RESEARCH DOSSIER
Property Damage Causation and Expert Evidence
Abstract. Physical damage causation requires the reconstruction of a mechanism connecting an event, asset condition, and observed damage. Expert evidence is the documented and testable basis for that reconstruction. Insurance coverage adds a further legal question about the precise contract, applicable law, proof, exclusions, conditions, and indemnity. This international research monograph examines these connected inquiries by comparing primary forensic and engineering standards, official hazard data, U.S. and English expert-evidence rules, selected insurance causation statutes and judgments, and international claims supervision. It develops a source-grounded synthesis for fire, flood, wind, hail, building-envelope, structural, and equipment failures; examines preservation and destructive testing, method validation, uncertainty, competing hypotheses, and institutional independence; and analyzes disputed coverage and causal scope. An explicitly hypothetical case and calculation illustrate the difference between diagnostic evidence and temporal association. The work contributes an original comparative interpretation and critical assessment of known methods and doctrines, without claiming a new authorial theory or empirical effect estimate.
Keywords: Property damage; property insurance; forensic engineering; expert evidence; physical causation; fire origin and cause; water intrusion; flood; hail; wind; concurrent causes; proximate cause; chain of custody; NFPA 921; ASTM; ISO 21043; expert reliability; claims governance; forensic uncertainty; policy exclusions.
Introduction
A damaged insured building can exhibit roof displacement, damp insulation, stained ceiling panels, corroded electrical connections, and structural cracking during a single inspection. The existence of these conditions does not establish when they began, which event or mechanism produced them, whether one incident accounts for all manifestations, or whether the insurance contract responds. The analytical problem has several stages. An event must be located in time and space; an exposure mechanism must be reconstructed; a connection to the specific asset and damage must be supported; and the resulting facts must then be assessed under policy wording and governing law. Confusing those stages converts proximity into proof, regional meteorological records into building-specific impact findings, or an engineering observation into an unwarranted legal coverage conclusion.
The scientific object is technical determination of physical-damage cause in property-insurance claims. The unit of analysis is an asset or component with an identified baseline and observation period, situated within an incident, claim, contract, and jurisdiction. Damage to a roof panel, corrosion of a pipe, and heat alteration of structural steel call for different measurements and competing explanations. Economic loss, physical damage, covered insured loss, legal liability, reserve, and actual settlement must be separately identified. Such distinctions are essential for comparability as well as fairness.
This study examines primary technical standard records, statutes, judgments, official weather and flood documentation, national and international insurance supervisory materials, and the scientific foundations of inferential testing. Its method is critical documentary synthesis, comparative legal analysis, physical-mechanism reconstruction, counterfactual evaluation, and transparent hypothetical calculation. Official descriptions of commercial standards establish edition and stated scope; complete paid normative clauses have not been independently inspected, and no claim of clause-by-clause certification follows from a publisher description. Legal propositions are attributed to the exact forum and instrument. No claim-specific testing or unpublished insurer data is represented as observed.
The guiding scientific question is how a technically reliable causation opinion is obtained, how its strength and limits should be communicated, and how it can be used without distorting legal coverage decisions. The inquiry considers corroborating evidence and alternatives throughout. A scientifically legitimate finding may be “cause undetermined” where the investigation cannot distinguish material hypotheses. That finding is more informative than a categorical attribution unsupported by observations.
Definitions, Historical Context, and Primary Source Foundations
Physical, causal, and insured losses
Physical damage is an observable or measurable adverse change in the material state, mechanical integrity, function, or performance of property. Typical traces include fracture, deformation, loss of watertightness, thermal degradation, corrosion, moisture-induced deterioration, contamination, and component malfunction. The baseline matters: a cracked material photographed after a storm may have cracked earlier; corrosion can precede an electrical incident; a flood may aggravate an already vulnerable finish. An observed condition is evidence of its presence when observed, while determining its cause and timing requires separate inference.
A causal hypothesis specifies an event or condition, an operative mechanism, and a chain connecting that mechanism to observed damage. The hypothesis has greater explanatory force when independent observations discriminate it from plausible alternatives. Merely naming a peril gives a classification, not the physical explanation. “Storm damage” may encompass impact, pressure differential, wind uplift, driven rain, flying debris, or combinations, each producing different predicted traces. Likewise, a broken mechanical component may be the endpoint of overload, fatigue, progressive corrosion, manufacturing defect, installation error, maintenance failure, or event-induced shock.
The insurance word “loss” often refers to distinct quantities: damage to a component, cost of repair, replacement expense, reduction in asset value, business interruption, indemnifiable contractual loss, or cash already paid. Measurements of each require explicit denominators. A repair proposal may represent prudent construction work without establishing that every item was caused by the alleged insured event. A policy may require particular damage, causation, location, period, and terms before an indemnity is payable.
Origins of formal causation and expert scrutiny
The history of insurance causation combines commercial doctrines of insured perils with engineering and forensic investigation. In United Kingdom marine insurance, section 55 of the Marine Insurance Act 1906 expressly addresses losses proximately caused by insured perils, subject to the statute and policy. Its importance is as a legally defined setting for proximate cause; it is not a universal code for modern land-based property contracts. Later judgments such as the UK Supreme Court's Cendor MOPU decision (2011) examine how marine insured perils and inherent characteristics of property interact in a concrete dispute.
Expert reliability became a developed evidentiary question in the United States with Daubert v. Merrell Dow (1993), which analyzed reliability and relevance of scientific evidence under Rule 702. Kumho Tire v. Carmichael (1999) recognized that the judicial reliability inquiry extends to engineering and specialized technical knowledge and should be adapted to the particular question. In England and Wales, Civil Procedure Rule 35 gives experts an overriding duty to assist the court within their expertise. These systems use different procedures while sharing concern for justified expert opinions.
Contemporary forensic standards make the sequence of evidence handling more explicit. ISO 21043-1:2025 establishes forensic vocabulary. The published ISO 21043-2:2018 covers recognition, recording, collection, transport, and storage of items, with a draft replacement under development as of November 2026. ISO 21043-3:2025 concerns analysis, ISO 21043-4:2025 concerns interpretation against propositions, and ISO 21043-5:2025 concerns reporting. The published scope descriptions directly support these distinctions, although binding application depends on the assignment, national rules, and organizational adoption.
Sources and evidence hierarchy
An expert report may rely on a direct measured observation, an independent contemporaneous record, a forensic experiment, a calibrated laboratory result, witness testimony, an authoritative hazard dataset, a technical standard, or general academic knowledge. Each source supports different propositions. A reliable weather station can establish conditions at its station, while a roof-specific damage opinion still needs site-level exposure and material evidence. A standard can establish a method's general recognized purpose; it cannot establish that a specific investigation properly followed that method without inspection of the investigation record. A statute establishes law within its scope, not a physical mechanism.
The publisher descriptions for ASTM E1188-23, ASTM E860-22, and ASTM E620-18 establish expectations respectively concerning evidence collection and preservation, examination with potential alteration of items, and reporting of technical opinions. These texts are useful at multiple investigation stages. The source audit should record origin, collection method, event date, measurement resolution, author, edition, access status, and relevant limitations. Multiple online copies of one original witness account do not amount to multiple independent observations.
Comparative research must also preserve the difference between a published normative instrument and its local implementation. A voluntary standard does not become automatically binding by being internationally recognized. A U.S. model act does not become enacted law in all states. The credibility of a technical report is related to method suitability, evidence completeness, and sound interpretation rather than the number of recognizable logos attached to its bibliography.
Engineering Cause Investigation
Frame the causal question
A useful investigation begins with a testable question. “Did the wind event create a new opening at this roof junction?” defines mechanism, asset, and time more precisely than “Was this storm damage?” The investigator documents the asset, loss notification, alleged event time, inspection time, known pre-loss condition, and any subsequent emergency repairs. Questions should be sufficiently open to accommodate alternative causes. An assignment to “confirm hail damage” contains an evidentiary conclusion before the site has been examined.
The first record set includes property plans, as-built information, permits, product data, prior inspections, maintenance and repair logs, contractor photographs, incident reports, sensor traces, security footage, witness accounts, and site-specific environment data. The team separates evidence observed personally from evidence provided by others. It records inaccessible spaces, unsafe conditions, and items already replaced. A scene investigation photographs wide context before close-up features, records orientation and scale, and compares damaged components with similar unaffected materials. This yields information about spatial patterns and possible baseline conditions.
Preservation is itself a causal consideration: emergency tarping, extinguishment, pumping, demolition, sample extraction, or contractor removal may alter the physical traces to be interpreted. The report should state what changed, when, and by whom. A later expert cannot recreate the exact initial configuration simply from a photograph that excludes the critical fasteners or water path.
Fire and explosion
Fire investigation considers origin, ignition, fuel, growth, ventilation, spread, suppression, and post-fire disturbance. These are sequential but interacting processes. Burned material at one location is not necessarily evidence that ignition began there; later ventilation, flashover, collapse, firefighting, and changes in contents can determine where the greatest damage appears. Electrical arc marks may arise before ignition, during fire exposure, or during system failure, depending on circuit and temporal conditions.
The official NFPA 921 development record identifies the 2024 Guide for Fire and Explosion Investigations as current, with three tentative interim amendments issued in April 2025. The NIST OSAC standard record separately confirms the edition. A scientific account therefore identifies the applicable edition and amendments and reconstructs the relation of patterns to fire dynamics. The presence of an electrical device in an estimated origin region does not alone demonstrate electrical ignition; the analysis must identify source, energy, fuel, operating state, competing causes, and supporting or inconsistent traces.
The investigator should formulate viable ignition hypotheses, search for predicted observations, and attempt to disconfirm them. Where evidence has been destroyed, an undetermined cause can be the appropriate conclusion. Failure to demonstrate an accidental ignition mechanism does not itself prove intentional ignition. In insurance disputes, the expertise and evidence supporting a claim of deliberate fire need separate examination from a mechanical reconstruction of fire origin and spread.
Water entry, flood, and plumbing escape
Flooding, external rainwater penetration, sewer backup, groundwater, and internal plumbing leaks can create visually similar damage. A cause investigation should locate the source, route, entry point, elevation, time course, and the property materials affected. A nearby river gauge cannot by itself prove that floodwater crossed the subject threshold, nor can a wet ceiling alone distinguish a leaking pipe from a roof penetration. Ground datum, building penetrations, drainage pathways, rainfall chronology, plumbing pressure, sediment, and contamination patterns may discriminate causes.
The U.S. Geological Survey's high-water mark field manual explains locating and preserving marks and attaching uncertainty to measurements. Some marks are transient, distorted by local flow, or altered after the event. A competent investigator records the physical datum, quality of the mark, surrounding context, and sources of measurement uncertainty. Traces of water position do not automatically reveal where water originated.
Moisture meters and thermal cameras also require careful interpretation. A meter reading is instrument- and material-dependent; a thermal anomaly represents temperature difference rather than an automatic diagnosis of water identity and source. Water can travel within cavities before emerging far from the entry. Material staining may accumulate across multiple events. The discovery date should never be substituted uncritically for the beginning of an intrusion. A cause opinion can identify a confirmed water path while properly qualifying the precise initiating event.
Wind, hail, and the building envelope
A wind or hail claim raises separate questions of regional event, local exposure, physical impact, damage mechanism, and timing. The National Weather Service describes preliminary local storm reports as information received from multiple reporters that is later integrated into storm records. NOAA's Storm Events data disclaimer cautions that some information can be externally reported or unverified. The NOAA Severe Weather Data Inventory distinguishes automatically inferred storm signatures from confirmed conditions and warns that data gaps do not establish absence of severe weather. These limits matter particularly for one building or one roof slope.
Wind assessment may require uplift load paths, fastener spacing and pullout, substrate integrity, pressure differences, roof geometry, condition of parapets, directional evidence, and comparative examination of adjacent areas. A distant station's peak gust is not necessarily the load experienced by a roof edge. Hail analysis requires distinguishing new impact morphology from weathering, manufacturing features, foot traffic, preexisting deterioration, biological growth, and maintenance artifacts. A mark that is compatible with hail does not, without more, establish impact date or hail size at that roof.
A report should compare exposed and sheltered surfaces, document the pattern of allegedly fresh damage, identify material vulnerabilities, and seek independent event and site evidence. The inference must be limited by inspection access and validated material-specific knowledge. Regional catastrophe models and radar indications assist reconstruction; they cannot alone substitute for the individual property examination.
Structural and mechanical failures
Structural cracks, deformed connections, rusted fasteners, broken shafts, seized bearings, tripped electrical protection, and failed seals are endpoints requiring mechanism-specific analysis. A broken shaft may follow overload, cyclic fatigue, misalignment, lubrication failure, manufacturing flaw, thermal expansion, or a distinct external incident. The nearest visible break is not necessarily the initiating cause. Evidence may include fracture surfaces, metallurgical analysis, lubricant samples, thermal records, vibration trends, electrical fault logs, load calculations, prior repairs, and original design assumptions.
The ASTM E2018-24 property condition guide describes a voluntary baseline property condition assessment process. Such a condition assessment can help establish an asset baseline, but it answers a different question from the cause of a specific failure. Original construction drawings describe design intention; an expert may need to compare them with the installed condition and actual service history.
An old defect can increase susceptibility to acute event loading without being sufficient to generate the specific damage at the specific time. Conversely, an observed storm may be temporally adjacent to a failure that would have occurred under normal operating conditions. The scientific task is to identify necessary and contributing conditions and, where possible, test whether the observed signature fits one mechanism better than competing accounts.
Timeline and counterfactual reasoning
A causal timeline distinguishes the baseline, initiating load, propagation, first detection, emergency intervention, repair, and later evidence collection. The question “Would this damage have occurred at this time without the reported wind?” is distinct from “Would the roof eventually have needed replacement without the wind?” The latter concerns long-run condition and economics; the former is event causation. Counterfactual claims are supported by engineering models and evidence, not simply by rearranging dates.
An effective report gives each hypothesis a mechanism, expected observations, actual observations, test results, contrary facts, and a residual uncertainty statement. The investigation should not force one exclusive explanation if independent physical mechanisms jointly contributed. The eventual legal treatment of concurrent causes is analyzed later and should not distort the expert's technical description.
Published forensic engineering on competing forces
The applied forensic engineering literature supplies a direct property-loss example. In Hinckley's wind-versus-water study (2006), hurricane damage attribution is examined through storm dynamics, applied forces, observed damage patterns, and debris geometry, drawing on post-Katrina Louisiana experience. This is a conference study based on particular catastrophe conditions; its contribution is a set of plausible ways to distinguish or recognize simultaneous wind and water action, rather than a universally validated method for all storms and structures. The finding supports examining mechanisms together and preserving evidence that separates them. It does not establish that an investigator can derive one cause from a single image.
The source also exposes an institutional asymmetry: evidence can be lost during urgent reconstruction, while the insurance dispute may require a degree of causal separation beyond what surviving traces support. An appropriate engineering opinion can identify the probable sequence and describe uncertainty without turning incomplete evidence into a falsely precise wind-to-water allocation.
Forensic and Expert Standards
Collection and provenance
Evidence handling connects the observed site to the material examined and the conclusion reported. Photographs require location, orientation, scale, capture time, preservation of originals, and editing history where relevant. Samples require unique identifiers, original position, collector, packaging, custody changes, storage conditions, and recorded transfer to laboratory. A laboratory result about an untraceable specimen may be analytically correct but uninformative about the disputed property.
The ASTM evidence collection practice expressly emphasizes identification and traceability to the incident. The published ISO 21043-2 process scope encompasses collection, transportation, and storage. An auditable file documents observations before interventions, changes in custody, and unavoidable evidence gaps. It should also separate a firsthand observation from a witness's account and an interpretive proposition. Reprinting a third-party photograph within a new expert report does not convert its original authorship into direct observation.
Examination and destructive testing
Testing may alter or destroy material that another expert would need to inspect. ASTM's E860-22 examination practice addresses procedures and safeguards when examination, disassembly, or testing can change the condition of evidence. A technically justified sequence begins with documentation and appropriate nondestructive examination, identifies what destructive work would discriminate, and records the planned method, material consumed, retained specimens, and test results. Where preservation obligations arise in prospective litigation, counsel should separately assess the forum's procedural requirements.
A piece of flashing detached without its original fasteners may prevent the investigation from reconstructing load direction. A switch dismantled before photographing contact alignment may eliminate an important distinction between original malfunction and thermal alteration. A flood sample introduced into unsuitable packaging can lose its diagnostic chemical value. These consequences explain why original condition, intervention, and post-test status must be separately documented.
Competence and laboratory quality
Property damage investigation is interdisciplinary. A roof-envelope engineer, fire investigator, metallurgist, electrical engineer, chemist, meteorologist, and quantity surveyor may have different relevant expertise. The report identifies who actually performed each measurement or test and whether the interpreter is qualified to interpret it. An expert's high standing in one discipline does not automatically extend to another. Instrument calibration, proficiency, laboratory method validation, and technical oversight address different sources of reliability.
The revised ISO/IEC 17020:2026, published in March with a June correction, addresses competence, impartiality, and consistent operation of inspection bodies. ISO/IEC 17025:2017, confirmed by ISO as current, sets competence requirements for testing and calibration laboratories. Their scopes differ, as do the benefits of compliance and accreditation. Neither furnishes a universal guarantee that any particular causal inference is correct. Accreditation is always tied to the accredited activities and authorized scope.
The ISO forensic standards further distinguish analysis requirements, interpretation against alternatives, and reporting of forensic work. A laboratory may correctly identify a residue without identifying the event that deposited it. A structural calculation may be mathematically correct without verifying historical loading. Technical opinion requires a valid inferential connection between observed results and the question asked.
Expert reports, transparency, and independence
A scientific expert report should contain the precise question, engagement scope, material facts and instructions, relied-on records, qualifications, methods, measurements, alternatives, limitations, and conclusions. ASTM E620-18 describes the scope of formal technical-opinion reporting. The English Practice Direction 35 requires specified qualifications, relied-on material, material instructions, who conducted tests, competing opinion ranges, qualifications, and a statement of truth in covered litigation. It is a jurisdiction-specific litigation rule, but its enumerated disclosures are an instructive scientific quality check.
Expert independence is expressed through the reasoning rather than only a label in a résumé. Instructions can be framed neutrally, evidence contradicting the preferred conclusion preserved, alternatives tested, and material changes in opinion communicated. A claim described in the referral as suspicious may induce anchoring, even though a fraud allegation has no mechanical significance for a particular fracture. A well-designed workflow withholds irrelevant allegations during initial technical inspection and introduces genuinely relevant history at the appropriate stage.
The Forensic Science Regulator's 2025 statutory code applies in England and Wales to specified forensic science activities associated with criminal investigation and proceedings, effective October 2, 2025. Its quality governance is a useful comparative reference without making it automatically applicable to civil insurance adjustments. Method validity, documentation, and impartial interpretation retain their own scientific value regardless of mandatory adoption.
Expert admissibility and reliability
The amended U.S. federal Rule 702 requires proponents to demonstrate by a more-likely-than-not threshold that expert testimony is relevant, founded on sufficient facts or data, and based on reliable principles and their reliable application to the case. The 2023 amendment emphasizes the court's duty at the admissibility stage. The Daubert and Kumho decisions explain why testing, error rates, and methodological fit are significant without establishing a rigid checklist for every engineering investigation. Other jurisdictions and U.S. state tribunals can apply different rules.
An expert may competently measure corrosion and yet lack evidence to date it. An accurate photograph may show an opening without establishing the initiating load. A fire investigator may properly identify a likely origin but remain unable to establish ignition. A report that states these limits is more reliable than one that presents inference as observation. Where several expert opinions conflict, the technical question is which evidentiary premises or methods generate the difference and what further discriminating test, if any, can resolve it. The single joint expert provision gives one procedural approach in England and Wales but does not substitute agreement for empirical validity.
The NIST evidential statistics program emphasizes interpretation of uncertainty and statistical strength. An instrument's measurement uncertainty, a model's parameter uncertainty, a lost baseline, and an expert's subjective selection of a mechanism are different uncertainty sources. A unitary “95 percent certain” statement without a defined statistical model, calibrated data, and clear hypothesis set can mislead. Qualified reporting should identify which observations remain ambiguous and why.
Scientific literature on reasoning reliability
Peer-reviewed studies of forensic decision-making provide relevant evidence about interpretation as distinct from test execution. Cooper and Meterko's systematic review of cognitive bias (2019) reviewed 29 studies across 14 forensic disciplines and documented effects of confirmation bias, with recommended safeguards including reduced exposure to irrelevant contextual information and blinded review. Those disciplines and study designs are heterogeneous; the review does not quantify error rates in property insurance investigations. Its relevance is methodological: expert objectivity must be supported by workflow and evidence, rather than asserted from professional title alone.
Georgiou, Morgan, and French's interdisciplinary uncertainty review (2020) develops ways to conceptualize, evaluate, and communicate uncertainty in forensic science. It helps explain why raw measurement uncertainty, inference among mechanisms, and communication to legal decision-makers are different problems. The review's toolkits do not make every empirical uncertainty measure available for a damaged building, and the present dossier applies its distinctions analytically.
Taroni and colleagues' Bayesian networks study (2004) discusses constructing probabilistic networks for forensic evaluation of evidence. Its value for property causation is the insistence on explicit conditional relationships, alternative propositions, and the risks of inadequately constructed probabilistic structures. It provides a methodological precedent for the hypothetical calculation below, not estimated probabilities for any particular roof, fire, or flood. These studies supplement normative standards with published research about cognitive interpretation, uncertainty, and probabilistic structure.
Proof and Disputed Coverage
Coverage is a separate legal determination
An engineering opinion about the cause of damage answers a question about the physical world. A coverage opinion answers a question about the application of legally operative words to established facts. The second question ordinarily requires the insuring agreement, policy period, insured locations and property, declarations, endorsements, definitions, exclusions, exceptions to exclusions, conditions, valuation clauses, deductibles, and applicable law. A documented wind-caused tear can support technical event attribution without resolving a roof limitation or a wind/hail deductible. A documented water pathway can establish how material became wet without resolving whether water entered through an excluded flood mechanism or an insured plumbing event.
The distinction is especially important for policies employing terms such as “direct physical loss or damage,” “sudden and accidental,” “named peril,” “ensuing loss,” “wear and tear,” “faulty workmanship,” or “concurrent causation.” Their interpretation varies with actual wording and jurisdiction. An expert retained for engineering cause investigation should present findings in physically meaningful language, preserving the difference between established observations, reasoned causal opinions, and questions of contract interpretation. A report that says “covered storm damage” without explaining the insuring agreement silently assumes legal work beyond mechanical inspection.
A practical file should preserve discrete claims: the reported peril occurred somewhere in the region; the subject property experienced particular conditions; those conditions caused a specific material mechanism; that mechanism caused specific damage; the damage occurred during the policy period; the loss falls within the operative grant; exclusions, exceptions, and limits have particular consequences. These are linked propositions, and a failure in one does not automatically disprove all the preceding ones.
The NAIC's homeowners claim settlement guidance explains the U.S. adjustment context, including damage assessment, deductibles, and potentially staged payments. That guidance describes a consumer-facing process; it does not determine the applicable coverage of an individual claim. Technical damage extent must also be distinguished from replacement-cost valuation, actual cash value, depreciation, remediation expenses, code-driven upgrades, and policy sublimits. Physical attribution is an input to settlement rather than a synonym for settlement.
Multiple causes in one physical loss
The causal structure of property damage is often conjunctive. Heavy rain can enter because wind recently removed flashing. Rain can also enter through flashing that had been defective for years. A fire may be ignited by an electrical fault made more likely by corrosion, then spread farther because of compartment construction. A wall can fail under a high load that alone might not have destroyed a correctly built wall. Such accounts require physical inquiry into both trigger and vulnerability, including their timing and necessity.
The historical Marine Insurance Act 1906, section 55 addresses proximate causation for marine insurance under its defined terms, while the Cendor MOPU judgment (2011) analyzes perils of the seas and inherent vice in the context of the particular marine claim. The doctrinal use of “proximate” cannot simply be reduced to the last physical event in a chronological chain. Nor should the marine result be automatically imputed to every modern homeowners or commercial property policy. Contract and jurisdiction control.
Anti-concurrent-causation wording can attempt to specify how an excluded cause operates when another cause contributes. The particular effectiveness and interpretation of such wording are matters for applicable law. Causal investigation should therefore retain all physically supportable contributors rather than forcing one peril to make a claims decision easier. The report can distinguish an initiating event, a necessary preexisting condition, an aggravating condition, and a later intervention. Contractual consequences can then be assessed through the legally relevant concepts without erasing the scientific record.
In Australia, section 54 of the Insurance Contracts Act 1984 addresses defined situations involving acts and omissions after a contract is entered into, including causal contribution and prejudice. The relevant current legislative compilation is dated October 8, 2026. It is a specific statutory rule requiring attention to its own conditions; it is not a universal physical theory or a rule that every excluded peril becomes insured. An international causation dossier should preserve those distinctions when comparing jurisdictions.
Evidentiary burdens and the meaning of probability
Legal burdens concern which party must establish a proposition under a governing standard of proof. Expert uncertainty concerns the quality, completeness, relevance, and discriminatory force of technical evidence. The existence of a legal burden cannot be used as a laboratory calibration or as evidence that a particular mechanism did or did not occur. In some jurisdictions, contract claims distinguish the claimant's proof of a loss within an insuring grant from an insurer's proof of a specific exclusion; other procedural and substantive regimes allocate questions differently. The correct allocation requires the actual law and policy rather than a generic insurer practice manual.
Suppose the roof has documented new tears after a storm, while the insurer has documented preexisting weathering. These facts can coexist. Whether wear merely heightened vulnerability, was itself sufficient to produce the tears, or produced a separable older component of damage is an engineering question. Whether the wear-and-tear exclusion applies to resulting damage is contractual and jurisdictional. A technical report should identify the observed tearing, material condition, test results, dates, and viable mechanisms. Its language should be usable by both parties even if their interpretation of a contractual exclusion differs.
Under U.S. federal Rule 702's admissibility requirements, a party offering an expert must satisfy a judge that the specified reliability and relevance conditions are met more likely than not. This standard concerns admissibility in the relevant litigation, not a numerical percentage that an engineer must attach to the physical event. A causal expert can explain the relative support for competing hypotheses and the absence of decisive data without pretending that a claims coverage determination has a universal numerical cutoff.
The NAIC model laws catalogue includes instruments addressing unfair claims settlement practices. These are model laws to be considered against enacted state versions and specific market conduct requirements. The IAIS Insurance Core Principles, particularly ICP 19, offer international supervisory principles regarding conduct, including claims and fair treatment. The IAIS 2025 consumer-treatment paper addresses the lifecycle from product design through claims and complaints. Such norms encourage disciplined reasoning, communication, and fair process, but do not create one global rule of proof.
Damaged area, necessary repair, and indemnifiable scope
Technical cause and technical repair scope are interdependent yet distinguishable. If a wind event tore a small membrane section, localized restoration may be physically possible; alternatively, assembly constraints, waterproofing continuity, safe access, or manufacturing requirements may make larger repair technically necessary. Neither conclusion follows solely from the area visibly torn. An expert can inspect repairability, material compatibility, fastening systems, weatherproofing details, and the reasonable condition of adjacent components. Claims valuation then applies insured-value definitions, depreciation, limits, and contract-specific treatment.
Cost documentation requires a common scope. A contractor's quote may encompass emergency temporary work, tear-off and disposal, repairs to damaged substrate, replacement of unaffected but technically integrated material, code-related alterations, and remediation of preexisting deficiencies. Each item requires explanation. The expert should identify which tasks repair physically attributed damage, which are necessary incidentals, which address unrelated baseline condition, and which require legal consideration. Duplicate line items and overlapping subcontractor allocations should be resolved before totals are interpreted.
A claim may involve joint physical contributors. For example, a newly damaged roof opening can lead to wet insulation, while longstanding moisture has already weakened adjacent sheathing. Where the condition of the sheathing before the event cannot be independently established, allocating restoration expense by an asserted precise historical percentage may exceed the evidence. A qualified range, scenario analysis, or nonallocation finding can be more scientifically defensible. The insurer's obligation to pay an amount remains dependent on contract and law.
Legal systems and expert processes
The principal comparative sources demonstrate the importance of jurisdictional labeling:
| Domain | Source and specific question | Transfer limit |
|---|---|---|
| U.S. federal evidence | Rule 702, Daubert, Kumho: admission of specialized testimony and method fit | Does not automatically determine state insurance law or state-court rules |
| England and Wales civil proceedings | CPR Part 35 and Practice Direction 35: expert duty, disclosure, procedural control | Applies to covered proceedings, not every private claim interview |
| United Kingdom marine insurance | Marine Insurance Act section 55 and Cendor MOPU: proximate cause and marine insured perils | Marine facts and wording cannot silently become universal property rules |
| Australian insurance contract law | Insurance Contracts Act section 54: specified post-contract acts and causal contribution | Requires statute-specific conditions and operative compilation |
| International supervision | IAIS ICP and fair-treatment paper: fair claims handling and conduct standards | National implementation and actual domestic rules remain relevant |
| U.S. market conduct | NAIC model acts and consumer guidance: models and claims education | Requires checking state enactment and contract terms |
The comparative outcome is not a single legal rule. It is an evidence requirement: an international claim analysis should retain jurisdiction, forum, law and policy edition, incident date, mechanism, and procedural stage as separate data. Causal categories stripped of these attributes invite erroneous cross-market comparison.
Empirical Verification, Uncertainty, and Competing Explanations
The hypothesis-to-observation relationship
Investigation is most informative when it specifies what observable traces would differ among competing hypotheses. A wet ceiling is predicted by multiple sources of water, so its presence has limited power to distinguish roof entry from plumbing failure. A newly deformed metal cap, a reliably dated pre-event image showing it intact, and a matching directional load signature may be substantially more discriminating. Evidence is valuable because of the contrast it draws between hypotheses, not merely because it resembles the favored account.
For a fire, a proposed electrical ignition mechanism predicts an energized component in an ignition-capable state, an available combustible, and a plausible origin and spread sequence; each must be confronted with possible post-ignition damage to the electrical system. For flood, an inundation account predicts a feasible pathway and elevation, compatible marks or sediments, and an event chronology matching the site's hydrology. For hail, fresh impact signatures must be compared with site-specific exposure, material characteristics, and alternative causes of surface marks. The evidentiary question differs in every case, even when all are recorded as property claims.
The official NOAA Storm Events disclaimer and Severe Weather Data Inventory limitations illustrate why absence of a report or radar signature cannot be interpreted as guaranteed absence of a hazard at one site. Sources have measurement resolution and reporting processes. The USGS high-water mark manual similarly treats the quality and uncertainty of individual marks as scientifically consequential. Reliable use depends on respecting what the originating data actually measure.
An explicitly hypothetical property investigation
Consider a two-story commercial building with upper-floor water staining first photographed after a reported convective storm. The claimant alleges that wind displaced a parapet cap and permitted rain to enter. A prior maintenance note indicates aging flashing near the same parapet. The owner has also reported intermittent plumbing leaks in an upper service room. The dates, observations, numbers, and outcomes in this case are fictional teaching assumptions; no real insured building, claimant, insurer, laboratory, or measured dataset is represented.
Three physical hypotheses are initially viable. Under the wind-opening hypothesis, the event mechanically displaced or tore the cap or flashing, creating a new route from rainfall to insulation. Under the preexisting-envelope hypothesis, an older flashing opening admitted rain during the storm without any new wind damage. Under the plumbing hypothesis, an internal release wetted the same ceiling independently of the storm. Each can produce water staining. Only some traces have discriminating value: consistent fresh fastener pull-out, reliable before-and-after photographs, water routing from the outside, plumbing pressure and flow records, rain chronology, sample distribution, or a controlled leak-location examination performed without destroying the scene.
Assume the first available inspection finds displaced cap metal, corrosion at fasteners, and saturated insulation, but the original cap attachment was removed during emergency work. The cap's post-removal position cannot reveal its original load direction, and the specific fasteners no longer have an intact in situ relationship to the assembly. Corrosion evidences age-related vulnerability without showing whether wind created a new opening. Saturated insulation establishes wetting, not water origin. If no plumbing tests or pre-loss image audit were performed, the evidence remains compatible with several mechanisms. An affirmative, unqualified statement that wind caused all damage would exceed these observations.
The investigation record can identify tests that may change the conclusion: preservation and authentication of contractor photographs; retrieval of removed fasteners if retained; comparison with intact comparable assemblies; local weather and roof geometry analysis; water pathway mapping; and examination of available plumbing histories. The investigator should also state whether any test is now impossible because of completed repairs. The costs and safety of tests affect practical choices; they do not convert an unperformed test into support for a favored hypothesis.
Quantifying illustrative diagnostic evidence
Bayes' rule can show the conceptual importance of the likelihood of an observation under alternatives. Let denote the hypothetical wind-opening hypothesis. For teaching purposes alone, assume , , and , treating and its complement as exhaustive. Then:
Every value in this calculation is invented and illustrative; none is estimated from roof claims, meteorological measurements, or engineering tests. Its meaning is conditional on the proposed hypotheses, prior, and supposed evidentiary performance. The result is not a probability of coverage, not a legal decision threshold, and not a calibrated expert confidence in the fictional case. If an overlooked plumbing cause matters, the two-hypothesis representation may be inadequate. If the observation is equally probable under both accounts, posterior odds stay at prior odds. The example illustrates why a distinct diagnostic observation can matter more than a large collection of compatible but nondiscriminating signs.
An observed repair quotation can also illustrate a different form of uncertainty. Suppose a fictional contractor proposes $120,000 for work to roof and interior components, and a separate condition survey values $20,000 of deferred maintenance. Simple subtraction to yield $100,000 of covered loss is invalid if the scope overlaps, the work required to restore damage includes some older elements, or policy terms treat code work differently. A technical allocation would require item-level estimates, repair methodology, baseline condition, and clear causal attribution; legal indemnity would further require the contract and governing law. No payment conclusion follows from those stylized numbers.
Negative evidence and incompletely observed scenes
Absence is diagnostic only when the proposed mechanism should leave a detectible trace and the investigation would likely have found it. Failure to see hail marks on a roof slope never inspected says almost nothing. The absence of arc patterns in a severely consumed circuit may reflect destruction rather than a demonstrably de-energized state. An intact fastener at one nearby location may inform construction variation but cannot recreate a lost fastener from the failure site.
The same principle applies to laboratory tests. A negative chemical result depends on sampling position, collection conditions, sensitivity, degradation, interference, and controls. A moisture reading depends on material, instrument, depth, and calibration. An expert reporting negative evidence should document expected trace, sampling coverage, detection capability, and why the observed absence bears on the cause.
Post-event missingness is often produced by claims operations. Emergency mitigation removes materials; a selected inspection photographs visually severe features but omits unaffected comparators; samples are taken after cleanup. Consequently, the analyzed record may be a nonrandom selection of the scene. This can induce apparent convergence in later expert reports because both use the same incomplete image set. A transparent file must document lost specimens and records, and the alternative explanations they might have distinguished.
Sources of uncertainty
Measurement uncertainty concerns the precision and accuracy of quantities such as moisture content, force, temperature, rain depth, water elevation, or material dimensions. Model uncertainty concerns structural boundary conditions, fire-development representations, event reconstruction, or damage functions. Historical uncertainty arises from poorly documented baseline condition, event timing, and prior repairs. Interpretive uncertainty arises when an observation is compatible with several mechanisms. Legal uncertainty arises from contract language, governing law, and factfinder determinations. These uncertainties need separate treatment rather than being aggregated into a poorly defined percentage.
An engineering model may fit observed failure precisely while relying on parameters selected after the failure was known. Post-event fit is weaker evidence for causal prediction than independent validation. A wind fragility result calibrated to a building class can be informative at the portfolio level without establishing what happened to a single roof. A moisture model can identify a plausible pathway but may be unreliable where actual cavity geometry is unknown. The researcher should disclose data availability, sensitivity to assumptions, and independent corroboration.
A crucial distinction also separates agreement among experts from correctness. Multiple experts using the same defective source image may agree and still be wrong. Divergent opinions can expose disputed assumptions worthy of testing. Independent review is useful when the reviewer is given the underlying record, can inspect competing hypotheses, and has a meaningful opportunity to reach a different finding. The ISO forensic interpretation standard emphasizes propositions and interpretation; the ISO reporting standard emphasizes transparent communication. Their normative scopes support rigorous distinction among analytical results and conclusions without automatically validating a case.
Bias, incentives, and report correction
Claims investigations may be commissioned by parties with opposed financial interests. Independence can be operationalized by recording all instructions, preserving original data, explicitly seeking contrary evidence, and explaining why a preferred cause withstands comparison. A referral identifying a claim as fraudulent before physical inspection may introduce irrelevant contextual bias. A referral insisting on “storm causation” can similarly predispose an expert against maintenance-related alternatives. The scientific remedy is a neutral causal question, a predeclared list of plausible explanations where practicable, and a transparent decision record.
When later evidence changes the opinion, the original report should remain identifiable, with a revised document describing what changed and why. A new test should not erase a previously recorded uncertainty; it should resolve or reclassify it with the additional evidence cited. A report must also distinguish what has been directly tested from what remains an inference from technical literature. This enables efficient scrutiny by another expert or court.
Empirical boundaries of this research
The present dossier is documentary and analytical. It has not obtained individual insurer claim microdata, physical specimens, proprietary adjuster notes, original fire scene images, laboratory datasets, or a representative sample of claim litigation. No causal estimate of the effect of a particular forensic protocol on claim outcomes is asserted. No frequency of erroneous expert attribution or proportion of excluded losses is estimated. The numerical illustration is an exact algebraic calculation under expressly hypothetical assumptions. Publisher descriptions of paid NFPA, ASTM, and ISO standards support verified edition and general scope, not a declaration that every technical clause was read or independently validated.
Integrative Comparative Scientific Synthesis
The scientific relation among events, mechanisms, and claims
The reviewed sources support an integrated account of how physical and institutional knowledge is produced. Engineering identifies mechanisms; forensic standards organize collection, testing, interpretation, and reporting; evidentiary rules scrutinize the relation between method and conclusion; insurance contracts determine legal consequences; supervisors specify conduct and fairness expectations. Their proper relation requires explicit transitions. Event evidence supports site exposure only to the extent that location and measurement permit. Exposure supports damage causation only to the extent that material traces and mechanisms connect them. Damage supports a contractual claim only where the correct coverage grant, conditions, exclusions, and legal rules are applied.
An event can be certain and an asset-specific mechanism uncertain. Physical damage can be certain and its onset uncertain. Two physical causes may be established while legal selection of a policy-relevant cause is disputed. A laboratory measurement can be reliable while the sample's link to the building is incomplete. These are independent states of evidence. Treating them as separable permits a review to determine exactly what additional work would affect a conclusion.
Method choice follows the disputed mechanism
Fire investigation involves origin analysis and fire dynamics; meteorological reconstruction requires spatial and temporal interpretation of observations; flood study may require high-water marks, source pathways, and elevation; wind and hail assessment requires material response and site exposure; structural failure requires load paths, materials, and connections; equipment failure requires mechanical and operating histories. A universal requirement to “conduct expert testing” is too indeterminate. The correct question is which test, with what performance and controls, could distinguish which causal accounts.
Cross-domain forensic safeguards still matter: original records, specimen traceability, operator competence, transparency of interpretation, and qualified reporting. ASTM's preservation and examination practices and the ISO forensic process offer a common procedural language even where a particular assignment has no mandatory adoption. The presence of a formally attractive chain-of-custody form cannot compensate for samples taken from the wrong component, nor can high laboratory precision compensate for an incomplete source attribution.
How premature causal closure distorts coverage
Premature causal closure means terminating alternative-mechanism testing before sufficiently discriminating evidence has been considered. It occurs when a claim is labeled “wear and tear” because corrosion is visible, or when every post-storm defect is labeled “wind damage” because the reported event occurred recently. Both interpretations collapse an evidentiary hypothesis into a conclusion. The appropriate scientific method is to state the candidate mechanisms, their expected traces, and what the observations actually distinguish. The resulting opinion can remain qualified without becoming operationally useless.
This matters financially. A categorical causal label may cause claims scope to expand to unrelated repairs or contractually covered repairs to be excluded because a preexisting condition is also present. A revised analysis should separate event-induced damage, baseline defects, necessary integrated restoration, and items whose indemnifiability depends on terms. Specific causal and cost line items are more informative than a single assertion of total storm loss or total deterioration.
Legal doctrine remains jurisdiction-specific
The selected U.S. federal and English expert evidence authorities emphasize methodological reliability, disclosure, and proper expert duties in their respective procedural systems. The marine insurance statute, its judicial interpretation, and Australian insurance contract legislation show how different substantive legal tests assign meaning to causal evidence. These sources do not establish one international legal definition of covered cause. The shared research lesson is the necessity to record jurisdiction, relevant policy wording, procedural stage, event date, and contract period with each expert finding.
An international claim dataset that records only “cause = hail” or “cause = wear” without coding physical mechanisms, simultaneous contributors, and legal context is ill-suited to compare coverage outcomes. It can combine different meanings into one nominal field. Proper research would separate a cause established by observation, a cause inferred by an expert, a cause found by a tribunal, and a cause recorded as the insurer's claims classification.
Evidence readiness and proportional investigation
Not every initial notice of loss warrants destructive testing or litigation-level reconstruction. A rapid emergency assessment can accurately identify immediate safety risks and provisional damage scope while explicitly leaving cause unresolved. A disputed high-value claim may justify controlled material testing, independent expert review, and a more complete local event reconstruction. The scientific requirement common to both is proportionality of conclusion to evidence, not identical investigation cost.
A useful report states what decision it is intended to inform and which new evidence could change it. A proposed test has high investigative value when its plausible outcomes would separate competing mechanisms and thereby affect the technical answer. A test that will merely repeat an already observed symptom may add cost without increasing causal discrimination. Recognizing test value makes evidence preservation more consequential: early loss of decisive components can turn a testable question into permanently uncertain reconstruction.
Original research contribution
The author's original contribution is a comparative synthesis of engineering mechanism analysis, evidentiary provenance, expert reliability, insurance causation doctrine, and claims governance around a single incident-specific question. The synthesis demonstrates where commonly conflated conclusions require separate sources, how physically joint causes should remain visible to legal analysis, and why site-level exposure cannot be equated with regional event reporting. It identifies practical audit questions and prospective empirical research needs without giving a new framework a proprietary name or asserting priority for concepts already familiar in the respective disciplines.
The source-specific findings support a direct institutional conclusion: a technically robust and procedurally defensible claim file is one in which another competent examiner can trace a conclusion backward through method, measurement, specimen or document provenance, causal alternative, event, and baseline condition, while a decision-maker can separately trace payment or denial forward through applicable wording and law. That traceability also permits a principled undetermined finding when sufficient discrimination is unavailable.
Further empirical investigation
Future research can use deidentified claims linked to contemporaneous site photographs, maintenance histories, local hazard records, physical test results, independent expert opinions, contract forms, legal jurisdictions, and outcomes. Studies should distinguish validation against a known experimental cause from agreement among reviewers. Case selection should be recorded because claims requiring outside experts are often disputed or unusually complex and are not representative of all property damage. Prospective preservation interventions could be compared on the availability of diagnostic material, but any claimed effect would require defined endpoints, comparators, and evidence.
The existing published standards provide research questions; they do not supply missing field results. High-quality studies would preregister physical classifications, specify competing explanations, identify diagnostic measurement performance, record the actual event and asset units, and publish uncertainty. Particular attention should be paid to allegations of fraud, post-loss contractor intervention, catastrophic scarcity of inspectors, and the availability of independent review. Those questions are proposed research topics rather than asserted empirical findings of this dossier.
Conclusion
Property damage causation is established through a traceable connection between observed changes, a physically plausible mechanism, the relevant event chronology, and documented treatment of alternative explanations. Expert evidence is scientifically useful when collection and provenance are reliable, testing is suitable to the exact question, expertise is appropriate, conclusions reflect rather than exceed the observations, and limits are disclosed. The presence of a regional hazard is distinct from its operation on a specific property; an established physical cause is distinct from an insured cause under a particular legal instrument.
NFPA, ASTM, and ISO documents support a disciplined process for origin and cause investigation, evidence preservation, examination, competence, interpretation, and reporting within their respective scopes. Official storm and high-water mark guidance establishes the need for source-aware site attribution. U.S. and English rules demonstrate procedural scrutiny of experts, while the selected UK and Australian legislation illustrates the dependency of the legal consequences of physical causes on actual statutory and contractual texts. IAIS and NAIC sources situate reliable evidence and communication within fair claims governance.
The research conclusion is that each transition from event to site exposure, exposure to mechanism, mechanism to damage, damage to repair scope, and repair scope to contractual indemnity must be supported by evidence appropriate to that proposition. Where missing records prevent discrimination among physically plausible alternatives, an accurately qualified or undetermined technical conclusion remains a valid scientific outcome. Its contractual and procedural consequences are then determined using the applicable insurance and legal framework. This monograph presents source-grounded interpretation rather than a new doctrinal system or an unmeasured empirical claim.
References
ASTM International. (2018). ASTM E620-18: Standard Practice for Reporting Opinions of Scientific or Technical Experts. https://store.astm.org/e0620-18.html
ASTM International. (2022). ASTM E860-22: Standard Practice for Examining and Preparing Items That Are or May Become Involved in Criminal or Civil Litigation. https://store.astm.org/e0860-22.html
ASTM International. (2023). ASTM E1188-23: Standard Practice for Collection and Preservation of Information and Physical Items by a Technical Investigator. https://store.astm.org/e1188-23.html
ASTM International. (2024). ASTM E2018-24: Standard Guide for Property Condition Assessments: Baseline Property Condition Assessment Process. https://store.astm.org/standards/e2018
Australian Government, Federal Register of Legislation. (2026). Insurance Contracts Act 1984, compilation of October 8, 2026; section 54. https://www.legislation.gov.au/C2004A02944/2026-10-08/2026-10-08/text/original/epub/OEBPS/document_1/document_1.html
Cooper, G. S., & Meterko, V. (2019). Cognitive bias research in forensic science: A systematic review. Forensic Science International, 297, 35–46. https://doi.org/10.1016/j.forsciint.2019.01.016
Federal Rules of Evidence. (2023 amendment). Rule 702: Testimony by Expert Witnesses. Cornell Legal Information Institute reproduction and amendment notes. https://www.law.cornell.edu/uscode/text/28a/courtrules-Evid/article-VII/courtrule-702
Georgiou, N., Morgan, R. M., & French, J. C. (2020). Conceptualising, evaluating and communicating uncertainty in forensic science: Identifying commonly used tools through an interdisciplinary configurative review. Science & Justice, 60(4), 313–336. https://pubmed.ncbi.nlm.nih.gov/32650934/
Hinckley, J. M. (2006). Guidelines for Determination of Wind versus Water Causation of Hurricane Damage. Forensic Engineering 2006, 195–209. American Society of Civil Engineers. https://ascelibrary.org/doi/10.1061/40853%28217%2917
International Association of Insurance Supervisors. (n.d.). Insurance Core Principles, including ICP 19: Conduct of Business. https://www.iais.org/icp-online-tool/
International Association of Insurance Supervisors. (2025, July 7). IAIS publishes Application Paper on fair treatment of a wide range of consumers. https://www.iais.org/2025/07/iais-publishes-application-paper-on-fair-treatment-of-a-wide-range-of-consumers/
International Organization for Standardization. (2017). ISO/IEC 17025:2017: General requirements for the competence of testing and calibration laboratories. https://www.iso.org/standard/66912.html
International Organization for Standardization. (2018). ISO 21043-2:2018: Forensic sciences—Part 2: Recognition, recording, collecting, transport and storage of items. https://www.iso.org/standard/72041.html
International Organization for Standardization. (2025a). ISO 21043-1:2025: Forensic sciences—Part 1: Vocabulary. https://www.iso.org/standard/87128.html
International Organization for Standardization. (2025b). ISO 21043-3:2025: Forensic sciences—Part 3: Analysis. https://www.iso.org/standard/72040.html
International Organization for Standardization. (2025c). ISO 21043-4:2025: Forensic sciences—Part 4: Interpretation. https://www.iso.org/standard/72039.html
International Organization for Standardization. (2025d). ISO 21043-5:2025: Forensic sciences—Part 5: Reporting. https://www.iso.org/standard/73896.html
International Organization for Standardization. (2026). ISO/IEC 17020:2026: Conformity assessment—Requirements for bodies performing inspection, third edition, corrected June 2026. https://www.iso.org/standard/17020
National Association of Insurance Commissioners. (n.d.). Model Laws, including unfair claims settlement practice models. https://content.naic.org/model-laws
National Association of Insurance Commissioners. (2022, May 22). What You Should Know About Settling a Homeowners Insurance Claim. https://content.naic.org/article/what-you-should-know-about-settling-homeowners-insurance-claim
National Fire Protection Association. (2024, with interim amendments in 2025). NFPA 921: Guide for Fire and Explosion Investigations; development and current edition record. https://www.nfpa.org/codes-and-standards/nfpa-921-standard-development/921
National Institute of Standards and Technology. (n.d.). NFPA 921:2024 Guide for Fire and Explosion Investigations; OSAC standards library record. https://www.nist.gov/standard/1556
National Institute of Standards and Technology. (n.d.). Evidential Statistics: Forensic Science Research. https://www.nist.gov/evidential-statistics
National Oceanic and Atmospheric Administration, National Centers for Environmental Information. (n.d.). Storm Events Database: Frequently Asked Questions and Data Disclaimer. https://www.ncdc.noaa.gov/stormevents/faq.jsp
National Oceanic and Atmospheric Administration, National Centers for Environmental Information. (n.d.). Severe Weather Data Inventory. https://www.ncei.noaa.gov/maps/swdi/
National Weather Service. (n.d.). Preliminary Local Storm Reports: Explanation. https://www.weather.gov/bgm/helpLocalStormReports
Supreme Court of the United Kingdom. (2011, February 1). Global Process Systems Inc v. Syarikat Takaful Malaysia Berhad [2011] UKSC 5 (The Cendor MOPU). https://www.supremecourt.uk/cases/uksc-2010-0006
Supreme Court of the United States. (1993). Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579. https://www.law.cornell.edu/supremecourt/text/509/579
Supreme Court of the United States. (1999). Kumho Tire Co. v. Carmichael, 526 U.S. 137. https://www.law.cornell.edu/supremecourt/text/526/137
Taroni, F., Biedermann, A., Garbolino, P., & Aitken, C. G. G. (2004). A general approach to Bayesian networks for the interpretation of evidence. Forensic Science International, 139(1), 5–16. https://pubmed.ncbi.nlm.nih.gov/14687767/
United Kingdom, Civil Procedure Rules. (current online edition). Part 35: Experts and Assessors. https://www.justice.gov.uk/courts/procedure-rules/civil/rules/part35
United Kingdom, Civil Procedure Rules. (current online edition). Practice Direction 35: Experts and Assessors. https://www.justice.gov.uk/courts/procedure-rules/civil/rules/part35/pd_part35
United Kingdom, Forensic Science Regulator. (2025). Forensic Science Activities: Statutory Code of Practice, Version 2. https://www.gov.uk/government/publications/forensic-science-activities-statutory-code-of-practice-version-2/forensic-science-activities-statutory-code-of-practice-version-2-accessible
United Kingdom, Parliament. (1906). Marine Insurance Act 1906, section 55: Included and Excluded Losses. https://www.legislation.gov.uk/ukpga/Edw7/6/41/section/55
U.S. Geological Survey. Feaster, T. D., & Koenig, T. A. (2017). Field Manual for Identifying and Preserving High-Water Mark Data. Open-File Report 2017-1105. https://pubs.usgs.gov/publication/ofr20171105