Skip to content

Compound-Hazard Stacking

The planning failure in which co-occurring hazards interact through shared drivers, infrastructure, or response resources, so their joint impact exceeds — and the available capacity falls below — what a plan built from independent single-hazard analyses predicts.

Core Idea

Compound-hazard stacking is the planning failure in which multiple hazards co-occur and interact — sharing underlying drivers, sharing infrastructure, or competing for the same response resources — so that their joint impact substantially exceeds what a plan built from independent single-hazard analyses would predict, and the response capacity available against the joint event is smaller than the sum of the capacities that would have been available to each hazard in isolation.

The mechanism has three structural pieces. First, hazard catalogues are conventionally constructed one hazard at a time, and emergency plans, resource allocations, and infrastructure designs are optimised against those standalone scenarios. Second, hazards share coupling pathways that violate the independence assumption: common drivers (a heat wave simultaneously drives wildfire ignition risk and grid stress through elevated air-conditioning demand); shared infrastructure (flood mitigation that depends on pumping stations that lose power during a flood-plus-outage event; backup generators sited in basements that floods reach); shared response resources (the same emergency medical personnel, shelter capacity, command staff, and supply-chain bandwidth are simultaneously demanded by two or more events). Third, the realised compound event reveals the interaction terms: the joint physical effects exceed the sum of the marginal effects because one hazard degrades the mitigations designed against another, and the response capacity available is below the sum of the standalone capacities because the same resources cannot be in two places at once.

The pattern is documented across major compound events: at Fukushima Daiichi, the earthquake mitigation infrastructure — seawall height, backup-power siting — was designed against the marginals and the joint tsunami exceeded the seawall while the earthquake damaged the backup-power systems the plant needed to respond; during the 2020 hurricane season, shelter capacity was constrained by pandemic distancing requirements and medical surge capacity was consumed by COVID patients simultaneously with storm evacuees; in recurring California events, heat waves drive grid stress that triggers public-safety power shutoffs as wildfire mitigation, removing air-conditioning during the heat emergency. The intervention family across these is the same in shape: build joint-occurrence scenarios rather than per-hazard playbooks, stress-test capacity under compound conditions, and explicitly map inter-hazard couplings through shared drivers, shared infrastructure, and shared response resources.

Structural Signature

Sig role-phrases:

  • the per-hazard catalogue — hazards enumerated, planned, and resourced one at a time, each optimized against its standalone worst case
  • the independence assumption — the load-bearing premise the catalogue silently makes: that hazards occur and can be planned for independently, so a plan is sound if it covers each marginal
  • the shared-driver coupling — a common underlying condition raising the probability of two hazards at once (a heat wave driving wildfire ignition and grid stress together)
  • the shared-infrastructure coupling — the mitigation for one hazard is the asset at risk for another (basement generators meant for outages reached by floods)
  • the shared-resource coupling — the same crews, shelters, command staff, and supply bandwidth demanded by two events simultaneously, unable to be in two places
  • the impact super-additivity — the realized joint physical effect exceeds the marginal sum because one hazard degrades another's defenses
  • the capacity sub-additivity — available response capacity falls below the sum of standalone capacities (the term standalone analysis never sees, since each plan books the full resource as uncontested)
  • the probability undercount — the joint probability is understated because reassuring marginals were computed as if occurrence were independent
  • the coupling-keyed intervention family — the diagnostic (single-hazard plans degrade under compound conditions in unpredicted ways) and the matched fixes: build "while-also" scenarios, stress-test capacity jointly, and map the three coupling pathways edge by edge

What It Is Not

  • Not a sequential cascade. A cascade is failures propagating in series — one triggering the next down a chain; compound-hazard stacking is hazards co-occurring and interacting simultaneously. The surprise is in the interaction terms of concurrent stressors, not in a chain of consequences, which is why the analysis frame is "while-also" scenarios rather than a propagation tree.
  • Not a single black-swan rare event. It is not one undercounted tail hazard but the co-occurrence of several, whose joint probability marginals understate because the hazards share drivers. The reassuring single-hazard odds are exactly what hide the coupled risk.
  • Not merely additive (just "worse") impact. The non-additivity has two separable parts, and the second is the one standalone analysis never sees: physical impacts exceed the marginal sum (one hazard degrades another's defenses), and available response capacity falls below the sum of standalone capacities (the same crews and shelters cannot be in two places). Summing single-hazard impacts misses both, and over-counts capacity besides.
  • Not the failure of any one hazard plan. Each single-hazard playbook can be individually sound; the failure lives in the blind spot between independently correct plans, where shared infrastructure and shared resources silently consume the margins. Drilling each plan harder in isolation does not surface the coupling.
  • Not the synergy_and_antagonism prime itself. The substrate-general non-additive-interaction pattern is carried by synergy_and_antagonism; compound-hazard stacking is its operationalizing instantiation in the hazard substrate, adding the three named coupling pathways (shared drivers, infrastructure, response resources) and the compound-scenario intervention family. The mechanism travels under the parent's name, not this idiom.
  • Not risk pooling. Risk pooling assumes independence to aggregate exposures and reduce variance; compound-hazard stacking is the failure of that independence assumption. Pooling that ignores shared drivers underestimates correlated joint loss exactly where this concept applies.

Scope of Application

Compound-hazard stacking lives across the emergency-management, disaster-risk-reduction, and critical-infrastructure-planning subfields concerned with co-occurring hazards; its reach is bounded by that domain — a hazard catalogue resourced one hazard at a time, with couplings that violate the independence assumption. The substrate-general non-additive-interaction pattern travels under synergy_and_antagonism (its toxicology, medicine, finance, and reliability cousins are co-instances of that parent), not under this idiom, so they stay out of this map.

  • Hurricane plus pandemic — the 2020 season, where shelter capacity was constrained by distancing requirements and medical surge capacity was consumed by COVID patients simultaneously with storm evacuees.
  • Heat plus grid stress plus wildfire — California's coupled pattern, where heat drives grid stress that triggers public-safety power shutoffs as wildfire mitigation, removing air-conditioning during the heat emergency.
  • Earthquake plus tsunami plus nuclear release — Fukushima Daiichi, where seawall height and backup-power siting were designed against the marginals and the joint event exceeded both.
  • Flood plus power outage — Hurricane Sandy, where flooded pumping stations and basement generators revealed flood mitigation depending on the power infrastructure the flood took down.
  • Pandemic plus economic crisis plus civil unrest — the 2020 simultaneity, where response institutions designed for any one were stretched across three sharing personnel and governance bandwidth.

Clarity

Naming compound-hazard stacking exposes the load-bearing assumption that a hazard-by-hazard catalogue silently makes: that hazards occur and can be planned for independently, so that a plan is sound if it covers each worst case on its own. The concept makes the independence assumption visible as the thing that fails, and pins down how it fails through three named coupling pathways — shared drivers, shared infrastructure, shared response resources. With those in hand a planner can ask the sharper question the per-hazard playbook never poses: not "are we ready for a flood, and separately for an outage?" but "what happens when these occur together — does the mitigation for one become the vulnerability for another, and can the same crews, shelters, and command staff be in two places at once?" The planning frame shifts from a list of "or" scenarios to a matrix of "while-also" ones.

It also sharpens exactly which kind of non-additivity is in play and where the surprise hides, distinguishing this failure from neighbours it is easy to collapse into. The joint event is worse than the sum in two separable senses the concept keeps apart: the physical impacts exceed the marginal sum because one hazard degrades another's defences, and — the part standalone analysis never sees at all — the response capacity available falls below the sum of the standalone capacities because resources are shared. And because the coupling is through common drivers, the joint probability is itself undercounted by marginals that look reassuring. Holding this distinct from a sequential cascade (where failures propagate in series rather than co-occur) and from a single rare event tells the practitioner precisely where to look: at the interaction terms and the shared-resource contention that sit in the blind spot between independently sound plans.

Manages Complexity

Catastrophe planning, done hazard by hazard, faces a combinatorial blow-up the moment hazards are allowed to co-occur: a catalogue of dozens of standalone hazards generates an unmanageable number of pairs, triples, and higher combinations, each a candidate joint scenario with its own physics and its own resource draw, and no planner can build a bespoke playbook for every "while-also" the world might produce. Compound-hazard stacking tames that explosion not by enumerating combinations but by naming the three — and only three — pathways through which any two hazards can interact: a shared driver (one underlying condition raises the probability of both, as a heat wave drives wildfire ignition and grid stress together), shared infrastructure (the mitigation for one is the asset at risk for another, as basement generators meant for outages are reached by floods), and shared response resources (the same crews, shelters, command staff, and supply bandwidth are demanded by both at once). The analyst stops asking the unanswerable "what does each of the thousands of hazard combinations do?" and instead screens each pair against three coupling questions, which collapses a combinatorial scenario space into a small, fixed checklist applied edge by edge.

The compression has a definite branch structure because the three pathways partition the surprise into separable, trackable quantities, and the joint outcome reads off which couplings are present. Where no pathway links two hazards, the marginal, independent plan is sound and additive estimation holds; where a coupling exists, the concept says exactly where and how the additive plan will fail. A shared-infrastructure edge predicts that the joint physical impact exceeds the marginal sum, because one hazard degrades the other's defences; a shared-resource edge predicts that the available response capacity falls below the sum of the standalone capacities, because the same crews cannot be in two places — and this is the term standalone analysis never sees, since each single-hazard plan books the full resource as if uncontested; a shared-driver edge predicts that the joint probability itself is undercounted, because reassuring marginals were computed as if occurrence were independent. So the planner tracks three edge-types on a hazard graph and reads off three distinct failure terms (impact super-additivity, capacity sub-additivity, probability undercount) rather than re-deriving the joint behaviour of each compound event from scratch.

This is what lets the concept convert an open-ended worry — "are we ready for everything at once?" — into a bounded, structured analysis whose unit is the coupling-bundle rather than the individual hazard. Because the couplings are few and named, the planning artefact shifts from a list of independent worst cases to a matrix of co-occurrences in which only the coupled cells need deep work, and the same three edge-types tell the planner where to intervene: break or harden a shared-infrastructure dependency, pre-stage or de-conflict a shared resource, recompute a joint probability where a common driver is found. The high-dimensional question "how bad is every possible simultaneous disaster, and are we resourced for it?" thereby collapses to a low-dimensional one in three coupling pathways and the three interaction terms they generate, with the diagnostic — single-hazard plans degrading under compound conditions in ways those plans never predicted — pointing straight at the coupled edges where the planning margins were silently consumed.

Abstract Reasoning

Compound-hazard stacking licenses a set of reasoning moves by which a disaster planner exposes the non-additivity that single-hazard analysis hides, all grounded in screening hazard pairs for three named couplings. The foundational move is edge-screening on a hazard graph. Rather than attempting to enumerate every joint scenario, the planner treats the hazard catalogue as a graph and asks of each pair only three questions: do they share a driver, share infrastructure, or share response resources? The reasoning runs from the presence or absence of these edges to whether the independence assumption holds — where no pathway links two hazards the marginal plan is sound and additive estimation is safe, and where an edge exists the planner knows the additive plan will fail and exactly where. This converts a combinatorial scenario space into a small, fixed checklist applied edge by edge.

A second move is predicting the specific failure term from the coupling type. The planner reasons from each edge-type to a distinct, named consequence: a shared-infrastructure edge predicts that the joint physical impact exceeds the marginal sum, because one hazard degrades the defences designed against the other (the basement generator meant for an outage is reached by the flood); a shared-resource edge predicts that available response capacity falls below the sum of the standalone capacities, because the same crews and shelters cannot be in two places at once; a shared-driver edge predicts that the joint probability itself is undercounted, because reassuring marginals were computed as if occurrence were independent. The three edge-types thus map to three separable interaction terms — impact super-additivity, capacity sub-additivity, probability undercount — which the planner reads off the graph rather than re-deriving for each compound event.

A third move is diagnostic recovery of the missing capacity term. The planner reasons that standalone analysis structurally cannot see the resource-contention term, because each single-hazard plan books the full resource as if uncontested, so a portfolio of individually sound plans systematically over-counts available capacity. From the observation that single-hazard plans degrade under compound conditions in ways those plans never predicted, the planner infers that a shared-resource edge has silently consumed the margins, and looks specifically at where the same crews, command staff, and supply bandwidth were promised to more than one event. The hidden state — the over-subscribed resource — is inferred from the gap between summed standalone capacity and realized joint capacity.

A fourth move is interventionist action keyed to the coupling found. Because the couplings are few and named, the planner reasons from each to its matched remedy: break or harden a shared-infrastructure dependency so one hazard's mitigation no longer falls to another; pre-stage or de-conflict a shared resource so it is not double-promised; recompute a joint probability wherever a common driver is found. The predicted effect of each is to remove the corresponding interaction term, so only the coupled cells of the co-occurrence matrix need deep work and effort is directed precisely at the edges where margins were lost rather than spread across all hazard combinations.

A fifth move is boundary-drawing against neighboring patterns to fix the analysis frame. The planner distinguishes compound stacking from a sequential cascade (failures propagating in series rather than co-occurring) and from a single rare event (one hazard, undercounted in isolation), reasoning that only simultaneous, coupled hazards generate the interaction terms and shared-resource contention at issue. Placing the situation correctly tells the planner to build "while-also" scenarios and stress-test capacity under joint conditions rather than chase a chain of consequences or a single tail event — directing the planning frame to the blind spot between independently sound plans where the compound surprise actually lives.

Knowledge Transfer

Within emergency management, disaster-risk reduction, and critical-infrastructure planning the compound-hazard-stacking construct transfers as mechanism, because the same three-coupling screen (shared drivers, shared infrastructure, shared response resources) and the three interaction terms it predicts (impact super-additivity, capacity sub-additivity, probability undercount) apply to any set of co-occurring hazards. The edge-screening move, the failure-term prediction, the recovery of the hidden capacity-contention term, and the coupling-keyed interventions all carry without translation across the documented compound events — hurricane-plus-pandemic (2020 shelter and surge constraints), heat-plus-grid-stress-plus-wildfire (California's coupled shutoffs), earthquake-plus-tsunami-plus-nuclear-release (Fukushima's marginally designed seawall and backup power), flood-plus-power-outage (Sandy's basement generators and flooded pumping), and pandemic-plus-economic-crisis-plus-civil-unrest (shared governance bandwidth). These share one structure — a hazard catalogue resourced one hazard at a time, with couplings that violate the independence assumption — so the analysis transfers literally, and the within-field migrations the construct already underwrites (correlated-tail-risk methods crossing from finance into climate compound-event modeling; inter-system dependency mapping crossing from critical-infrastructure protection into pandemic preparedness; the FEMA/IPCC shift from all-hazards checklists to compound scenarios) are genuine because the coupled-stressor structure is common.

Beyond disaster-risk fields the honest reading is case (B) — a more general mechanism recurs across domains, while "compound-hazard stacking" stays an emergency-management idiom. The underlying structural pattern — non-additive interaction of simultaneous stressors — is synergy_and_antagonism applied to the hazard/risk substrate, and it genuinely recurs elsewhere as co-instances under other names: multi-stressor interaction in toxicology, comorbidity in medicine, correlated tail risk in finance and insurance, joint failure in reliability engineering. What travels across those domains is that synergy pattern (composed with correlated-risk and shared-resource-preemption structure), not "compound-hazard stacking" as a unit; stripped of the hazard vocabulary the construct is synergy + correlated risk + resource contention, all already housed in the parent primes. So when the cross-domain lesson is wanted — simultaneous stressors interact non-additively, so summing marginal analyses understates joint impact and overstates joint capacity — what should carry is synergy_and_antagonism (with cascade distinguished as the sequential sibling, black_swan_high_impact_low_probability_events as the single-rare-event neighbor, and risk_pooling as the independence-assuming counterpart). The construct's irreducible contribution is the operationalizing layer that makes synergy actionable in its home substrate — the three named coupling pathways and the compound-scenario / capacity-stress-testing / coupling-mapping intervention family — which is emergency-management furniture that does not and should not travel under its own name, even as the synergy mechanism beneath it does (see Structural Core vs. Domain Accent).

Examples

Canonical

Fukushima Daiichi, 11 March 2011, is the textbook compound event. The magnitude-9.0 Tōhoku earthquake and the tsunami it spawned were not independent: they arrived together, from one source. The plant's defenses had been sized against the marginals — a seawall roughly 5.7 m high and emergency diesel generators sited low, several in basements. The tsunami run-up reached well past 13 m, overtopping the wall, and the same water then flooded the generators and switchgear meant to power cooling after the quake. With the earthquake having already cut the offsite grid, the loss of backup power caused station blackout and three core meltdowns. Neither a "big earthquake" playbook nor a "big wave" playbook, run separately, exposed this: the wave defeated the barrier designed for one hazard and drowned the mitigation reserved for the other.

Mapped back: Sizing the seawall and generators hazard-by-hazard is the per-hazard catalogue resting on the independence assumption; the common tectonic source is the shared-driver coupling; the flood reaching the very generators meant to answer the quake is the shared-infrastructure coupling, producing the impact super-additivity (one hazard degraded the other's defense). The joint quake-plus-tsunami-plus-blackout outcome vastly exceeded any marginal plan — the exact failure the coupling-keyed intervention family (build "while-also" scenarios) is meant to forestall.

Applied / In Practice

The 2020 Atlantic hurricane season forced U.S. emergency managers to run their playbooks inside the COVID-19 pandemic. When Hurricane Laura drove mass evacuations along the Gulf Coast in August 2020, the standard congregate-shelter response collided with pandemic distancing: shelters had to cut occupancy, and agencies switched to hotel rooms and spaced facilities. Simultaneously, the hospital surge capacity a hurricane assumes it can draw on was already partly consumed by COVID patients, and the same public-health staff were double-tasked. The independence each plan assumed — "a hurricane plan may book the region's shelter and medical capacity as available" — simply did not hold.

Mapped back: The hurricane plan and the pandemic plan were each a per-hazard catalogue entry built on the independence assumption. Shelter beds and hospital surge capacity became the shared-resource coupling — the same beds and clinicians demanded by both events at once — yielding the capacity sub-additivity that neither standalone plan could see, since each booked the full resource as uncontested. The remedy adopted (de-conflicting the shared resource via non-congregate sheltering) is exactly the coupling-keyed intervention family keyed to a shared-resource edge.

Structural Tensions

T1: Per-hazard tractability versus systematic blindness (the decomposition that plans and blinds). Building a hazard catalogue one hazard at a time is not laziness — it is what makes catastrophe planning tractable at all, letting a planner optimise a defensible worst-case response to each threat and audit it in isolation. But the very decomposition that makes each plan buildable is what manufactures the blind spot: independence is assumed the moment hazards are separated, and the interaction terms live precisely between the individually sound plans, where no single-hazard analysis is looking. The feature cuts both ways — decompose and you can plan; decompose and you cannot see the coupling. Drilling any one plan harder in isolation deepens the very frame that hides the compound surprise. Diagnostic: Is the residual risk one that a single-hazard plan could catch by being made more thorough, or does it live only in the interaction between two plans that are each already sound?

T2: Three-pathway compression versus coupling completeness (a fixed checklist against an open world). The construct's power is that it replaces an unmanageable enumeration of hazard pairs, triples, and higher combinations with exactly three coupling questions — shared driver, shared infrastructure, shared response resource — applied edge by edge. That compression is what makes compound analysis finite and actionable. But a fixed three-pathway screen is a bet that every consequential interaction falls into one of three named channels, and a coupling that runs through none of them — a behavioural feedback, a legal-liability entanglement, a reputational contagion linking two responses — could pass the screen clean while still producing non-additive joint impact. The tension is between a checklist small enough to use and a coupling space possibly richer than three edges. Trust the three pathways and you may miss a fourth; abandon them and you are back in the combinatorial blow-up. Diagnostic: Does the suspected interaction route through a shared driver, shared infrastructure, or shared resource — or through some coupling the three-pathway screen is not built to detect?

T3: Compound-robustness slack versus lean single-hazard efficiency (redundancy that is waste until it isn't). The matched remedies — de-conflict a shared resource, harden a shared-infrastructure dependency, pre-stage duplicate capacity — all amount to holding slack: crews, generators, shelter beds, and command bandwidth that are not double-promised across co-occurring events. But that slack is, under the far more common independent-hazard conditions, straightforwardly inefficient: capacity sitting idle, redundant infrastructure maintained against a joint event that may not arrive for decades. The feature is double-edged — the reserve that saves the compound response is overhead that a cost-audit will flag every year it goes unused. Provisioning for stacking trades everyday efficiency for tail resilience, and there is no setting where both are maximised at once. Diagnostic: Is the proposed reserve justified by the probability and severity of the coupled event, or is it capacity that a single-hazard accounting would correctly retire as idle?

T4: Physical super-additivity versus hidden capacity sub-additivity (where the real surprise hides). The construct names two separable non-additivities, and they are not symmetric in visibility. Impact super-additivity — one hazard degrading another's defences — is at least in principle discoverable by a diligent single-hazard analyst who imagines the flood reaching the generator. But capacity sub-additivity is structurally invisible to standalone analysis, because each single-hazard plan books the full resource as uncontested, so a portfolio of individually correct plans systematically over-counts total capacity with no term ever flagging the double-count. The tension is that the more dangerous surprise is the one no isolated plan can even represent: summed capacity looks additive and reassuring right up to the moment two events demand the same crews. Focusing on the vivid physical interaction can leave the quieter accounting error — promised capacity that does not exist — entirely unexamined. Diagnostic: Has anyone summed the standalone capacity commitments to check whether the same crews, beds, or command staff are promised to more than one simultaneous event?

T5: Compound stacking versus cascade versus rare event (getting the frame right). The construct earns its keep partly by what it excludes: it is not a sequential cascade (failures propagating in series), and not a single undercounted black-swan tail event. Those neighbours demand different analyses — a propagation tree for the cascade, better tail estimation for the rare event — and mislabelling sends the planner down the wrong one. But the boundary cuts both ways: a real event often is several of these at once (Fukushima was a shared-source co-occurrence that then cascaded into station blackout), and insisting on one clean category can suppress the parts that belong to the others. The frame is a tool for directing attention, yet reality does not respect its partitions, so the discipline of classification competes with the messiness of events that stack and cascade and sit in a tail. Diagnostic: Are the hazards co-occurring and interacting simultaneously (stacking), propagating one-then-the-next (cascade), or a single undercounted extreme (rare event) — and does the situation demand more than one of these frames at once?

T6: Autonomy versus reduction (an emergency-management idiom or the synergy prime). "Compound-hazard stacking" is a fully operationalized construct in situ — the three named coupling pathways plus the compound-scenario, capacity-stress-testing, coupling-mapping intervention family make synergy actionable in the hazard substrate in a way no abstract prime does. But the portable content is not the idiom: stripped of hazard vocabulary the construct is synergy_and_antagonism applied to risk, composed with correlated-risk and resource-contention structure, and that is what recurs as multi-stressor interaction in toxicology, comorbidity in medicine, and correlated tail risk in finance — co-instances of the parent, not of "compound-hazard stacking." The tension is between an idiom that earns its keep by making synergy operable in emergency management and the recognition that its cross-domain cargo already belongs to synergy_and_antagonism (with cascade the sequential sibling and risk_pooling the independence-assuming counterpart). Diagnostic: Resolve toward synergy_and_antagonism (with its correlated-risk and resource-contention companions) when asking what carries the lesson to other domains; toward named compound-hazard stacking when operationalizing joint-hazard planning in situ.

Structural–Framed Character

Compound-hazard stacking sits at the framed-leaning band of the spectrum — an emergency-management idiom for a planning failure, deeply bound to the practice of hazard cataloguing, though built on a genuinely physical non-additivity at its core. On evaluative_weight it is intermediate: the construct is diagnostic rather than moral, but it names a failure (joint impact exceeding, and capacity falling below, what independent single-hazard plans predict), so a negative valence is built in even as it stays an analytical account of where the additive plan breaks. On human-practice-bound it splits: the physical super-additivity (a tsunami overtopping a seawall and drowning the backup power meant for the quake) happens in the world regardless of observers, but "compound-hazard stacking" as a planning failure is constituted by the human practice of per-hazard cataloguing and its silent independence assumption — the failure is in the plans, not in nature. Institutional_origin is pronounced for the named construct: it is an emergency-management and disaster-risk-reduction idiom, its three named coupling pathways and its compound-scenario / capacity-stress-testing / coupling-mapping intervention family all FEMA/IPCC-lineage furniture. On vocab_travels the hazard vocabulary (per-hazard catalogue, shared-driver/infrastructure/resource couplings, capacity sub-additivity) is pinned to disaster-risk fields; carried elsewhere the same phenomenon appears as multi-stressor interaction, comorbidity, or correlated tail risk. Import_vs_recognize is bimodal: across documented compound events (Fukushima, 2020 hurricane-plus-pandemic, California heat-grid-wildfire) the same mechanism is recognized as one structure, but beyond disaster-risk the recurrence belongs to the parent, and the idiom itself does not travel.

The portable structural skeleton is synergy_and_antagonism — the non-additive interaction of simultaneous stressors — composed with correlated-risk structure (a shared driver making joint occurrence more likely than independent marginals imply) and resource-contention structure (the same capacity double-promised to concurrent events); the composition is genuinely multi-part, since the construct's two signature terms, impact super-additivity and capacity sub-additivity, come from different companions (synergy for the first, resource contention for the second). That skeleton is what compound-hazard stacking instantiates, and it is what recurs as multi-stressor interaction in toxicology, comorbidity in medicine, and correlated tail risk in finance — with cascade the sequential sibling, risk_pooling the independence-assuming counterpart, and black_swan the single-rare-event neighbor — while the construct's own contribution, the operationalizing layer that makes synergy actionable (the three named coupling pathways and the compound-scenario intervention family), is emergency-management furniture that stays home. Its character: a practice-constituted, disaster-risk planning-failure idiom resting on a real physical non-additivity, structural in the synergy-plus-correlated-risk-plus-resource-contention skeleton it instantiates, framed in the hazard-cataloguing apparatus that pins the "compound-hazard stacking" name to emergency management.

Structural Core vs. Domain Accent

This section decides why compound-hazard stacking is a domain-specific abstraction and not a prime, and it carries the case for its domain-specificity too — with the wrinkle that its skeleton is genuinely composite rather than single.

What is skeletal (could lift toward a cross-domain prime). Strip the hazard vocabulary and a thin relational structure survives, and it is multi-part: simultaneous stressors interact non-additively, so that a plan built by summing independent single-stressor analyses understates the joint impact and overstates the joint capacity. Two distinct portable pieces underlie the two signature surprises. The non-additive interaction of co-occurring stressors is synergy_and_antagonism — the source of impact super-additivity, where one stressor degrades another's defences. The over-count of available capacity is a separate resource-contention structure — the same reserve promised to two demands that cannot both be met — and the under-count of joint probability is a correlated-risk structure, a shared driver making co-occurrence likelier than independent marginals imply. That composite skeleton is genuinely substrate-portable, which is exactly why the construct instantiates synergy_and_antagonism composed with correlated-risk and resource-contention, and why multi-stressor interaction in toxicology, comorbidity in medicine, correlated tail risk in finance, and joint failure in reliability engineering are co-instances of it. But it is the core the construct shares, not what makes compound-hazard stacking distinctive.

What is domain-bound. Everything that makes the construct compound-hazard stacking in particular is the operationalizing layer that makes synergy actionable in the hazard substrate, and it is emergency-management furniture. The per-hazard catalogue and its silent independence assumption; the three named coupling pathways (shared driver, shared infrastructure, shared response resources) that discretise the interaction space; the three named failure terms (impact super-additivity, capacity sub-additivity, probability undercount); and the compound-scenario / capacity-stress-testing / coupling-mapping intervention family are FEMA/IPCC-lineage apparatus. The decisive test: stripped of hazard vocabulary the construct is just synergy + correlated risk + resource contention, and its home instances — Fukushima's overtopped seawall and drowned backup power, the 2020 hurricane-plus-pandemic shelter squeeze, California's heat-grid-wildfire coupling — lose their referents; a toxicologist or a reliability engineer meets the same composite mechanism under entirely different instrumentation. Remove the hazard-cataloguing frame and what remains is the general non-additive-interaction pattern, not this idiom.

Why this does not clear the prime bar. A prime's vocabulary travels and its transfer is recognition of the same mechanism, not analogy. Compound-hazard stacking's transfer is bimodal. Within emergency management, disaster-risk reduction, and critical-infrastructure planning it travels as mechanism — the three-coupling screen and the three interaction terms apply intact across every documented compound event, and the within-field method migrations (correlated-tail-risk methods crossing from finance into climate compound-event modeling, dependency mapping crossing into pandemic preparedness) are genuine because the coupled-stressor structure is common. Beyond disaster-risk fields the recurrence belongs to the parent, not the idiom: toxicology's multi-stressor interaction and finance's correlated tail risk are co-instances of synergy_and_antagonism, not of "compound-hazard stacking." When the cross-domain lesson is wanted — simultaneous stressors interact non-additively, so summed marginals understate joint impact and overstate joint capacity — it is carried by synergy_and_antagonism (with correlated-risk and resource-contention companions, cascade the sequential sibling, risk_pooling the independence-assuming counterpart, and black_swan the single-rare-event neighbor). The cross-domain reach belongs to those parents; the coupling-pathway and compound-scenario cargo that makes it "compound-hazard stacking" stays home in emergency management.

Relationships to Other Abstractions

Local relationship map for Compound-Hazard StackingParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Compound-HazardStackingDOMAINPrime abstraction: Synergy and Antagonism — is part ofSynergy andAntagonismPRIMEPrime abstraction: Correlated Capacity Demand — is a kind ofCorrelatedCapacity DemandPRIME

Current abstraction Compound-Hazard Stacking Domain-specific

Parents (2) — more general patterns this builds on

  • Compound-Hazard Stacking is a kind of Correlated Capacity Demand Prime

    Compound-Hazard Stacking is the emergency-planning species of Correlated Capacity Demand, where co-occurring hazards jointly load shared response capacity sized under independence.

  • Compound-Hazard Stacking is part of Synergy and Antagonism Prime

    A positive interaction effect—joint harm above the specified independent-hazard baseline—is an internal constituent of Compound-Hazard Stacking.

Hierarchy paths (2) — routes to 2 parentless roots

Not to Be Confused With

  • Sequential cascade. Failures propagating in series — one hazard triggering the next down a chain (a quake causes a blackout, which causes a cooling failure). Compound-hazard stacking is hazards co-occurring and interacting simultaneously; the surprise is in the interaction terms of concurrent stressors, not a propagation tree. Real events can do both (Fukushima co-occurred, then cascaded). Tell: do the failures happen one-then-the-next along a chain (cascade), or at the same time through shared drivers, infrastructure, or resources (stacking)? Analysis frame: propagation tree versus "while-also" matrix.

  • Black-swan / single rare tail event. One undercounted extreme hazard in isolation. Compound-hazard stacking is the co-occurrence of several hazards whose joint probability is undercounted because they share drivers — the reassuring single-hazard marginals are exactly what hide the coupled risk. Tell: is the concern one under-estimated extreme in isolation (black swan), or several ordinary hazards whose simultaneous occurrence is under-estimated because their marginals were treated as independent (stacking)?

  • Risk pooling. A technique that assumes independence to aggregate exposures and reduce variance (insurance spreading uncorrelated losses). Compound-hazard stacking is the failure of that independence assumption — pooling that ignores shared drivers underestimates correlated joint loss exactly where this concept bites. They are opposites: one relies on independence, the other diagnoses its breakdown. Tell: is independence being assumed to smooth aggregate risk (pooling), or is the violation of independence through shared coupling the whole problem (stacking)?

  • Common-cause failure. The reliability-engineering phenomenon where a single shared cause defeats supposedly redundant components at once (one fire disables both primary and backup pumps). This is essentially the shared-driver / shared-infrastructure coupling of stacking seen in a components-and-redundancy substrate — a close cousin, a co-instance of the same synergy-plus-correlated-risk parent under different vocabulary. Tell: is the object redundant components failing together from one root cause (common-cause failure), or multiple hazards co-occurring and overwhelming a response system through shared drivers/infrastructure/resources (compound-hazard stacking)?

  • Systemic / correlated risk. The finance-and-network notion that exposures move together, so a shock hits many positions at once and diversification fails. This captures the probability-undercount and correlation aspect of stacking but in a financial/network substrate, and lacks stacking's distinctive capacity sub-additivity (shared response resources double-promised). Tell: is the concern that exposures co-move so joint loss is under-estimated (correlated/systemic risk), or specifically that co-occurring hazards also overwhelm shared response capacity neither plan booked (stacking)?

  • synergy_and_antagonism (the parent / umbrella). The substrate-neutral non-additive-interaction pattern stacking instantiates (composed with correlated-risk and resource-contention structure). This umbrella carries the cross-domain reach — multi-stressor interaction in toxicology, comorbidity in medicine, correlated tail risk in finance, joint failure in reliability — while compound-hazard stacking adds the three named coupling pathways and the compound-scenario intervention family. Tell: is the setting a hazard catalogue with shared-driver/infrastructure/resource couplings (stacking), or any non-additive interaction of simultaneous stressors in any substrate (the synergy parent)?

Neighborhood in Abstraction Space

Compound-Hazard Stacking sits in a sparse region of the domain-specific corpus (73rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (309 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-07-12