Green Buildings May Be Fragile: New Framework Reveals Hidden Carbon Cost of Disruption

When a flood swallows a ground floor, when an earthquake buckles a shear wall, or when a heat wave pushes a living room past the threshold of habitability, a building stops doing the one thing it was designed to do: provide shelter. Yet according to a new review published in the Journal of Industrial Ecology, the dominant tool for measuring the environmental performance of buildings is structurally incapable of seeing that failure. Life cycle assessment, or LCA, the accounting framework used to tally the carbon and resource burdens of everything from a concrete foundation to a curtain wall, quietly assumes that a building delivers its full service continuously across its entire life. A new study by Fernanda Cruz Rios of Drexel University argues that this assumption is not a minor technicality but a blind spot large enough to distort how the industry judges which buildings are truly green.

The study set out to answer a deceptively simple question: how does the science of building LCA actually handle resilience, the capacity of a structure to withstand hazards and recover its function afterward? Rios conducted a systematic review of 40 original LCA studies spanning four hazard types: heat, flood, wind, and seismic events. Each study was mapped against a diagnostic grid built from two established resilience frameworks. The first, from Bruneau and colleagues, describes resilience through four dimensions: robustness, the strength to withstand a shock; redundancy, the presence of substitutable systems; rapidity, the speed of restoration; and resourcefulness, the capacity to mobilize people and materials for recovery. The second, from Rehak and colleagues, divides the resilience timeline into four phases: preparation, system response, recovery, and adaptation.

The results of that mapping are strikingly lopsided. Nearly every study in the review clustered in a single cell of the grid: robustness during the preparation and recovery phases. In practice, this means researchers have become adept at asking how hazard-resistant design choices change the initial bill of materials, and at converting predicted damage into repair quantities that are added to the inventory as extra embodied carbon. Seismic studies, aided by the mature FEMA P-58 performance-based earthquake engineering pipeline, are the most internally consistent. Heat studies, powered by building energy simulation with future climate projections, form a second coherent camp. Flood studies are the most fragmented, with four distinct integration mechanisms, from depth-damage functions to insurance-based input-output conversion, coexisting without cross-fertilization. Wind studies, only three in total, borrow from both the seismic and insurance modeling traditions.

What the review found missing is more consequential than what it found present. The rapidity dimension appeared only as partial evidence in two seismic papers that treated time-to-repair as an output but never translated recovery duration into environmental flows. The redundancy and resourcefulness rows of the evidence map were entirely empty: no study modeled backup systems as a design variable, evaluated the environmental cost of built-in spare capacity, or asked what emissions would be avoided by systems that prevent damage escalation. The system response phase, the chaotic period during and immediately after a hazard when a building is operating below its functional threshold, was nearly absent across all hazard groups. In other words, the environmental cost of the disruption itself, the emergency fuel, the temporary shelter, the evacuation transport, falls outside every LCA system boundary in the reviewed literature.

The root of the problem, Rios argues, lies in the functional unit, the quantified description of service against which all environmental flows are normalized. In standard practice, that unit is static: one square meter of office space for fifty years, for example. This definition assumes the building delivers its service completely and continuously, an assumption that hazard exposure systematically violates. If a building loses functionality for six months after a flood, dividing its total life cycle impacts by the full fifty-year service life, as if those six months of failed delivery had never occurred, overstates its environmental efficiency and understates the cost of its fragility. The review found that not a single one of the 40 studies defined the functional unit in terms of sustained service delivery through disruption events.

To close this gap, the paper introduces a resilience-extended LCA framework, or RES-LCA, built on a reformulation of the functional unit as time-integrated functionality. Expressed mathematically as the sum over time of the functionality level Q multiplied by the duration of each time step, the new unit captures the cumulative service actually delivered over the building lifetime under all conditions, including disruption. When a building performs perfectly, the formulation reduces exactly to the conventional one. The difference appears only when service fails, which is precisely when resilience needs to be measured. Under this reformulation, the four resilience dimensions acquire distinct signatures in the functionality curve: robustness governs the magnitude of the initial drop, rapidity the slope of recovery, redundancy the residual service maintained during the disruption, and resourcefulness the shape and duration of the recovery trajectory.

The framework also extends the LCA inventory in two directions. A proposed system response module captures the environmental cost of compensatory services deployed while a building operates below functional threshold, flows that currently escape every assessment. An adaptation module captures the embodied impacts of post-event improvement works, embedding the bounce-forward logic in which a rebuilt building outperforms its pre-disaster baseline. Meanwhile, the often-optional Module D, normally limited to credits for reuse and recycling, is repositioned as a system-wide account of the avoided burdens generated by resilience investments. The approach builds on earlier work by Pizzol, who showed that representing resilience requires reformulating the functional unit and expanding the system boundary, and applies that insight within the existing EN 15978 modular structure that practitioners already use.

An illustrative example shows how the reformulation can flip conventional conclusions. Consider two identical residential buildings in a coastal flood zone, one at grade and one elevated on a concrete podium whose upfront embodied carbon is 30 percent higher. Both experience a single 100-year flood. The elevated building suffers no damage; the at-grade building spends two months nearly non-functional and four more months partially functional, requiring repairs and generating displacement-related emissions. Under conventional LCA, the elevated design appears environmentally preferable only if the avoided repair carbon exceeds the podium premium. Under RES-LCA, the compensatory service emissions of the disrupted building enter the calculation, and for a serious flood event, any non-zero displacement burden is enough to tip the balance toward the resilient design. The framework, in effect, makes the environmental value of resilience visible for the first time.

Perhaps the most provocative implication is what this means for the long-assumed trade-off between resilience and sustainability. The review confirms empirically that the tension is real in current practice: timber has the lowest embodied carbon but the highest flood vulnerability, green roofs reduce operational impacts but increase seismic repair flows, and passive cooling lowers energy use but raises overheating risk in extremes. Yet the study argues that this trade-off is, in part, an artifact of a measurement system that undercounts the environmental costs of service disruption. When environmental performance is defined in terms of service actually delivered, resilient buildings cease to look like costly deviations from efficiency and instead emerge as systems that deliver more reliable service per unit of environmental burden. Sustainability and resilience, long treated as competing agendas, become aligned.

The author acknowledges significant limitations. Data on compensatory service flows, phased recovery materials, and adaptation timelines are scarce and inconsistently reported, and no standardized method exists for quantifying building-level resourcefulness. The framework addresses only the technical dimension of resilience, leaving community-scale interdependencies and the social and organizational domains of the TOSE framework for future work, and the functionality trajectories it requires carry substantial uncertainty under a changing climate. Even so, the core message stands: as climate hazards intensify and billions in public infrastructure funding become tied to resilience criteria, the tools used to certify environmental performance must learn to count what happens when buildings fail. A green building that cannot keep its occupants sheltered through a disaster, the study suggests, is not as green as the spreadsheet claims.

Subject of Research: Integrating hazard resilience into building life cycle assessment through time-integrated functionality

Article Title: Resilience in building life cycle assessment: a critical review and framework for time-integrated functionality

Article References: Rios, F. C. (2026). Resilience in building life cycle assessment: a critical review and framework for time-integrated functionality. Journal of Industrial Ecology. https://doi.org/10.1007/s44498-026-00152-7

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00152-7

Keywords: life cycle assessment, building resilience, embodied carbon, functional unit, seismic hazard, flood, heat waves, wind hazards, sustainability, climate adaptation, redundancy, EN 15978