The study, published in the Journal of Industrial Ecology, focuses on the Advanced Engineering Building at the University of Queensland’s St Lucia campus, a 21,000-square-metre facility whose annual operations were assessed for the 2024 reporting year. Rather than treating the building as a static snapshot, the researchers adopted what they call a prospective, attributional, location-based workflow. The functional unit was deliberately simple: one building-year of operations. That framing allowed the team to align the entire inventory with metered annual activity, including main-meter electricity readings, rooftop photovoltaic inverter logs, and waste audits reconciled against monthly hauler records. The result is an assessment grounded not in assumptions but in audited, verifiable data streams.

Technically, the workflow rests on two complementary open-source platforms. The primary calculations were performed in openLCA version 2.5, using the Australian Life Cycle Inventory database, AusLCI version 1.42, and the Intergovernmental Panel on Climate Change 2013 characterisation method for 100-year global warming potential. Two named electricity datasets anchored the scenario design: one representing the current Queensland low-voltage mix and another representing a projected Queensland-2050 mix. Waste flows were mapped to a municipal landfill dataset and a composting proxy for diverted organics, with a landfill emission factor of 0.833 kilograms of carbon dioxide equivalent per kilogram and a composting factor of 0.046. Climate change characterisation factors were held fixed while parameter uncertainty was propagated on the inventory side, a standard choice in probabilistic life cycle assessment.

The second layer of the workflow is where the methodological innovation becomes most visible. A lightweight metamodel built in Brightway2, an open-source Python framework for life cycle assessment, reproduced the same audited foreground algebraically, enabling paired scenario contrasts using common random numbers and a screening sensitivity analysis. Monte Carlo simulation ran with 2,000 iterations per scenario in openLCA and 3,000 in Brightway2, with the Brightway2 seed fixed at 2025 to guarantee reproducibility. Common random numbers matter here because ignoring dependence between scenarios can bias variance estimates and mis-rank sensitivities; by sampling the same random draws in paired runs, the researchers could resolve differences as small as a few tonnes of carbon dioxide equivalent, differences that independent sampling would have obscured within its own precision limits.

The headline result is stark. Under the current Queensland grid, the building’s mean annual operational footprint was 2.151 million kilograms of carbon dioxide equivalent, equivalent to an intensity of 102.4 kilograms per square metre per year. Swapping in the Queensland-2050 electricity background cut that to roughly 1.136 million kilograms, an intensity of 54.1 kilograms per square metre per year, a reduction of about 1.0 million kilograms annually. The diversion scenario, by contrast, moved the needle by only about 5,300 kilograms per year under either electricity background. Contribution analysis confirmed the asymmetry: purchased electricity contributed about 2.1 million kilograms of the annual total, while all waste-related terms combined amounted to roughly ten thousand kilograms, two orders of magnitude smaller.

Sensitivity analysis sharpened the picture further. Using standardised regression coefficients, the team found that energy use intensity dominated the variance in annual emissions under both electricity backgrounds, with a coefficient of approximately 0.95 and a coefficient of determination near 1.00, indicating an almost perfectly linear model. The photovoltaic self-use fraction was a distant second at roughly minus 0.06, while total waste mass and the diversion rate registered coefficients close to zero. In plain terms, the single most powerful lever available to building operators is reducing the energy the building consumes per square metre, followed by maximising on-site solar consumption, with waste adjustments trailing far behind at current volumes of around 50 tonnes per year.

The authors are careful to note that the small waste effect does not mean waste management is unimportant. The audited waste stream at this building is simply small relative to its electricity demand, so even a substantial improvement in diversion yields limited climate leverage. Higher diversion remains valuable for circularity, stewardship, and regulatory compliance, and the accounting choices were deliberately conservative: diverted non-organic materials received no avoided-burden credit in the baseline, and rooftop photovoltaic generation was treated as subtract-only in the diagnostic metamodel. Sensitivity runs that added a photovoltaic life-cycle factor of 0.045 kilograms per kilowatt-hour changed annual totals by less than one percent, and even more generous recycling credits would not have altered the ranking of levers.

What distinguishes this workflow from a conventional assessment is its emphasis on auditability and reuse. The openLCA project was version controlled, dataset versions were archived, and high-precision scenario exports storing means, standard deviations, and fifth-to-ninety-fifth percentile ranges were saved for every scenario. Data quality was screened with a simplified pedigree approach covering temporal, geographical, and technological representativeness: metered electricity data rated as high quality, while audit-based waste data rated as moderate, which informed the wider uncertainty distributions assigned to waste parameters. The researchers describe the result not as a cyber-physical digital twin but as an auditable lifecycle-scenario layer, one that links a fixed, audited foreground to named, versioned background datasets so that the same analysis can be rerun for subsequent reporting years or transferred to comparable institutional buildings.

The findings arrive at a moment when the buildings and construction sector accounts for roughly one third of global final energy use and energy-related carbon dioxide emissions, making it a central battleground for Paris-aligned mitigation. Previous studies of passive houses in Northern Ireland and the United Kingdom have reported even larger emission declines of 58 to 70 percent under decarbonising grids, but those examined low-energy dwellings and, in some cases, whole-life trajectories. The Queensland study’s somewhat smaller figure reflects its gate-to-gate operational scope and a fixed audited foreground. The direction, however, is consistent: future electricity backgrounds can materially alter building-level results, and static baselines risk biasing the estimated value of interventions on long-lived assets.

For campus operators, energy procurement teams, and policy stakeholders, the practical message is unambiguous. Within this boundary, the largest near-term climate gains come from lowering energy use intensity, through measures such as recommissioning and efficiency-targeted performance management, and from securing a lower-carbon electricity supply, rather than from marginal increases in diversion alone. The scenario structure also supports procurement and retrofit planning by expressing operational risk as distributions rather than single values, and by allowing decision-makers to test how grid decarbonisation and efficiency improvements move individual buildings toward organisation-wide targets. The authors caution that their conclusions apply to an electricity-dominated institutional building under an attributional, gate-to-gate scope and should be transferred cautiously, and that other environmental impact categories may respond differently as grids decarbonise. But the workflow itself, with its transparent provenance, fixed seeds, and publicly archived data on Zenodo under a Creative Commons licence, offers a template that any university, and indeed any large institution, could adopt to make its annual carbon accounting both rigorous and repeatable.

Subject of Research: Prospective operational life cycle assessment of energy and waste systems in a university building

Article Title: Prospective operational life cycle assessment of energy and waste systems in a university building: an auditable workflow using openLCA and Brightway2

Article References: Wang, S., & Halog, A. (2026). Prospective operational life cycle assessment of energy and waste systems in a university building: an auditable workflow using openLCA and Brightway2. Journal of Industrial Ecology, 30(4), 2159-2171. https://doi.org/10.1007/s44498-026-00147-4

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00147-4

Keywords: life cycle assessment, prospective LCA, university building, grid decarbonisation, waste diversion, greenhouse gas emissions, openLCA, Brightway2, uncertainty analysis, energy use intensity, AusLCI, campus sustainability

The conventional route to forward-looking characterisation factors runs through Integrated Assessment Models, or IAMs. These are vast coupled simulations of the economy, energy system, land use and climate that researchers use to sketch coherent scenarios of how the world might evolve. IAMs have powered most prospective life cycle assessments to date, including a well-known 2022 study that generated future water scarcity factors using output from the IMAGE model. But IAMs come with baggage. They are computationally expensive, they bundle similar technologies and regions into coarse aggregate categories, and their deterministic pathways can clash with the fine-grained spatial resolution that life cycle assessment demands. The characterisation factors they produce often do not line up neatly with the geographic units used in background inventory databases such as ecoinvent. Engberg and his collaborators hypothesised that a data-driven alternative, one that learns directly from historical records rather than from chains of socio-economic assumptions, could sidestep some of these frictions.

Their target variable was the AWARE factor, the consensus method adopted by the water-focused life cycle initiative WULCA, which expresses the relative availability of water in a region as the ratio between a consumption-weighted world average of available water and the water available in that specific basin. A high AWARE value means water is scarce; the factor can be read as the surface-time equivalent needed to generate one cubic meter of unused water in that place. The official AWARE dataset covers only a single year, 2010, which is far too thin to train a machine learning model. So the team rebuilt the factor from scratch, month by month, from 1960 to 2016, using the global freshwater model WaterGAP v2.2d. For each of roughly 9,700 watersheds they computed an availability-minus-demand quantity that subtracts human water consumption and environmental flow requirements from natural runoff, then normalised it against the world average. Environmental water requirements were weighted according to the scheme of Pastor and colleagues, protecting low-flow months more stringently than high-flow ones. The same boundary conditions as the original method, capping factors at 100 and flooring them at 0.01, were applied throughout.

Before any forecasting began, the team checked whether their reconstructed historical factors were trustworthy. Compared against the published 2010 AWARE values, their WaterGAP-based calculations showed a correlation of 0.81, high but not perfect, largely because the original method relied on an older, inaccessible version of the hydrological model. More telling was the comparison with the IMAGE-based factors of the earlier IAM study, which correlated at only 0.53 with the original values. Many basins that were genuinely water-rich appeared substantially scarcer in the IAM-derived dataset. This discrepancy matters: it suggests that the choice of underlying model and scenario machinery can shift characterisation factors dramatically, an uncomfortable truth for a field that treats these numbers as authoritative multipliers.

With the historical time series in hand, the researchers turned to model selection. They benchmarked several algorithms, including the gradient boosting methods XGBoost and LightGBM, alongside deep learning architectures designed for time series. The winner was N-Beats, a neural network architecture built from stacks of fully connected blocks that decompose a signal into interpretable trend and seasonality components. N-Beats is what forecasters call a global model: rather than fitting one model per basin, as classical approaches like ARIMA would require across thousands of series, it learns shared temporal patterns across all basins simultaneously. The deployed network used 24 stacks of two blocks each, with 136 units per layer, a learning rate of 0.001 and a batch size of 1024. Training was disciplined with early stopping, which halts the process when validation loss stalls for five consecutive epochs, and a learning rate scheduler that halves the rate whenever performance plateaus. The input window spanned 72 months, six full annual cycles, long enough to capture hydrological seasonality without dragging in outdated patterns that invite overfitting.

Performance was measured with the symmetric mean absolute percentage error, or sMAPE, a metric that penalises over- and under-forecasting equally and ranges from 0 to 200 percent. On the 1,463 largest basins, those spanning at least three half-degree grid cells, the N-Beats model achieved a median sMAPE of about 25 percent, and more than 60 percent of basins came in below 30 percent. Some basins were forecast with remarkable precision, errors near 3 percent, while the worst performers exceeded 50 percent and occasionally approached 78 percent. XGBoost and LightGBM fared worse, with maximum errors above 100 percent and a larger share of poorly predicted basins. On raw accuracy, the deep learning approach clearly outclassed both its statistical ancestors and its tree-based rivals.

But accuracy on monthly values tells only half the story, and it is the half that matters less. When the team smoothed the forecasts into 30-month moving averages to reveal long-term trajectories, cracks appeared. The models tracked seasonal oscillations faithfully, yet their multi-year trends often diverged substantially from the validation data. In the Volga basin, the gap between forecast and reality at the end of the validation period reached roughly 20 cubic meters of world-equivalent water per cubic meter consumed. Even some basins with seemingly excellent sMAPE scores showed poor alignment in their long-term behaviour, a reminder that a model can nail the seasonal rhythm while missing the underlying melody. Attempts to forecast annual values directly produced unsatisfactory results altogether.

The final deployment, trained on the full 1960 to 2016 record, projected AWARE factors out to 2032. According to the forecasts, more than 70 percent of basins will show lower water scarcity in 2030 than in 2010, while roughly 400 basins that already had factors above 60 were predicted to climb toward nearly 100, the ceiling of the scale. Comparing these projections with the IMAGE-based factors for 2030 revealed substantial disagreement in many regions, particularly near the equator, in eastern North America and in western Australia, while South America, Europe and Central Africa showed closer agreement between the two methods. The authors are careful to note that the comparison is imperfect, since the two approaches rest on different hydrological models and the IAM version embeds explicit socio-economic pathways, in this case a middle-of-the-road scenario, whereas the machine learning forecasts simply extrapolate past dynamics with no policy or demographic assumptions baked in.

That absence of assumptions is both the method’s defining feature and its Achilles heel. Time series forecasting, by construction, cannot anticipate structural breaks: a new water management regulation, a leap in irrigation efficiency, a wave of desalination plants, or an unprecedented drought all lie invisible to a model trained only on historical scarcity values. The authors are refreshingly blunt about the implications. Machine learning forecasts, they conclude, are not a standalone alternative to IAMs; they excel at capturing seasonal patterns but cannot reliably predict the long-term structural trends that matter most for prospective life cycle assessment. They also flag a subtle but significant finding for the field at large: the world-average availability term that anchors the AWARE method shifted by about 30 percent between 2010 and 2011, which alone moves every basin’s factor by the same magnitude. Even without any machine learning, that observation argues for treating single-year characterisation factors with caution and for embedding historical trends into impact assessment.

Still, the study opens doors that IAMs cannot reach. The approach could be extended to other characterisation factors where local historical time series exist, such as land use change, and the authors point to covariates from climate model ensembles like ISIMIP and CMIP, variables such as surface temperature, groundwater recharge and leaf area index, as a route to forecasts that respond to climate signals rather than merely repeating them. Transfer learning offers another avenue: a pretrained network could be adapted by practitioners with local data to generate factors for regions the global model never saw. The recently updated AWARE 2.0 method, with improved consumption and environmental flow calculations, may provide cleaner training data for follow-up work. For now, the message is one of calibrated optimism. Deep learning can render the seasonal pulse of the planet’s watersheds with impressive fidelity, and it does so at a spatial resolution that scenario models struggle to match. What it cannot yet do is tell us where the current of change is heading when the change itself has never happened before. Bridging that gap, between pattern recognition and genuine foresight, is the next challenge for anyone hoping to forecast the future state of the environment.

Subject of Research: Machine learning forecasting of water scarcity characterisation factors for prospective life cycle assessment

Article Title: Forecasting future states of the environment with machine learning: a case study on water scarcity

Article References: Engberg, N., Blanco, C. F., Barbarossa, V., Bakker, C., Balkenende, R., Lian, J. Z., & Sprecher, B. (2026). Forecasting future states of the environment with machine learning: a case study on water scarcity. Journal of Industrial Ecology, 30(4), 1285-1298. https://doi.org/10.1007/s44498-026-00010-6

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00010-6

Keywords: machine learning, water scarcity, life cycle assessment, AWARE, time series forecasting, N-Beats, characterisation factors, integrated assessment models, WaterGAP, hydrology, prospective LCA, deep learning