia Segura settlements and the Ramats de Foc blocks are ALL inside L_0 at 2026. What is creditable is only the forward increment - raising compliance beyond its 2026 level and extending the treated network beyond its 2026 extent. THE COMPLIANCE QUANTITY, CORRECTED AT ITERATION 1: the iteration-0 text called it 'roughly 40% non-compliance, second-hand and unverified'. It is neither. The T1 primary h_sgifr_activity_report_2024 (AGIF, Relatorio de Atividades do SGIFR 2024) records that the GNR signalled 10,256 breaches in 2024, monitored all of them and found about 40% not regularised after notification, so voluntary compliance with fuel management reached 60% - and that 40% is the non-regularisation rate among already-detected breaches, not a population rate, which the page's own caveat 1 names as the wrong denominator. The 60% is itself bounded: the GNR's 2024 enforcement footprint was 991 freguesias, 32% of the mainland surface and 2 million hectares of forest space, so it describes the inspected subset. This is the same source phase_05_adoption.yaml quotes to set THIS category's y0 anchor of 0.03. Effectiveness was cut 0.20 -> 0.08, lower bound 0.00, and this cell does not walk it back. Turco 2016 is a CEILING on the whole historical stock, shared with cat_h_aerial_suppression_surge_capacity, never the anchor. CONFIDENCE M: three genuine axes and a T1 regulator statement of the physical mechanism support the DIRECTION; the magnitude is Phase IV.5's signed-off analyst inference, so M and not H - no source in this pilot quantifies an insured-loss outcome for any mechanism, and H on a positive mitigation effect is unavailable anywhere in this matrix. THE LIVE ATTRIBUTION GAP THIS CELL DOES NOT PAPER OVER: I015 (the Portuguese fuel-management duty) is assigned here on its LANDSCAPE and SETTLEMENT-BAND limb, but all three of I015's own records describe a BUILDING-PERIMETER measure, and its dossier states in terms that the 50 m / 100 m distances are not on the page. The risk that this category and cat_v_defensible_space_clearance describe the SAME building-perimeter act is LIVE, and the 0.55/0.45 addressable-share partition between them is a deliberate conservative device with a known downward bias, not a derived geometry. AND THE DEPENDENCY IS NOT ONLY IN THE SHARES: wiki/categories/cat_h_landscape_fuel_management.md states as a binding Phase IV.5 instruction that this category and cat_v_defensible_space_clearance must be driven from a single Portuguese compliance series and must NOT be treated as independent in adoption. The matrix_schema.md section 6 Pi-product nevertheless multiplies their two effective factors as though they were independent, so a misread of that one series would propagate to both adoption anchors at once. That is a known and unmodelled breach of the v0.7 cross-category-independence assumption (section 8), disclosed here rather than left implicit."},"I4":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no independent operational channel. A near-miss worth naming so it is not mistaken for one: RAPCA's field-verification staff and Aldeia Segura's local safety officers are PUBLIC administrative costs incurred to run the programmes, not insurer operating costs, and they are costs incurred rather than avoided. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No insurer handling cost or expense movement attributed to fuel treatment. Searched: h_sgifr_fuel_management, h_grazing_fuel_reduction, h_rapca_remp_corroboration, h_aldeia_segura, h_pau_costa_foundation, r_claims_visual_intelligence."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","ngo_reports","regulator_filings"],"distinct_sources":7,"total_disclosed":0.0,"unknown_count":5,"reason":"3 evidence axes across 7 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_h_landscape_fuel_management.md"},{"id":"cat_h_fire_management_programmes_and_statutes","name":"National and Regional Fire-Management Programmes and Statutes","channel":"H","channel_label":"Hazard","tier":"confirmed","tier_rationale":"TIER MET (issue #52, 2026-09-23): three source-type axes — regulator_filings, academic_studies and ngo_reports. academic_studies: Bacciu et al. 2025 (iForest), a peer-reviewed review of Italian wildfire governance — the Law 353/2000 AIB planning framework and the recent restructuring of the governance system — set against Portugal's post-2017 AGIF reform. ngo_reports: the EEA Climate-ADAPT case study (March 2026) on Portugal's Integrated Rural Fire Management System, describing AGIF's cross-governmental coordination mandate; third-party, though it overlaps the AGIF mandate page on the regulator axis. Both evidence the governance mechanism; neither gives an effect size. PHASE X CONSEQUENCE, RETAINED PENDING LEAD REVIEW: the §5 `emerging` confidence cap (confidence: L on every active cell), applied while this row carried a declared tier deficit, is kept; lifting it is a Phase X judgement for the leads, not a consequence of meeting the axis bar. THE ORIGINAL AXIS: regulator_filings, which is genuinely strong and on-mechanism — AGIF's institutional mandate page (planning, strategic coordination and evaluation of the whole SGIFR), the Andalusian annual INFOCA plans published as separately downloadable instruments, the Greek AEGIS programme pages and its ESPA budget item, and Legge 353/2000 as published by the Italian Parliament. WITHDRAWN AT ITERATION 2 (critic M4): academic_studies is withdrawn on an internal contradiction — this category's effectiveness is defined as the GOVERNANCE INCREMENT ONLY, explicitly excluding the operational capabilities these programmes fund, yet Turco et al. 2016 attributes its measured decline to \"an increased effort in fire management and prevention\", i.e. precisely those excluded operational capabilities. A source cannot evidence a category and simultaneously evidence the thing that category defines itself as excluding. The second academic citation, the JRC/EFFIS 2024 advance report, was cited on the composer's own admission that it is a DENOMINATOR — a burned-area series against which execution is a proxy — and a denominator is not a discussion of the mechanism. industry_reports is withdrawn because its single citation, the INFOCA satellite-tracking trade item, is about a region adopting OPERATIONAL DETECTION TECHNOLOGY, which is the mechanism of cat_h_detection_forecasting_decision_support. WHY NOT `emerging`: the EUR 2.1 bn AEGIS figure would clear §4.2(b) arithmetically, but it is an ALL-HAZARD civil- protection capital programme, not a wildfire budget (phase_04_venture_inventory.md §1 says so in terms), and §6 contemplates pro-rating a programme budget by programme-area share where a line item is available. The Axis-2 line-item amount has NOT been retrieved. Clearing a rigor gate with an un-pro-rated all-hazard sovereign figure would be precisely the inflation this iteration is correcting elsewhere, so the `emerging` route is REFUSED. Retrieving the Axis-2 line item was the precondition for revisiting; revisited 2026-09-23 (issue #52): the programme's published axis split has no wildfire-only axis, and the wildfire-only line items located are aircraft and helicopter purchases — the operational capability this category excludes — so the route stays refused, and the category instead meets `confirmed` on three axes. Stated plainly: the governance increment has no effect-size evidence of any kind — the axes establish that the mechanism is real and independently studied, not how large it is — only a disclaimed-double-count argument for why it must be small.","definition":"The state instruments that set, fund, legally compel and evaluate the level of wildfire-management capability in a jurisdiction: national integrated rural-fire management systems and their planning cascades, annually re-authorised regional civil-protection plans, multi-billion national capital programmes financed from EU recovery instruments, and the framework statutes that make prevention obligations and land-use constraints enforceable. The mechanism is HAZARD and it is institutional rather than technological: in this geography the level of detection, fuel treatment and suppression available in any given year is set by a budget decision and a plan approval, not by the diffusion of a product. SECOND, SEPARATE LIMB — NOT PART OF THE GOVERNANCE MECHANISM (critic J2): the category also carries Italy's Legge 353/2000, whose mitigating effect runs through ARSON DETERRENCE — freezing the land-use classification of burned land for fifteen years and banning construction on it for ten removes the development incentive behind deliberate ignition. That is a distinct causal pathway, not a governance increment, and its magnitude is unquantified in every source in this pilot. Phase X must treat the arson-deterrence limb as `evidence_gap: true` rather than as a component of this category's effectiveness value.","cluster_rationale":"All six ventures are state instruments whose mechanism is to determine, by plan, budget or statute, how much wildfire-management capability exists in a jurisdiction and whether it operates as a system. I011 (AGIF) plans, coordinates and evaluates Portugal's rural-fire system; I014 (SGIFR/PNGIFR) is the planning cascade that specifies where fuel is managed and which fuel-break networks are built; I012 and I013 are the annually re-approved Spanish regional plans that fund treatments, lookout networks and aerial means; I017 (AEGIS) is the Greek capital programme that buys the equipment. WEAKEST SEAM IN THE LATTICE, flagged deliberately for the critic: I016 (Legge 353/2000) is included here on the ground that it is Italy's framework statute for wildfire and its mitigating limb operates at the same institutional level — legal compulsion that changes how many ignitions occur — but its proximate mechanism is arson deterrence through land-use freeze, which is not the same act as planning and funding a fire-management system. It was not stood up as its own category because a single-venture category with one source-type axis and zero disclosed funding tiers `insufficient` under §4.3, and moving it to considered_but_excluded would orphan it against validate.py's no-orphan rule. See considered_but_excluded, \"Statutory arson deterrence and post-fire land-use constraint\". Note that the arson-deterrence pathway is now named as a SEPARATE, separately-unquantified limb in `definition` and is carried to Phase X as `evidence_gap: true` rather than as part of the effectiveness value (critic J2). DOUBLE-COUNT CONTROL: the capabilities these programmes procure are scored elsewhere and this category is scored on the governance increment only, with addressable_share held to 0.40 and effectiveness cut to 0.04 at iteration 2 for that reason plus baseline netting. INVESTMENT CAVEAT (critic S6): the EUR 2.1 bn is I017 AEGIS, which phase_04_venture_inventory.md §1 records as \"all-hazard, not a wildfire budget\"; it is NOT pro-rated because the Axis-2 line item has not been retrieved, it is reported for transparency only, and it is expressly NOT used as a rigor gate — see tier_rationale for why the `emerging` route was refused. Boundary vs cat_h_landscape_fuel_management: I015 (the Portuguese fuel-management obligation) and I018 (Aldeia Segura) are the executed physical treatments delivered under the SGIFR cascade and sit there, not here; this category carries the cascade, not the clearance. Boundary vs cat_v_defensible_space_clearance: that category is the building-perimeter duty on the private owner; this one is the state's own planning and funding apparatus.","members":["I011","I012","I013","I014","I016","I017"],"member_names":["AGIF - Agencia para a Gestao Integrada de Fogos Rurais","Plan INFOCA (Andalusia)","Plan INFOCAM (Castilla-La Mancha)","SGIFR - Sistema de Gestao Integrada de Fogos Rurais (PNGIFR / PNA)","Legge 353/2000 and the municipal catasto of burned areas (Italy)","AIGIS national civil protection programme (Greece)"],"effectiveness":{"value":0.04,"lower":0.0,"upper":0.12,"unit":null},"effectiveness_definition":"ANALYST-INFERRED, set low to prevent double-counting, and BASELINE-NETTED — cut at iteration 2 from 0.10 (critic M1). Two constraints compress this value. FIRST, double-count control: the value covers the GOVERNANCE INCREMENT ONLY — coherent planning, execution monitoring, annual re-authorisation — over and above the operational capabilities these programmes fund, which are scored in cat_h_detection_forecasting_decision_support, cat_h_aerial_suppression_surge_capacity and cat_h_landscape_fuel_management. The aircraft, sensors and fuel treatments bought with the AEGIS EUR 2.1 bn and with INFOCA's annual budget must NOT be credited twice. The arson-deterrence limb of Legge 353/2000 is a separate pathway and is NOT inside this value — carry it as `evidence_gap: true` (see definition). SECOND, baseline netting per phase_00_input.yaml#fit_check.rationale: L_0 is anchored at base_year 2026, so AGIF, the SGIFR cascade, the annually re-authorised INFOCA and INFOCAM plans and Legge 353/2000 as they operate TODAY are all inside L_0. The counterfactual is the 2026 governance system as it stands, and only the forward increment over it — further execution-monitoring maturity, the remaining AEGIS delivery schedule — is creditable. Turco et al. 2016 is NOT used as an anchor here; it attributes its decline to the operational capabilities this category explicitly disclaims. No source quantifies the governance increment at all. Upper 0.12 reflects Portugal's post-2017 system rebuild, whose headline effectiveness evidence (the reported prevention-spend shift and ignition decline) is currently UNCITED in this pilot and is the second-highest-priority outstanding retrieval; lower 0.00 reflects Italy, where the national instrument is a framework law plus a shared aircraft pool with no standing national prevention budget, and reflects that a governance increment over an already-functioning 2026 system may be nil. PHASE V INSTRUCTION: anchor adoption on the forward increment (y0 near zero), not on the existence of the programmes, which is already in L_0.","effectiveness_sources":[{"slug":"h_programmes_portugal","title":"AGIF - Agencia para a Gestao Integrada de Fogos Rurais | Homepage","tier":"T1","publisher":"agif","url":"https://www.agif.pt/pt","wayback_url":"https://web.archive.org/web/https://www.agif.pt/pt/","snapshot_hash":"bead6193993ccc698ace2f0b13f112f3f5629869b406efe20c69ce449dad4323","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/bead6193993ccc698ace2f0b13f112f3f5629869b406efe20c69ce449dad4323.html","publication_date":null,"fetched_at":"2026-07-31T13:02:09Z","page_exists":true},{"slug":"h_programmes_spain","title":"Planes Anuales para la Prevencion, Vigilancia y Extincion de Incendios Forestales de la Comunidad Autonoma de Andalucia","tier":"T1","publisher":"junta_de_andalucia","url":"https://www.juntadeandalucia.es/medioambiente/portal/areas-tematicas/medio-forestal/incendios-forestales/plan-infoca-marco-referencia-frente-incendios-forestales/planes-anuales-prevencion-vigilancia-extincion-incendios-forestales-andalucia","wayback_url":"https://web.archive.org/web/20231113130112/https://www.juntadeandalucia.es/medioambiente/portal/areas-tematicas/medio-forestal/incendios-forestales/plan-infoca-marco-referencia-frente-incendios-forestales/planes-anuales-prevencion-vigilancia-extincion-incendios-forestales-andalucia","snapshot_hash":"73648a6f2ba23f4408dec8c8219dc20ba6490861ce75f62e65aaf14f97cb7283","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/73648a6f2ba23f4408dec8c8219dc20ba6490861ce75f62e65aaf14f97cb7283.html","publication_date":null,"fetched_at":"2026-07-31T13:01:00Z","page_exists":true},{"slug":"h_aegis_espa_budget","title":"Parousiasi tou Programmatos Politikis Prostasias 'Aigis' (ESPA news item)","tier":"T1","publisher":"espa_greece","url":"https://www.espa.gr/el/Pages/NewsFS.aspx?item=1754","wayback_url":"https://web.archive.org/web/20260212230245/https://www.espa.gr/el/Pages/NewsFS.aspx?item=1754","snapshot_hash":"81abb93e0fdb448d015bd04ada2ecc121ffee213cad95e3dcba0189d2bdb4b7a","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/81abb93e0fdb448d015bd04ada2ecc121ffee213cad95e3dcba0189d2bdb4b7a.html","publication_date":"2024-04-25","fetched_at":"2026-07-31T13:11:09Z","page_exists":true},{"slug":"h_programmes_italy","title":"Legge 21 novembre 2000, n. 353 - Legge-quadro in materia di incendi boschivi","tier":"T1","publisher":"parlamento_italiano","url":"https://www.parlamento.it/parlam/leggi/00353l.htm","wayback_url":"https://web.archive.org/web/https://www.parlamento.it/parlam/leggi/00353l.htm","snapshot_hash":"050cf572a9e77e52e2bd82a93f06b0445472688b5a8d7a177e2ffac18d3e3338","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/050cf572a9e77e52e2bd82a93f06b0445472688b5a8d7a177e2ffac18d3e3338.html","publication_date":"2000-11-30","fetched_at":"2026-07-31T13:06:42Z","page_exists":true}],"addressable_share":0.4,"axes_hit":["regulator_filings","academic_studies","ngo_reports"],"citations":[{"axis":"regulator_filings","wiki_source_id":"h_programmes_portugal","span":"AGIF's mission is the planning, strategic coordination and evaluation of the SGIFR, integrating public policy effects on vegetation-fuel accumulation, population behaviour and the activity of system agents, with prevention and suppression monitoring."},{"axis":"regulator_filings","wiki_source_id":"h_programmes_spain","span":"Andalusia publishes the Plan Anual para la Prevención, Vigilancia y Extinción de Incendios Forestales as separate downloadable instruments for 2024 and 2025 — INFOCA is re-authorised and re-scoped every single year."},{"axis":"regulator_filings","wiki_source_id":"h_aegis_espa_budget","span":"The AEGIS equipment programme is EUR 2.1 bn, financed from the Recovery Fund, ESPA and the European Investment Bank, covering aerial and ground response means, coordination equipment and early-warning and prevention systems."},{"axis":"regulator_filings","wiki_source_id":"h_programmes_italy","span":"Legge 353/2000 art. 10: burned woodland and pasture may not change land-use classification for at least fifteen years; construction is banned for ten; any sale contract within fifteen years that fails to recall the constraint is void."},{"axis":"regulator_filings","wiki_source_id":"h_programmes_greece","span":"The AEGIS national civil-protection programme is structured in four axes — infrastructure upgrade, early-warning and prevention systems, support and coordination equipment, and operational response means."},{"axis":"academic_studies","wiki_source_id":"bacciu_2025_italian_wildfire_governance_holistic","span":"In recent years, Italy has significantly restructured its wildfire governance system, highlighting both the strengths and limitations of shifting paradigms in fire management in Europe."},{"axis":"ngo_reports","wiki_source_id":"climate_adapt_2026_portugal_irfms_governance","span":"The strategic coordination and monitoring of the new system is under the responsibility of AGIF (Agency for Integrated Rural Fire Management), that facilitates coordination of public policies, programs and initiatives related to the implementation of IRFMS."}],"total_disclosed":2100000000.0,"unknown_count":5,"adoption":{"y0":{"value":0.02,"lower":0.0,"upper":0.06,"unit":null},"y3":{"value":0.22,"lower":0.08,"upper":0.4,"unit":null},"y5":{"value":0.4,"lower":0.18,"upper":0.58,"unit":null},"y10":{"value":0.55,"lower":0.25,"upper":0.7,"unit":null}},"saturation_cap":0.7,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced. National and regional fire-management programmes are budget and legal instruments; none of the four countries links plan approval, programme funding or statutory compliance to an insurance premium, a rate relativity or an underwriting condition. THE ONE INSTRUMENT THAT COMES CLOSEST IS OUT OF GEOGRAPHY AND IS NOT A PREMIUM: France's obligation legale de debroussaillement carries a contractual price through a DEDUCTIBLE ON INDEMNITY, which is (a) an indemnity term, not a premium term, and (b) France, excluded by phase_00_input.yaml.out_of_scope. It is recorded on cat_v_defensible_space_clearance as a comparator, and in neither place as a premium channel. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No programme- or statute-linked premium effect exists in ES/PT/IT/GR. Searched: h_programmes_portugal, h_programmes_spain, h_programmes_italy, h_programmes_greece, h_aegis_espa_budget, h_agif_institutional_report_2024, h_sgifr_activity_report_2024, h_infoca_2026_budget, insurance_journal_2026_marsh_europe_wildfire."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. State fire-management budgets and framework statutes do not touch insurer asset allocation, duration or investment return. The nearest adjacency in the corpus is sovereign fiscal stress from disaster expenditure (eea_eucra_2024 records EU solidarity funds critically threatened by oversubscription), which is a public-finance observation, is not wildfire-attributable, and is not an insurer investment-income pathway. Searched: h_aegis_espa_budget, h_programmes_greece, h_programmes_portugal, eea_eucra_2024. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.984,"confidence":"L","sources":["wiki/sources/h_programmes_portugal.md","wiki/sources/h_programmes_spain.md","wiki/sources/h_programmes_greece.md","wiki/sources/h_aegis_espa_budget.md","wiki/sources/h_programmes_italy.md","yaml-ref:phase_04_5_categories.yaml#cat_h_fire_management_programmes_and_statutes.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_h_fire_management_programmes_and_statutes.addressable_share"],"rationale":"value 0.984 = 1 - (effectiveness 0.04 x addressable_share 0.40), Phase IV.5 parameters unchanged. THE GOVERNANCE INCREMENT HAS NO EVIDENCE OF ANY KIND - only a disclaimed-double-count argument for why it must be small - and this cell repeats that rather than softening it. The one honest axis, regulator_filings, is strong and on-mechanism: AGIF's mandate to plan, coordinate strategically and evaluate the whole SGIFR, integrating fuel accumulation, population behaviour and agent activity; Andalusia's INFOCA annual prevention/surveillance/extinction plans published as separately downloadable instruments for 2024 and 2025, i.e. re-authorised and re-scoped every year; the Greek AEGIS four-axis programme; and Legge 353/2000 as published by the Italian Parliament. WHAT THIS CELL SCORES AND WHAT IT MUST NOT: the GOVERNANCE INCREMENT ONLY - coherent planning, execution monitoring, annual re-authorisation - over and above the operational capabilities these programmes fund, which are scored in cat_h_detection_forecasting_decision_support, cat_h_aerial_suppression_surge_capacity and cat_h_landscape_fuel_management. The aircraft, sensors and fuel treatments bought with AEGIS's EUR 2.1 bn and INFOCA's annual budget are NOT credited twice; addressable_share is held to 0.40 for that reason, and h_aegis_espa_budget is cited for transparency and is expressly NOT pro-rated. BASELINE NETTING: AGIF, the SGIFR cascade, the annually re-authorised INFOCA and INFOCAM plans and Legge 353/2000 as they operate today are all inside L_0; only forward execution-monitoring maturity and the remaining AEGIS delivery schedule are creditable. Effectiveness was cut 0.10 -> 0.04, lower bound 0.00. TURCO 2016 IS EXPLICITLY NOT USABLE HERE, on an internal contradiction Phase IV.5 found and acted on: Turco attributes its decline to 'increased effort in fire management and prevention', i.e. precisely the operational capabilities this category defines itself as EXCLUDING. A source cannot evidence a category and simultaneously evidence what that category excludes. CONFIDENCE L DESPITE THE `confirmed` LABEL: declared tier deficit - one axis, and the `emerging` route on AEGIS's EUR 2.1 bn was REFUSED because AEGIS is an all-hazard civil-protection capital programme rather than a wildfire budget and the Axis-2 line item has not been retrieved. SEPARATE, SEPARATELY-UNQUANTIFIED LIMB, EXCLUDED FROM THIS VALUE AND CARRIED AS AN EXPLICIT SUB-LIMB GAP: Legge 353/2000's arson-deterrence pathway (land-use classification frozen fifteen years, construction banned ten, sale contracts void if they fail to recall the constraint) removes the development incentive behind deliberate ignition. That is a distinct causal pathway, its magnitude is unquantified in every source in this pilot, and Phase IV.5 directs that it be treated as an evidence gap rather than folded into this category's effectiveness. It is NOT inside the 0.984. The v0.7 cell shape has no sub-cell in which to set evidence_gap for one limb of an otherwise sourced cell, so it is named here; the critic should decide whether that is sufficient."},"I4":{"value":1.0,"confidence":"L","sources":["wiki/sources/r_italy_post_fire_land_constraint.md"],"rationale":"NEUTRAL, EVIDENCE GAP - AND ONE DOCUMENTED OPERATIONAL BURDEN RUNS THE OTHER WAY. Legge 353/2000 art. 10 bans construction on burned land for ten years and freezes its land-use classification for fifteen, which can make in-place reinstatement legally unperformable in Italy and push settlement toward cash indemnity with a long additional-living-expense tail - a LENGTHENING of claim life, not a shortening. v0.7 mitigation cells cannot express amplification (matrix_schema.md v0.7 section 8), so this is recorded in the rationale rather than scored, and it is COUNTED ONCE: the operational bite of the same statute is carried on cat_r_physical_restoration_capacity's I4, and it must not be counted in both places. On the saving side there is nothing. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No insurer operating-cost effect of fire-management programmes or statutes. Searched: h_programmes_portugal, h_programmes_spain, h_programmes_italy, h_programmes_greece, h_agif_institutional_report_2024, h_sgifr_activity_report_2024, r_italy_post_fire_land_constraint. Nothing on the saving side; the one documented effect is an amplification a v0.7 cell cannot carry."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","ngo_reports","regulator_filings"],"distinct_sources":7,"total_disclosed":2100000000.0,"unknown_count":5,"reason":"3 evidence axes across 7 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_h_fire_management_programmes_and_statutes.md"},{"id":"cat_h_utility_corridor_ignition_prevention","name":"Utility-Corridor Ignition Prevention","channel":"H","channel_label":"Hazard","tier":"emerging","tier_rationale":"EMERGING, not confirmed. Two source-type axes only. industry_reports: Endesa's own disclosure of the EUR 105 m 2026 Campana de Verano (36% up year on year, 17,337 distribution-centre inspections and 11,900 km of overhead line reviewed since January), the Conecta Industria trade account of the same campaign naming the FuVeX long-range drone fleet, E-REDES's statutory fuel-strip obligation along 28,000 km of overhead line in forest, and Overstory's vegetation-intelligence proposition. regulator_filings: the Municipio de Satao notice publishing E-REDES's 2025 execution of the secondary fuel-management strip network under the Portuguese rural-fire statute — municipal publication being itself the compliance step. MISSING AXES: academic_studies and ngo_reports. No peer-reviewed study, JRC/EEA technical report, or NGO report in this pilot's corpus discusses European utility-caused wildfire ignition or its prevention, and no European utility ignition statistic has been located at all. The category qualifies as `emerging` under §4.2 on the strength of EUR 105 m disclosed deployed capital — the largest disclosed PRIVATE prevention spend in the pilot — which is above the EUR 50 m gate. Per §5, every active mitigation cell in this category MUST carry confidence: L in Phase X. HOW THIN THE GATE IS — three facts the reader needs (critic S5). (1) The EUR 105 m is the WHOLE summer campaign, not a wildfire budget: h_utility_ignition_prevention's own caveat records that it covers reliability, storm hardening and demand-peak preparation as well as vegetation control. (2) The wildfire-attributable share is therefore UNKNOWN; the frequently quoted \"EUR 19.8 m ringfenced for forest/vegetation control\" split is NOT in the snapshot and must not be attributed to it, and pressing for the vegetation-control share is I021's standing next_step. §6 contemplates pro-rating a programme budget by programme-area share where a line item is available — none is, so the figure enters un-pro-rated. If any credible split emerges the category almost certainly drops below EUR 50 m and re-tiers to the declared-deficit pattern. (3) There is a construction asymmetry: Overstory's USD 43 m Series B was deliberately excluded from this same total because no dated FX pair was verifiable (inventory §2.2), so the gate is cleared by the one un-pro-rated programme number while a genuine equity round sits outside it. The tier is not changed on these grounds — the figure is admissible in form under §6 — but the leads should know the gate is cleared by EUR 55 m of headroom on a number whose relevant share is unmeasured.","definition":"Removal of the electrical-network ignition pathway before a fire starts: annual grid-operator campaigns that inspect overhead lines and clear the vegetation corridor around them, statutory fuel-management strips cut along high- and medium-voltage lines at locations fixed in municipal forest-defence plans, long-range drone and LiDAR/thermographic inspection that finds hot spots, conductor anomalies and vegetation encroachment before they arc, and machine-learning vegetation classification that re-ranks where existing pruning crews work. The mechanism is HAZARD and it is upstream of every other category in this channel: an ignition that never happens produces no loss to detect, suppress or defend against.","cluster_rationale":"All four ventures act on one ignition cause — the electricity distribution corridor — and on the two pathways by which it starts fires: vegetation contact and equipment failure. I021 is the funded annual campaign that does the physical clearing and inspection across most of Spain's highest-hazard regions; I029 is the legally compelled Portuguese equivalent, cutting strips at locations fixed in municipal plans; I028 supplies the long-range drone inspection sorties that find the defects before they arc; I030 supplies the vegetation classification that re-ranks where the pruning crews go. I028 and I030 are inputs into I021's and I029's programmes rather than independent mitigations — I030's own inventory entry directs that it must not be scored as an independent mitigation but treated as a multiplier on utility vegetation programmes — and they are co-clustered for exactly that reason: their contribution to loss reduction is inseparable from the corridor programme. Boundary vs cat_h_landscape_fuel_management: the linear strips here exist to isolate an ignition source from the fuel bed around a conductor; landscape fuel management exists to create discontinuity in the path of a fire that has already started elsewhere. The two are geometrically distinct (a corridor along 28,000 km of overhead line versus a treated network and settlement band) and their addressable shares do not overlap: this category's 0.10 is scoped to the utility-caused ignition share of insured loss, which the landscape category does not claim. Boundary vs cat_h_detection_forecasting_decision_support: drone thermography here finds an overheating conductor before ignition; that category detects a fire after ignition.","members":["I021","I028","I029","I030"],"member_names":["e-distribucion (Endesa) Campana de Verano","FuVeX long-range inspection drones","E-REDES fuel-management strips along distribution lines","Overstory"],"effectiveness":{"value":0.5,"lower":0.2,"upper":0.8,"unit":null},"effectiveness_definition":"ANALYST-INFERRED, and note that the high central value is CONDITIONAL — it is the fractional reduction in insured loss arising from the utility-ignition cause only, which is what addressable_share (0.10) scopes. Where a powerline ignition is prevented, the loss it would have caused is prevented in full, so conditional effectiveness is high by construction; the uncertainty sits almost entirely in the addressable share, not here. The bracket reflects that corridor clearance and predictive inspection address the vegetation-contact and equipment-failure pathways but not conductor slap in extreme wind, and that I030's contribution is an allocation improvement that only bites if the utility acts on the ranking. NO European study attributes a share of wildfire ignitions or insured loss to electrical infrastructure; the only quantified utility evidence in the pilot is the Technosylva/PG&E \"99% reduction in acres impacted\" and \"68% decrease in reportable ignitions\", which is US, vendor-published, bundled with public-safety power shutoff and vegetation management, and explicitly NOT portable — it is recorded here only to state that it was considered and refused.","effectiveness_sources":[{"slug":"h_utility_ignition_prevention","title":"Endesa eleva a 105 millones de euros la inversion en control forestal, mantenimiento e inspeccion de la red de cara al verano","tier":"T3","publisher":"endesa","url":"https://www.endesa.com/es/prensa/sala-de-prensa/noticias/eficiencia-energetica/medioambiente/105-millones-euros-control-forestal-mantenimiento-inspeccion-red-verano","wayback_url":"https://web.archive.org/web/20260802072326/https://www.endesa.com/es/prensa/sala-de-prensa/noticias/eficiencia-energetica/medioambiente/105-millones-euros-control-forestal-mantenimiento-inspeccion-red-verano","snapshot_hash":"360a574c130eb9c0324aac9bed622b993345e3b88d221bb8c167e4c843664be3","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/360a574c130eb9c0324aac9bed622b993345e3b88d221bb8c167e4c843664be3.html","publication_date":"2026-06-10","fetched_at":"2026-07-31T13:12:48Z","page_exists":true},{"slug":"h_endesa_campaign_trade_press","title":"Endesa: 105 millones en red con drones, IA y control forestal","tier":"T2","publisher":"conecta_industria","url":"https://www.conectaindustria.es/articulo/sostenibilidad/endesa-105-millones-red-drones-ia-incendios/20260611111039014584.html","wayback_url":"https://web.archive.org/web/https://www.conectaindustria.es/articulo/sostenibilidad/endesa-105-millones-red-drones-ia-incendios/20260611111039014584.html","snapshot_hash":"2c08587f88c068e8ed2473c0ab9b60f1823bc86b6ccc6d0ad32bfb44e1a59c8b","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/2c08587f88c068e8ed2473c0ab9b60f1823bc86b6ccc6d0ad32bfb44e1a59c8b.html","publication_date":"2026-06-11","fetched_at":"2026-07-31T13:44:11Z","page_exists":true},{"slug":"h_eredes_fuel_strips","title":"Gestao da Vegetacao - E-REDES","tier":"T3","publisher":"e_redes","url":"https://www.e-redes.pt/pt-pt/sustentabilidade/nos-e-o-planeta/gestao-da-vegetacao","wayback_url":"https://web.archive.org/web/https://www.e-redes.pt/pt-pt/sustentabilidade/nos-e-o-planeta/gestao-da-vegetacao","snapshot_hash":"03d4b27c7776939cdb9aee069b668c87b349fd160b8df5dbef41236aee47bf0f","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/03d4b27c7776939cdb9aee069b668c87b349fd160b8df5dbef41236aee47bf0f.html","publication_date":null,"fetched_at":"2026-07-31T13:12:53Z","page_exists":true},{"slug":"h_eredes_municipal_execution","title":"Execucao de Faixa de Gestao de Combustivel de 2025 a cargo da E-REDES","tier":"T1","publisher":"municipio_de_satao","url":"https://www.cm-satao.pt/noticia/execucao-de-faixa-de-gestao-de-combustivel-de-2025-a-cargo-da-e-redes","wayback_url":"https://web.archive.org/web/https://www.cm-satao.pt/noticia/execucao-de-faixa-de-gestao-de-combustivel-de-2025-a-cargo-da-e-redes","snapshot_hash":"fa9e91d8b9343ffb03a033181f98bb152d52277d45596f8cac7f6c93013d217d","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/fa9e91d8b9343ffb03a033181f98bb152d52277d45596f8cac7f6c93013d217d.html","publication_date":"2024-12-17","fetched_at":"2026-07-31T13:55:33Z","page_exists":true}],"addressable_share":0.1,"axes_hit":["industry_reports","regulator_filings"],"citations":[{"axis":"industry_reports","wiki_source_id":"h_utility_ignition_prevention","span":"The Campana de Verano budget is close to EUR 105 m, 36% above the previous year, funding inspection, preventive maintenance and management of the vegetation surrounding electricity lines to minimise fire risk around network infrastructure; 17,337 distribution-centre inspections and 11,900 km of high- and medium-voltage overhead line reviewed since the start of the year."},{"axis":"industry_reports","wiki_source_id":"h_endesa_campaign_trade_press","span":"Seventeen FuVeX long-range drones with digital cameras, LiDAR and thermography detect hot spots and anomalies invisible to the eye at up to 10 km range, feeding a network digital twin — the shift from reactive to predictive maintenance of critical infrastructure."},{"axis":"industry_reports","wiki_source_id":"h_eredes_fuel_strips","span":"E-REDES operates 179,000 km of overhead distribution network, 28,000 km of it in forest, and has a duty under the national forest-defence system to constitute fuel-management strips along high- and medium-voltage lines at strategic locations defined in municipal forest-defence plans, with landowners obliged to permit access."},{"axis":"industry_reports","wiki_source_id":"h_overstory_vegetation","span":"Machine-learning vegetation intelligence directs pruning effort to minimise risk; vegetation management grew from 50% of overhead maintenance budgets in 2015 to 75% in 2020 — the allocation problem this venture addresses."},{"axis":"regulator_filings","wiki_source_id":"h_eredes_municipal_execution","span":"Municipio de Satao notice: E-REDES will execute the 2025 fuel-management actions on the strips associated with high- and medium-voltage lines inscribed in the applicable planning instruments and forming part of the Secondary Fuel-Management Strip Network, in compliance with the rural-fire management statute."}],"total_disclosed":105000000.0,"unknown_count":3,"adoption":{"y0":{"value":0.1,"lower":0.03,"upper":0.2,"unit":null},"y3":{"value":0.35,"lower":0.15,"upper":0.55,"unit":null},"y5":{"value":0.5,"lower":0.25,"upper":0.7,"unit":null},"y10":{"value":0.7,"lower":0.35,"upper":0.85,"unit":null}},"saturation_cap":0.85,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced, AND THE COMMERCIALLY OBVIOUS ADJACENT CHANNEL IS OUT OF SCOPE. Utility-corridor ignition prevention is funded by grid operators out of network maintenance budgets and by statutory obligation; no source ties it to a property premium. The adjacency a later reader will reach for - utility WILDFIRE LIABILITY and insurer subrogation recovery - is excluded by phase_00_input.yaml.out_of_scope ('Wildfire liability and subrogation exposure (e.g. utility ignition liability), which is a financial/litigation claim cause'), and must not be reintroduced here as a property-premium effect. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No property-premium or underwriting effect attributed to corridor clearance or predictive line inspection in ES/PT/IT/GR. Searched: h_utility_ignition_prevention, h_endesa_campaign_trade_press, h_eredes_fuel_strips, h_eredes_municipal_execution, h_overstory_vegetation, h_ree_overstory_deployment, insurance_journal_2026_marsh_europe_wildfire. NOTE FOR ANY LATER READER: the utility-liability and subrogation channel is closed by SCOPE, not by absence of evidence - do not 'rediscover' it."},"I2":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Grid vegetation-management campaigns, drone inspection sorties and ML vegetation classification do not touch insurer asset allocation, duration or investment return. A utility's own credit quality is not an insurer investment-income channel in this pilot and would in any case run through the excluded liability pathway. Searched: h_utility_ignition_prevention, h_endesa_campaign_trade_press, h_eredes_fuel_strips, h_overstory_vegetation. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). The finding is a structural NEGATIVE - that no channel exists at this loss scale - and not a weak positive; see the confidence note at the end of this rationale. No evidence_gap flag: this is a finding, not a hole. CONFIDENCE SET TO L AT ITERATION 1, DEPARTING FROM THE OTHER TEN I2 CELLS AND MATCHING THIS CATEGORY'S DOSSIER. wiki/categories/cat_h_utility_corridor_ignition_prevention.md prescribes, row by row, confidence L on I1, I2 and I4 for this tier: emerging category. The schema's emerging hard cap (matrix_schema.md section 4) reaches only active cells, so H on a neutral cell is not a validator violation and the composer's 'confidence in the negative' reading is coherent - but the rendered sheet 1 would show an H beside a tier column reading 'emerging', which is exactly the rigor signal the cap exists to suppress, and a reader comparing dossier to matrix would read the divergence as drift. The finding itself is unchanged and remains a structural negative, not a weak one. Note the deliberate inconsistency this creates and why it is accepted: ten other I2 cells carry H on a verbatim-identical structural-negative finding. The divergence is a rendered-sheet-optics judgement, not an evidence judgement - the finding here is exactly as strong as the other ten - and it follows the dossier rather than the composer's own uniform convention."},"I3":{"value":0.95,"confidence":"L","sources":["wiki/sources/h_utility_ignition_prevention.md","wiki/sources/h_endesa_campaign_trade_press.md","wiki/sources/h_eredes_fuel_strips.md","wiki/sources/h_eredes_municipal_execution.md","wiki/sources/h_overstory_vegetation.md","yaml-ref:phase_04_5_categories.yaml#cat_h_utility_corridor_ignition_prevention.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_h_utility_corridor_ignition_prevention.addressable_share"],"rationale":"[READ THIS WITH THE NUMBER - AGGREGATE CONTEXT, repeated on each of the five rows that dominate the product so it survives being read in isolation.] The matrix total is a 22.9% claims reduction at t=10 (total_factor_I3 = 0.770928, rebased on the observed L_0 per issue #53). NOT ONE VENTURE, PROGRAMME OR STATUTE IN THIS PILOT'S POOL PUBLISHES AN AUDITED INSURED-LOSS-REDUCTION FIGURE ANYWHERE IN SPAIN, PORTUGAL, ITALY OR GREECE, so that 22.9% is a structured analyst inference over sourced mechanisms, not a measured effect. TWO BASES ARE QUOTED BELOW AND THEY ARE NOT INTERCHANGEABLE: 'marginal' is the share of the 22.9% that disappears if the named rows are removed, and marginal shares do NOT sum to 100% across rows; 'log' is each row's share of the log-reduction, which does sum. Cells at confidence: L carry 71.4% marginal / 73.9% log. The two tier: emerging rows carry 25.7% marginal / 28.3% log, and removing both moves the headline from 22.9% to 17.0%. THIS ROW CONTRIBUTES 10.5% MARGINAL / 11.8% LOG. Full decomposition in loop_metadata.aggregation_disclosure, which does not render - see loop_metadata.disclosure_rendering_note. || value 0.950 = 1 - (effectiveness 0.50 x addressable_share 0.10), Phase IV.5 parameters unchanged. READ THE TWO FACTORS TOGETHER OR THE CELL IS MISREAD: the 0.50 is a CONDITIONAL effectiveness - where a powerline ignition is prevented, the loss it would have caused is prevented in full, which is why it is high relative to every other effectiveness in the matrix. THE UNCERTAINTY CLAIM MADE AT ITERATION 0 IS WITHDRAWN. That text asserted, inheriting the sentence from wiki/categories/cat_h_utility_corridor_ignition_prevention.md section 5, that essentially all of the uncertainty sits in the addressable share rather than in the effectiveness. The pilot's own parameter record does not support it: phase_04_5_categories.yaml gives this effectiveness value 0.50 with lower 0.20 and upper 0.80 - a fourfold spread and the widest effectiveness bracket in the matrix, which on its own moves this cell across 0.92 to 0.98 - while the addressable share is stored as a bare float 0.10 with no lower, no upper, no confidence, no sources and no basis field, because v0.7 records addressable_share as a scalar (matrix_schema.md v0.8.2 promotes it to a bracketed, sourced, basis-tagged parameter precisely for this reason). The correct statement is therefore: the one factor the pilot DID bracket is the effectiveness and it is bracketed widely; the addressable share carries no recorded uncertainty at all, which is not the same as carrying none; and neither factor is sourced to a European ignition-attribution statistic, because none exists. THE UPSTREAM-MOST MECHANISM IN THE PILOT, WITH THE THINNEST QUANTIFICATION OF ITS ADDRESSABLE SHARE. An ignition that never happens produces no loss to detect, suppress or defend against. Four on-mechanism deployment sources: Endesa's Campana de Verano at close to EUR 105 m, 36% above the previous year, with 17,337 distribution-centre inspections and 11,900 km of high- and medium-voltage overhead line reviewed since the start of the year; seventeen FuVeX long-range drones with cameras, LiDAR and thermography detecting hot spots at up to 10 km feeding a network digital twin; E-REDES's 179,000 km of overhead distribution network, 28,000 km of it in forest, under a statutory duty to constitute fuel-management strips at locations fixed in municipal forest-defence plans; and the Municipio de Satao notice publishing E-REDES's 2025 execution of those strips, municipal publication being itself the compliance step. Overstory supplies the ML vegetation classification that re-ranks where existing pruning crews work - an ALLOCATION improvement that only bites if the utility acts on the ranking. THE NUMBER THAT DOES NOT EXIST: no European study attributes any share of wildfire ignitions or insured loss to electrical infrastructure. The 0.10 addressable share is therefore an analyst estimate and is the weakest quantity in this row. BASELINE NETTING, DISCHARGED UPSTREAM AND STATED HERE AT ITERATION 2 - this was the only active cell in the matrix that never named its netting position, and the omission mattered because this row spans two limbs sitting on OPPOSITE sides of L_0. The statutory fuel-strip limb is already executing in 2026 - E-REDES's 179,000 km of overhead network with 28,000 km in forest, under a standing duty, with the Municipio de Satao notice evidencing 2025 execution - and is therefore INSIDE L_0 by construction, exactly as the aerial and landscape-fuel rows are. The predictive-inspection limb is genuinely new: seventeen FuVeX long-range drones at one Spanish DSO, plus Overstory's ML vegetation ranking. The netting is discharged not in the effectiveness but UPSTREAM, in the adoption term - and precisely: it is not the y0 VALUE that discharges it, because y0 enters no computation (matrix_schema.md section 8, the trajectory does not enter the math). What discharges it is that the WHOLE Phase V anchor SERIES for this row, y0 through y10, is denominated over the predictive-inspection layer rather than over corridor management as a whole - the y0 note reads 'Seventeen aircraft at one Spanish DSO is a first-fleet deployment, not an installed base', and the y10 of 0.70 is the projected reach of that same predictive layer, not of the statutory strip network. So the installed strips are excluded by the DENOMINATOR of the adoption series, not by subtracting a base-year value. AND THE DEEPER GROUND, ADDED AT ITERATION 4 because 'a first-fleet deployment, not an installed base' is an assertion about the fleet rather than about L_0: phase_02_baseline.yaml#method_notes builds L_0 from the EIOPA/CATDAT insured mean over 1980-2024, CPI-adjusted with a window midpoint around 2002, cross-checked against a 2025 economic-damage times insured-share estimate. A predictive drone-inspection layer first flying in summer 2026 therefore cannot be inside L_0 at all - not because it is small, but because it postdates every year of experience L_0 is built from. That is why this row credits its full y10 of 0.70, and it is a stronger ground than the first-fleet argument. SUPERSEDED IN ITS CONTRAST, 2026-09-21, ISSUE #53: this passage used to add that cat_v_defensible_space_clearance credited y10 minus y0, and priced subtracting y0 here as well at a move from 24.3% to 23.9%. Both figures predate the rebasing and neither is current. Every row now credits its full y10, and the base-year overlap is removed by f(0) in L(t) = L_0 (1+g)^t f(t)/f(0), so there is no contrast left to draw between this row and that one. The argument above still does work, but a different job: it is why f(0) for this row is close to 1 (y0 = 0.10, f(0) = 0.995000) and therefore why rebasing costs this row almost nothing, rather than a reason to exempt it from a subtraction no row makes any more. Subtracting y0 here on top of the rebasing would move the headline from 22.9% to 22.5%; it is not done, because f(0) has already done it. That is why the 0.50 conditional effectiveness is NOT walked back here despite the row containing an inside-L_0 limb, and it is stated so that a reader does not mistake the silence for an exemption. REFUSED AND RECORDED AS REFUSED so it is not resurrected: the Technosylva/PG&E '99% reduction in acres impacted' and '68% decrease in reportable ignitions' figures are US, vendor-published, bundled with public-safety power shutoff and vegetation management, and explicitly NOT portable. CONFIDENCE L IS A HARD CAP: tier is `emerging`, cleared on EUR 105 m disclosed, and that gate is thin in three stated ways - the EUR 105 m is the WHOLE summer campaign covering reliability, storm hardening and demand-peak preparation as well as vegetation control; the wildfire-attributable share is unknown and the often-quoted 'EUR 19.8 m ringfenced for forest/vegetation control' split is NOT in the snapshot and is not attributed to it; and Overstory's USD 43 m Series B was excluded from the same total for want of a verifiable dated FX pair."},"I4":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no INSURER operational channel. The documented cost movements in this category are the UTILITY's: Endesa's campaign budget up 36% year on year, and the shift from reactive to predictive maintenance. A grid operator's operating cost is not an insurer's operating cost. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No claims-handling or expense effect on an insurer attributed to corridor ignition prevention. Searched: h_utility_ignition_prevention, h_endesa_campaign_trade_press, h_eredes_fuel_strips, h_eredes_municipal_execution, h_overstory_vegetation, r_claims_visual_intelligence."}},"tier_check":{"computed":"emerging","declared":"emerging","agrees":true,"axes_hit":["industry_reports","regulator_filings"],"distinct_sources":5,"total_disclosed":105000000.0,"unknown_count":3,"reason":"2 axes and 105,000,000 disclosed investment (floor 50,000,000)"},"wiki_page":"wiki/categories/cat_h_utility_corridor_ignition_prevention.md"},{"id":"cat_h_detection_forecasting_decision_support","name":"Wildfire Detection, Fire-Danger Forecasting and Suppression Decision Support","channel":"H","channel_label":"Hazard","tier":"emerging","tier_rationale":"EMERGING — retiered at iteration 2 from `confirmed` on critic finding M3. TWO source-type axes plus the investment gate. industry_reports: OroraTech's Hellenic Fire System announcement (4x4 m hotspots, latency in minutes, 100% of Greek territory), the Allianz Seguros x Satellites on Fire launch (detection 35 minutes faster than NASA reference systems, up to 30% more fires detected), Dryad's Silvanet smouldering-phase gas sensing, and the Globalstar/Technosylva INFOCA deployment reported in the International Fire and Safety Journal. regulator_filings: the CORDIS fact sheets for FIRE-RES (GA 101037419), SILVANUS (GA 101037247) and FirEUrisk (GA 101003890) with their EU-contribution line items, plus the Greek AEGIS Axis-2 \"early-warning systems and prevention means\" allocation. WHY NOT `confirmed`: the academic_studies axis claimed at iteration 1 does not survive §4.1 and both its citations are withdrawn. The JRC/EFFIS advance report documents Rapid Damage Assessment — post-event burned-area mapping, which is the mechanism of cat_r_post_event_damage_establishment, not detection and not fire-danger forecasting. The JRC PESETA IV wildfire card is a two-page climate-impact brief whose extractable content is WUI exposure projections and burned-area context; it says nothing about forecasting as a mitigation mechanism. No NGO-report axis exists either. THE GATE: §4.2(b) is cleared on total_disclosed_eur EUR 55,669,561 >= EUR 50 m, and unlike the utility category's whole-campaign figure every component here is a §6-admissible disclosed amount — Dryad EUR 5.6 m company funding, and FIRE-RES/SILVANUS EUR 39,798,517 plus FirEUrisk EUR 10,271,044 as regulator-funded- pilot grant sizes taken from the CORDIS EU-contribution line items. Per §5, EVERY active mitigation cell in this category MUST carry confidence: L in Phase X. That is the correct signal for the pilot's largest H row: it rests on vendor and programme evidence, not on research. Three of the four industry-axis sources are vendor-published (T3) and establish deployment, not loss reduction.","definition":"Information systems that shorten the interval between an ignition (or a forecast fire-danger condition) and effective, correctly-aimed suppression action — thermal- infrared and SAR satellite constellations, satellite-feed fusion with push alerting, ground gas-sensor and camera networks, physics-based fire-behaviour simulation, and pan-European fire-danger forecasting. The loss-reduction mechanism is a HAZARD mechanism: a larger share of ignitions is suppressed while still small enough that no built asset is reached, and where a fire does run, suppression is pre-positioned on the flank that threatens insured structures rather than distributed evenly around the perimeter. The category acts on the number and size of fires, not on the damage a fire does to a building it reaches.","cluster_rationale":"All ten ventures act on the same causal step: produce information earlier or better, so that suppression starts sooner or is aimed better, so that fewer ignitions become fires large enough to reach insured property. Tech stacks differ completely — thermal- infrared CubeSats (I001), SAR (I008), geostationary/polar feed fusion with push alerting (I003), an insurer's own policyholder-alerting layer built on that feed (I004), LoRaWAN gas sensors detecting the smouldering phase (I005), forest sensors plus 360-degree cameras on satellite backhaul (I006), physics-based spread simulation (I002), and pan-European fire-danger forecasting and EU research platforms (I025, I026, I027) — but the mechanism is identical, which is exactly the case the convention says must be co-clustered. I004 is the insurer peer template for the mechanism and is clustered by mechanism, not by actor type. TWO CROSS-CATEGORY DOUBLE-COUNT CONTROLS, restated at CATEGORY level because v0.7 has no per-venture mitigation cells (validate.py: \"v0.7 ventures carry only descriptive context\"; the analytical unit is the category, so an instruction phrased at venture level is un-executable). (1) I006 also sells MAK 5, a long-term retardant that converts standing vegetation into a months-durable low-flammability strip. That limb is a fuel treatment whose effect belongs to cat_h_landscape_fuel_management's mechanism. THIS category's effectiveness is scored on the detection limb only and must not be uplifted for I006's retardant product; cat_h_landscape_fuel_management likewise must not be uplifted for I006, which is not among its constituents. (2) I002's KatRisk arm serves portfolio loss modelling, an R-channel adjacency. That limb is NOT credited in this category's effectiveness, and cat_r_post_event_damage_establishment must not be uplifted for I002 either, which is not among its constituents. DETECTION-VERSUS-FORECASTING BOUNDARY, stated because it is the weakest internal seam: post-ignition detection (I001, I003, I004, I005, I006, I008) and pre-ignition danger forecasting plus in-event behaviour simulation (I002, I025, I026, I027) are sub-mechanisms of one chain. CORRECTED AT ITERATION 2 (critic M8): it is not true that neither sub-cluster could stand alone. A forecasting cluster of I002+I025+I026+ I027 carries EUR 39,798,517 + EUR 10,271,044 = EUR 50,069,561 and clears the EUR 50 m emerging gate by EUR 69,561, so it COULD stand alone as `emerging`. The binding constraint is only the DETECTION sub-cluster: two axes and EUR 5.6 m disclosed makes it `insufficient`, and §4.3's disposal route would orphan I001, I003, I004, I005, I006 and I008 against validate.py's no-orphan rule. The leads are therefore being asked to accept one 10-venture row, whose single effectiveness blends two sub-mechanisms with very different adoption dynamics, in order not to orphan six detection ventures. The cost is much reduced by the `emerging` retier: the whole row is capped at confidence L regardless. PHASE V INSTRUCTION: do not anchor adoption for this row from a single S-curve — the forecasting limb (EFFIS, closed Horizon platforms) and the detection limb (2026-launch satellite and sensor systems) sit at very different points at y0. Boundary vs cat_h_aerial_suppression_surge_capacity: that category is the physical delivery of water and retardant onto the fire; this category decides where and when it is delivered. Boundary vs cat_h_fire_management_programmes_and_statutes: that category is the budget and legal instrument that procures capability; this category is the capability. Boundary vs cat_v_vulnerability_assessment_targeting: that category scores the susceptibility of individual buildings given a fire; this category acts on whether the fire reaches them.","members":["I001","I002","I003","I004","I005","I006","I008","I025","I026","I027"],"member_names":["OroraTech GmbH","Technosylva SL","Satellites on Fire","Allianz Seguros early wildfire detection initiative (Spain)","Dryad Networks GmbH","Pyro Fire Extinction S.L.","ICEYE Oy","EFFIS - European Forest Fire Information System","FIRE-RES and SILVANUS Horizon innovation actions","FirEUrisk"],"effectiveness":{"value":0.12,"lower":0.03,"upper":0.3,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. BASELINE-NETTED per phase_00_input.yaml#fit_check.rationale: \"the Phase X mitigation term must count only *new* adaptation above the already-embedded historical suppression baseline\". L_0 is anchored at base_year 2026, so EFFIS-grade pan-European fire-danger forecasting, the national detection and dispatch chains, and the INFOCA/INFOCAM operational technology already deployed are INSIDE L_0 and are not credited here. The counterfactual is the 2026 detection-and-decision-support system as it stands, and the credited effect is only the forward increment over it — the minutes-latency satellite and smouldering-phase sensor layer that is genuinely new (OroraTech's Greek constellation launched 2026; the Allianz alerting initiative is in a summer-2026 validation phase; Dryad has no named Southern-European end customer). Value cut from 0.15 to 0.12 at iteration 2 for this reason. No source in the pilot measures it. The inference chain: (a) the only quantified deltas available are latency and detection-count deltas (35 min faster, up to 30% more fires detected — Allianz/Satellites on Fire, vendor-stated, validation phase not yet reported); (b) these are inputs, not loss outcomes, and the conversion from detection latency to avoided insured loss is unmeasured anywhere in Europe; (c) the upper bound is the fraction of Southern-European insured loss plausibly originating in fires extinguishable at first attack, the lower bound the possibility that loss is dominated by wind-driven runs no detection speed changes. PHASE V INSTRUCTION: because effectiveness is stated as a forward increment, the adoption anchors for this category must be set on that increment — y0 near zero — and NOT on the installed base of detection coverage, or the 2026 baseline is re-embedded and counted twice.","effectiveness_sources":[{"slug":"h_ororatech_greece_system","title":"Greece Launches World's First National Wildfire Satellite System","tier":"T3","publisher":"ororatech","url":"https://ororatech.com/resources/news-blog/greece-launches-world-s-first-national-wildfire-satellite-system","wayback_url":"https://web.archive.org/web/20260729062920/https://ororatech.com/resources/news-blog/greece-launches-world-s-first-national-wildfire-satellite-system","snapshot_hash":"6c3204987f7d704c3790f3e94b2466aeecd782c3b2facc3431d03d089eb37a1f","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/6c3204987f7d704c3790f3e94b2466aeecd782c3b2facc3431d03d089eb37a1f.html","publication_date":"2026-05-04","fetched_at":"2026-07-31T12:52:26Z","page_exists":true},{"slug":"h_satellites_on_fire_allianz","title":"Allianz launches in Spain an early wildfire detection system using artificial intelligence and satellite technology","tier":"T2","publisher":"allianz","url":"https://www.allianz.com/en/mediacenter/news/articles/260720-early-wildfire-detection-system-launch-in-spain.html","wayback_url":null,"snapshot_hash":"dc95dc1cc70f078216204eeea44c4e5ec06766a23e183008c7725801c9187cd1","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/dc95dc1cc70f078216204eeea44c4e5ec06766a23e183008c7725801c9187cd1.html","publication_date":"2026-07-20","fetched_at":"2026-07-31T13:09:02Z","page_exists":true},{"slug":"h_detection_ventures","title":"Ultra Early Wildfire Detection | Dryad Networks","tier":"T3","publisher":"dryad_networks","url":"https://www.dryad.net/","wayback_url":"https://web.archive.org/web/https://www.dryad.net/","snapshot_hash":"4175e91bc9bc2137790c73c1ed8562e737267e95b64f0c095e3dc03c92d5d884","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/4175e91bc9bc2137790c73c1ed8562e737267e95b64f0c095e3dc03c92d5d884.html","publication_date":null,"fetched_at":"2026-07-31T12:58:04Z","page_exists":true},{"slug":"h_effis_hazard_baseline","title":"EFFIS - European Forest Fire Information System","tier":"T1","publisher":"jrc","url":"https://forest-fire.emergency.copernicus.eu/","wayback_url":"https://web.archive.org/web/https://forest-fire.emergency.copernicus.eu/","snapshot_hash":"233ba811264caf38577d38e30e5a0e94164739c88dcf4befca9ab35ff02d4f39","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/233ba811264caf38577d38e30e5a0e94164739c88dcf4befca9ab35ff02d4f39.html","publication_date":null,"fetched_at":"2026-07-31T13:06:44Z","page_exists":true}],"addressable_share":0.7,"axes_hit":["industry_reports","regulator_filings"],"citations":[{"axis":"industry_reports","wiki_source_id":"h_ororatech_greece_system","span":"Four dedicated wildfire monitoring satellites for Greece deliver data directly to Greek fire services with latency measured in minutes and the ability to identify hotspots as small as 4 by 4 metres across 100% of Greek territory."},{"axis":"industry_reports","wiki_source_id":"h_satellites_on_fire_allianz","span":"Allianz Seguros with Satellites on Fire: the technology reduces average detection times by 35 minutes compared with reference systems such as NASA's and detects up to 30% more fires; summer 2026 is an initial implementation phase to validate scope and effectiveness."},{"axis":"industry_reports","wiki_source_id":"h_detection_ventures","span":"Silvanet fourth-generation sensors combine VOC, CO and PM2.5 sensing to detect fires as early as the smouldering phase, over off-grid solar-powered LoRaWAN mesh networks."},{"axis":"industry_reports","wiki_source_id":"h_technosylva_infoca_deployment","span":"Technosylva software combined with satellite tracking deployed for Andalusia's INFOCA to improve field communication and decision-making; 550 devices deployed in the region by 2022, with Castilla-La Mancha using the technology since 2015."},{"axis":"regulator_filings","wiki_source_id":"h_horizon_research_projects","span":"FIRE-RES, GA 101037419: total cost EUR 21,541,422.51, EU contribution EUR 19,896,326.62, coordinated from Catalonia, closed 30 November 2025 — a holistic extreme-wildfire management approach."},{"axis":"regulator_filings","wiki_source_id":"h_horizon_fireurisk","span":"FirEUrisk, GA 101003890: EU contribution EUR 10,271,044, Portuguese-coordinated, closed 31 March 2025 — expanding wildland-fire risk assessment to previously uncovered risk factors and producing measures to reduce current risk conditions."},{"axis":"regulator_filings","wiki_source_id":"h_aegis_espa_budget","span":"AEGIS Axis 2 funds early-warning systems and prevention means — a national AI fire database, fire-detection systems, meteorological stations and radar, and a 112 upgrade — inside a EUR 2.1 bn state programme."}],"total_disclosed":55669561.0,"unknown_count":7,"adoption":{"y0":{"value":0.05,"lower":0.01,"upper":0.12,"unit":null},"y3":{"value":0.22,"lower":0.1,"upper":0.35,"unit":null},"y5":{"value":0.35,"lower":0.15,"upper":0.52,"unit":null},"y10":{"value":0.55,"lower":0.25,"upper":0.75,"unit":null}},"saturation_cap":0.8,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":["wiki/sources/insurance_journal_2026_marsh_europe_wildfire.md","wiki/sources/h_satellites_on_fire_allianz.md","wiki/sources/v_insurer_wildfire_initiatives.md"],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced, and the one insurer-operated instance in this category sharpens the point rather than softening it. Allianz Seguros runs the pilot's only insurer-side detection deployment (I004, with Satellites on Fire, monitoring more than 50 million hectares per h_satellites_on_fire_allianz; the 15 July 2026 launch date is carried only by v_insurer_wildfire_initiatives as a scout bundle claim and is cited to that page rather than to the Allianz release). ATTRIBUTION CORRECTED AT ITERATION 1: the operational count is NOT in the Allianz release, whose quantitative content is the 35-minutes-faster figure, the up-to-30%-more-fires figure and the >50 million hectares figure. It is in insurance_journal_2026_marsh_europe_wildfire (T3), verbatim, as a news article quoting Allianz's global head of P&C - as of 20 July the system had generated 417 wildfire alerts, triggered 73 response activations and enabled notifications to thousands of customers. Iteration 0 attributed it to the Allianz page and called it 'its published outputs'; both were wrong, and the same inconsistency exists upstream, where phase_09_evidence.yaml's I1 stanza for this category lists only h_satellites_on_fire_allianz while its I4 stanza lists only insurance_journal_2026_marsh_europe_wildfire for the identical datum - the source pages settled it in favour of the I4 stanza. THAT UPSTREAM CORRECTION HAS SINCE LANDED and this sentence is updated at iteration 4 to record it rather than to leave a discharged to-do standing: phase_09_evidence.yaml's I1 stanza now carries its own 'ATTRIBUTION CORRECTION (2026-08-02)' paragraph and cites both pages, with the split spelled out in the relevance strings - h_satellites_on_fire_allianz marked as NOT carrying the 417/73 counts, and insurance_journal_2026_marsh_europe_wildfire marked as their actual source. No action remains for synthesis; the provenance history is kept because it explains why this cell cites three pages for one datum. THE POINT STRENGTHENS UNDER THE CORRECT ATTRIBUTION: those are ACTIVITY COUNTS, and the cited page's own caveat characterises them as a vendor-reported operational count with no baseline, no denominator and no loss-avoided estimate. No premium credit, rate relativity, discount, underwriting-appetite change or retention figure is attached to them anywhere, and Allianz itself frames summer 2026 as an implementation-and-validation phase whose scope and effectiveness are not yet reported. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No detection- or forecasting-linked premium discount, rate credit, capacity restoration or underwriting-appetite statement exists for ES/PT/IT/GR. Searched: h_satellites_on_fire_allianz, h_satellites_on_fire_platform, v_insurer_wildfire_initiatives, insurance_journal_2026_marsh_europe_wildfire, h_ororatech_greece_system, h_detection_ventures, h_technosylva_infoca_deployment. The closest available datum is the Allianz activity count (417 alerts / 73 activations), which is activity, not premium."},"I2":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Satellite constellations, gas-sensor meshes, fire-behaviour simulation and pan-European danger forecasting do not touch an insurer's asset allocation, duration or investment return by any pathway. Searched: h_ororatech_greece_system, h_satellites_on_fire_allianz, h_detection_ventures, h_technosylva_infoca_deployment, h_horizon_research_projects, h_effis_hazard_baseline. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). The finding is a structural NEGATIVE - that no channel exists at this loss scale - and not a weak positive; see the confidence note at the end of this rationale. No evidence_gap flag: this is a finding, not a hole. CONFIDENCE SET TO L AT ITERATION 1, DEPARTING FROM THE OTHER TEN I2 CELLS AND MATCHING THIS CATEGORY'S DOSSIER. wiki/categories/cat_h_detection_forecasting_decision_support.md prescribes, row by row, confidence L on I1, I2 and I4 for this tier: emerging category. The schema's emerging hard cap (matrix_schema.md section 4) reaches only active cells, so H on a neutral cell is not a validator violation and the composer's 'confidence in the negative' reading is coherent - but the rendered sheet 1 would show an H beside a tier column reading 'emerging', which is exactly the rigor signal the cap exists to suppress, and a reader comparing dossier to matrix would read the divergence as drift. The finding itself is unchanged and remains a structural negative, not a weak one. Note the deliberate inconsistency this creates and why it is accepted: ten other I2 cells carry H on a verbatim-identical structural-negative finding. The divergence is a rendered-sheet-optics judgement, not an evidence judgement - the finding here is exactly as strong as the other ten - and it follows the dossier rather than the composer's own uniform convention."},"I3":{"value":0.916,"confidence":"L","sources":["wiki/sources/h_ororatech_greece_system.md","wiki/sources/h_satellites_on_fire_allianz.md","wiki/sources/h_detection_ventures.md","wiki/sources/h_technosylva_infoca_deployment.md","wiki/sources/h_effis_hazard_baseline.md","yaml-ref:phase_04_5_categories.yaml#cat_h_detection_forecasting_decision_support.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_h_detection_forecasting_decision_support.addressable_share"],"rationale":"[READ THIS WITH THE NUMBER - AGGREGATE CONTEXT, repeated on each of the five rows that dominate the product so it survives being read in isolation.] The matrix total is a 22.9% claims reduction at t=10 (total_factor_I3 = 0.770928, rebased on the observed L_0 per issue #53). NOT ONE VENTURE, PROGRAMME OR STATUTE IN THIS PILOT'S POOL PUBLISHES AN AUDITED INSURED-LOSS-REDUCTION FIGURE ANYWHERE IN SPAIN, PORTUGAL, ITALY OR GREECE, so that 22.9% is a structured analyst inference over sourced mechanisms, not a measured effect. TWO BASES ARE QUOTED BELOW AND THEY ARE NOT INTERCHANGEABLE: 'marginal' is the share of the 22.9% that disappears if the named rows are removed, and marginal shares do NOT sum to 100% across rows; 'log' is each row's share of the log-reduction, which does sum. Cells at confidence: L carry 71.4% marginal / 73.9% log. The two tier: emerging rows carry 25.7% marginal / 28.3% log, and removing both moves the headline from 22.9% to 17.0%. THIS ROW CONTRIBUTES 14.8% MARGINAL / 16.6% LOG. Full decomposition in loop_metadata.aggregation_disclosure, which does not render - see loop_metadata.disclosure_rendering_note. || value 0.916 = 1 - (effectiveness 0.12 x addressable_share 0.70), the Phase IV.5 parameters carried through unchanged per the matrix_schema.md v0.8.1 identity. MECHANISM (Hazard): information produced earlier or aimed better, so that more ignitions are suppressed while still small enough that no built asset is reached and so that suppression is pre-positioned on the flank threatening insured structures. Four on-mechanism deployment sources exist. OroraTech's Hellenic Fire System: four dedicated wildfire satellites delivering data to Greek fire services with latency in minutes and 4x4 m hotspot resolution across 100% of Greek territory. Allianz Seguros x Satellites on Fire: detection an average of 35 minutes faster than reference systems such as NASA's, and up to 30% more fires detected - vendor-stated, and in a validation phase not yet reported. Dryad's Silvanet: VOC/CO/PM2.5 sensing detecting fires in the smouldering phase over off-grid LoRaWAN mesh. Technosylva for Andalusia's INFOCA: 550 devices by 2022, Castilla-La Mancha since 2015. THE CONVERSION THAT DOES NOT EXIST: every one of those figures is a LATENCY DELTA or a DETECTION-COUNT DELTA. The conversion from detection latency to avoided insured loss is unmeasured anywhere in Europe, by anyone, in any source in this corpus - which is why effectiveness sits at 0.12 rather than at anything the vendor deltas would suggest. BASELINE NETTING, BINDING: EFFIS-grade pan-European fire-danger forecasting, the national detection and dispatch chains and the installed INFOCA/INFOCAM operational technology are INSIDE L_0 at base_year 2026 (h_effis_hazard_baseline is cited as a BASELINE MARKER - it evidences what is already in L_0, not the forward increment this cell scores). Only the genuinely new layer is creditable: OroraTech's Greek constellation launched 2026, the Allianz initiative in summer-2026 validation, Dryad with no named Southern-European end customer. Phase IV.5 cut effectiveness 0.15 -> 0.12 on exactly that ground and this cell does not walk it back. CONFIDENCE L IS A HARD CAP, not a judgement: tier is `emerging` (two source-type axes, gate cleared on EUR 55,669,561 disclosed), per matrix_schema.md v0.7 section 4. DOUBLE-COUNT CONTROLS: I006's MAK 5 retardant limb belongs to cat_h_landscape_fuel_management's mechanism and does not uplift this row; I002's KatRisk portfolio-modelling limb belongs to the R channel and does not uplift this row. INTERNAL SEAM carried forward for the critic: this row blends post-ignition detection (I001, I003, I004, I005, I006, I008) with pre-ignition danger forecasting and in-event simulation (I002, I025, I026, I027), which sit at very different points at y0 and are averaged into a single adoption anchor of 0.55."},"I4":{"value":1.0,"confidence":"L","sources":["wiki/sources/h_satellites_on_fire_allianz.md","wiki/sources/insurance_journal_2026_marsh_europe_wildfire.md"],"rationale":"NEUTRAL, EVIDENCE GAP, AND THE ONLY QUANTIFIED DATUM POINTS THE OTHER WAY. Running a detection-and-alerting layer is itself an operating activity: Allianz's 417 alerts and 73 response activations in the first five days of operation are work performed, not work avoided, and no source states that the programme reduced any insurer cost. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No claims-handling, adjuster-deployment or expense effect attributed to earlier wildfire detection or better fire-danger forecasting. Searched: h_satellites_on_fire_allianz, h_satellites_on_fire_platform, h_ororatech_greece_system, h_detection_ventures, h_technosylva_infoca_deployment, r_claims_visual_intelligence, insurance_journal_2026_marsh_europe_wildfire. Nothing on the saving side."}},"tier_check":{"computed":"emerging","declared":"emerging","agrees":true,"axes_hit":["industry_reports","regulator_filings"],"distinct_sources":7,"total_disclosed":55669561.0,"unknown_count":7,"reason":"2 axes and 55,669,561 disclosed investment (floor 50,000,000)"},"wiki_page":"wiki/categories/cat_h_detection_forecasting_decision_support.md"},{"id":"cat_v_noncombustible_envelope_fabric","name":"Non-Combustible Permanent Building-Envelope Fabric","channel":"V","channel_label":"Vulnerability","tier":"confirmed","tier_rationale":"TIER MET (issue #52, 2026-09-23): three source-type axes — regulator_filings, ngo_reports and academic_studies. academic_studies: Papathoma-Köhle et al. 2022 (Scientific Reports) build a peer-reviewed wildfire vulnerability index from the Mati (Greece, 2018) damage record and find roof material the building attribute most associated with fire-induced damage, with structural type close behind — envelope fabric as a loss driver inside the pilot's geography. It is an association, not an effect size, and does not isolate EN 13501-1 classes. PHASE X CONSEQUENCE, RETAINED PENDING LEAD REVIEW: the §5 `emerging` confidence cap (confidence: L on every active cell), applied while this row carried a declared tier deficit, is kept; lifting it is a Phase X judgement for the leads, not a consequence of meeting the axis bar. THE ORIGINAL TWO AXES. regulator_filings, which IS legitimate and on-mechanism here because EN 13501-1's A1 / A2-s1,d0 classes are an affirmative classification of non-combustibility, which is precisely this category's mechanism: Normattiva's Legge 353/2000 showing no chapter of the Italian framework addresses construction, materials or building assemblies; the verified negative that Spain's Codigo Tecnico de la Edificacion DB-SI contains no WUI provision; the record that EN 13501-1 defines A1 and A2-s1,d0 while testing flame spread and heat release and NOT wind-driven ember ingress; and the SGIFR legal framework within which Despacho 8591/2022 sits. ngo_reports: Headwaters Economics with CAL FIRE, an independent non-profit supplying the only analysis-derived cost of constructing and retrofitting to wildfire resistance located anywhere in the pilot. WITHDRAWN AT ITERATION 2 (critic M6): industry_reports is withdrawn. Its single citation was the manufacturer bundle, whose evidentiary content is expressly that NONE of the five manufacturers makes a wildfire or ember-exposure claim — the page is headed \"THE LOAD-BEARING NEGATIVE\". A source whose content is \"this source does not discuss the mechanism\" cannot be an axis hit for that mechanism under §4.1; counting it inverted the rigor test. v_noncombustible_roofing_ materials is retained in effectiveness.sources and throughout the prose as what it actually is — a NEGATIVE CONTROL ON MARKETED FIRE PERFORMANCE, and the load-bearing evidence for the addressable_share of 0.05 — but it is not an axis. Investment EUR 0: all six constituents were checked; they are established manufacturers and a regulatory instrument, and none discloses venture-level funding. Read EUR 0 here as the opposite of immaturity — the significance of this category is its installed base.","definition":"The permanent outer fabric of the building — roof covering, facade skin and the regulatory instruments that prescribe them — insofar as it is inherently non-combustible and therefore does not ignite under ember rain or contribute fuel to an external flame path. In Southern Europe this is overwhelmingly clay tile and natural slate, installed for aesthetic, thermal and durability reasons that have nothing to do with fire, supplied at industrial scale by manufacturers who market no fire performance whatsoever. The mechanism is VULNERABILITY: given a fire that reaches the building, a non-combustible envelope does not become the ignition point. The category also carries Portugal's Despacho 8591/2022, the only WUI-specific building standard located in any of the four pilot countries and the only instrument in scope whose mechanism acts on the structure rather than on the vegetation around it — and whose substantive requirements this pilot has not read.","cluster_rationale":"All six ventures act on the permanent fabric of the building envelope. I050, I051 and I052 supply the inherently non-combustible roof covering — clay tile from Alicante and Valencia, natural slate from Ourense — that dominates the pilot geography; I051 and I053 and I054 extend the same non-combustible skin onto the facade plane through ventilated slate, fibre-cement rainscreen and stone-wool board; I047 is the only regulatory instrument located in the four countries that prescribes protection measures on the structure itself rather than on the vegetation around it. Three manufacturers of the same product class are carried rather than one because depth of supply is what makes the addressable-share finding structural rather than incidental. THE ADDRESSABLE-SHARE FINDING, which is the whole point of this category for Phase X: Southern-European roof stock is already substantially A1 non-combustible as an accident of vernacular clay-tile and slate construction, and the manufacturers market no fire performance at all. Addressable share on EXISTING stock is therefore compressed toward zero and is set at 0.05. The DIRECTION of that finding is well supported by the manufacturer evidence and by the absence of any fire proposition in their marketing; the MAGNITUDE is unmeasured — no Declaration of Performance or EN 13501-1 classification has been obtained for any Southern-European tile line, so the A1 attribution is an inference, and the 0.05 is an analyst's estimate of the residual combustible share, not a measured figure. I054 carries a further caveat that only premium lines hold the non-combustibility class, so specification detail decides whether any mitigation is obtained at all. Boundary vs cat_v_retrofit_ember_path_protection: this category is the permanent fabric, already largely in place and effectively un-upgradeable on existing stock; that category is the retrofittable residual — the vent openings the non-combustible covering does not close, and applied surface treatments. That distinction is the analytically important one in this channel and is why the two are not merged. Boundary vs cat_v_defensible_space_clearance: fabric versus vegetation. Boundary vs cat_v_vulnerability_assessment_targeting: I049 (WUIVIEW) models how these assemblies fail but produces vulnerability estimates, not vulnerability reductions, and is scored there.","members":["I047","I050","I051","I052","I053","I054"],"member_names":["Portugal — Despacho n.º 8591/2022, building protection measures against rural-fire spread (RJ-SCIE, DL 220/2008)","La Escandella","Cupa Pizarras","Tejas Borja (EDILIANS group)","EQUITONE (Etex Group)","Rockpanel / Rockwool non-combustible façade lines"],"effectiveness":{"value":0.25,"lower":0.05,"upper":0.5,"unit":null},"effectiveness_definition":"ANALYST-INFERRED, and note that the binding constraint for this category is addressable_share (0.05), not effectiveness. Effectiveness here is the fractional reduction in expected insured loss for a building whose envelope is upgraded FROM a combustible covering TO an A1/A2-s1,d0 covering, given a fire that reaches it. No source measures it for wildfire exposure. Direction is well supported — a non-combustible surface cannot become the ignition point — but every European classification available (EN 13501-1 A1 and A2-s1,d0) tests flame spread and heat release, NOT wind-driven ember ingress, so no test result in existence speaks to the actual failure mode. The upper bound reflects the case where surface ignition is the dominant loss pathway; the lower bound reflects the far likelier case that loss enters through openings, vents and adjacent combustibles rather than through the roof plane itself, in which case upgrading the plane changes almost nothing. The IBHS survival figures are NOT used here: they are a vegetation and Zone-0 effect and attaching them to a materials category would be exactly the misattribution the inventory warns against.","effectiveness_sources":[{"slug":"v_noncombustible_roofing_materials","title":"Southern-European non-combustible roofing and cladding manufacturers — La Escandella (clay tile), with Cupa Pizarras (natural slate) and EQUITONE/Etex (fibre cement) recorded as bundle claims; the vernacular-baseline evidence","tier":"T3","publisher":"roofing_manufacturer_bundle","url":"https://www.laescandella.com/es/","wayback_url":"https://web.archive.org/web/20260526082609/https://www.laescandella.com/es/","snapshot_hash":"6d20c5a1a7a20c3c8b25cf04c46b3559b1c430eab0f5c0978b57e0d0102aff23","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/6d20c5a1a7a20c3c8b25cf04c46b3559b1c430eab0f5c0978b57e0d0102aff23.html","publication_date":null,"fetched_at":"2026-07-31T12:58:09Z","page_exists":true},{"slug":"v_wui_building_standards_gap","title":"WUI building standards in Southern Europe — Legge-quadro 353/2000 art. 1 as the Italian framework-and-delegation instrument, with the Portuguese, Spanish and EN-standard gaps recorded as bundle claims","tier":"T1","publisher":"normattiva_italy","url":"https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:legge:2000-11-21;353","wayback_url":"https://web.archive.org/web/20260314021546/https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:legge:2000-11-21;353","snapshot_hash":"caf6ec2ec0da6f3027d6451ca3c7cc33d05dcc698547e8ccb81a32b65312c8f9","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/caf6ec2ec0da6f3027d6451ca3c7cc33d05dcc698547e8ccb81a32b65312c8f9.html","publication_date":"2000-11-30","fetched_at":"2026-07-31T13:06:09Z","page_exists":true},{"slug":"v_retrofit_cost_evidence","title":"Headwaters Economics with CAL FIRE — the cost of retrofitting a home for wildfire resistance, and construction costs for a wildfire-resistant home","tier":"T2","publisher":"headwaters_economics","url":"https://headwaterseconomics.org/natural-hazards/retrofitting-home-wildfire-resistance/","wayback_url":"https://web.archive.org/web/20260516231219/https://headwaterseconomics.org/natural-hazards/retrofitting-home-wildfire-resistance/","snapshot_hash":"030398ed8975a61398dfc3f782264900ad9ca63f76d440d7259d9f8a008a6a11","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/030398ed8975a61398dfc3f782264900ad9ca63f76d440d7259d9f8a008a6a11.html","publication_date":"2024-06-26","fetched_at":"2026-07-31T13:06:05Z","page_exists":true}],"addressable_share":0.05,"axes_hit":["ngo_reports","regulator_filings","academic_studies"],"citations":[{"axis":"regulator_filings","wiki_source_id":"v_wui_building_standards_gap","span":"EN 13501-1 under Delegated Regulation (EU) 2016/364 defines A1 and A2-s1,d0 as the non-combustibility benchmark but tests flame spread and heat release, NOT wind-driven ember ingress; Legge 353/2000's chapters cover forecasting, prevention, active suppression, administrative functions and finance, and none addresses construction, materials or building assemblies; Spain's CTE DB-SI contains no WUI or vegetation-fire-exposure provision."},{"axis":"regulator_filings","wiki_source_id":"v_defensible_space_portugal","span":"Portugal administers prevention obligations through named legal instruments under the SGIFR Quadro Legal / Regulamentar — the regime within which Despacho 8591/2022, the only located WUI-specific building standard in the four countries, sits unread."},{"axis":"ngo_reports","wiki_source_id":"v_retrofit_cost_evidence","span":"Headwaters Economics with CAL FIRE: developers are already beginning to incorporate wildfire-resistant materials and methods into new homes without a significant impact on building costs, while the open question is how to upgrade millions of existing homes — the new-build versus existing-stock distinction that sets this category's addressable share."},{"axis":"academic_studies","wiki_source_id":"papathoma_kohle_2022_wildfire_vulnerability_index","span":"Roof material showed the strongest association, partially confirming previous studies [23] suggesting that the roof is one of the weakest points of a building."}],"total_disclosed":0.0,"unknown_count":6,"adoption":{"y0":{"value":0.02,"lower":0.0,"upper":0.05,"unit":null},"y3":{"value":0.06,"lower":0.02,"upper":0.12,"unit":null},"y5":{"value":0.1,"lower":0.03,"upper":0.2,"unit":null},"y10":{"value":0.2,"lower":0.05,"upper":0.32,"unit":null}},"saturation_cap":0.35,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced, and the market structure explains why. The Southern-European roof stock is substantially non-combustible as an ACCIDENT of vernacular clay-tile and slate construction, installed for aesthetic, thermal and durability reasons, and the manufacturers market NO fire performance whatsoever - so there is no product claim for an insurer to price against and no certification an underwriter could require. No European wildfire building-certification scheme exists to attach a premium to; IBHS's Wildfire Prepared Home is named in the corpus precisely as the benchmark Southern Europe has no counterpart to. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No premium credit or underwriting condition tied to A1/A2-s1,d0 envelope classification in ES/PT/IT/GR. Searched: v_noncombustible_roofing_materials, v_wui_building_standards_gap, v_retrofit_cost_evidence, v_ibhs_home_hardening_evidence, v_tecnifuego_spain_negative_control, v_insurer_wildfire_initiatives."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Roof-covering and facade manufacturing, and the standards that classify them, do not touch insurer asset allocation, duration or investment return. Searched: v_noncombustible_roofing_materials, v_wui_building_standards_gap, v_retrofit_cost_evidence. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.9875,"confidence":"L","sources":["wiki/sources/v_wui_building_standards_gap.md","wiki/sources/v_noncombustible_roofing_materials.md","wiki/sources/v_retrofit_cost_evidence.md","yaml-ref:phase_04_5_categories.yaml#cat_v_noncombustible_envelope_fabric.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_v_noncombustible_envelope_fabric.addressable_share"],"rationale":"value 0.9875 = 1 - (effectiveness 0.25 x addressable_share 0.05), Phase IV.5 parameters unchanged. THE BINDING CONSTRAINT IS THE ADDRESSABLE SHARE, NOT THE EFFECTIVENESS, and the cell is near-neutral for that reason rather than because the mechanism is weak. MECHANISM: a non-combustible outer plane does not become the ignition point under ember rain or external flame. THE ADDRESSABLE-SHARE FINDING, which is the whole analytical point of the category: Southern-European roof stock is already substantially A1 as an ACCIDENT of vernacular clay-tile and slate construction, so the addressable share on EXISTING stock compresses to 0.05 - and Headwaters Economics with CAL FIRE draws the distinction the share rests on, that new-build to wildfire-resistant standards adds little cost while upgrading existing stock is the open problem. THE TEST THAT DOES NOT TEST THE FAILURE MODE, the load-bearing finding here: EN 13501-1 under Delegated Regulation (EU) 2016/364 defines A1 and A2-s1,d0 as the non-combustibility benchmark but tests FLAME SPREAD AND HEAT RELEASE, NOT wind-driven ember ingress - so no test result in existence speaks to the actual wildfire failure mode for a European roof. THE REGULATORY VOID, verified: Legge 353/2000's chapters cover forecasting, prevention, active suppression, administrative functions and finance, and NONE addresses construction, materials or building assemblies; Spain's Codigo Tecnico de la Edificacion DB-SI contains no WUI or vegetation-fire-exposure provision. Portugal's Despacho 8591/2022 is the only WUI-specific building standard located in any of the four countries and THIS PILOT HAS NOT READ ITS SUBSTANTIVE REQUIREMENTS. WHAT THE MAGNITUDE RESTS ON, honestly: the DIRECTION of the addressable-share compression rests on manufacturer evidence and on the ABSENCE of any fire proposition in their marketing; the MAGNITUDE is unmeasured, because no Declaration of Performance or EN 13501-1 classification has been obtained for any Southern-European tile line. The A1 attribution is an inference. CONFIDENCE L DESPITE THE `confirmed` LABEL: Phase IV.5 records a declared tier deficit (two axes, EUR 0 disclosed investment) and directs Phase X to apply the section 5 `emerging` cap. v_noncombustible_roofing_materials is cited as a NEGATIVE CONTROL on marketed fire performance - its evidentiary content is that none of the five manufacturers makes a wildfire or ember-exposure claim - and was withdrawn at Phase IV.5 as an axis hit, because a source whose content is 'this source does not discuss the mechanism' is not an axis for that mechanism; it is not reinstated here. THE IBHS FIGURES ARE NOT USABLE HERE: they are a vegetation and Zone-0 effect and attaching them to a materials category would be exactly the misattribution the inventory warns against. BASELINE NETTING, STATED AT ITERATION 3: the vernacular non-combustible roof stock is already installed and is inside L_0 by construction - that is the whole reason addressable_share is compressed to 0.05, so the netting is carried in the SHARE rather than in the adoption series. What the adoption series then counts is conversion of the residual combustible envelope, denominated over that residue; the Phase V y0 note records that with no code trigger anywhere in the geography such conversion 'is happening only incidentally at base year'."},"I4":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no operational channel. Building fabric does not change how a claim is handled, and no source in the corpus attributes an insurer handling or inspection cost to envelope classification. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No claims-operations or survey-cost effect of non-combustible construction. Searched: v_noncombustible_roofing_materials, v_wui_building_standards_gap, v_retrofit_cost_evidence, r_claims_visual_intelligence."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","ngo_reports","regulator_filings"],"distinct_sources":4,"total_disclosed":0.0,"unknown_count":6,"reason":"3 evidence axes across 4 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_v_noncombustible_envelope_fabric.md"},{"id":"cat_v_retrofit_ember_path_protection","name":"Retrofit Ember- and Heat-Path Protection for Existing Buildings","channel":"V","channel_label":"Vulnerability","tier":"confirmed","tier_rationale":"TIER MET (issue #52, 2026-09-23): three source-type axes — industry_reports, ngo_reports and academic_studies. academic_studies: Ribeiro et al. 2020 (Fire), the CEIF/ADAI field survey of structures hit by the Pedrógão Grande fire complex (Portugal, 2017), attribute about 61% of structure damage to ember deposition at weak spots and name unscreened vents in older houses as a firebrand entry point — the retrofit target on the existing stock. It attributes damage paths; it does not measure the loss avoided by retrofitting. PHASE X CONSEQUENCE, RETAINED PENDING LEAD REVIEW: the §5 `emerging` confidence cap (confidence: L on every active cell), applied while this row carried a declared tier deficit, is kept; lifting it is a Phase X judgement for the leads, not a consequence of meeting the axis bar. THE ORIGINAL TWO AXES. industry_reports: the ember-resistant vent bundle (BrandGuard's opening-by-opening range, Vulcan's retrofit mesh-and-intumescent kit for an existing opening, O'Hagin's Fire&Ice profile-specific vents for clay-tile, concrete-tile and natural-slate roofs), the ECOFIRE exterior-retardant page, and Tecnifuego as the negative control. ngo_reports: Headwaters Economics with CAL FIRE, whose USD 2,000-15,000 bracket for simple effective actions names installing flame- and ember-resistant vents first. WITHDRAWN AT ITERATION 2 (critic M5): regulator_filings is withdrawn. Its citation established the ABSENCE of regulatory coverage — no European counterpart to ASTM E2886, no CE marking held by any located maker, nothing to certify against. §4.1's rigor heuristic treats an axis hit as affirmative third-party attention to the mechanism, so counting a documented regulatory void as a regulator-filings hit inverts the test. Compounding it, v_wui_building_standards_gap's own caveat 5 records that the EN-gap claim is second-hand and consequential and should be re-verified against the Delegated Regulation and the harmonised standards list before use. The standards gap is retained throughout the prose as a finding — it is the actionable object in this category — but it is not an axis. Investment EUR 0: all three constituents were checked and none discloses funding; there is no European manufacturer, no CE marking and no EN test method, so there is no European market for capital to have entered. This remains the category where the pilot's evidence is thinnest relative to its analytical importance.","definition":"Measures that can be applied to an EXISTING building without replacing its permanent fabric, to close or protect the residual pathways by which fire enters a structure whose envelope is already non-combustible: ember-resistant and intumescent vents and retrofit screening kits for roof-void, eave, gable, soffit and foundation openings, and applied water-based retardant coatings that raise the heat and flame load a treated exterior surface can absorb during the passage of a fire front. The mechanism is VULNERABILITY, and it is the complement of the non-combustible-fabric category: embers entering the roof void through unscreened vents ignite the structure without any flame contact, which is precisely the ignition pathway that Southern Europe's vernacular clay-tile and slate roof covering does not close.","cluster_rationale":"All three ventures share the property that decides whether any structural mitigation can reach meaningful adoption in this geography within a ten-year horizon: they act on EXISTING buildings without replacing building fabric. I055 makes ember-resistant vents in exactly the clay-tile, concrete-tile and natural-slate roof profiles that dominate the four countries; I056 adds the only located retrofit kit that screens an existing vent opening rather than replacing the unit; I057 is an applied coating that raises the surface's tolerance of radiant and flame load and is removed by weather rather than built in. The Southern-European wildfire problem is overwhelmingly an existing-stock problem, and retrofittability is therefore the clustering variable here rather than the physical form of the measure. THE SEAM, flagged for the critic: an ember-screened vent (a permanent hardware change closing an opening) and a sprayed retardant (a temporary surface treatment applied in anticipation of a front) are not the same physical act, and a purist reading would separate them. They are carried together because both are applied to existing stock without altering the permanent envelope, because neither clears the three-axis bar alone, and because neither has any disclosed funding, so a split would produce two `insufficient` categories and orphan three inventory ventures. Conceded to the critic (J1): the shared property is a DEPLOYMENT property, not a mitigation mechanism, which strains §2's definition of a category; the merge is accepted only under the orphan trap. PHASE V INSTRUCTION, BINDING, because the two limbs have completely different adoption dynamics: a vent retrofit is a ONE-OFF CAPITAL ACT with a permanent effect, while ECOFIRE is a PER-EVENT APPLICATION requiring warning time and a person to apply it. A single adoption anchor for this row is wrong. Anchor on the VENT LIMB and hold the applied-coating limb at an EVIDENCE GAP — its only documented wildfire use was July 2026 with results undocumented. Boundary vs cat_v_noncombustible_envelope_fabric: that category is the permanent plane, already largely non-combustible and effectively fixed on existing stock; this category is the residual pathway that plane leaves open and the only part of the structural mechanism with real headroom. Boundary vs cat_v_defensible_space_clearance: that category removes the fuel that produces the embers; this one stops the embers that arrive anyway. A property can and should do both, and the two must be scored on separate rows because clearance is compelled by statute in three of four countries while ember screening is not sold in Europe at all.","members":["I055","I056","I057"],"member_names":["O'Hagin Fire&Ice profile-specific roof vents","Vulcan Vents","ECOFIRE (Simon Rack / Ecofire Global Distributor)"],"effectiveness":{"value":0.2,"lower":0.05,"upper":0.45,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. Fractional reduction in expected insured loss for a building whose vent openings are screened or replaced with ember-resistant units and, where applied, whose exterior surfaces are pre-treated, given a fire that reaches the parcel. No source measures it for any of the three ventures. The only traceable independent effectiveness attribution in this mechanism family — up to 96% of radiant and 92% of convective heat blocked by aluminized structure wrap, attributed to Takahashi 2019 — belongs to a product excluded from the pool for want of Southern-European deployment evidence, and the underlying paper was not retrieved by this pilot; it is recorded to show what was considered, not used. ECOFIRE's INIA-CSIC certification states efficacy greater than water for the sample analysed only; the vendor's separate 30-to-40-times-water claim is contradicted by the fact-check of that same study and is NOT used anywhere. The upper bound assumes vent-path ignition is a large share of structure loss where the roof plane itself does not ignite; the lower bound reflects that no European field evidence exists, that ember screening carries a measurable ventilation penalty in hot Mediterranean roof spaces (net free area 81% / 75% / 71% for quarter-, eighth- and sixteenth-inch mesh), and that ECOFIRE's first documented wildfire use was July 2026 with results undocumented.","effectiveness_sources":[{"slug":"v_ember_resistant_vents","title":"Ember-resistant vents — BrandGuard Vents, with Vulcan Vents and O'Hagin recorded as bundle claims; the category with no European supplier and no European standard","tier":"T3","publisher":"ember_vent_manufacturer_bundle","url":"https://www.brandguardvents.com/","wayback_url":"https://web.archive.org/web/20260731130821/https://www.brandguardvents.com/","snapshot_hash":"117a2906179e164c10a3a718b5e26bc1a68f83cfe7c03e1ff80024c281628d62","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/117a2906179e164c10a3a718b5e26bc1a68f83cfe7c03e1ff80024c281628d62.html","publication_date":null,"fetched_at":"2026-07-31T12:58:21Z","page_exists":true},{"slug":"v_exterior_sprinkler_gel","title":"Exterior fire-retardant gels, wraps and sprinklers — ECOFIRE (Spain), with Simon Rack, Firezat and the Tecnifuego negative control recorded as bundle claims","tier":"T3","publisher":"ecofire_global_distributor","url":"https://ecofireglobaldistributor.com/","wayback_url":"https://web.archive.org/web/20260728074243/https://ecofireglobaldistributor.com/","snapshot_hash":"d761299ed856f37b6b5f141395d344fe58dd8937ba9191947483cfa4bf318024","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/d761299ed856f37b6b5f141395d344fe58dd8937ba9191947483cfa4bf318024.html","publication_date":null,"fetched_at":"2026-07-31T13:11:50Z","page_exists":true},{"slug":"v_retrofit_cost_evidence","title":"Headwaters Economics with CAL FIRE — the cost of retrofitting a home for wildfire resistance, and construction costs for a wildfire-resistant home","tier":"T2","publisher":"headwaters_economics","url":"https://headwaterseconomics.org/natural-hazards/retrofitting-home-wildfire-resistance/","wayback_url":"https://web.archive.org/web/20260516231219/https://headwaterseconomics.org/natural-hazards/retrofitting-home-wildfire-resistance/","snapshot_hash":"030398ed8975a61398dfc3f782264900ad9ca63f76d440d7259d9f8a008a6a11","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/030398ed8975a61398dfc3f782264900ad9ca63f76d440d7259d9f8a008a6a11.html","publication_date":"2024-06-26","fetched_at":"2026-07-31T13:06:05Z","page_exists":true}],"addressable_share":0.15,"axes_hit":["industry_reports","ngo_reports","academic_studies"],"citations":[{"axis":"industry_reports","wiki_source_id":"v_ember_resistant_vents","span":"O'Hagin's product list includes Clay Tile Vents, Concrete Tile Vents and Natural Slate Vents, with the Fire&Ice line offered as profile-specific vents for concrete, clay and synthetic tile roofs — precisely the roof profiles that dominate the Spanish, Portuguese, Italian and Greek building stock. Vulcan markets a retrofit product that installs mesh screening and honeycomb intumescent coating on any existing vent on the house."},{"axis":"industry_reports","wiki_source_id":"v_exterior_sprinkler_gel","span":"ECOFIRE is marketed for dwelling protection and as a firebreak agent — Viviendas: proteccion preventiva y actuacion rapida ante conatos en viviendas y su entorno — with the vendor's 30-to-40-times-water efficacy claim contradicted by the fact-check of its own INIA-CSIC study, whose citable framing is efficacy greater than water for the sample analysed only."},{"axis":"industry_reports","wiki_source_id":"v_tecnifuego_spain_negative_control","span":"NEGATIVE CONTROL: Tecnifuego's mainstream Spanish autoproteccion guidance mentions no exterior sprinklers, water curtains or fire gels at all — the Spanish fire-protection industry does not recognise this category."},{"axis":"ngo_reports","wiki_source_id":"v_retrofit_cost_evidence","span":"Headwaters Economics with CAL FIRE: retrofitting costs between USD 2,000 and USD 15,000 for simple and effective actions such as installing flame- and ember-resistant vents, placing metal flashing along a deck, keeping gutters clean or using noncombustible mulch; a full retrofit to the highest level could cost nearly USD 100,000 but is in many cases not necessary."},{"axis":"academic_studies","wiki_source_id":"ribeiro_2020_pedrogao_grande_structure_impact","span":"About 61% of the structures (Table 10) were damaged because of the deposition of burning embers, or firebrands, in one or more weak spots."}],"total_disclosed":0.0,"unknown_count":3,"adoption":{"y0":{"value":0.01,"lower":0.0,"upper":0.03,"unit":null},"y3":{"value":0.03,"lower":0.0,"upper":0.08,"unit":null},"y5":{"value":0.06,"lower":0.01,"upper":0.15,"unit":null},"y10":{"value":0.15,"lower":0.02,"upper":0.28,"unit":null}},"saturation_cap":0.3,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced, and structurally there is nothing for a premium to attach to: there is no European manufacturer of ember-resistant vents, no CE marking held by any located maker, and no EN test method - so there is no European market, no certification and no verifiable installed state an underwriter could price. The corpus does contain a policyholder COST figure (Headwaters Economics with CAL FIRE: USD 2,000-15,000 for simple effective actions, naming ember-resistant vents first; up to nearly USD 100,000 for a full retrofit) - that is what the OWNER pays, not what the insurer charges, and it is US-derived. Tecnifuego is a negative control: mainstream Spanish self-protection guidance mentions no exterior sprinklers, water curtains or fire gels at all. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No premium credit, verification scheme or underwriting condition attached to vent screening or exterior pre-treatment in ES/PT/IT/GR. Searched: v_ember_resistant_vents, v_exterior_sprinkler_gel, v_retrofit_cost_evidence, v_tecnifuego_spain_negative_control, v_property_risk_scoring_ventures, v_insurer_wildfire_initiatives."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Vent hardware and applied retardant coatings do not touch insurer asset allocation, duration or investment return. Searched: v_ember_resistant_vents, v_exterior_sprinkler_gel, v_retrofit_cost_evidence. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.97,"confidence":"L","sources":["wiki/sources/v_ember_resistant_vents.md","wiki/sources/v_exterior_sprinkler_gel.md","wiki/sources/v_retrofit_cost_evidence.md","wiki/sources/v_tecnifuego_spain_negative_control.md","yaml-ref:phase_04_5_categories.yaml#cat_v_retrofit_ember_path_protection.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_v_retrofit_ember_path_protection.addressable_share"],"rationale":"value 0.970 = 1 - (effectiveness 0.20 x addressable_share 0.15), Phase IV.5 parameters unchanged. THE MECHANISM WITH THE LARGEST HEADROOM AND THE THINNEST EVIDENCE IN THE PILOT. Embers entering the roof void through unscreened vents ignite a structure without any flame contact - precisely the pathway Southern Europe's clay-tile and slate covering does NOT close, which is why this category and cat_v_noncombustible_envelope_fabric are kept apart. Product evidence exists and is on-mechanism: O'Hagin's Fire&Ice line offers profile-specific vents for concrete, clay and synthetic tile and natural slate roofs - exactly the profiles that dominate the four countries - and Vulcan markets a retrofit kit installing mesh screening and honeycomb intumescent coating on any existing vent. NO SOURCE MEASURES EFFECTIVENESS FOR ANY OF THE THREE CONSTITUENTS. THREE NUMBERS THAT LOOK USABLE AND ARE NOT, recorded so nobody reaches for them: (1) the aluminized structure-wrap attribution (up to 96% radiant and 92% convective heat blocked, Takahashi 2019) belongs to a product EXCLUDED from the pool for want of Southern-European deployment evidence, and the underlying paper was never retrieved by this pilot; (2) ECOFIRE's vendor claim of 30-40 times the efficacy of water is CONTRADICTED by the fact-check of its own INIA-CSIC study, whose only citable framing is efficacy greater than water for the sample analysed only; (3) the ventilation penalty is real and cuts AGAINST the mechanism - net free area of 81% / 75% / 71% for quarter-, eighth- and sixteenth-inch mesh in hot Mediterranean roof spaces. THE BINDING LIMB SPLIT, AND A SUB-LIMB GAP CARRIED INSIDE AN OTHERWISE SOURCED CELL: the two limbs are different acts with different dynamics - a vent retrofit is a ONE-OFF CAPITAL ACT with a permanent effect, ECOFIRE is a PER-EVENT APPLICATION requiring warning time and a person to apply it. This value is anchored on the VENT LIMB. The applied-coating limb is held at an evidence gap on Phase IV.5's explicit instruction - its first documented wildfire use was July 2026 with results undocumented - and is NOT inside the 0.970. As with the arson limb on cat_h_fire_management_programmes_and_statutes, the v0.7 cell shape has no sub-cell in which to flag it, so it is named here. CONFIDENCE L DESPITE THE `confirmed` LABEL: declared tier deficit (two axes, EUR 0 disclosed), section 5 `emerging` cap applied. WITHDRAWN AND NOT REINSTATED: the regulator_filings axis, because its citation established the ABSENCE of regulatory coverage - the standards gap remains the actionable finding of this category but it is not affirmative evidence of the mechanism. Tecnifuego is cited as the negative control that bounds adoption from above: the Spanish fire-protection industry does not recognise exterior pre-treatment at all. BASELINE NETTING, STATED AT ITERATION 3, and it is close to vacuous here: there is no installed base to net out. The Phase V y0 note records that 'a product that cannot be certified against anything cannot be specified into a European contract, so base-year adoption is effectively nil' - no European manufacturer, no CE marking, no EN test method - so essentially none of this mechanism is inside L_0 and the whole series is a forward increment."},"I4":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no operational channel. Vent hardware and applied coatings do not change how a claim is handled. One direction-of-travel note the reader should not misread: an applied coating requires warning time and a person to apply it, which is an operational cost incurred by the policyholder or a contractor, not a saving to the insurer. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No claims-handling or inspection-cost effect of vent screening or exterior pre-treatment. Searched: v_ember_resistant_vents, v_exterior_sprinkler_gel, v_retrofit_cost_evidence, r_claims_visual_intelligence."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","ngo_reports"],"distinct_sources":5,"total_disclosed":0.0,"unknown_count":3,"reason":"3 evidence axes across 5 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_v_retrofit_ember_path_protection.md"},{"id":"cat_v_defensible_space_clearance","name":"Statutory Defensible-Space Clearance at the Building Perimeter","channel":"V","channel_label":"Vulnerability","tier":"confirmed","tier_rationale":"Three source-type axes documented. regulator_filings: the Portuguese SGIFR portal with the graded 0–2 m / 2–10 m geometry and the DL 82/2021 art. 3 duty, the Hellenic Fire Service publication of the preventive-fire-protection KYA with the national property register and the Pyrosvestiki Diataxi 19/2024 fine regime, the Spanish MITECO portal with the Diputacio de Barcelona and Catalan statutory scope recorded as bundle claims, the Georisques OLD dataset for the French comparator, and Normattiva's text of Legge 353/2000 art. 1 establishing that Italy delegates operative clearance geometry to the regions. academic_studies: the FIREPRIME / WUIVIEW record (UCPM GA 101140381 and the UPC-led performance-based WUI simulation published in the International Journal of Wildland Fire, June 2023), which is the only European research instrument that scores nearby vegetation alongside construction at household level. industry_reports: Tecnifuego, the Spanish fire-protection industry association that collaborates with the administration on legislation, functioning here as a NEGATIVE control — its mainstream autoproteccion guidance specifies no clearance distances at all — and the IBHS Wildfire Prepared Home programme as the benchmark certification scheme Southern Europe has no counterpart to. No NGO-report axis. Investment intensity EUR 0 disclosed: every constituent is a statutory duty or a public programme with no venture-level funding figure.","definition":"Legally compelled or programme-delivered maintenance of a low-fuel band immediately around a building or an urbanisation, so that a fire arriving at the parcel boundary meets a maintained clearance instead of continuous shrub. The mechanism is VULNERABILITY: the fire still occurs and still arrives, but the radiant and ember load reaching the building envelope is reduced, and the fireline intensity at the point where wildland fire meets insured property is lowered. The instruments differ in how compulsion is applied — a graded statutory profile with municipal substitution and cost recovery, a filed self-protection plan as a condition of occupation, a declaration recorded in a national compliance register backed by fines and a custodial sanction, a provincial public-works programme that executes the strips itself, or a framework that delegates the geometry to regions and therefore prescribes nothing — but the physical act is one act.","cluster_rationale":"All six ventures compel or execute the same physical act — maintaining a cleared band between the vegetation and the building — and differ only in the instrument of compulsion and in the geometry each jurisdiction prescribes. I043 (Portugal) is the graded 0–2 m / 2–10 m duty with municipal substitution; I044 (Greece) adds a national property-level compliance register that makes each property's mitigation state an addressable record rather than an inference; I042 (Catalonia) makes a filed self-protection plan a condition of occupation for urbanizations in the 500 m forest band; I041 (Diputacio de Barcelona) is the provincial programme that physically executes the perimeter strips and is the only Southern-European source of executed-hectares throughput; I045 (Italy) defines the WUI in law but sets no national numeric distance, and is carried here as the negative member of the cluster because a category cannot be scored across four countries while omitting the one where the duty does not bite; I046 (France) is outside the pilot geography and is carried as the comparator that shows the only located instance of defensible-space compliance carrying a contractual price, through a deductible on indemnity. THE DISJOINTNESS CONTROL WITH cat_h_landscape_fuel_management: I043 here carries ONLY the band dimensioned to the building envelope; I015 there carries ONLY the landscape fuel-management network and the 100 m settlement band. They are two mechanisms of one legal instrument, kept disjoint by geometry — no square metre appears in both addressable shares — but they share ONE compliance rate and ONE enforcement machinery, so their adoption anchors must be driven from a single Portuguese compliance series and Phase X must not credit the same clearance twice. Boundary vs cat_v_noncombustible_envelope_fabric: this category changes what is around the building, that one changes what the building is made of; a compliant parcel with a combustible roof and a non-compliant parcel with a clay-tile roof are different risks and must be scored on different rows. Boundary vs cat_v_vulnerability_assessment_targeting: that category decides WHICH parcels get cleared when clearance capacity is smaller than the obligated area; this category is the clearance itself.","members":["I041","I042","I043","I044","I045","I046"],"member_names":["Diputació de Barcelona — Pla de Prevenció d'Incendis Forestals en Urbanitzacions","Generalitat de Catalunya — Llei 5/2003 + Decret 123/2005 urbanisation self-protection regime","Portugal — building-perimeter fuel-management duty (DL 82/2021 art. 3, current standard Despacho 675/2026)","Greece — mandatory plot clearing and the Εθνικό Μητρώο Προληπτικής Πυροπροστασίας property register","Italy — DL 120/2021 wildland-urban-interface regime with regional delegation","France — Obligations Légales de Débroussaillement (comparator regime)"],"effectiveness":{"value":0.3,"lower":0.1,"upper":0.55,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. Fractional reduction in expected insured wildfire property loss, given a fire reaching the parcel, at full compliance with the graded clearance profile. The pilot's closest quantification is the 2021 IBHS + ZestyAI finding that an ember-resistant buffer cuts a structure's risk of igniting from a wildfire in half and that clearing perimeter vegetation can nearly double a structure's likelihood of surviving. THAT FINDING IS USED HERE AND ONLY HERE, because it is a VEGETATION and Zone-0 effect; it must never be attached to any product category. PROVENANCE OF THE ANCHOR, stated in full at iteration 2 (critic S2): it is US-derived, single-study and insurer-funded; it is NOT validated on Mediterranean fuels, building spacing or the Southern-European settlement pattern; NEITHER FIGURE IS IN EITHER SNAPSHOT — both are carried from the scout dossier, un-snapshotted in this pilot; and ZestyAI co-authored the study AND SELLS A WILDFIRE RISK SCORE, so the study is not independent of one of its commercial beneficiaries. That is why the central value (0.30) sits well below what \"nearly double the survival likelihood\" would imply and the bracket runs from 0.10 to 0.55. The lower bound reflects Italy, where no national numeric clearance distance binds a private owner at all; the upper bound reflects Portugal's fully executed graded profile with canopies held 5 m from the structure. NOT baseline-netted downward as the three H categories were: unlike suppression and fuel-break capacity, compliance with the perimeter duty is NOT near-saturated in 2026 — Portugal reports roughly 40% non-compliance and Italy prescribes no numeric distance at all — so most of this category's effect is genuinely forward-available headroom rather than embedded history. Phase V nonetheless carries the 2026 compliance level as the y0 anchor.","effectiveness_sources":[{"slug":"v_ibhs_home_hardening_evidence","title":"IBHS Wildfire Prepared Home — US home-hardening certification programme and the IBHS wildfire research programme (ibhs.org), the benchmark Southern Europe has no counterpart to","tier":"T2","publisher":"ibhs","url":"https://wildfireprepared.org/","wayback_url":"https://web.archive.org/web/20260720081747/https://wildfireprepared.org/","snapshot_hash":"6a930f3e14483ce95338ca263b10b634053616f6edb48b21f58a456305d0e8dc","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/6a930f3e14483ce95338ca263b10b634053616f6edb48b21f58a456305d0e8dc.html","publication_date":null,"fetched_at":"2026-07-31T12:54:07Z","page_exists":true},{"slug":"v_defensible_space_portugal","title":"Portugal — SGIFR integrated rural-fire management system portal (with DL 82/2021 fuel-management duties and the Aldeia Segura settlement programme recorded as bundle claims)","tier":"T1","publisher":"sgifr_portugal","url":"https://www.sgifr.gov.pt/","wayback_url":"https://web.archive.org/web/20260731130845/https://sgifr.gov.pt/","snapshot_hash":"f6ff390cb1e69487673004b3c602e69cd6b2216bef5e649c1700134a8bfdf8fa","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/f6ff390cb1e69487673004b3c602e69cd6b2216bef5e649c1700134a8bfdf8fa.html","publication_date":null,"fetched_at":"2026-07-31T13:07:21Z","page_exists":true},{"slug":"v_defensible_space_greece","title":"Greece — Hellenic Fire Service announcements on the preventive fire-protection KYA, mandatory plot clearing and the National Registry of Preventive Fire Protection Measures","tier":"T1","publisher":"hellenic_fire_service","url":"https://www.fireservice.gr/el","wayback_url":"https://web.archive.org/web/20260718212906/https://www.fireservice.gr/el","snapshot_hash":"00afe1ccaf00c0de6f247ecfef4804bd9fd9402f236c2d924ae03d9079a2f943","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/00afe1ccaf00c0de6f247ecfef4804bd9fd9402f236c2d924ae03d9079a2f943.html","publication_date":"2026-04-25","fetched_at":"2026-07-31T13:09:12Z","page_exists":true},{"slug":"v_france_old_insurance_linkage","title":"France — Obligations Légales de Débroussaillement (Géorisques national OLD zoning database), the comparator regime that wires defensible-space compliance to the insurance contract","tier":"T1","publisher":"georisques_france","url":"https://www.georisques.gouv.fr/donnees/bases-de-donnees/obligations-legales-de-debroussaillement","wayback_url":"https://web.archive.org/web/20260716142104/https://www.georisques.gouv.fr/donnees/bases-de-donnees/obligations-legales-de-debroussaillement","snapshot_hash":"3e3229d6cc1b642005901d328c746751ed7f4ae2dea8fb44f298dd3b3407fb38","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/3e3229d6cc1b642005901d328c746751ed7f4ae2dea8fb44f298dd3b3407fb38.html","publication_date":null,"fetched_at":"2026-07-31T12:59:37Z","page_exists":true}],"addressable_share":0.45,"axes_hit":["industry_reports","academic_studies","regulator_filings"],"citations":[{"axis":"regulator_filings","wiki_source_id":"v_defensible_space_portugal","span":"Decreto-Lei 82/2021 art. 3 imposes fuel management of 50 m around isolated buildings and 100 m around settlements, annually, with municipal substitution and cost recovery where the owner defaults; the graded interface prescribes elimination of all surface fuels in the 0-2 m band, removal or thinning of shrubs and trees in the 2-10 m band, and canopies kept at least 5 m from the building. [Scout-verified bundle claim; diariodarepublica.pt serves a JavaScript-only shell to this pilot's fetcher.]"},{"axis":"regulator_filings","wiki_source_id":"v_defensible_space_portugal","span":"ADOPTION CEILING, RECORDED SEPARATELY: the scout dossier records approximately 40% non-compliance with the fuel-management duty in 2024. This is a distinct bundle claim from the statutory prescription above, it is second-hand and un-snapshotted, the source page directs that it enter Phase X as a wide bracket rather than a point estimate, and the venture inventory ranks its retrieval #4 of the highest-value outstanding items and calls it the single variable that decides whether the largest V-channel mechanism bites at all."},{"axis":"regulator_filings","wiki_source_id":"v_defensible_space_greece","span":"The Joint Ministerial Decision governs preventive fire protection of properties, the clearing obligations for plots and other uncovered spaces, and the operation of the National Registry of Preventive Fire Protection Measures for Properties; enforcement runs through an administrative-fine regime anchored in Law 4662/2020."},{"axis":"regulator_filings","wiki_source_id":"v_defensible_space_programmes_es","span":"Diputacio de Barcelona's Pla de Prevencio d'Incendis Forestals en Urbanitzacions reported for 2025: 73 delimitation maps, 78 drafted perimeter-strip projects, 61 executed strip works, 309 hectares of protection strips executed and 266 municipalities supported; the Catalan statutory scope is the 500 m band around urbanizations in forest terrain under Llei 5/2003 and Decret 123/2005."},{"axis":"regulator_filings","wiki_source_id":"v_france_old_insurance_linkage","span":"The French state's risk portal states that debroussaillement is the best means of protecting one's dwelling from a fire and of limiting its spread; the obligation is compulsory in classified fire-risk areas across 48 departements and its perimeter is published as national open geospatial data at parcel scale."},{"axis":"regulator_filings","wiki_source_id":"v_wui_building_standards_gap","span":"Legge 353/2000 art. 1 comma 3 delegates implementation to the regions — the statutory mechanism behind the finding that Italy sets no national numeric clearance distance on private owners."},{"axis":"academic_studies","wiki_source_id":"v_fireprime_wuiview_projects","span":"FIREPRIME's household assessment scores factors spanning both structure and surroundings — construction materials, design, nearby vegetation and the storage of flammable items — under UCPM GA 101140381, coordinated from CERTEC-UPC Barcelona."},{"axis":"industry_reports","wiki_source_id":"v_tecnifuego_spain_negative_control","span":"NEGATIVE CONTROL: Tecnifuego, the Spanish fire-protection industry association that collaborates with the administration on legislation and standards, publishes mainstream autoproteccion guidance that specifies NO clearance distances at all."},{"axis":"industry_reports","wiki_source_id":"v_ibhs_home_hardening_evidence","span":"The IBHS Wildfire Prepared Home is a certification scheme with a published standard, a decision-tree process and designation that transfers with the property — the benchmark instrument Southern Europe has no counterpart to. Its widely quoted 'half the ignition risk' and 'nearly double the survival likelihood' figures are a vegetation and Zone-0 effect and are attached here to clearance, never to a product."}],"total_disclosed":0.0,"unknown_count":6,"adoption":{"y0":{"value":0.3,"lower":0.18,"upper":0.45,"unit":null},"y3":{"value":0.42,"lower":0.25,"upper":0.58,"unit":null},"y5":{"value":0.5,"lower":0.3,"upper":0.66,"unit":null},"y10":{"value":0.65,"lower":0.35,"upper":0.78,"unit":null}},"saturation_cap":0.8,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":["wiki/sources/v_france_old_insurance_linkage.md"],"rationale":"NEUTRAL, EVIDENCE GAP - and this is the closest thing to a premium channel in the whole pilot, which is why the gap is stated rather than filled. France's obligation legale de debroussaillement is the only located instance anywhere of defensible-space compliance carrying a contractual price, and the price is a DEDUCTIBLE ON INDEMNITY, not a premium credit; France is excluded by phase_00_input.yaml.out_of_scope and I046 is carried as a COMPARATOR, not an in-scope member. Greece's National Registry of Preventive Fire Protection Measures makes each property's compliance state an addressable record backed by administrative fines - the precondition an insurer would need in order to price compliance - but no Greek insurer is documented as using it and no such use is asserted on the source page. Inventing one would manufacture exactly the channel Phase IX exists to prevent. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No premium credit, rate relativity or underwriting condition tied to clearance compliance in Spain, Portugal, Italy or Greece. Searched: v_defensible_space_portugal, v_defensible_space_greece, v_defensible_space_programmes_es, v_france_old_insurance_linkage, v_insurer_wildfire_initiatives, v_property_risk_scoring_ventures, v_tecnifuego_spain_negative_control, insurance_journal_2026_marsh_europe_wildfire. Nothing in-geography. The French deductible-on-indemnity design is the model that does not yet exist here."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Statutory clearance duties, municipal substitution regimes and provincial strip-execution programmes do not touch insurer asset allocation, duration or investment return. Searched: v_defensible_space_portugal, v_defensible_space_greece, v_defensible_space_programmes_es, v_france_old_insurance_linkage. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.865,"confidence":"M","sources":["wiki/sources/v_defensible_space_portugal.md","wiki/sources/v_defensible_space_greece.md","wiki/sources/v_defensible_space_programmes_es.md","wiki/sources/v_ibhs_home_hardening_evidence.md","wiki/sources/h_sgifr_activity_report_2024.md","wiki/sources/v_wui_building_standards_gap.md","wiki/sources/v_fireprime_wuiview_projects.md","wiki/sources/v_tecnifuego_spain_negative_control.md","yaml-ref:phase_04_5_categories.yaml#cat_v_defensible_space_clearance.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_v_defensible_space_clearance.addressable_share"],"rationale":"[READ THIS WITH THE NUMBER - AGGREGATE CONTEXT, repeated on each of the five rows that dominate the product so it survives being read in isolation.] The matrix total is a 22.9% claims reduction at t=10 (total_factor_I3 = 0.770928, rebased on the observed L_0 per issue #53). NOT ONE VENTURE, PROGRAMME OR STATUTE IN THIS PILOT'S POOL PUBLISHES AN AUDITED INSURED-LOSS-REDUCTION FIGURE ANYWHERE IN SPAIN, PORTUGAL, ITALY OR GREECE, so that 22.9% is a structured analyst inference over sourced mechanisms, not a measured effect. TWO BASES ARE QUOTED BELOW AND THEY ARE NOT INTERCHANGEABLE: 'marginal' is the share of the 22.9% that disappears if the named rows are removed, and marginal shares do NOT sum to 100% across rows; 'log' is each row's share of the log-reduction, which does sum. Cells at confidence: L carry 71.4% marginal / 73.9% log. The two tier: emerging rows carry 25.7% marginal / 28.3% log, and removing both moves the headline from 22.9% to 17.0%. THIS ROW CONTRIBUTES 17.4% MARGINAL / 19.4% LOG. Full decomposition in loop_metadata.aggregation_disclosure, which does not render - see loop_metadata.disclosure_rendering_note. || [ADOPTION FOR THIS ROW IS THE PHASE V y10, THE SAME AS EVERY OTHER ROW. THE EARLIER INSTRUCTION NOT TO 'FIX' IT BACK IS WITHDRAWN AND IS QUOTED BELOW AS HISTORY.] This row's `adoption` is 0.65, which is Phase V's y10 unchanged, and its `adoption_trajectory` carries the same compliance series ending at 0.65. The base-year overlap that the y0 of 0.30 represents is removed by the model rather than by the anchor: L(t) = L_0 (1+g)^t f(t)/f(0), so f(0) divides out whatever is already deployed in 2026, which is exactly the part sitting inside the observed L_0. Subtracting y0 from the endpoint as well removes it twice, and is the linear approximation of that ratio besides, so it under-credits. CURRENT ARITHMETIC FOR THIS ROW: f(y10) = 1 - 0.65 x (1 - 0.865) = 0.912250 and f(0) = 1 - 0.30 x (1 - 0.865) = 0.959500, so the row enters the product as 0.912250 / 0.959500 = 0.950756, inside total_factor_I3 = 0.770928 and a 22.9% headline. WHAT A LATER READER MUST NOT DO, AND NOTE THE SIGN HAS FLIPPED: re-setting this row's adoption to the withdrawn increment of 0.35 while rebasing is in force would net the 2026 compliance level twice and move the headline from 22.9% down to 19.5%. The danger the capitalised instruction guarded against now runs the other way. || ERRATUM, 2026-09-21, ISSUE #53. KEPT RATHER THAN DELETED BECAUSE THE ELICITATION IS PART OF THE AUDIT TRAIL. At iterations 3 and 4 this cell carried adoption 0.35 = Phase V y10 0.65 minus y0 0.30, and told readers in capitals not to restore it. The ground was phase_00_input.yaml, which binds Phase X to count 'only *new* adaptation above the already-embedded historical suppression baseline', read together with a claim that ten other rows discharged that rule inside their Phase V y0 while this row alone was denominated as a compliance STATE. On that arithmetic the row's effective factor moved 0.912250 -> 0.952750, total_factor_I3 moved 0.725226 -> 0.757423, and the headline fell from 27.5% to 24.3%, a 3.22-point move. ALL OF THOSE FIGURES ARE SUPERSEDED. The Phase 0 rule itself is not withdrawn: it is discharged by the f(0) divisor instead of by an anchor subtraction. The premise about the other ten rows did not survive either, because phase_05_adoption.yaml was audited on 2026-09-21 and every one of the twelve anchors is a total deployed share of that category's base, none of them a delta, this row's included. THE TWO-SIDED RISK RECORDED AT ITERATION 4 LOSES ITS ARITHMETIC AND KEEPS ITS SUBSTANCE: phase_02_baseline.yaml#method_notes builds L_0 from the EIOPA/CATDAT insured mean over 1980-2024 with a window midpoint around 2002, while Portugal's DL 82/2021, Greece's 2024 National Registry and its 2026 KYA, and Portugal's Despacho 675/2026 all postdate most of that window. So part of the y0 = 0.30 compliance level that f(0) now divides out is itself post-baseline adaptation which L_0 never observed, and on that reading f(0) nets slightly too much. No source in this pilot decomposes the 0.30 by vintage, so no correction is proposed and none is implied. This remains a leads' call, and wiki/categories/cat_v_defensible_space_clearance.md records the same tension and leaves it open. THE DOSSIER CONFLICT THAT DROVE THE ORIGINAL SUBTRACTION IS NOW MOOT: this category's Phase IV.5 dossier instructed Phase V to carry the 2026 compliance level as y0 and declined baseline netting on a headroom argument, that compliance is not near-saturated. Under rebasing both positions hold at once, because carrying the compliance state as y0 is precisely what f(0) needs in order to net. FOR A SPREADSHEET READER: main_matrix and the Adoption curves sheet now both show 0.65 for this row. The divergence an earlier draft of this cell warned about is gone, and the v0.8.2 adoption_denominator field it called for is no longer needed on this ground. || value 0.865 = 1 - (effectiveness 0.30 x addressable_share 0.45), Phase IV.5 parameters unchanged. THE LARGEST V-CHANNEL MECHANISM, THE ONLY ONE IN THE PILOT WITH A QUANTIFIED EFFECTIVENESS ANCHOR - AND THE ANCHOR IS BORROWED. Three genuine on-mechanism axes. regulator_filings: the Portuguese SGIFR portal with the graded profile (all surface fuels eliminated 0-2 m, shrubs and trees removed or thinned 2-10 m, canopies held at least 5 m from the structure) under the DL 82/2021 art. 3 duty with municipal substitution and cost recovery; the Hellenic Fire Service's preventive-fire-protection KYA with the National Registry and the Pyrosvestiki Diataxi 19/2024 fine regime anchored in Law 4662/2020; Diputacio de Barcelona's 2025 results against the Catalan statutory 500 m band - 73 delimitation maps, 78 drafted perimeter-strip projects, 61 executed works, 309 ha of protection strips, 266 municipalities, ALL FIVE OF WHICH ARE MARKED BUNDLE AND SECOND-HAND AND ARE NOT LOAD-BEARING HERE. They are carried from the scout dossier, not from the MITECO snapshot: diba.cat is lawfully robots-excluded from this pilot's fetcher, so no snapshot stands behind any of them, and v_defensible_space_programmes_es's own caveat 1 says in terms 'Do not cite it for hectares, budgets, or compliance rates' - its evidentiary value is the ABSENCE of a Spanish building-standard track, which is the only thing it is relied on for in this cell. The same caveat is the one this composer applied at iteration 1 to demote cat_v_vulnerability_assessment_targeting from M to L, and consistency requires marking it here rather than reading the same page two ways. The figures are corroborative of executed throughput and re-verification is pending; the M confidence on this cell rests on the genuine T1 statutory sources (v_defensible_space_portugal, v_defensible_space_greece, v_wui_building_standards_gap) and the 0.30 effectiveness rests on the already-disclosed IBHS/ZestyAI anchor, not on any Diputacio hectare; and Legge 353/2000 art. 1 comma 3 establishing that Italy DELEGATES clearance geometry to the regions and therefore sets no national numeric distance on a private owner - which is what sets the lower bound. academic_studies: FIREPRIME / WUIVIEW, the only European research instrument scoring nearby vegetation alongside construction at household level. industry_reports: Tecnifuego as a NEGATIVE CONTROL - the Spanish fire-protection industry association's mainstream autoproteccion guidance specifies NO clearance distances at all. THE ANCHOR AND ITS FULL PROVENANCE, stated because this cell leans on it: the 2021 IBHS + ZestyAI finding that an ember-resistant buffer cuts a structure's ignition risk in half and that clearing perimeter vegetation can nearly double its survival likelihood is USED HERE AND ONLY HERE, because it is a VEGETATION and Zone-0 effect and must never be attached to any product category. It is US-derived, single-study and insurer-funded; it is NOT validated on Mediterranean fuels, building spacing or the Southern-European settlement pattern; NEITHER FIGURE IS IN EITHER SNAPSHOT - both are carried from the scout dossier, un-snapshotted in this pilot; and ZestyAI co-authored the study AND SELLS A WILDFIRE RISK SCORE, so it is not independent of a commercial beneficiary. That is why the central 0.30 sits well below what 'nearly double the survival likelihood' would imply and why the Phase IV.5 bracket runs 0.10-0.55. NOT BASELINE-NETTED DOWNWARD, and this is the one place in the pilot where that is correct - RE-ARGUED AT ITERATION 1 ON THE SOURCED QUANTITY, because the iteration-0 text rested it on a figure it wrongly called unretrieved and read against the wrong denominator. The T1 primary is h_sgifr_activity_report_2024 (AGIF, Relatorio de Atividades do SGIFR 2024), and what it says is: the GNR signalled 10,256 non-compliance situations in 2024, down 28% on 2023, monitored all of them, and found about 40% were not regularised after notification - so 'o grau de cumprimento voluntario da gestao de combustivel atingiu 60%'. THE 40% IS THE NON-REGULARISATION RATE AMONG ALREADY-DETECTED BREACHES, NOT A POPULATION RATE; that page's caveat 1 says in terms that using it as a population-level penetration cap is the wrong denominator. And the 60% is not national either: the GNR's 2024 enforcement footprint was 991 freguesias covering 32% of the mainland surface and 2 million hectares of forest space, so 60% is compliance INSIDE the inspected third. Phase V already reasons from exactly this and blends Portugal to roughly 0.45 nationally and the four countries to a y0 of 0.30 (phase_05_adoption.yaml#cat_v_defensible_space_clearance.anchors.y0). THE EXEMPTION SURVIVES ON THOSE FIGURES AND NOT ON THE MISREAD ONE: two thirds of mainland Portugal sits outside the inspection footprint entirely, Italy prescribes no numeric distance binding a private owner at all, and the four-country y0 is 0.30 against a y10 of 0.65 - so compliance with the perimeter duty is demonstrably NOT near-saturated in 2026 and most of this category's effect is genuinely forward-available headroom rather than embedded history, unlike suppression and fuel-break capacity. CONFIDENCE M: T1 statutory geometry in three of four countries plus a quantified (if borrowed) effectiveness anchor support direction and rough magnitude; H is unavailable because no Southern-European insured-loss outcome is measured anywhere. THE LIVE OVERLAP, restated: the risk that I015 and I043 describe the same building-perimeter act is live, and the 0.45/0.55 addressable-share partition against cat_h_landscape_fuel_management is a conservative device, not a geometry. SHARED-SERIES DEPENDENCY, DISCLOSED: wiki/categories/cat_h_landscape_fuel_management.md makes it a binding Phase IV.5 instruction that this category and cat_h_landscape_fuel_management be driven from a SINGLE Portuguese compliance series, so their adoption anchors are not independent - while the matrix_schema.md section 6 Pi-product multiplies their effective factors as though they were. That is the same disclosure this cell already makes about the addressable-share overlap, and it is the second known breach of the v0.7 cross-category independence assumption. WHAT WOULD STILL MOVE THIS ROW: not the compliance figure, which is now T1 and bounded, but the enforcement footprint - the 60% describes 32% of mainland Portugal, and no source in this pilot measures compliance in the other 68% or in Spain, Italy or Greece."},"I4":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - though this is the category where an independent operational channel is most conceivable, which is why the gap is stated explicitly. Greece's National Registry of Preventive Fire Protection Measures makes each property's compliance state a record an insurer could in principle query, which would speed underwriting and post-event triage. NO SOURCE SAYS ANY INSURER DOES SO, and inventing that use would be exactly the manufactured channel Phase IX exists to prevent. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No claims-handling, underwriting-cost or inspection-cost effect attributed to clearance compliance or to the Greek registry. Searched: v_defensible_space_greece, v_defensible_space_portugal, v_defensible_space_programmes_es, v_france_old_insurance_linkage, v_property_risk_scoring_ventures, r_claims_visual_intelligence. The Greek registry is the artefact a future source would have to show an insurer actually querying."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","regulator_filings"],"distinct_sources":8,"total_disclosed":0.0,"unknown_count":6,"reason":"3 evidence axes across 8 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_v_defensible_space_clearance.md"},{"id":"cat_v_vulnerability_assessment_targeting","name":"Wildfire Vulnerability Assessment, Scoring and Targeting","channel":"V","channel_label":"Vulnerability","tier":"confirmed","tier_rationale":"Three source-type axes documented. academic_studies: the FIREPRIME and WUIVIEW record — UCPM GA 101140381 for the household self-assessment scoring construction materials, design, nearby vegetation and flammable-item storage, and the EUR 572,015.58 UPC-led WUIVIEW project whose performance-based fire-simulation workbench for European building systems was published in the International Journal of Wildland Fire in June 2023, with ADAI, Fundacio Pau Costa, Universita di Bologna, ARMINES and RISE as partners. industry_reports: Mitiga Solutions' asset- and portfolio-level physical-climate-risk product, SilvIA Earth's parcel-level wildfire score sold explicitly to insurers for pricing and to administrations for deciding where clearance and inspection go, Faura's score-to-retrofit-to-verify loop and ZestyAI's wildfire model as the US operating template, and Fidelidade's ICCC national risk map. regulator_filings: the Spanish MITECO portal establishing that the national framing of wildfire is coordination-plus-suppression with no building-standard block, which is the regulatory vacuum this assessment layer is trying to fill. No NGO-report axis. Investment intensity EUR 1,245,713.58, entirely EU research grants; well below the EUR 50 m emerging gate, but the gate is irrelevant because the category clears `confirmed` on axes.","definition":"The analytical layer that identifies which buildings, parcels or portfolio segments are vulnerable and therefore where mitigation effort, inspection capacity and pricing attention should go — performance-based structural fire simulation of European building assemblies under modelled WUI exposure, household self-assessment scoring of construction and surroundings, satellite-derived parcel-level risk scores, and insurer-built national vulnerability and loss-curve maps. The mechanism is VULNERABILITY but it is INDIRECT: none of these ventures changes a building or a hectare. They reduce loss only to the extent that someone else acts on their output, and the binding constraint they address is that clearance and inspection capacity in all four countries is far smaller than the obligated area, so targeting decides how much of the available mitigation lands where it matters.","cluster_rationale":"All five ventures produce an assessment or a score and stop there — none of them clears a hectare, screens a vent or changes a roof. I049 (WUIVIEW) simulates how a specific European assembly fails under modelled WUI exposure; I048 (FIREPRIME) puts the same question to the householder as a scored questionnaire and prescribes measures; I059 (SilvIA Earth) scores the parcel from satellite fuel state and sells the ranking both to insurers for pricing and to administrations for deciding where clearance work is sent; I058 (Mitiga) does the same at asset and portfolio level for owners and underwriters; I060 (Fidelidade) is the peer-insurer instance, building national hazard, exposure, vulnerability and loss-curve layers. They are clustered on the mechanism they share — information that directs someone else's mitigation or pricing decision — and NOT on tech stack, which spans fire-physics simulation, a mobile questionnaire, satellite remote sensing and geospatial loss modelling. THE SEAM, flagged for the critic: I058's own inventory entry states that its pathway is hazard and exposure screening and that it carries no building-materials, ember- vulnerability, defensible-space or retrofit content, which makes its fit with the structure-focused I048/I049 arguable; it is nonetheless a V-channel venture in the inventory and its mechanism is targeting, so it sits here rather than in an H-channel forecasting category it could not join without violating the no-cross-channel rule. Boundary vs cat_v_defensible_space_clearance: that category is the clearance; this one decides which parcels receive it when capacity is smaller than the obligated area. Boundary vs cat_v_noncombustible_envelope_fabric and cat_v_retrofit_ember_path_protection: those categories change the building; this one only says which building to change. Boundary vs cat_h_detection_forecasting_decision_support: that category forecasts where fire will be and routes suppression; this one scores how badly a given structure would fare if it got there.","members":["I048","I049","I058","I059","I060"],"member_names":["FIREPRIME — European Programme for Wildfire-Prepared Communities","WUIVIEW — Wildfire Vulnerability Assessment Tool (UPC-led)","Mitiga Solutions","SilvIA Earth","Fidelidade — ICCC national wildfire risk map (Portugal)"],"effectiveness":{"value":0.08,"lower":0.0,"upper":0.2,"unit":null},"effectiveness_definition":"ANALYST-INFERRED, with a lower bound of exactly zero because the pilot cannot exclude that this category reduces no insured loss at all. Fractional reduction in expected insured loss attributable to better targeting of a fixed mitigation capacity, over and above the same capacity applied without targeting. The only MEASURED effect of any constituent is a distribution effect: agents given Fidelidade's detailed risk sheets sold materially more adequate cover, an increase of nearly 60% in sales of adequate coverage. That CLOSES A PROTECTION GAP — it raises insured loss by raising insured exposure — and is explicitly NOT a vulnerability reduction; it must not be entered as effectiveness anywhere in Phase X. WUIVIEW produces vulnerability estimates, not vulnerability reductions. FIREPRIME assesses and recommends but does not inspect, designate or verify, so the behaviour change it can produce is unenforced. Mitiga's pathway is hazard and exposure screening with no retrofit recommendation or verification loop. SilvIA is pre-traction with no named client. The upper bound reflects the case where targeting materially raises the yield of a capacity-constrained clearance programme; the central 0.08 is a deliberately small allocation of that gain.","effectiveness_sources":[{"slug":"v_fireprime_wuiview_projects","title":"FIREPRIME — European Programme for Wildfire-Prepared Communities (UCPM-2023-KAPP, GA 101140381), with the WUIVIEW performance-based WUI simulation workbench recorded as a bundle claim","tier":"T2","publisher":"fireprime_consortium","url":"https://fireprime.eu/","wayback_url":"https://web.archive.org/web/20260515070457/https://fireprime.eu/","snapshot_hash":"b4b5ecb1247052f24c1c47c751c365a1677261fe16b1a56250761de5ed3e1ca7","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/b4b5ecb1247052f24c1c47c751c365a1677261fe16b1a56250761de5ed3e1ca7.html","publication_date":null,"fetched_at":"2026-07-31T12:58:00Z","page_exists":true},{"slug":"v_iberian_risk_scoring_ventures","title":"Iberian wildfire risk-scoring ventures — Mitiga Solutions (Barcelona), with SilvIA Earth (Madrid) recorded as a bundle claim; hazard and exposure screening, not structure hardening","tier":"T3","publisher":"iberian_risk_scoring_bundle","url":"https://www.mitigasolutions.com/","wayback_url":"https://web.archive.org/web/20260624112402/https://www.mitigasolutions.com/","snapshot_hash":"9be758bc424238c200df261731d8b5808ecc31ff3fa71f3bf3f97c0df514a93a","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/9be758bc424238c200df261731d8b5808ecc31ff3fa71f3bf3f97c0df514a93a.html","publication_date":null,"fetched_at":"2026-07-31T12:59:43Z","page_exists":true},{"slug":"v_insurer_wildfire_initiatives","title":"Insurer-run wildfire initiatives in the pilot geography — Fidelidade (Portugal) as the snapshotted primary, with the ICCC national wildfire risk map and Allianz Seguros España recorded as bundle claims","tier":"T3","publisher":"fidelidade","url":"https://www.fidelidade.pt/PT/particulares/Paginas/default.aspx","wayback_url":"https://web.archive.org/web/20260731154102/https://www.fidelidade.pt/PT/particulares/Paginas/default.aspx","snapshot_hash":"f1acf0ff535067c43fc401472806efeb37c0b96dd866bc422f59eb4dc8635eee","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/f1acf0ff535067c43fc401472806efeb37c0b96dd866bc422f59eb4dc8635eee.html","publication_date":null,"fetched_at":"2026-07-31T13:11:23Z","page_exists":true}],"addressable_share":0.5,"axes_hit":["industry_reports","academic_studies","regulator_filings"],"citations":[{"axis":"academic_studies","wiki_source_id":"v_fireprime_wuiview_projects","span":"WUIVIEW was a EUR 572,015.58 UPC-led project building a performance-based fire-simulation workbench for European building systems under WUI exposure, with ADAI, Fundacio Pau Costa, Universita di Bologna, ARMINES and RISE; its Wildfire Vulnerability Assessment Tool was published in the International Journal of Wildland Fire in June 2023. FIREPRIME's household tool assesses construction materials, design, nearby vegetation and flammable-item storage under UCPM GA 101140381."},{"axis":"industry_reports","wiki_source_id":"v_iberian_risk_scoring_ventures","span":"Mitiga Solutions delivers asset- and portfolio-level physical-climate-risk simulation whose unit of analysis is the asset's exposure, not the building's susceptibility; SilvIA Earth sells a parcel-level wildfire risk score to insurers to price policies parcel by parcel from satellite data rather than from historical tables."},{"axis":"industry_reports","wiki_source_id":"v_property_risk_scoring_ventures","span":"Faura's proposition is the score-to-retrofit-to-verify loop rather than risk selection — creating better risks, turning homeowners into active risk managers and verifying mitigation before losses occur — the US operating template for the loop that does not exist in the pilot geography."},{"axis":"industry_reports","wiki_source_id":"v_insurer_wildfire_initiatives","span":"Fidelidade launched Portugal's first national wildfire risk map at 100x100 m resolution covering hazard, building and resident exposure, vulnerability and loss curves, with results expected around 2028–2029; its sole quantified outcome so far is a distribution effect — agents given detailed risk sheets increased sales of adequate coverage by nearly 60%."},{"axis":"regulator_filings","wiki_source_id":"v_defensible_space_programmes_es","span":"Spain's national wildfire portal is organised around coordination, prevention/preparation/extinction and statistics, with no building-standard or construction-requirement block anywhere — the regulatory vacuum in which a private assessment and scoring layer is operating."}],"total_disclosed":1245713.58,"unknown_count":3,"adoption":{"y0":{"value":0.05,"lower":0.01,"upper":0.12,"unit":null},"y3":{"value":0.2,"lower":0.08,"upper":0.35,"unit":null},"y5":{"value":0.32,"lower":0.12,"upper":0.5,"unit":null},"y10":{"value":0.5,"lower":0.2,"upper":0.68,"unit":null}},"saturation_cap":0.7,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":["wiki/sources/v_insurer_wildfire_initiatives.md","wiki/sources/v_iberian_risk_scoring_ventures.md"],"rationale":"NEUTRAL 1.00, AND DELIBERATELY NOT FLAGGED AS AN EVIDENCE GAP - the only measured insurer-side effect in the entire pilot sits here, and it points the WRONG WAY for a v0.7 mitigation cell. Fidelidade's programme produced one quantified outcome: agents given detailed risk sheets increased sales of ADEQUATE COVERAGE by nearly 60%. That is a DISTRIBUTION effect. It CLOSES A PROTECTION GAP - it raises insured exposure and therefore raises both premium revenue and insured loss - and Phase IV.5 refuses it as effectiveness in terms. Two consequences bind this cell. (1) v0.7 mitigation cells are mitigation-only, 0 < value <= 1.0 (matrix_schema.md v0.7 section 8), so a premium-INCREASING effect is not expressible in this cell at all; it belongs to Phase II/IX exposure work and to the Phase VII recalibration triggers. (2) SYMMETRY: this same lattice sets cat_r_public_recovery_substitution to neutral because an insured-share DECREASE is equally inadmissible. BOTH DIRECTIONS OF THE SAME QUANTITY ARE EXCLUDED, and admitting one of them silently would bias the forecast. The remaining constituents supply underwriting-relevant capability with no priced outcome: SilvIA Earth sells a parcel-level wildfire score to insurers to price policies parcel by parcel from satellite data rather than historical tables, but is pre-traction with no named client; Mitiga delivers asset- and portfolio-level physical-climate-risk simulation; Faura's score-to-retrofit-to-verify loop is named explicitly as the US operating template for a loop that DOES NOT EXIST in the pilot geography. WHAT WAS SEARCHED FOR AND NOT FOUND: any executed premium credit, rate relativity or verified-mitigation discount in ES/PT/IT/GR. Searched: v_insurer_wildfire_initiatives, v_iberian_risk_scoring_ventures, v_property_risk_scoring_ventures, v_fireprime_wuiview_projects, insurance_journal_2026_marsh_europe_wildfire. This is a named exclusion, not an unexplained absence, which is why no evidence_gap flag is set. ONE-SIDEDNESS DISCLOSED AT ITERATION 1 (full statement in loop_metadata.one_sided_by_construction): this pilot documents three wrong-way effects - Fidelidade's +60% adequate-coverage uplift here, Legge 353/2000's ten-year rebuild ban on two I4 cells, and public recovery substitution's insured-share movement - and all three run AGAINST the insurer. v0.7 mitigation cells are bounded at 1.0 and can express none of them. The refusals are symmetric and correct, but the consequence is that the matrix's net t=10 picture is one-sided by construction rather than by evidence, and Fidelidade's +60% is the ONLY measured insurer-side effect in the entire pilot."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Vulnerability scoring, WUI fire simulation and national risk mapping improve underwriting and targeting information but do not touch insurer asset allocation, duration or investment return. The closest adjacency - Mitiga's asset- and portfolio-level physical-climate-risk product - is sold for EXPOSURE screening of physical assets, which is an underwriting and risk-management input; its own inventory record states it carries no building-materials, ember-vulnerability, defensible-space or retrofit content, and it is not an investment-portfolio product in this pilot's evidence. Searched: v_iberian_risk_scoring_ventures, v_property_risk_scoring_ventures, v_fireprime_wuiview_projects, v_insurer_wildfire_initiatives. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.96,"confidence":"L","sources":["wiki/sources/v_fireprime_wuiview_projects.md","wiki/sources/v_iberian_risk_scoring_ventures.md","wiki/sources/v_insurer_wildfire_initiatives.md","wiki/sources/v_defensible_space_programmes_es.md","yaml-ref:phase_04_5_categories.yaml#cat_v_vulnerability_assessment_targeting.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_v_vulnerability_assessment_targeting.addressable_share"],"rationale":"value 0.960 = 1 - (effectiveness 0.08 x addressable_share 0.50), Phase IV.5 parameters unchanged. AN INDIRECT MECHANISM WITH A PHASE IV.5 LOWER BOUND OF EXACTLY ZERO: the pilot cannot exclude that this category reduces no insured loss at all. Three genuine on-mechanism axes. academic_studies: WUIVIEW, a EUR 572,015.58 UPC-led project building a performance-based fire-simulation workbench for European building systems under WUI exposure with ADAI, Fundacio Pau Costa, Universita di Bologna, ARMINES and RISE, published in the International Journal of Wildland Fire in June 2023; and FIREPRIME's household tool scoring construction materials, design, nearby vegetation and flammable-item storage. industry_reports: Mitiga, SilvIA Earth, Faura, and Fidelidade's ICCC national wildfire risk map at 100x100 m resolution covering hazard, building and resident exposure, vulnerability and loss curves. regulator_filings: Spain's national wildfire portal, organised around coordination, prevention/preparation/extinction and statistics with NO building-standard block anywhere - the regulatory vacuum this private assessment layer operates in. WHAT THE MECHANISM ACTUALLY IS, and why the value is small: NONE of these ventures changes a building or a hectare. They reduce loss only to the extent that someone else acts on their output. The binding constraint they address is real - clearance and inspection capacity in all four countries is far smaller than the obligated area, so targeting decides how much of the available mitigation lands where it matters - but the gain is an ALLOCATION gain over the same capacity applied untargeted, and no source measures it. FOUR REASONS THE LOWER BOUND IS ZERO, each traceable: WUIVIEW produces vulnerability ESTIMATES, not vulnerability reductions; FIREPRIME assesses and recommends but does not inspect, designate or verify, so the behaviour change is unenforced; Mitiga's pathway is hazard and exposure screening with no retrofit recommendation or verification loop; SilvIA is pre-traction with no named client. Fidelidade's ICCC results are expected around 2028-2029 and do not exist yet. EXPLICITLY EXCLUDED FROM THIS CELL: the +60% adequate-coverage sales uplift, which is a protection-gap movement and is handled - and refused - in this category's I1. CONFIDENCE LOWERED M -> L AT ITERATION 1, on the pilot's own void-is-not-an-axis rule. The M rested on v_defensible_space_programmes_es being a T1, but that page is cited here for 'the regulatory vacuum this private assessment layer operates in' - and its own caveat 1 states that it is a topic hub whose 'evidentiary value here is the absence of a building-standard track, not any number', with caveat 6 adding that it evidences delivery of vegetation clearance and not structural hardening. phase_04_5_categories.yaml applies exactly this rule twice: it withdrew the industry_reports axis from cat_v_noncombustible_envelope_fabric because 'a source whose content is \"this source does not discuss the mechanism\" cannot be an axis hit for that mechanism ... counting it inverted the rigor test', and withdrew regulator_filings from cat_v_retrofit_ember_path_protection because 'counting a documented regulatory void as a regulator-filings hit inverts the test'. Both withdrawals sent those categories to L on every active cell. Applying the same rule here removes the regulator_filings axis and leaves one T2 that this rationale itself says produces vulnerability ESTIMATES rather than reductions, plus two T3 vendor pages. M would have remained formally legal under matrix_schema.md section 5 because the T2 is present, but sustaining it would have required this category to be treated differently from two others stripped on identical grounds. The value is unchanged: 0.960 is 1 - (0.08 x 0.50) from signed-off Phase IV.5 parameters and no confidence tag bears on it. BASELINE NETTING, STATED AT ITERATION 3: the Phase V y0 note records that WUIVIEW and FIREPRIME 'are research instruments, not deployed routing capacity, so almost no mitigation capacity is score-routed at base year', which is what puts y0 at 0.05. The series is denominated over mitigation capacity actually routed by a score, not over the existence of scoring tools, so the 2026 research base is outside the credit."},"I4":{"value":1.0,"confidence":"L","sources":["wiki/sources/eiopa_protection_gap_technical_description.md"],"rationale":"NEUTRAL, EVIDENCE GAP - A PLAUSIBLE CHANNEL THIS PILOT CANNOT EVIDENCE. In principle a parcel-level score reduces underwriting and post-event triage effort, and EIOPA's finding that wildfire is assessed by EXPERT JUDGEMENT in all thirty countries (where earthquake and flood are Modelled) establishes that there is a manual expert workload for such a layer to displace. But no source connects the two: none of the five constituents publishes an insurer operating-cost effect, Fidelidade's only quantified outcome is a distribution effect, and SilvIA has no named client whose costs could have moved. Any genuine I4 benefit here would be THE OPERATIONAL SHADOW OF I3 - fires that never reach insured property, and buildings that survive, generate no claims to adjust - which is the I3 mechanism seen from the operations side, not an independent channel. Scoring it would double-count the reduction already taken in this category's I3. Value held at 1.00.","evidence_gap":true,"gap_reason":"No underwriting-cost, inspection-cost or claims-handling effect attributed to wildfire scoring or vulnerability assessment. Searched: v_iberian_risk_scoring_ventures, v_property_risk_scoring_ventures, v_insurer_wildfire_initiatives, v_fireprime_wuiview_projects, eiopa_protection_gap_technical_description, r_claims_visual_intelligence. The EIOPA expert-judgement workload is the quantity a future source would have to show being displaced."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","regulator_filings"],"distinct_sources":5,"total_disclosed":1245713.58,"unknown_count":3,"reason":"3 evidence axes across 5 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_v_vulnerability_assessment_targeting.md"},{"id":"cat_r_physical_restoration_capacity","name":"Physical Restoration and Reinstatement Capacity","channel":"R","channel_label":"Recovery","tier":"confirmed","tier_rationale":"TIER MET (issue #52, 2026-09-23): three source-type axes — industry_reports, regulator_filings and academic_studies. academic_studies: Döhrmann, Gürtler & Hibbeln 2013 (Zeitschrift für die gesamte Versicherungswissenschaft) find econometrically that a higher construction-sector capacity restrains post-catastrophe demand surge, while more repair work and more claims per event amplify it — repair capacity acting on claim severity, which is this category's mechanism. GEOGRAPHY AND PERIL GAP: US events, mostly storms and floods, no wildfire; the mechanism transfers, the magnitudes do not. PHASE X CONSEQUENCE, RETAINED PENDING LEAD REVIEW: the §5 `emerging` confidence cap (confidence: L on every active cell), applied while this row carried a declared tier deficit, is kept; lifting it is a Phase X judgement for the leads, not a consequence of meeting the axis bar. The original two axes are: industry_reports — BELFOR Italia's nine operational branches with 24/7 call-out and the RED ALERT and P.I.A. pre-loss programmes, Polygon Espana's restauracion integral de danos por incendio for insurers across Spain and Portugal, and Grupo Beer's contents decontamination and asbestos capability; and regulator_filings — Eurostat's residential construction-price series, the official measure of the rebuild cost that repair-over-replace acts against, and Normattiva's Legge 353/2000 art. 10, which makes reinstatement in place unlawful for ten years where the building ban binds. MISSING AXIS: ngo_reports. phase_04_venture_inventory.md §2.7 found no academic study and no NGO report on post-event wildfire restoration in these markets; the 2026-09-23 search confirms that for these markets and for wildfire, so the academic axis rests on the mechanism studied in US catastrophe repair markets, not on Southern-European wildfire data. See revision_log entry 1.","definition":"The physical capacity that converts a fire-damaged building and its contents back into a usable asset, and the contractual arrangements that guarantee access to that capacity when demand spikes — in-place restoration of fire, smoke, soot, corrosion and fire-suppression water and foam damage instead of replacement; contents salvage and decontamination; regulated asbestos removal that otherwise stalls a rebuild in older stock; and pre-loss retainers that contract priority response before any event occurs. The mechanism is RECOVERY: every square metre restored instead of rebuilt is indemnity not paid, and every day saved on drying, decontamination and reinstatement is additional-living-expense or business-interruption not accrued.","cluster_rationale":"All four ventures perform or guarantee access to the same physical act: returning a fire-damaged asset to use rather than replacing it, and doing so faster. I078 and I080 are the structural restoration contractors covering Italy and Iberia respectively; I081 is the contents-salvage and decontamination specialist whose regulated asbestos capability unblocks rebuilds in the older rural and peri-urban stock most exposed at the Iberian wildland-urban interface; I079 is the pre-loss retainer that converts latent capacity into contracted priority response, and is clustered with the contractors rather than treated as a commercial instrument because without capacity there is nothing for a retainer to reserve. Geography is complementary rather than duplicative (Italy from I078 and I079, Spain and Portugal from I080 and I081), which matters because in a mass-claim wildfire the operative constraint is branch density and surge capacity, not technique. Boundary vs cat_r_post_event_damage_establishment: that category establishes what is owed; this one performs the works that discharge it — sequential mechanisms on the same claim, and a Phase X composer must not credit cycle-time compression twice by counting the faster damage record and the faster physical works as one saving. Boundary vs cat_r_public_recovery_substitution: that category shifts cost off the private policy; this one lowers the cost that remains on it.","members":["I078","I079","I080","I081"],"member_names":["BELFOR Italia","BELFOR RED ALERT and P.I.A. pre-loss programmes","Polygon España","Grupo Beer"],"effectiveness":{"value":0.15,"lower":0.05,"upper":0.3,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. Fractional reduction in total post-event insured cost — rebuild plus temporary accommodation plus business interruption — from restoring rather than replacing and from compressing the reinstatement timeline. No source in the pilot quantifies it. The constituent evidence is capability evidence only: branch counts, 24/7 response, service scope and group-level project counts. Wildfire- specific capability and mass-claim surge capacity are UNEVIDENCED for every one of the four ventures — the mechanism documented is generic structural fire restoration. The upper bound assumes a large share of wildfire property claims are smoke, soot and suppression-water damage amenable to in-place restoration rather than total structural loss; the lower bound reflects that in a regional mass-claim wildfire the binding constraint is capacity rather than technique, that restoration capacity without a retainer is allocated first-come-first-served while secondary damage compounds in the queue, and that in Italy the Legge 353/2000 art. 10 building ban can make reinstatement in place legally unperformable for up to ten years, pushing settlement toward cash indemnity with a long ALE tail.","effectiveness_sources":[{"slug":"r_restoration_contractors","title":"BELFOR Italia — post-fire restoration services, Italian branch network and pre-loss retainer programmes","tier":"T3","publisher":"belfor","url":"https://www.belfor.com/it/it/","wayback_url":"https://web.archive.org/web/20260607073729/https://www.belfor.com/it/it/","snapshot_hash":"a245c673fae7169551723558c855e395f9ae328f9db39766d9a085b70fd89f51","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/a245c673fae7169551723558c855e395f9ae328f9db39766d9a085b70fd89f51.html","publication_date":null,"fetched_at":"2026-07-31T13:10:33Z","page_exists":true},{"slug":"r_restoration_iberian_contractors","title":"Iberian post-loss restoration contractors — Polygon España and Grupo Beer","tier":"T3","publisher":"polygon_group","url":"https://www.polygongroup.com/es-ES/","wayback_url":"https://web.archive.org/web/20260115092248/https://www.polygongroup.com/es-es/","snapshot_hash":"1f8bdce5c71b5c39fc0a45a16cf214a5bc404b6ae6c6da5a24d96d9a7a5bba0c","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/1f8bdce5c71b5c39fc0a45a16cf214a5bc404b6ae6c6da5a24d96d9a7a5bba0c.html","publication_date":null,"fetched_at":"2026-07-31T13:10:35Z","page_exists":true},{"slug":"eurostat_construction_prices_2026","title":"Residential buildings construction price +48% since 2015","tier":"T1","publisher":"eurostat","url":"https://ec.europa.eu/eurostat/web/products-eurostat-news/w/wdn-20260625-1","wayback_url":"https://web.archive.org/web/20260629064417/https://ec.europa.eu/eurostat/web/products-eurostat-news/w/wdn-20260625-1","snapshot_hash":"76c9a7a2e828c5b8bad0fad9eea8a40eea9fbfc4e2f961c609de20591d5e5e00","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/76c9a7a2e828c5b8bad0fad9eea8a40eea9fbfc4e2f961c609de20591d5e5e00.html","publication_date":"2026-06-25","fetched_at":"2026-07-31T12:18:29Z","page_exists":true}],"addressable_share":0.55,"axes_hit":["industry_reports","regulator_filings","academic_studies"],"citations":[{"axis":"industry_reports","wiki_source_id":"r_restoration_contractors","span":"BELFOR Italia offers the full post-fire chain — emergency response and salvage, cleaning and remediation, restoration and reconstruction — across nine operational branches with a 24/7 emergency line, plus two pre-loss programmes, RED ALERT and P.I.A., the latter aimed at small and medium enterprises across the national territory."},{"axis":"industry_reports","wiki_source_id":"r_restoration_iberian_contractors","span":"Polygon Espana performs restauracion integral de danos por incendio covering fire, smoke and fire-suppression damage for insurers across Spain and Portugal from a Madrid base with 24/7 response; Grupo Beer performs contents decontamination and recovery after a loss plus asbestos removal across Spain and Portugal."},{"axis":"regulator_filings","wiki_source_id":"eurostat_construction_prices_2026","span":"Residential building construction prices are up 48% since 2015 across the EU — the official rebuild-cost level against which repair-over-replace economics are set and the reason cycle-time and technique choices move indemnity materially."},{"axis":"regulator_filings","wiki_source_id":"r_italy_post_fire_land_constraint","span":"Legge 353/2000 art. 10 consolidated text: construction on burned land is banned for ten years and the land-use classification is frozen for fifteen — the legal constraint that can make in-place reinstatement unperformable in Italy and push settlement to cash indemnity."},{"axis":"academic_studies","wiki_source_id":"dohrmann_2013_demand_surge_econometric","span":"In contrast, a higher capacity of the construction sector has a restraining effect on Demand Surge."}],"total_disclosed":0.0,"unknown_count":4,"adoption":{"y0":{"value":0.05,"lower":0.0,"upper":0.12,"unit":null},"y3":{"value":0.15,"lower":0.05,"upper":0.28,"unit":null},"y5":{"value":0.25,"lower":0.08,"upper":0.4,"unit":null},"y10":{"value":0.4,"lower":0.12,"upper":0.58,"unit":null}},"saturation_cap":0.65,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced. Restoration contractors and pre-loss retainers sit entirely on the claims side. BELFOR's RED ALERT and P.I.A. pre-loss programmes are the nearest thing to a priced instrument in the category, but they are contracted between the RESTORER and the insured business, not expressed as an insurance premium term, and no source states otherwise. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No premium credit, rate relativity or underwriting condition attached to a restoration retainer or to contracted surge capacity in ES/PT/IT/GR. Searched: r_restoration_contractors, r_restoration_iberian_contractors, eurostat_construction_prices_2026, r_italy_post_fire_land_constraint, insurance_journal_2026_marsh_europe_wildfire."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Structural restoration, contents decontamination and regulated asbestos capability do not touch insurer asset allocation, duration or investment return. Searched: r_restoration_contractors, r_restoration_iberian_contractors, eurostat_construction_prices_2026. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.9175,"confidence":"L","sources":["wiki/sources/r_restoration_contractors.md","wiki/sources/r_restoration_iberian_contractors.md","wiki/sources/eurostat_construction_prices_2026.md","wiki/sources/r_italy_post_fire_land_constraint.md","yaml-ref:phase_04_5_categories.yaml#cat_r_physical_restoration_capacity.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_r_physical_restoration_capacity.addressable_share"],"rationale":"[READ THIS WITH THE NUMBER - AGGREGATE CONTEXT, repeated on each of the five rows that dominate the product so it survives being read in isolation.] The matrix total is a 22.9% claims reduction at t=10 (total_factor_I3 = 0.770928, rebased on the observed L_0 per issue #53). NOT ONE VENTURE, PROGRAMME OR STATUTE IN THIS PILOT'S POOL PUBLISHES AN AUDITED INSURED-LOSS-REDUCTION FIGURE ANYWHERE IN SPAIN, PORTUGAL, ITALY OR GREECE, so that 22.9% is a structured analyst inference over sourced mechanisms, not a measured effect. TWO BASES ARE QUOTED BELOW AND THEY ARE NOT INTERCHANGEABLE: 'marginal' is the share of the 22.9% that disappears if the named rows are removed, and marginal shares do NOT sum to 100% across rows; 'log' is each row's share of the log-reduction, which does sum. Cells at confidence: L carry 71.4% marginal / 73.9% log. The two tier: emerging rows carry 25.7% marginal / 28.3% log, and removing both moves the headline from 22.9% to 17.0%. THIS ROW CONTRIBUTES 10.0% MARGINAL / 11.3% LOG. Full decomposition in loop_metadata.aggregation_disclosure, which does not render - see loop_metadata.disclosure_rendering_note. || value 0.9175 = 1 - (effectiveness 0.15 x addressable_share 0.55), Phase IV.5 parameters unchanged. MECHANISM PLAUSIBLE AND ECONOMICALLY LARGE; EVIDENCE IS CAPABILITY EVIDENCE ONLY, AND WILDFIRE-SPECIFIC CAPABILITY IS UNEVIDENCED FOR ALL FOUR CONSTITUENTS. Every square metre restored instead of rebuilt is indemnity not paid, and every day saved on drying, decontamination and reinstatement is ALE or BI not accrued. What the sources actually establish: BELFOR Italia offers the full post-fire chain - emergency response and salvage, cleaning and remediation, restoration and reconstruction - across nine operational branches with a 24/7 line, plus the RED ALERT and P.I.A. pre-loss programmes; Polygon Espana performs restauracion integral de danos por incendio for insurers across Spain and Portugal from a Madrid base; Grupo Beer performs contents decontamination and regulated asbestos removal that otherwise stalls rebuilds in older stock. Those are BRANCH COUNTS, RESPONSE HOURS AND SERVICE SCOPE - capability, not effectiveness. NO SOURCE IN THE PILOT QUANTIFIES the restore-versus-replace saving or the timeline compression, and the documented mechanism is GENERIC STRUCTURAL FIRE RESTORATION, with wildfire-specific capability and mass-claim surge capacity unevidenced for every one of the four ventures. TWO DOCUMENTED CONSTRAINTS THAT SET THE LOWER BOUND: in a regional mass-claim wildfire the binding constraint is CAPACITY rather than technique, and capacity without a retainer is allocated first-come first-served while secondary damage compounds in the queue; and in Italy, Legge 353/2000 art. 10's ten-year building ban and fifteen-year land-use freeze can make in-place reinstatement legally UNPERFORMABLE, pushing settlement to cash indemnity with a long ALE tail. That statute is the second of the pilot's two documented wrong-way effects; v0.7 cannot express amplification, so it is recorded here and its operational bite is carried once, on this category's I4. THE COST LEVEL THIS ACTS AGAINST: Eurostat records EU residential construction prices up 48.2% since 2015 - with Italy (+17.0%) and Greece (+17.3%) the SLOWEST risers in the EU, a material qualification on how much repair-over-replace economics move in two of the four pilot countries. CONFIDENCE L DESPITE THE `confirmed` LABEL: declared tier deficit (two axes, EUR 0 disclosed), section 5 `emerging` cap applied. NO-DOUBLE-COUNT: cat_r_post_event_damage_establishment establishes what is owed and this category performs the works that discharge it - sequential mechanisms on the same claim, and cycle-time compression is not counted in both. BASELINE NETTING, STATED AT ITERATION 3 - this row was not cut for netting at Phase IV.5 and carries the second-highest effectiveness in the matrix, while describing a capability that plainly exists in 2026: BELFOR, Polygon and Grupo Beer are all operating today. The generic structural-fire restoration base is therefore inside L_0. The exclusion is discharged in the adoption denominator, not in the effectiveness: the Phase V y0 note records that 'no source in this pilot measures retainer penetration or wildfire surge capacity for any of the four constituents, so the base-year increment over the existing generic structural-fire restoration base is judged near nil', and the series is denominated over wildfire-specific and retainer-backed surge capacity rather than over restoration capacity as such. What is credited is the forward increment in that capacity, not the existing branch network."},"I4":{"value":1.0,"confidence":"L","sources":["wiki/sources/r_italy_post_fire_land_constraint.md"],"rationale":"NEUTRAL, EVIDENCE GAP - DIRECTIONALLY PLAUSIBLE, WHOLLY UNQUANTIFIED, AND ONE DOCUMENTED EFFECT RUNS THE OTHER WAY. A shorter reinstatement timeline shortens the open-claim lifecycle and therefore the adjuster supervision and contractor oversight hours per claim, and the pre-loss retainer (BELFOR RED ALERT / P.I.A.) is the instrument that would deliver it by converting latent capacity into contracted priority response. NO SOURCE QUANTIFIES ANY OF THAT for any of the four constituents. Running the other way, and documented: Legge 353/2000 art. 10 can make Italian reinstatement legally unperformable for up to ten years, which LENGTHENS claim life and raises supervision cost - an amplification v0.7 cannot express in a mitigation cell, so it belongs in this rationale. Same finding as on cat_h_fire_management_programmes_and_statutes I4; COUNTED ONCE, here, where it bites operationally. NO-DOUBLE-COUNT also against this category's own I3 (cycle-time compression is scored there) and against cat_r_post_event_damage_establishment's I4 (adjuster-hours are scored there). Value held at 1.00.","evidence_gap":true,"gap_reason":"No insurer claims-operations, adjuster-hours or expense figure attributed to restoration capacity or retainers. Searched: r_restoration_contractors, r_restoration_iberian_contractors, r_italy_post_fire_land_constraint, r_claims_visual_intelligence, eurostat_construction_prices_2026. Nothing on the saving side."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","regulator_filings"],"distinct_sources":5,"total_disclosed":0.0,"unknown_count":4,"reason":"3 evidence axes across 5 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_r_physical_restoration_capacity.md"},{"id":"cat_r_post_event_damage_establishment","name":"Post-Event Damage Establishment and Claims Quantification","channel":"R","channel_label":"Recovery","tier":"confirmed","tier_rationale":"Three source-type axes documented — and this is the ONLY R-channel category in the pilot that reaches three. regulator_filings: the Copernicus EMS on-demand mapping portal (free of charge, authorised-requester gated, 908 rapid-mapping and 149 recovery activations at fetch, with four live Iberian and Greek wildfire activations listed), the EMS service-component page, and EIOPA's technical description recording that wildfire risk is assessed by expert judgement for all thirty countries in its protection-gap dashboard while earthquake and flood are modelled. industry_reports: the CEMS national-node bundle (e-GEOS, Planetek Hellas, Telespazio Iberica), Bdeo's claimant-side visual-intelligence product, and Reask's product line with its announced-but-unreleased European wildfire view. academic_studies: the JRC/EFFIS advance report, which documents the Rapid Damage Assessment product chain and the land-cover and threshold rules that determine what a mapped burned area actually contains. Investment intensity EUR 0 disclosed across all seven ventures: five are public-service or public-contract entities, and the two commercial ventures disclose no European funding figure. Confirmed tier does not require investment evidence.","definition":"Establishing, faster and more objectively, what was damaged and what it will cost — satellite-derived burned-area and damage-grading delineation delivered to national civil-protection authorities free of charge within hours, the recovery-phase cartography that follows over weeks to months, the industrial primes and national nodes that actually produce those products, claimant-side smartphone capture that classifies damage and estimates repair cost without a physical inspection, and calibrated event reconstruction used to set reserves and buy matched reinsurance. The mechanism is RECOVERY and it is purely economic: none of it changes the physical severity of the fire. It reduces loss-adjustment expense, settlement delay and the additional-living-expense and business-interruption tails that accrue while the damage footprint is contested.","cluster_rationale":"All seven ventures answer the same question faster and more objectively — what burned, how badly, and what is owed — and they are clustered on that mechanism rather than on actor type or tech stack, which here spans an EU public service, its industrial prime and two national delivery nodes, a smartphone-capture claims platform and a physics-based catastrophe model. I071 and I072 are the emergency- and recovery-phase limbs of the same free EU service, counted separately by the service itself; I073, I074 and I075 are the contractors that actually produce those outputs for Italy, Spain and Greece; I076 is the claimant-side capture layer; I077's own stated recovery effect is that a calibrated post-event footprint lets a carrier reserve and settle faster, which is this mechanism exactly. THE STRUCTURAL FINDING THAT MUST DRIVE PHASE X: this layer is occupied by a FREE PUBLIC GOOD, and the correct model is DISPLACEMENT, NOT GREENFIELD PENETRATION. Copernicus EMS and its national nodes deliver post-event damage assessment at zero cost to the user, and every commercial post-catastrophe imagery player is verifiably absent from these four markets — Nearmap and Betterview do not fly Europe, Cape Analytics and ZestyAI are US-only, Vexcel names only Germany and the UK, and ICEYE's wildfire product is gated to the US and Australia. Adoption for the imagery limb is therefore already high at y0 and is not a diffusion curve; there is no commercial revenue pool to model in it. The only headroom is the CLAIMANT-SIDE LAST MILE that the public service does not reach — per-policy triage and claimant capture — where I076 is the sole identified occupant, with the caveat on its own page that its demonstrated depth is motor and generic home peril and no wildfire capability is evidenced. The recovery-phase limb is thinner than the emergency-phase one (149 activations against 908), which is where a commercial recovery-analytics offer could still find room and where nothing currently sits. Boundary vs cat_r_physical_restoration_capacity: this category establishes what is owed; that one performs the works that discharge it. Boundary vs cat_r_public_recovery_substitution: this category makes the loss cheaper to settle; that one shifts part of the loss off the private policy altogether. Boundary vs cat_h_detection_forecasting_decision_support: I008 (ICEYE) also classifies per-building damage, but its verified Southern-European footing is the sovereign Hazard-monitoring contract and it is scored there; no ICEYE damage-assessment effect may be credited in this category.","members":["I071","I072","I073","I074","I075","I076","I077"],"member_names":["Copernicus EMS Rapid Mapping","Copernicus EMS Risk and Recovery Mapping","e-GEOS","Telespazio Ibérica","Planetek Hellas","Bdeo","Reask"],"effectiveness":{"value":0.06,"lower":0.01,"upper":0.15,"unit":null},"effectiveness_definition":"ANALYST-INFERRED. Fractional reduction in total R-channel insured cost — loss-adjustment expense plus settlement-delay costs plus the ALE and BI tail — attributable to faster, authoritative establishment of the damage footprint and per-claim severity. The only quantified figures available are vendor-stated and unaudited: Bdeo's handling time from three weeks to five days and inspection from five days to five minutes, plus repair costs reduced up to 16% and fraud detection increased up to 15%. Those are motor and generic home-peril figures with no wildfire capability evidenced, and they are cycle-time and expense metrics, not insured-loss metrics; they set the shape of the upper bound and nothing more. The central value is small because the R channel here is an expense channel, not a damage channel, and because the imagery layer is already occupied at zero cost — the marginal gain is on the residue, not on the whole. Lower bound is near zero for the case where authoritative public mapping already reaches every relevant claim. BASELINE NETTING, made explicit at iteration 2 (critic J4): the Copernicus imagery limb is ALREADY AT OR NEAR FULL ADOPTION at y0 — the service is free, gated only to authorised civil-protection requesters, and it produced four Iberian and Greek wildfire activations inside the fetch window — so it is already embedded in L_0 under the same phase_00_input.yaml#fit_check.rationale constraint applied to the H channel, and no diffusion of it is creditable. The 0.06 is scoped to the forward increment, which lives almost entirely in the claimant-side last mile. PHASE V INSTRUCTION: this row CANNOT take a single adoption anchor — the imagery limb sits at roughly 1.0 at y0 while the claimant-side limb sits at roughly 0, and averaging them produces a number that describes neither.","effectiveness_sources":[{"slug":"r_copernicus_ems_rapid_mapping","title":"Copernicus EMS On-Demand Mapping portal — activation register and free-of-charge statement","tier":"T1","publisher":"copernicus_ems","url":"https://mapping.emergency.copernicus.eu/","wayback_url":"https://web.archive.org/web/20260719210639/https://mapping.emergency.copernicus.eu/","snapshot_hash":"4297203d1a9607090d4362d44a5e07886204a1b8c124cce6208b165df775d96a","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/4297203d1a9607090d4362d44a5e07886204a1b8c124cce6208b165df775d96a.html","publication_date":null,"fetched_at":"2026-07-31T13:10:05Z","page_exists":true},{"slug":"r_claims_visual_intelligence","title":"Bdeo — visual-intelligence claims automation, product scope and vendor performance claims","tier":"T3","publisher":"bdeo","url":"https://bdeo.io/","wayback_url":"https://web.archive.org/web/20260709020237/https://bdeo.io/","snapshot_hash":"9003b5bb1f57740474ff94330c6a38f84d69f31775786c23441dd0796994ad31","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/9003b5bb1f57740474ff94330c6a38f84d69f31775786c23441dd0796994ad31.html","publication_date":null,"fetched_at":"2026-07-31T13:10:19Z","page_exists":true},{"slug":"jrc_effis_annual_report_2024","title":"Advance report on Forest Fires in Europe, Middle East and North Africa 2024","tier":"T1","publisher":"jrc","url":"https://data.effis.emergency.copernicus.eu/effis/reports-and-publications/effis-related-publications/advance_report_on_forest_fires_in_europe_middle_east_and_north_africa_2024.pdf","wayback_url":"https://web.archive.org/web/https://data.effis.emergency.copernicus.eu/effis/reports-and-publications/effis-related-publications/advance_report_on_forest_fires_in_europe_middle_east_and_north_africa_2024.pdf","snapshot_hash":"3140c958a6f34d47f9c15f4a94a81ebecbc1290f6d8967017a8d0f5e6f256e62","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/3140c958a6f34d47f9c15f4a94a81ebecbc1290f6d8967017a8d0f5e6f256e62.pdf","publication_date":"2025-03-21","fetched_at":"2026-07-31T12:17:12Z","page_exists":true}],"addressable_share":0.7,"axes_hit":["industry_reports","academic_studies","regulator_filings"],"citations":[{"axis":"regulator_filings","wiki_source_id":"r_copernicus_ems_rapid_mapping","span":"The service provides free-of-charge satellite mapping in cases of natural hazards; activation counters at fetch were 908 rapid-mapping, 88 preparedness and 149 recovery; activation is restricted to authorised requesters — national civil-protection authorities, not claimants or insurers. Wildfire activations listed at fetch: Crete (EMSR907), Castilla y Leon and Norte Portugal (EMSR908), Province of Leon (EMSR906), Plana Baixa (EMSR905)."},{"axis":"regulator_filings","wiki_source_id":"r_copernicus_ems_wildfire_activations","span":"The service portal presents Copernicus EMS as a set of components: On-Demand Mapping, Wildfires, Floods (EFAS and GloFAS), Droughts (EDO and GDO) and Exposure Mapping. Wildfires is a standing named component, not an ad-hoc activation category. A wildfire activation headed 'Wildfires in Castilla y Leon, Spain and Norte, Portugal' was carried as current on the portal, dated 30 Jul 2026 — a single activation spanning two of the pilot's four countries."},{"axis":"regulator_filings","wiki_source_id":"eiopa_protection_gap_technical_description","span":"EIOPA records that wildfire risk is assessed by expert judgement for all thirty countries in the protection-gap dashboard, where earthquake and flood are modelled — the supervisory confirmation that no European wildfire catastrophe model exists in any member state."},{"axis":"industry_reports","wiki_source_id":"r_cems_national_nodes","span":"e-GEOS is the industrial prime whose newsroom confirms an active operational role in CEMS Rapid Mapping delivery; Planetek Hellas signed its Copernicus EMS contract in March 2023 as the Greek node yet carries no wildfire-mapping content at fetch; Telespazio Iberica publishes post-fire satellite imagery for Spain without asserting a formal node role."},{"axis":"industry_reports","wiki_source_id":"r_claims_visual_intelligence","span":"Bdeo turns smartphone image and video capture into damage detection, repair-cost estimation and automated claims decision, with vendor-stated unaudited figures of handling from three weeks to five days and inspection from five days to five minutes; lines of business named are motor, home claims and fleet — no wildfire and no peril-specific property line."},{"axis":"industry_reports","wiki_source_id":"r_european_wildfire_cat_models","span":"Reask's named products are tropical-cyclone only (DeepCyc, LiveCyc, Metryc); the European wildfire item on the homepage is an article, not a released model — evidence of stated commercial interest, not of availability."},{"axis":"academic_studies","wiki_source_id":"jrc_effis_annual_report_2024","span":"EFFIS Rapid Damage Assessment methodology: where a mapped fire includes agricultural or urban areas those land covers are included in the event's area estimate, and per-land-cover information is provided for each mapped event — the product rules that determine what a published burned-area footprint actually contains for claims purposes."}],"total_disclosed":0.0,"unknown_count":7,"adoption":{"y0":{"value":0.03,"lower":0.0,"upper":0.08,"unit":null},"y3":{"value":0.15,"lower":0.05,"upper":0.28,"unit":null},"y5":{"value":0.25,"lower":0.1,"upper":0.42,"unit":null},"y10":{"value":0.45,"lower":0.15,"upper":0.62,"unit":null}},"saturation_cap":0.7,"mitigation":{"I1":{"value":1.0,"confidence":"L","sources":[],"rationale":"NEUTRAL, EVIDENCE GAP - no premium channel is evidenced. Faster, more objective damage establishment is a post-loss process; it could in principle improve retention through a better claims experience, but no source in this corpus asserts or measures that, and inferring it would manufacture a channel. Copernicus EMS is free of charge and gated to AUTHORISED REQUESTERS - national civil-protection authorities, not claimants or insurers - so there is not even a commercial relationship to price. The structural reason no such channel exists anywhere in this pilot: wildfire is not a separately rated line in Spain, Portugal, Italy or Greece - it is bundled inside multi-peril household and commercial fire cover - so there is no wildfire premium series for any mitigation to move. Value held at 1.00.","evidence_gap":true,"gap_reason":"No premium, retention or underwriting effect attributed to post-event mapping, claimant-side capture or event reconstruction in ES/PT/IT/GR. Searched: r_copernicus_ems_rapid_mapping, r_copernicus_ems_wildfire_activations, r_cems_national_nodes, r_claims_visual_intelligence, r_claims_tech_scope_negatives, r_european_wildfire_cat_models, insurance_journal_2026_marsh_europe_wildfire."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. Satellite damage mapping, claimant-side capture and event reconstruction do not touch insurer asset allocation, duration or investment return. The one adjacency worth naming and dismissing: faster reserving marginally changes the TIMING of cash outflow and therefore of investable float - a second-order effect, unquantified by any source, and immaterial at this pilot's loss scale. Searched: r_copernicus_ems_rapid_mapping, r_claims_visual_intelligence, r_european_wildfire_cat_models, eiopa_protection_gap_technical_description. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":0.958,"confidence":"L","sources":["wiki/sources/r_copernicus_ems_rapid_mapping.md","wiki/sources/r_copernicus_ems_wildfire_activations.md","wiki/sources/jrc_effis_annual_report_2024.md","wiki/sources/r_claims_visual_intelligence.md","wiki/sources/r_cems_national_nodes.md","wiki/sources/r_european_wildfire_cat_models.md","yaml-ref:phase_04_5_categories.yaml#cat_r_post_event_damage_establishment.effectiveness","yaml-ref:phase_04_5_categories.yaml#cat_r_post_event_damage_establishment.addressable_share"],"rationale":"value 0.958 = 1 - (effectiveness 0.06 x addressable_share 0.70), Phase IV.5 parameters unchanged. A REAL MECHANISM ALREADY OCCUPIED BY A FREE PUBLIC GOOD, SO THE CORRECT MODEL IS DISPLACEMENT RATHER THAN GREENFIELD PENETRATION, AND THE FORWARD INCREMENT LIVES ALMOST ENTIRELY IN THE CLAIMANT-SIDE LAST MILE. Three genuine on-mechanism axes. regulator_filings: Copernicus EMS provides free-of-charge satellite mapping for natural hazards, with activation counters at fetch of 908 rapid-mapping, 88 preparedness and 149 recovery, restricted to authorised requesters; Wildfires is a STANDING NAMED COMPONENT of the service, and a single activation dated 30 July 2026 spanned Castilla y Leon and Norte Portugal - two of the pilot's four countries. industry_reports: the CEMS national-node bundle (e-GEOS as industrial prime, Planetek Hellas as the Greek node with no wildfire-mapping content at fetch, Telespazio Iberica publishing post-fire imagery without asserting a formal node role); Bdeo's claimant-side smartphone capture; and Reask, whose named products are tropical-cyclone only and whose European wildfire item is an ARTICLE, not a released model. academic_studies: the JRC/EFFIS report documenting the Rapid Damage Assessment product chain and the land-cover and threshold rules that determine what a published burned-area footprint actually contains for claims purposes. WHAT IS AND IS NOT MEASURED: the only quantified figures are Bdeo's vendor-stated, unaudited handling 3 weeks -> 5 days, inspection 5 days -> 5 minutes, repair costs reduced up to 16% and fraud detection up to 15% - MOTOR AND GENERIC HOME figures, with NO wildfire capability evidenced anywhere on the page, and cycle-time and EXPENSE metrics rather than insured-loss metrics; they shape the upper bound and nothing more. BASELINE NETTING, EXPLICIT: the Copernicus imagery limb is already at or near FULL ADOPTION at y0 - free, gated only to authorised civil-protection requesters, four Iberian and Greek wildfire activations inside the fetch window - so it is embedded in L_0 and no diffusion of it is creditable. The 0.06 is scoped to the forward increment. CONFIDENCE LOWERED M -> L AT ITERATION 3, on the same coherence test this composer applied to cat_v_vulnerability_assessment_targeting at iteration 1. The M rested on three genuine T1 sources - but all three document the free Copernicus imagery layer, which THIS CELL THEN REMOVES FROM THE CREDIT as embedded in L_0 with no diffusion creditable. The creditable residue is the claimant-side last mile, and the only evidence for it is Bdeo: T3, vendor-stated, unaudited, motor and generic home, with no wildfire capability evidenced and no customer named. A T1 that evidences the excluded limb cannot carry confidence for the included one. M would have remained formally legal under matrix_schema.md section 5, which is a source-quality rule and the T1s are genuine; it is withdrawn because sustaining it would read the same evidence two ways in one file. The value is unchanged: 0.958 is 1 - (0.06 x 0.70) from signed-off Phase IV.5 parameters and no confidence tag bears on it. DOUBLE-COUNT CONTROL: I008 (ICEYE) also classifies per-building damage, but its verified Southern-European footing is the sovereign Hazard-monitoring contract and it is scored in cat_h_detection_forecasting_decision_support; no ICEYE damage-assessment effect is credited here. ADOPTION CAVEAT CARRIED FORWARD FOR THE CRITIC: Phase IV.5 states in terms that this row cannot take a single adoption anchor - the imagery limb sits near 1.0 at y0 while the claimant-side limb sits near 0, and the 0.45 y10 anchor in this matrix averages them."},"I4":{"value":1.0,"confidence":"L","sources":["wiki/sources/r_claims_visual_intelligence.md","wiki/sources/r_claims_tech_scope_negatives.md","wiki/sources/r_copernicus_ems_rapid_mapping.md"],"rationale":"NEUTRALISED TO 1.00 AND FLAGGED AS AN EVIDENCE GAP AT ITERATION 1. Iteration 0 carried this cell active at 0.979 = 1 - (an I4-specific effectiveness of 0.03 x addressable_share 0.70). Both halves of that construction are withdrawn. FIRST, THE DISPOSITION WAS WRONG. The iteration-0 rationale gave four reasons the cell could not carry a confident number - the Bdeo figures are vendor-published and unaudited, the named lines are motor, home and fleet with no wildfire and no peril-specific property line, no customer is named, and the adjacent vendor evidence is a clean negative in which Tractable's four modules are framed exclusively around vehicle damage with no property, home, building or wildfire peril named anywhere. That is precisely matrix_loop.md section 7's cannot-source condition, whose remedy is evidence_gap: true with a gap_reason, not an unflagged active value. Every other cell in this matrix where the composer reached that judgement is flagged; this one was not. SECOND, THE ARGUMENT WAS APPLIED TO ONE FACTOR AND NOT THE OTHER. The reason given for refusing to carry the I3 effectiveness of 0.06 into I4 - that doing so would assert the operational effect equals the claims effect - disqualifies the addressable share by exactly the same move, and the addressable share was carried across unchanged at 0.70. On matrix_schema.md v0.8.1's reading, which the iteration-0 rationale itself invoked, BOTH stored quantities are I3 quantities wearing no label; and v0.8.8 defines AS_c once per category against a share_of_damage denominator, while I4 is an operational-expense line and not damage. Halving one imported I3 quantity while importing the other whole is not a defensible split. THIRD, NO REPLACEMENT NUMBER IS OFFERED, DELIBERATELY. The 0.03 was arithmetic performed on an instruction in phase_09_evidence.yaml - 'keep I4 small relative to I3 rather than counting the same compression twice' - and an instruction about relative magnitude is not evidence for any particular ratio. 0.015 or any other fraction would be a second unsourced number replacing the first. WHAT SURVIVES AND IS UNCHANGED: this is still the only impact area outside I3 with any evidence at all in this pilot, and the only category where I4 is the PRIMARY mechanism rather than the shadow of I3 - the R channel here is an EXPENSE channel, reducing loss-adjustment expense, settlement delay and the ALE/BI tail that accrues while the damage footprint is contested, and Bdeo is the only genuine Southern-European commercial claims-tech venture in the pilot. The mechanism is real, sourced and directionally favourable; only its magnitude is unavailable. BASELINE NETTING ALSO STILL APPLIES: the free Copernicus layer already absorbs the authoritative-footprint cost at zero price to the user, so any creditable I4 increment is the claimant-side residue and not the whole. CONSEQUENCE FOR THE AGGREGATE: total_factor_I4 moves from 0.99055 to 1.000000, so the matrix now shows no operational-cost mitigation at all. That is the honest position, not a modelling choice - it says the pilot found a real mechanism and no number for it.","evidence_gap":true,"gap_reason":"Neither factor of the withdrawn 0.979 has an I4-scoped basis: the 0.03 effectiveness was a transparent allocation of an I3 effectiveness against a Phase IX instruction about relative magnitude, not a measurement, and the 0.70 addressable share is an I3 share_of_damage quantity applied to an operational-expense line. Substantively, NO WILDFIRE LOSS-ADJUSTMENT-EXPENSE, EXPENSE-RATIO, ADJUSTER-DEPLOYMENT OR CLAIMS-OPERATING-COST FIGURE EXISTS FOR SPAIN, PORTUGAL, ITALY OR GREECE ANYWHERE IN THIS CORPUS. Searched: r_claims_visual_intelligence (Bdeo - vendor-stated, unaudited, motor and generic home, no wildfire capability evidenced, no customer named), r_claims_tech_scope_negatives (Tractable - clean negative, vehicle damage only), r_copernicus_ems_rapid_mapping, r_copernicus_ems_wildfire_activations, r_cems_national_nodes, jrc_effis_annual_report_2024, eiopa_protection_gap_technical_description, reinsurance_news_attica_2018_insured_loss, moodys_europe_record_wildfire_season_2026. What a future source would have to supply is a wildfire-peril loss-adjustment expense per claim, or an audited cycle-time-to-expense conversion for a European property book."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","regulator_filings"],"distinct_sources":7,"total_disclosed":0.0,"unknown_count":7,"reason":"3 evidence axes across 7 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_r_post_event_damage_establishment.md"},{"id":"cat_r_public_recovery_substitution","name":"Public Post-Event Recovery Substitution and Backstops","channel":"R","channel_label":"Recovery","tier":"confirmed","tier_rationale":"TIER MET (issue #52, 2026-09-23): three source-type axes — regulator_filings, academic_studies and industry_reports. Note this category is scored NEUTRAL (effectiveness 0.0) on scope grounds, so no active cell should arise from it at all. industry_reports: the Geneva Association's February 2026 report on public-private insurance programmes states that the expectation of post-disaster government relief may suppress demand for insurance — the substitution mechanism itself, for built property, though global in scope with no Southern-European wildfire section. total_disclosed_eur EUR 0. regulator_filings: the SGIFR portal establishing Pos-Evento as one of six named, permanently mandated phases of Portugal's integrated rural-fire system, and the Kratiki Arogi portal consolidating Greek state housing assistance, household-goods compensation, business subsidies and property-tax relief with a dedicated wildfire section and event-by-event schemes. academic_studies: the EEA's European Climate Risk Assessment (EEA Report 01/2024), recording that the viability of EU solidarity funds is already critically threatened by oversubscription from events including wildfires, and that policy needs to promote insurance with resilience-enhancing provisions limiting stress on public finances after extreme events — the technical-report basis for treating public substitution as a determinant of insured share. Read that axis for what it is: a framing document with no wildfire-recovery numbers. WITHDRAWN AT ITERATION 2 (critic M11): industry_reports is withdrawn. Its sole citation, the Agroseguro pool page, evidences a scheme whose covered lines are agricultural, livestock, aquaculture and forestry — which phase_00_input.yaml.out_of_scope excludes from this pilot in terms (\"Agricultural, forestry and standing-timber loss … only damage to built property … is in scope\"), and I084's own record adds that Spain's Consorcio does not cover wildfire property damage. The axis therefore evidenced the mechanism operating ENTIRELY OUTSIDE the loss quantity this pilot forecasts. Agroseguro is retained as a constituent and as the reference architecture for what a working wildfire backstop looks like, but not as an axis. Investment EUR 0: all three were checked; none publishes a wildfire-specific budget, and the pilot explicitly refuses the circulating Greek 80/20 grant-loan split as unsourced. PHASE X CONSEQUENCE: if any cell in this category is ever made active against this composer's instruction, apply the §5 `emerging` cap — confidence: L.","definition":"Public and pooled capacity that absorbs part of the post-event cost before it reaches a private property policy — a standing state mandate for burned-land stabilisation, community support and reconstruction coordination; consolidated but discretionary state disaster aid created scheme by scheme after each event; and a state-backed co-insurance pool that indemnifies fire damage in agricultural, livestock, aquaculture and forestry lines. The mechanism is RECOVERY, and for an insurer it cuts two ways: it accelerates physical recovery, and it substitutes state capacity for private indemnity, suppressing the insured share of wildfire loss structurally rather than cyclically. RE-CAST AT ITERATION 2 (critic M10): this row is an EXPOSURE-MODEL / L_0-BRIDGE INPUT, NOT A MITIGATION. A change in the insured share is a protection-gap movement, and phase_00_input.yaml.out_of_scope confines the protection gap to the bridge used when converting economic-damage sources into insured terms — it does not belong in the mitigation term Pi_c (1 - M_c(t)). Scoring it as mitigation would also have made the lattice internally incoherent: cat_v_vulnerability_assessment_targeting correctly REFUSES Fidelidade's measured +60% adequate-coverage sales uplift on exactly this ground, so admitting insured-share DECREASES as mitigation while excluding insured-share INCREASES as anti-mitigation would be a one-sided treatment of the same physical quantity and would bias the forecast downward. The same quantity is now treated the same way in both places. The category is carried in the lattice ONLY because v0.7 cannot express a non-mitigating row and dropping it would orphan three inventory ventures; its effectiveness is set to 0.0 (I3 = 1.0, neutral) and its content belongs in Phase II/IX exposure work and in the Phase VII recalibration triggers.","cluster_rationale":"All three ventures act on the same variable: how much wildfire loss ever reaches a private property policy. I082 gives Portuguese post-event stabilisation, community support and reconstruction coordination a standing interinstitutional mandate; I083 consolidates Greek ex-post state aid across housing, contents, business and property tax; I084 is the only verified public backstop in the four countries that actually pays for fire, through a state-backed co-insurance pool. They differ in form — a programme mandate, discretionary aid, and a pooled indemnity scheme — and are clustered on mechanism, which is substitution of public or pooled capacity for private indemnity, together with the recovery acceleration that a standing mandate provides. THE SCOPE CAVEAT THAT SETS THE ADDRESSABLE SHARE: Agroseguro covers agricultural, livestock, aquaculture and forestry lines, NOT residential or commercial buildings, and Spain's Consorcio explicitly excludes wildfire property damage, so the pooled limb touches very little of this pilot's property scope; the Greek and Portuguese limbs are discretionary and unquantified. addressable_share is held to 0.25 for that reason and the whole category should be read as an exposure-model input rather than a mitigation opportunity — none of the three is contractable by VIG. Boundary vs cat_r_physical_restoration_capacity: that category lowers the cost that stays on the policy; this one moves cost off the policy entirely, and the two must not both be credited against the same euro of loss. Boundary vs cat_r_post_event_damage_establishment: that category is the evidence base on which any payer, public or private, acts; this one is who pays.","members":["I082","I083","I084"],"member_names":["SGIFR Pós-Evento (Portugal)","Κρατική Αρωγή (Greek state disaster aid)","Agroseguro"],"effectiveness":{"value":0.0,"lower":0.0,"upper":0.0,"unit":null},"effectiveness_definition":"SET TO ZERO ON SCOPE GROUNDS AT ITERATION 2 — this is a deliberate neutral, not a measurement (critic M10). The mechanism this category describes is a movement in the INSURED SHARE: public or pooled capacity absorbs cost that would otherwise fall on a private policy. phase_00_input.yaml.out_of_scope confines the protection gap to the L_0 bridge — \"the model forecasts insured-market loss; the protection gap enters only as a bridge when converting economic-damage sources into insured terms\" — so an insured-share movement is not admissible in the mitigation term. The iteration-1 value of 0.10 scored it as mitigation, which was both a scope breach and inconsistent with this lattice's own refusal of the Fidelidade +60% adequate-coverage sales uplift in cat_v_vulnerability_assessment_targeting on identical grounds. Both directions of the same quantity are now excluded from mitigation. WHAT WAS SEARCHED FOR AND NOT FOUND, so that the zero is a finding rather than a shrug: Portugal's SGIFR portal publishes no figures on post-fire recovery activity and the 21 August 2025 reconstruction-financing announcement has no primary source in this pilot; Greece's Kratiki Arogi publishes no benefit amounts, ceilings or eligibility thresholds and creates schemes event by event with discretionary scope, making any offset not merely large but UNPREDICTABLE — worse for pricing than a large but stable backstop; and Agroseguro, the only verified public backstop in the four countries that actually pays for wildfire, covers agricultural, livestock, aquaculture and forestry lines, with Spain's Consorcio excluding wildfire property damage, so residential and commercial building loss — this pilot's scope — sits outside any state consortium. WHERE THE CONTENT GOES INSTEAD: Phase II / Phase IX insured-share and exposure work, and the Phase VII recalibration trigger, since a single legislative act bringing wildfire into the Spanish Consorcio or the Italian CatNat mandate would move the insured share in a step — which is UPSIDE to insured loss, not downside.","effectiveness_sources":[{"slug":"r_state_recovery_portugal","title":"SGIFR — Portugal's integrated rural-fire management system and its named Pós-Evento phase","tier":"T1","publisher":"sgifr","url":"https://sgifr.gov.pt/","wayback_url":"https://web.archive.org/web/20260731130845/https://sgifr.gov.pt/","snapshot_hash":"29ad94a24d769044d575866509173d3f207a96c7d2cfd4894b3af06fdc669f9b","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/29ad94a24d769044d575866509173d3f207a96c7d2cfd4894b3af06fdc669f9b.html","publication_date":null,"fetched_at":"2026-07-31T13:10:40Z","page_exists":true},{"slug":"r_state_recovery_greece","title":"Κρατική Αρωγή (arogi.gov.gr) — Greek state disaster-recovery aid, wildfire section and benefit types","tier":"T1","publisher":"arogi_gov_gr","url":"https://arogi.gov.gr/","wayback_url":"https://web.archive.org/web/20260309163153/https://arogi.gov.gr/","snapshot_hash":"b1b6c99395ad0245637559605df98d6af058bfd5e37a94bca6554bb8a71409f9","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/b1b6c99395ad0245637559605df98d6af058bfd5e37a94bca6554bb8a71409f9.html","publication_date":null,"fetched_at":"2026-07-31T13:10:41Z","page_exists":true},{"slug":"r_agroseguro_public_backstop","title":"Agroseguro — Spain's combined agricultural insurance pool and its fire cover","tier":"T2","publisher":"agroseguro","url":"https://agroseguro.es/","wayback_url":"https://web.archive.org/web/20260724051114/https://agroseguro.es/","snapshot_hash":"81f1762104d2d3064ea26e7ee43eb28c994aea04ff7daf2a7932724e62971efd","snapshot_path":"pilots/2026-wildfire-southern-europe/wiki/snapshots/81f1762104d2d3064ea26e7ee43eb28c994aea04ff7daf2a7932724e62971efd.html","publication_date":null,"fetched_at":"2026-07-31T13:10:43Z","page_exists":true}],"addressable_share":0.25,"axes_hit":["academic_studies","regulator_filings","industry_reports"],"citations":[{"axis":"regulator_filings","wiki_source_id":"r_state_recovery_portugal","span":"The SGIFR process chain explicitly lists Pos-Evento as one of six operational phases, establishing post-fire recovery as a named, permanently resourced pillar of the Portuguese national fire system rather than an ad-hoc response; the portal itself publishes no figures on recovery activity."},{"axis":"regulator_filings","wiki_source_id":"r_state_recovery_greece","span":"Kratiki Arogi consolidates Greek government measures for support and restoration of disaster-affected areas — housing assistance, household-goods compensation, business subsidies and ENFIA property-tax exemptions — with event-specific schemes and their own application platforms, and no benefit amounts, ceilings or eligibility thresholds published."},{"axis":"academic_studies","wiki_source_id":"eea_eucra_2024","span":"EEA Report 01/2024: the viability of EU solidarity funds is already critically threatened as these have been oversubscribed by costly events such as floods and wildfires; policies need to promote insurance with resilience-enhancing provisions that limit the stress on public finances following extreme events, considering large differences in insurance penetration across Member States."},{"axis":"industry_reports","wiki_source_id":"geneva_association_2026_public_private_insurance_programmes","span":"Institutional factors: The expectation that governments will step in, or will be obliged, to provide post-disaster relief may suppress demand for insurance."}],"total_disclosed":0.0,"unknown_count":3,"adoption":{"y0":{"value":0.0,"lower":0.0,"upper":0.0,"unit":null},"y3":{"value":0.0,"lower":0.0,"upper":0.0,"unit":null},"y5":{"value":0.0,"lower":0.0,"upper":0.0,"unit":null},"y10":{"value":0.0,"lower":0.0,"upper":0.0,"unit":null}},"saturation_cap":0.0,"mitigation":{"I1":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 ON SCOPE GROUNDS, NOT FOR WANT OF SEARCHING - and therefore no evidence_gap flag. This category's whole mechanism is a movement in the INSURED SHARE: public or pooled capacity absorbing cost that would otherwise fall on a private policy. A change in insured share is a protection-gap movement, which phase_00_input.yaml.out_of_scope confines to the L_0 bridge ('the model forecasts insured-market loss; the protection gap enters only as a bridge when converting economic-damage sources into insured terms'). A smaller insured share means less premium as well as less loss; neither side is admissible in the mitigation term. Phase IV.5 set this category's effectiveness to 0.00 at iteration 2 for exactly this reason, and did so for SYMMETRY with its refusal of Fidelidade's +60% adequate-coverage uplift in cat_v_vulnerability_assessment_targeting. Admitting insured-share decreases as mitigation while excluding insured-share increases would be a one-sided treatment of one quantity and would bias the forecast downward. Confidence H is confidence in the scope decision, not in a measurement. WHAT WAS SEARCHED FOR AND NOT FOUND, appended at iteration 1 to discharge the matrix_schema.md v0.7 section 2 obligation that a neutral cell's rationale name the search and not only the argument: no premium, retention or underwriting effect of state recovery programmes or public backstops was found in any of r_state_recovery_portugal, r_state_recovery_greece, r_agroseguro_public_backstop or eea_eucra_2024. Agroseguro is the only one of the four that prices anything at all, and it prices agricultural, livestock, aquaculture and forestry lines that this pilot excludes."},"I2":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00 - NO PLAUSIBLE INVESTMENT CHANNEL, and this is a structural finding rather than an absent search. State disaster aid, standing post-event mandates and a state-backed agricultural co-insurance pool do not touch insurer asset allocation, duration or investment return. eea_eucra_2024's finding that EU solidarity funds are critically threatened by oversubscription is a PUBLIC-FINANCE observation, not an insurer investment-income pathway. Searched: r_state_recovery_portugal, r_state_recovery_greece, r_agroseguro_public_backstop, eea_eucra_2024. Nothing. The scale argument that closes it is the same one that closes baseline I2: Phase II puts the entire four-country insured wildfire market at EUR 130 m/yr central and EUR 350 m/yr in a record season, against insurer balance sheets orders of magnitude larger, so no loss path in this pilot's bracket alters asset allocation, duration structure or the yield on float. The two candidate indirect routes were considered and rejected at Phase IX: catastrophe-bond issuance (a fact about the buy side, not about a primary insurer's investment income) and utility or sovereign credit exposure (closed by phase_00_input.yaml.out_of_scope, which excludes wildfire liability and subrogation). Confidence H is confidence in the NEGATIVE - that no channel exists at this loss scale - not confidence in a number. No evidence_gap flag: this is a finding, not a hole."},"I3":{"value":1.0,"confidence":"H","sources":["wiki/sources/r_agroseguro_public_backstop.md","wiki/sources/r_state_recovery_portugal.md","wiki/sources/r_state_recovery_greece.md","wiki/sources/eea_eucra_2024.md"],"rationale":"NEUTRAL 1.00 BY DELIBERATE SCOPE DECISION - value 1.000 = 1 - (effectiveness 0.00 x addressable_share 0.25), so the identity holds trivially and the zero is a finding rather than a shrug. No evidence_gap flag: the search behind it is recorded. THE SCOPE GROUND: the mechanism is an insured-share movement, excluded from the mitigation term by phase_00_input.yaml.out_of_scope, and Phase IV.5 set effectiveness to 0.00 at iteration 2 for SYMMETRY with its refusal of Fidelidade's +60% adequate-coverage uplift in cat_v_vulnerability_assessment_targeting. Both directions of the same quantity are excluded; admitting one would bias the forecast. THE SCOPE GROUND IS REINFORCED BY WHAT THE SOURCES ACTUALLY SAY. (a) Agroseguro is the only verified public backstop in the four countries that actually PAYS for wildfire - and its covered lines are agricultural, livestock, aquaculture and forestry, which phase_00_input.yaml excludes in terms; Spain's Consorcio does not cover wildfire property damage at all, so residential and commercial building loss - this pilot's entire scope - sits OUTSIDE any state consortium in Spain. Phase IV.5 withdrew the industry_reports axis on exactly this ground. (b) Portugal's SGIFR portal establishes Pos-Evento as one of six permanently mandated phases but publishes NO figures on post-fire recovery activity, and the 21 August 2025 reconstruction-financing announcement has no primary source in this pilot. (c) Greece's Kratiki Arogi consolidates housing assistance, household-goods compensation, business subsidies and ENFIA property-tax relief but publishes NO benefit amounts, ceilings or eligibility thresholds and creates schemes EVENT BY EVENT with discretionary scope - making any offset not merely large but UNPREDICTABLE, which is worse for pricing than a large but stable backstop. The circulating Greek 80/20 grant-loan split is refused as unsourced and is not carried, not even as a bracket endpoint. WHERE THE CONTENT GOES INSTEAD: Phase II / Phase IX insured-share and exposure work, and the Phase VII recalibration trigger - a single legislative act bringing wildfire into the Spanish Consorcio or the Italian CatNat mandate would move the insured share in a STEP, which is UPSIDE to insured loss, not downside. Note also that this category's adoption is 0.00, so even an active cell would contribute nothing to the aggregation. Confidence H is confidence in the scope decision, not in a measurement."},"I4":{"value":1.0,"confidence":"H","sources":[],"rationale":"NEUTRAL 1.00, AND NO OPERATIONAL CHANNEL EVIDENCED EITHER WAY - so no evidence_gap flag, because the category is neutral on SCOPE grounds before evidence is reached. A standing post-event mandate could in principle accelerate physical recovery and so shorten claim life, but Portugal's SGIFR portal publishes no recovery-activity figures and Greece's Kratiki Arogi schemes are created event by event with no published thresholds, so there is nothing to measure. The category is in any case carried as an exposure-model input rather than as a mitigation - its mechanism is an insured-share movement, excluded from the mitigation term by phase_00_input.yaml.out_of_scope - and none of its three constituents is contractable by VIG. WHAT WAS SEARCHED FOR AND NOT FOUND: any insurer claims-operations or expense effect attributed to state recovery programmes or public backstops in ES/PT/IT/GR. Searched: r_state_recovery_portugal, r_state_recovery_greece, r_agroseguro_public_backstop, eea_eucra_2024, r_claims_visual_intelligence."}},"tier_check":{"computed":"confirmed","declared":"confirmed","agrees":true,"axes_hit":["academic_studies","industry_reports","regulator_filings"],"distinct_sources":4,"total_disclosed":0.0,"unknown_count":3,"reason":"3 evidence axes across 4 sources (needs 3 and 3)"},"wiki_page":"wiki/categories/cat_r_public_recovery_substitution.md"}],"excluded":[{"proposed_name":"Standalone commercial post-catastrophe aerial and satellite property intelligence","reason":"No admissible supply exists in the pilot geography, and the mechanism it would carry is already occupied at zero cost. Copernicus EMS Rapid Mapping and its national nodes deliver post-event damage delineation to authorised users free of charge, and every commercial post-catastrophe imagery player examined is verifiably absent from Spain, Portugal, Italy and Greece — Nearmap and Betterview do not fly Europe, Cape Analytics and ZestyAI are US-only, Vexcel names only Germany and the UK, and ICEYE's wildfire product is gated to the US and Australia. All of those entities sit in the inventory's own considered_but_excluded and are therefore not in the pool; standing the category up would create a matrix row with no constituent venture, which v0.7 forbids. The correct Phase X reading is displacement by a public good, not greenfield penetration, and it is carried inside cat_r_post_event_damage_establishment's cluster_rationale rather than as a separate row.\n","re_evaluation_trigger":"Re-evaluate when EITHER (a) a commercial provider publishes a named ES/PT/IT/GR wildfire damage-assessment engagement with an insurer or reconstruction authority, OR (b) Copernicus EMS recovery-phase activations for the four pilot countries are shown to be near zero for wildfire, confirming the thin recovery-phase layer as genuine white space, OR (c) ICEYE or an equivalent opens European commercial availability for its wildfire damage product. Next scheduled review: the quarterly venture scan.\n"},{"proposed_name":"Statutory arson deterrence and post-fire land-use constraint","reason":"Single-venture mechanism (I016, Legge 353/2000 and the municipal catasto of burned areas) whose proximate effect — freezing the land-use classification of burned land for fifteen years, banning construction for ten and voiding sale contracts that fail to recall the constraint — removes the development incentive behind deliberate ignition and is genuinely distinct from planning, funding, detection, suppression and fuel treatment. It cannot stand alone: one source-type axis (regulator_filings only, via Normattiva and the Italian Parliament), zero disclosed funding, and therefore `insufficient` under §4.3. Because §4.3's disposal route (considered_but_excluded) would orphan I016 against validate.py's no-orphan rule, the venture is carried inside cat_h_fire_management_programmes_and_statutes as Italy's framework statute, and the mechanism heterogeneity is declared in that category's cluster_rationale as the lattice's weakest seam. The instrument's post-fire limb was separately deduplicated by the inventory at I085 on the finding that it is neutral-to-loss-amplifying.\n","re_evaluation_trigger":"Re-evaluate when EITHER (a) any study quantifies the ignition reduction attributable to burned-area cadastre regimes — the magnitude is currently unquantified in every source — OR (b) a second pool venture appears with the same deterrence-by-land-use- constraint mechanism (a Spanish, Portuguese or Greek analogue would suffice), making a multi-venture category possible, OR (c) an academic or NGO source axis is added on arson deterrence in Mediterranean Europe.\n"},{"proposed_name":"Exterior structure pre-treatment (retardant gels, wraps and sprinklers) as a standalone category","reason":"Single admissible venture (I057, ECOFIRE) with one usable source-type axis and zero disclosed funding, so `insufficient` under §4.3. The evidence position is worse than thin: the vendor's 30-to-40-times-water efficacy claim is contradicted by the fact-check of its own INIA-CSIC study and is unusable; the only traceable independent figure in the mechanism family (up to 96% radiant and 92% convective heat blocked, Takahashi 2019) belongs to Firezat, which the inventory excluded for want of any named Southern-European deployment, and the underlying paper was not retrieved; IBHS itself states that limited methodical research exists on external sprinkler effectiveness; and Spain's own fire-protection industry association does not recognise the category at all. I057 is therefore carried inside cat_v_retrofit_ember_path_protection, whose shared property is applicability to existing stock without altering permanent fabric.\n","re_evaluation_trigger":"Re-evaluate when EITHER (a) an independent post-incident assessment compares treated and untreated buildings in a documented Spanish or Portuguese wildfire — ECOFIRE's first documented wildfire use was July 2026 with results undocumented — OR (b) the Takahashi 2019 paper is retrieved and its assemblies shown to be representative of Southern-European construction, OR (c) a European vendor brings a certified exterior pre-treatment product to market with a published test method.\n"},{"proposed_name":"Insurer-led wildfire prevention programmes as an actor-type category","reason":"Rejected on the clustering rule itself, not on evidence. Only two internal-insurer initiatives exist in the entire 64-venture pool — I004 (Allianz Seguros early detection) and I060 (Fidelidade ICCC risk map) — and the v0.7 convention clusters on mitigation MECHANISM, not on who operates it. Grouping them would put a detection-and-alerting mechanism and a vulnerability-mapping mechanism in one row solely because both are run by insurers, which is exactly the conflation §3.1 of the convention forbids. They are assigned by mechanism instead: I004 to cat_h_detection_forecasting_decision_support and I060 to cat_v_vulnerability_assessment_targeting. That the peer-insurer row is nearly empty is a finding in its own right — the discovery sweep specifically tested Generali Italia, Unipol, Mapfre, Interamerican, Ageas Portugal and Fundacion Mapfre and found no wildfire prevention programme at any of them — and it belongs in the synthesis report as white space, not in the matrix as a category.\n","re_evaluation_trigger":"Re-evaluate when three or more peer-insurer wildfire initiatives in the pilot geography share a single mitigation mechanism that no vendor category already carries — for example an insurer-operated inspect-and-verify loop coupled to a premium credit, which does not currently exist anywhere in Europe. Next review: on publication of the Allianz Seguros summer-2026 validation outcome or the first Fidelidade ICCC outputs.\n"},{"proposed_name":"European wildfire catastrophe modelling as a standalone category","reason":"No available supply. EIOPA's own technical description records that wildfire risk is assessed by expert judgement for all thirty countries in its protection-gap dashboard while earthquake and flood are modelled, and the pilot's verified negatives include Moody's RMS having no European wildfire model (US and Canada only) and Climate X's wildfire product page returning HTTP 404 at fetch. The single admissible venture, I077 (Reask), lists tropical-cyclone products only; its European wildfire view appears on the site as an article, not a released product, and committing to a category on the strength of an article headline is the failure mode the inventory exists to prevent. I077 is carried inside cat_r_post_event_damage_establishment on its stated recovery mechanism — a calibrated post-event footprint enabling faster reserving and settlement — which is evidenced, rather than on a model that is not.\n","re_evaluation_trigger":"Re-evaluate when EITHER (a) Reask or another vendor publicly releases a European wildfire model with stated resolution and validation geography, OR (b) a named European wildfire reinsurance placement using such a model is announced, OR (c) EIOPA moves wildfire from \"Expert judgement\" to \"Modelled\" for any of the four pilot countries in its protection-gap dashboard.\n"}],"revision_log":[{"iteration":1,"date":"2026-07-31","author":"category-clustering composer (out-of-band run against the resolved cluster_categories.py prompt; no ANTHROPIC_API_KEY configured on the run machine)","note":"METHODOLOGY DEFECT FOUND DURING THIS PASS — raised for the leads, not patched here. category_derivation.md §4.3 requires an `insufficient` category to be sent to considered_but_excluded, but tools/validate.py's no-orphan check (_validate_phase_04_5_categories_v07_extra, check 3) only accepts a venture as accounted-for if it is assigned to a category OR appears in the INVENTORY's considered_but_excluded. A category-level exclusion in this file therefore orphans its constituent ventures and hard-fails validation. Combined with the tier enum's two states (confirmed / emerging) and the fact that 57 of 64 ventures disclose no funding, the practical effect is that any mechanism whose ventures cannot be dropped MUST be labelled `confirmed` or `emerging` even where the evidence does not support it. This pass handled it in three ways, all declared in the affected records: merging mechanisms that cannot clear the three-axis bar alone (cat_h_detection_forecasting_decision_support, cat_v_retrofit_ember_path_protection); tiering `emerging` where disclosed capital genuinely clears the EUR 50 m gate (cat_h_utility_corridor_ignition_prevention); and carrying ONE category — cat_r_physical_restoration_capacity — at `confirmed` with an explicit two-axis deficit statement as the first line of its tier_rationale, plus an instruction that Phase X apply the §5 confidence: L cap to it as though it were emerging. Suggested v0.7.x fix: either add a third tier state that is renderable in the matrix, or extend validate.py's no-orphan check to accept ventures accounted for in this file's own considered_but_excluded with a venture-id list.\n"},{"iteration":1,"date":"2026-07-31","author":"category-clustering composer","note":"Recorded for the record: R-channel source-type coverage is structurally capped at two axes in this pilot for restoration and claims handling (phase_04_venture_inventory.md §2.7 — no academic study and no NGO report discusses post-event wildfire claims handling in ES/PT/IT/GR). Two of the three R categories nonetheless reach three axes by drawing on JRC/EFFIS technical reporting for the mapping product chain (cat_r_post_event_damage_establishment) and on the EEA European Climate Risk Assessment for public-finance substitution (cat_r_public_recovery_substitution). Both of those citations are load-bearing and a critic should test them first.\n"},{"iteration":2,"date":"2026-08-01","author":"category-clustering composer (out-of-band), applying the adversarial critic pass of 2026-08-01 — 12 MUST FIX, 9 SHOULD FIX, 5 judgement calls, 7 methodology-level items; critic verdict material_changes_recommended","note":"ACCEPTED IN FULL — all 12 MUST FIX, all 9 SHOULD FIX and all 5 judgement calls. NOTHING REJECTED. The three the composer had not self-nominated were checked against the repo before being applied and all three were confirmed. M1: phase_00_input.yaml #fit_check.rationale does bind this pilot to \"count only *new* adaptation above the already-embedded historical suppression baseline\", verbatim, and L_0 is anchored at base_year 2026. M2: I015's inventory record, its dossier and its primary source all describe a building-perimeter measure and the dossier states in terms that the 50 m / 100 m distances are NOT on the page. M11: phase_00_input.yaml.out_of_scope does exclude agricultural, forestry and standing-timber loss, which is Agroseguro's entire covered scope. (1) BASELINE NETTING (M1) — the largest change. Three H effectiveness centrals were re-anchored on the 2026 system as the counterfactual and fell hard: aerial suppression 0.25 -> 0.06 (lower 0.08 -> 0.00), landscape fuel management 0.20 -> 0.08 (lower 0.05 -> 0.00), programmes governance increment 0.10 -> 0.04 (lower 0.02 -> 0.00). The composer additionally volunteered detection/forecasting 0.15 -> 0.12, which the critic had not flagged, because EFFIS-grade forecasting and the national detection chains are equally inside L_0. The sub-baseline counterfactual \"materially thinner aerial and surge capacity\" is withdrawn. Turco et al. 2016 is retained in effectiveness.sources in three categories but only as a CEILING ON THE WHOLE HISTORICAL MITIGATION STOCK, never as the anchor for a forward increment. Each affected definition now cites phase_00_input.yaml#fit_check.rationale and carries a PHASE V INSTRUCTION that adoption anchors be set on the increment (y0 near zero) so the baseline is not re-embedded. The same netting is now stated explicitly for the Copernicus imagery limb (J4). (2) TIER HONESTY (M3, M4, M5, M6, M7, M11, M12) — the declared-deficit pattern is now applied consistently instead of to one row only. cat_h_detection_forecasting_decision_ support retiered `confirmed` -> `emerging` on the §4.2 gate (EUR 55,669,561, every component a §6-admissible disclosed amount), which is a better landing than defending a shaky academic axis. cat_h_fire_management_programmes_and_statutes was offered the same `emerging` route on AEGIS's EUR 2.1 bn and the composer REFUSED IT, applying the declared-deficit pattern instead: AEGIS is an all-hazard programme (inventory §1), §6 contemplates pro-rating by programme-area share, the Axis-2 line item has not been retrieved, and clearing a rigor gate with an un-pro-rated sovereign figure would be exactly the inflation this iteration is correcting elsewhere. That refusal is the one place the composer declined a critic-offered option, and it declines it in the conservative direction. Declared deficits now stand on five categories: cat_h_aerial_suppression_surge_capacity (1 axis), cat_h_fire_management_programmes_ and_statutes (1 axis), cat_v_noncombustible_envelope_fabric (2 axes), cat_v_retrofit_ember_path_protection (2 axes), cat_r_physical_restoration_capacity (2 axes, carried from iteration 1), plus cat_r_public_recovery_substitution (2 axes, and now neutral on scope). Axes withdrawn: two academic citations from detection; academic + industry from programmes; industry + academic from aerial; regulator from retrofit (an ABSENCE source — P4); industry from non-combustible fabric (a NEGATIVE CONTROL — P4); industry from public recovery (out of scope — M11). Every withdrawn source is retained in prose or in effectiveness.sources as what it actually is. (3) SCOPE (M10) — cat_r_public_recovery_substitution's effectiveness is set to 0.00 (I3 = 1.0, neutral) and the row is re-cast as an exposure-model / L_0-bridge input. An insured-share movement is a protection-gap movement, which phase_00_input.yaml confines to the L_0 bridge; and scoring it as mitigation while this same lattice refuses Fidelidade's +60% adequate-coverage uplift on identical grounds was a one-sided treatment of one quantity that biased the forecast downward. Both directions are now excluded from the mitigation term. (4) EVIDENCE ATTRIBUTION (M2) — the I015/I043 disjointness control is materially weakened and re-stated rather than defended. The cluster_rationale now says plainly that I015's own three records evidence only the building band, that the 100 m settlement band survives only as a second-hand bundle claim on a V-channel source page, that the split is defensible as a reading of DL 82/2021 but not as a reading of I015's record, and that Phase X must treat the risk that I015 and I043 describe the SAME act as live. Critic option (a) — a Phase IV correction to I015's summary, tags and sources — is ESCALATED TO THE LEADS as work Phase IV.5 has no authority to do. Per S1 the 0.55/0.45 partition is relabelled: it is not a derived geometry, it is a deliberate conservative device with a stated downward bias, adopted precisely because the limbs may overlap. (5) EXECUTABILITY (M8, M9, S7) — the detection merge's stated arithmetic was wrong and is corrected: the forecasting sub-cluster (I026 + I027 = EUR 50,069,561) DOES clear the gate by EUR 69,561 and could stand alone; only the detection sub-cluster is orphan-trapped. The I006 MAK 5 control is restated at CATEGORY level because v0.7 has no per-venture mitigation cells, and the same control is added for I002's KatRisk R-channel limb, which was unflagged. (6) SOURCING HYGIENE (S2, S3, S4, S5, S6, S8, S9) — the IBHS anchor now carries its full provenance including that neither figure is in either snapshot and that ZestyAI co-authored the study and sells a wildfire risk score; three gloss `span` values are replaced with source key_claim text; the DL 82/2021 prescription and the ~40% non-compliance figure are split into two citations with the latter marked second-hand; the utility `emerging` gate now states all three ways it is thin; AEGIS carries its pro-rating caveat; a confidence row is added to every dossier §5 and a confidence column to every §6; and every EUR 0 investment block now names what was checked. (7) JUDGEMENT CALLS — all five accepted with their asks. J1: the retrofit row now carries a binding Phase V instruction to anchor on the vent limb and hold the applied-coating limb at an evidence gap, and concedes that the shared property is a deployment property, not a mechanism. J2: arson deterrence is named as a separate, separately-unquantified limb and routed to Phase X as `evidence_gap: true` rather than as part of the 0.04. J3 and J4: both rows now state that a single adoption S-curve is wrong for them and why. HONEST AXIS COUNT PER CATEGORY after this pass, for the leads' one-view scan (M12): detection/forecasting 2 (emerging, gate cleared); aerial suppression 1 (deficit); programmes 1 (deficit); landscape fuel management 3 (confirmed); utility corridor 2 (emerging, gate cleared but thin); defensible space 3 (confirmed); non-combustible fabric 2 (deficit); retrofit ember path 2 (deficit); assessment and targeting 3 (confirmed); post-event damage establishment 3 (confirmed); physical restoration 2 (deficit); public recovery substitution 2 (deficit, and neutral on scope). Four categories stand on three genuine on-mechanism axes; two are `emerging` on the investment gate; six carry declared deficits. No category is silently thin. UNCHANGED, per the critic's own do-not-churn list: all 64 ventures assigned exactly once, no cross-channel members, every source id resolving, investment arithmetic, the IBHS quarantine, the refused-evidence discipline, the 1 - (E x AS) identity, and the displacement-by-public-good finding in cat_r_post_event_damage_establishment.\n"},{"iteration":2,"date":"2026-08-01","author":"category-clustering composer (out-of-band)","note":"METHODOLOGY-LEVEL DEFECTS FOR SIGN_OFF_CHECKLIST.md — recorded here for the coordinator to lift out, NOT fixed in-pilot (frozen-pilot rule; this pilot runs at its pinned hash 77b20a9f). P1: category_derivation.md §4.3 vs tools/validate.py check 3 — the categories file's own considered_but_excluded is never consulted, so an `insufficient` category orphans its ventures and hard-fails validation; independently confirmed by the critic against the source. Fix options, best first: extend check 3 to accept an `absorbed_venture_ids` list in this file; add a renderable third tier state; or allow a documented Phase IV.5 write-back into the inventory. P2: phase_04_5_categories.json's tier enum has two states while §2 and §4 define three — the proximate cause of the declared-deficit workaround, now used on six rows. P3: the `emerging` gate is investment-only, and 57 of 64 ventures in this pilot disclose nothing; the inventory predicted this at §2.2 and the category pass walked into it. In a domain delivered by statutes, state programmes and free public goods, \"deployed capital validates market belief\" is not a usable rigor signal — a non-monetary route (hectares under contract, km of line, activations, settlements enrolled) would fit this evidence base and this pilot has all of them. P4: §4.1 gives no rule on whether an ABSENCE source counts as an axis hit; this pass counted three such and has now withdrawn them all. Proposed rule: an axis is hit only by an affirmative claim about the mechanism; absences are findings recorded in tier_rationale and never in axes_hit. P5: venture_discovery.md's 8 axes vs sources.md's fixed 4 remain unreconciled and propagate into every axis judgement here. P6: v0.7 addressable_share is an unbracketed, unsourced, undefined-basis scalar with no statement of whether category shares may overlap — which is why the I015/I043 partition is unadjudicable and why the 0.05 on non-combustible fabric has nowhere to record its uncertainty; matrix_schema.md §v0.8 already fixes all four. P7: matrix_schema.md OQ-10 (AS_c carries no impact-area index, questionable for I1/I2) is open and every dossier's neutral I1/I2 is a workaround for it, not an answer.\n"}],"orphans":[],"thresholds":{"emerging_investment_floor":50000000,"confirmed_min_axes":3,"confirmed_min_sources":3,"shallow_catalog_min_deep_share":0.3,"discovery_min_axes_hit":4},"gates":[]}