Wind-farm design and operation are traditionally approached through economic objectives such as annual energy production and levelized cost of energy. The study notes that this focus can overlook environmental impacts. Wind energy has substantially lower greenhouse-gas emissions than fossil-fuel-based plants, but those emissions are not negligible when assessed over the entire life cycle.

The research therefore examines whether a carbon-related objective can be introduced alongside conventional objectives during design and control optimization. The analysis treats wind-farm performance as a combined design problem. One component is layout optimization, which determines turbine positions.

The study also considers co-design, in which layout and control strategies are optimized together rather than handled as separate, sequential tasks. This approach allows the analysis to compare the performance of a farm designed with attention to cost or energy against one designed with attention to carbon. Across the illustrative examples and the offshore cases, the main comparison was unexpectedly favorable.

Designs selected using a carbon-related objective achieved economic performance very similar to that obtained with conventional economic objectives. The reverse pattern was also reported: designs optimized using economic objectives performed similarly when assessed against the carbon-related objective. In the authors’ interpretation, these modeled results indicate that prioritizing carbon need not necessarily compromise economic performance, and vice versa.

The similarity in overall performance does not mean the two objectives generate identical wind farms. The researchers found that turbine positions differed depending on the selected objective and on whether layout and control were optimized jointly or sequentially. Thus, a carbon-focused design and an economics-focused design may reach comparable aggregate results while placing turbines in different locations.

The finding concerns compatibility between the objectives, not a single turbine arrangement that the study identifies as universally optimal. The offshore case studies were evaluated under varying site conditions. This was the study’s largest-scale application and provided an additional test of whether the carbon-oriented strategy could retain economic performance in a large wind-farm setting.

The reported comparison used the modeled performance measures of the optimized designs; it does not establish that a particular layout would have the same results in operation, nor does it extend beyond the sites and assumptions represented in the analysis. Instead, it shows that the two optimization approaches can produce similar economic performance in the cases examined, with the most visible difference appearing in the locations selected for turbines. The work is a preprint in Wind Energy Science Discussions , received on 30 September 2026 and posted for discussion on 20 October 2026.

At the time of the posting, it was under review for the journal Wind Energy Science . The findings should therefore be understood as preliminary research awaiting peer review and public discussion. The authors’ conclusion is bounded by the case studies, modeled objectives, site conditions and life-cycle assumptions used in the study: within those scenarios, carbon-related and conventional economic optimization achieved comparable economic performance, while yielding layouts with different turbine positions.

Subject of Research: Wind-farm layout and control optimization for carbon and economic objectives

Article Title: Is a cost-optimal wind farm also carbon-optimal? A layout optimization and control co-design study

Article References: Baricchio, M., Kainz, S., van Wingerden, J.-W., & Bottasso, C. L. (2026). Is a cost-optimal wind farm also carbon-optimal? A layout optimization and control co-design study. https://doi.org/10.5194/wes-2026-178

Image Credits: AI Generated

DOI: 10.5194/wes-2026-178

Keywords: wind energy, wind-farm optimization, carbon emissions, life-cycle assessment, offshore wind, turbine layout, flow control, levelized cost of energy, co-design, preprint research, cost-optimal, wind

The new study arrives at a moment when the economics of cable layout have become harder to ignore. As wind farms grow in capacity and are pushed into increasingly congested waters, the share of total expenditure tied up in the inter-array network rises accordingly. The authors point to figures suggesting that, in bottom-fixed offshore projects, the collection cables can represent on the order of eight to nine percent of the levelised cost of energy, a proportion large enough that even modest percentage savings in trenching and conductor expenditure translate into meaningful reductions in the overall cost of electricity delivered. That economic weight is precisely what motivates the multi-stage structure of the methodology, which treats the cable network not as an afterthought to turbine placement but as an optimisation problem in its own right, with its own distinct layers of spatial, topological, and electrical decision-making.

One of the more instructive aspects of the work is the way it separates the problem into stages that mirror how real projects are actually engineered. Before any routing algorithm can run, the physical site must be translated into a mathematical graph: nodes for turbines, substations, and candidate waypoints; edges for the corridors in which cable may plausibly be laid. The paper is careful about this preparatory step, distinguishing between hard-exclusion zones, where cable installation is simply prohibited, and soft-exclusion zones, where routing is possible but undesirable and therefore carries a cost. This distinction, the authors argue, is often glossed over in the literature, where spatial restrictions are either penalised loosely or handled only after a topology has already been fixed. By embedding terrain profiles and clearance requirements around fixed structures directly into the graph-construction stage, the methodology ensures that the routes considered by the optimiser are ones that could plausibly be built, rather than straight-line connections that later prove impossible.

The routing stage itself is formulated as a mixed-integer linear programme, a choice that brings both rigour and practical limitations. Mixed-integer formulations can guarantee optimality for the model as stated, which is a significant advantage over purely heuristic approaches such as genetic algorithms or ant colony methods that have featured prominently in earlier collector-system studies. Yet the combinatorial explosion inherent in network design means that the formulation must be carefully scoped. The authors restrict attention to radial topologies, in which turbines are arranged in string-like chains without bifurcations at turbine nodes, and they impose constraints on the maximum number of turbines per string. These restrictions reflect common industry practice, where radial strings simplify protection coordination, installation logistics, and maintenance planning, even though alternative topologies such as single loops or star configurations can offer higher redundancy at additional cost.

A notable contribution is the treatment of balance. In irregular sites, particularly onshore projects or offshore farms constrained by shipping lanes and other maritime activities, turbine positions are rarely uniform, and the natural clustering of turbines can produce strings of very different lengths feeding substations unevenly. The methodology includes a procedure for determining the minimum number of substation connections required under such irregular distributions, along with the ability to enforce a minimum number of turbines per string when the user requires it. This kind of balancing matters for reasons that go beyond aesthetics: uneven loading across substations affects transformer specification, installation sequencing, and the flexibility available to operators when individual strings must be taken offline for repair.

The electrical dimension of the problem is handled through conductor size selection, which the authors approach via a linear approximation of the power flow. Full nonlinear power flow calculations, while accurate, are computationally expensive when embedded inside an integer optimisation loop, so the linear approximation serves as a tractable surrogate that still captures the essential relationship between current loading, losses, and conductor capability. An interesting wrinkle in the formulation is the ability to limit the number of distinct conductor types used across the network. Standardising on fewer cable cross-sections simplifies procurement, reduces spare-part inventories, and eases installation logistics, since a single cable-laying vessel configuration can handle more of the campaign. But the paper shows that this standardisation comes at a hidden price: restricting the palette of conductor types may exclude the larger cross-sections that would otherwise be economically attractive on heavily loaded routes, and this exclusion in turn changes which routing configurations are feasible, ultimately altering the conductor types selected. The optimisation therefore navigates a genuine coupling between electrical standardisation and spatial topology, rather than treating them as separable sub-problems.

This coupling is what gives the comparison between the two solution strategies its analytical interest. In the integrated approach, routing and conductor sizing are solved together, allowing the optimiser to trade a longer trench against a cheaper conductor, or a shorter route against a more expensive cable, in a single coherent objective. In the sequential approach, the routing problem is solved first, and conductor sizes are then chosen for the fixed topology. The sequential decomposition is computationally lighter and, as the authors note, offers greater modelling flexibility when routing problems are heavily constrained, since the routing stage can be solved with techniques suited to difficult feasibility questions before the electrical sizing is layered on top. The evaluation, which uses current turbine ratings and the actual cable routings of existing onshore and offshore projects, finds that the integrated formulation generally achieves lower total costs, which is the expected outcome when a joint search space is explored without decomposition penalties. Yet the sequential approach remains valuable precisely because real projects are rarely free to redesign their topology from scratch; constrained sites, existing infrastructure, and phased development schedules often fix large parts of the layout in advance.

Perhaps the most practically significant finding is that routings which minimise trench length do not necessarily minimise total investment cost. This result cuts against a common simplification in early-stage design, where the shortest total cable length is treated as a reasonable proxy for the cheapest network. Cable cost per kilometre varies substantially with conductor cross-section, and a shorter route that forces higher loading on expensive large-section conductors, or that pushes the design toward more numerous strings with more substation connections, can easily cost more overall than a longer route using smaller cables. The paper’s demonstration that this effect interacts with the number of permitted conductor types adds a further layer: the standardisation constraint does not merely add cost directly, it reshapes the feasible routing space itself, so that the topology chosen under a restricted cable palette may differ fundamentally from the topology chosen when all cross-sections are available. For developers and engineering contractors, this suggests that trench-length minimisation and conductor standardisation should be evaluated jointly rather than sequentially as independent design rules of thumb.

The methodological choices also sit within a broader research landscape that the paper situates carefully. Earlier approaches to collector-system optimisation have drawn on minimum spanning tree and travelling salesman formulations, on metaheuristics of various kinds, and on mixed-integer programming with directed arc-pairs in which the selected direction of each candidate connection follows the flow of energy toward the substation rather than being fixed by geometry. Steiner nodes have been introduced to model detours around exclusion zones, and navigational meshes built from Delaunay triangulations have been used to construct feasible paths around obstacles. The present work distinguishes itself by insisting that the construction of the candidate graph, the enforcement of string-like radial structure, the balancing of turbines across strings and substations, and the clearance requirements around fixed structures be handled simultaneously and explicitly, rather than patched together through post-processing or iterative constraint addition.

For the wind industry, the practical implications are straightforward even if the mathematics is not. As sites move into deeper waters, more crowded sea areas, and terrain-challenging onshore locations, the cost of getting the collection system wrong grows. A methodology that can take a real site, with its real exclusions and its real terrain, and produce a cable network that is simultaneously buildable, electrically compliant, and near-optimal in cost, addresses a genuine gap between academic optimisation studies and the constraints that dominate actual project engineering. The open publication of the work, and its grounding in the routings of existing projects rather than purely synthetic cases, should make it easier for practitioners to assess how the approach would behave on their own site geometries, and to judge where the integrated and sequential strategies each deserve a place in the design workflow.

Subject of Research: Optimisation of wind farm inter-array cable routing and conductor sizing

Article Title: A multi-stage methodology for wind park inter-array cabling: graph preparation, routing, and sizing

Article References: Castro Valerio, B., Gebraad, P. M. O., Cheah-Mane, M., A. Lacerda, V., & Gomis-Bellmunt, O. (2026). A multi-stage methodology for wind park inter-array cabling: graph preparation, routing, and sizing. Wind Energy Science, 11(10), 3887-3917. https://doi.org/10.5194/wes-11-3887-2026

Image Credits: AI Generated

DOI: 10.5194/wes-11-3887-2026

Keywords: wind energy, offshore wind, cable routing, mixed-integer optimisation, conductor sizing, graph theory, wind farm design, levelised cost of energy, inter-array cables, electrical collection system, spatial constraints, optimisation

A new review published in Frontiers in Energy Research has now taken the most comprehensive look yet at how researchers actually model those costs. Engineers at the University of Wollongong in Australia systematically surveyed 99 peer-reviewed studies published mainly between 2000 and 2026, covering cost models for both fixed foundations such as jackets and monopiles and floating platforms including spar buoys, semi-submersibles and tension leg platforms. Their verdict is striking: the field’s favourite cost model, in two out of three cases, is also its crudest. Material costs appear in 98 percent of the studies reviewed, but 66.7 percent of studies rely on material cost alone, essentially equating the price of a platform with the price of its steel.

That weight-based approach has an obvious appeal. It requires nothing more than structural mass and a steel price, which makes it ideal for early-stage design studies and for comparing optimisation algorithms. The review documents how researchers have used it to pit genetic algorithms against particle swarm optimisation, colliding bodies optimisation, simulated annealing, ant colony methods and gradient-based sequential quadratic programming, all racing to shave tonnes off jacket legs and braces. It also underpins topology optimisation, where material is redistributed within a design space rather than resized member by member. But the simplicity comes at a price, because the mass of steel tells you almost nothing about what it actually costs to turn that steel into a finished structure bolted to the seabed.

Manufacturing is the most glaring omission. Processes such as blasting, cutting, welding and painting can account for more than 30 percent of the total production cost of welded steel structures, and for offshore wind turbines specifically, welding and bending make up 46 percent of manufacturing costs. Yet only 29 percent of the reviewed studies include manufacturing at all. The detailed models that do exist mostly descend from a time-based costing framework developed by Jármai and Farkas, which estimates the hours needed for preparation, assembly, tacking, welding, flattening, painting and cutting, and multiplies them by labour and equipment rates. Adaptations of this method have been applied to ring-stiffened cylindrical shells for platform columns, spar hulls and semi-submersibles, while other studies estimate weld volumes from the geometry of tubular joints or use activity-based costing borrowed from shipyards. Simpler studies multiply material cost by a manufacturing complexity factor, with values ranging from 1.1 for a spar buoy to 4.0 for a jacket relative to a monopile baseline.

Transport and installation fare even worse. These operations typically consume 10 to 16 percent of total CAPEX, yet they appear in only 16 and 20 percent of studies respectively. The review attributes this neglect to the sheer difficulty of modelling them: costs hinge on vessel day rates, weather windows, towing distances and mobilisation fees, all of which fluctuate with unsteady market conditions. The most rigorous approaches calculate vessel-based costs from day rates, travel time and weather downtime, distinguishing between floating platforms that can be towed to site by tugboats and fixed foundations that demand specialised barges and heavy-lift vessels. At the other extreme, some studies lump transport and installation into a single term pegged to structural mass or to turbine capacity, at rates such as 2.5 million euros per megawatt for spar platforms. Most strikingly, only 14.1 percent of the reviewed studies include all four CAPEX components together.

To make the comparison rigorous, the Australian team introduced a new quantitative framework combining cost model completeness, modelling detail and structural complexity, producing a composite score that ranges from zero to one. Plotted against the number of CAPEX components each study covers, the results reveal a field clustered in the bottom-left corner: most work considers material cost only, and even studies that cover more components rarely model them in full detail. The top-right corner, representing complete CAPEX coverage with fully detailed modelling of every component, is empty. Fewer than 5 percent of studies combine all four components with high detail scores, and none achieved a fully detailed treatment across the board.

The review also exposes asymmetries between fixed and floating technology. Offshore wind dominates the literature, appearing in 77 percent of the studies, and jackets and semi-submersibles attract the most attention because their complex welding and assembly make cost modelling more critical. Yet internal structures of floating platforms, such as stiffeners and bulkheads that are essential for structural integrity, show up in only 12 percent of CAPEX calculations, always alongside the hull, which itself appears in 96 percent of models. Mooring lines are considered in half of the floating platform studies and anchors in just 30 percent. On the fixed side, pile detail is included in only a small share of jacket cost models, and monopiles and tension leg platforms remain comparatively understudied despite their industrial importance.

Uncertainty adds another layer of neglect. Every CAPEX model, however detailed, ultimately depends on steel prices, labour rates and vessel day rates that swing with markets and geography, yet most studies treat these inputs as fixed constants. Only a handful of the reviewed works tested the consequences, using probabilistic distributions, bounded sensitivity ranges of around 10 percent, or one-at-a-time parameter sweeps. Across those that did, material and substructure costs emerged most often as the dominant driver of the final estimate. The review argues that robust optimisation, in which cost uncertainty is built directly into the design process rather than examined afterwards, remains largely unexplored for offshore support structures.

The recommendations that follow are straightforward. Future cost models should cover manufacturing, transportation and installation together rather than defaulting to steel mass, treat internal floating platform components as standard rather than optional, adopt detailed vessel-based transport and installation models especially for fixed foundations, model jacket piles explicitly, and extend established manufacturing costing methods beyond the dominant Jármai-Farkas lineage. The authors also call for cutting, assembly and coating to receive more consistent attention, since these operations shape real fabrication bills even if welding dominates.

Why does this matter beyond the journals? Steel jacket support structures alone can account for 19 percent of the total cost of an offshore wind farm, and as floating wind pushes into deeper waters with new platform concepts appearing every year, the economic penalty of optimising against an incomplete cost picture grows. A design that minimises tonnes of steel may unwittingly inflate welding hours, barge days or weather-exposed installation windows. By mapping exactly where the literature’s blind spots lie, this review gives designers and researchers a checklist for building cost models that reflect what an offshore project actually costs, a small change in method that could translate into large savings across a rapidly expanding global industry.

Subject of Research: CAPEX cost models for steel support structures of offshore energy platforms

Article Title: Cost-driven design and optimization of steel support structures for offshore energy platforms: a review of CAPEX cost models

Article References: Asgari Motlagh, A., McCarthy, T., & Ahmed, A. (2026). Cost-driven design and optimization of steel support structures for offshore energy platforms: a review of CAPEX cost models. Frontiers in Energy Research, 14, Article 1917780. https://doi.org/10.3389/fenrg.2026.1917780

Image Credits: AI Generated

DOI: 10.3389/fenrg.2026.1917780

Keywords: offshore wind, CAPEX, cost models, steel structures, jacket platforms, monopiles, semi-submersibles, floating wind, manufacturing costs, installation, structural optimisation, offshore energy

The work, published as a brief communication in the discussion-stage preprint series of the journal Wind Energy Science, addresses a stubborn gap in wind farm modeling. Full Reynolds-Averaged Navier-Stokes simulations, known as RANS, capture the physics of turbine wakes with respectable accuracy by solving averaged equations of fluid motion. But they are expensive. At the other end of the spectrum, engineering wake models run in milliseconds yet gloss over crucial details, particularly in the induction zone, the region upstream and around a rotor where the turbine begins decelerating and deflecting the incoming wind before it has even reached the blades. FM-RANS is designed to span that entire range in a single seamless calculation.

The key insight lies in how the new model handles pressure. Conventional parabolic RANS methods march the solution forward through the domain in the streamwise direction, which makes them dramatically faster than elliptic solvers that must iterate over the whole flow field simultaneously. The catch is that forward marching requires neglecting upstream pressure influences, and pressure gradients are precisely what shape the near-wake structure behind a turbine. In practice, modelers have patched this weakness with ad hoc near-wake corrections, bolting analytical fixes onto the numerics to fake the physics that was thrown away. Those corrections introduce discontinuities and tunable parameters that undermine confidence in the predictions.

Blondel and Bastankhah took a different route. Instead of discarding pressure physics and repairing the damage afterward, they coupled the parabolic marching framework with a recently developed analytical pressure-gradient formulation based on actuator disk theory. This analytical solution describes how the rotor imposes a pressure field on the flow, allowing the pressure gradient term to be reintroduced exactly where it matters, in the vicinity of the disk, without resorting to an expensive elliptic pressure solve. The result is a model that honors the physical mechanism of wake formation while retaining the raw speed of a forward-marching scheme.

The second pillar of the approach is turbulence closure, the mathematical recipe that tells the averaged equations how turbulent eddies mix momentum through the wake. Many fast wake models lean on oversimplified closures that cannot cope with the intense velocity gradients immediately downstream of a rotor, where wind speed can drop abruptly and shear is ferocious. FM-RANS instead adopts the k–ε–fp closure developed by van der Laan and colleagues, a variant of the classic two-equation turbulence model augmented with a correction term specifically calibrated for the steep gradients of turbine near-wakes. Choosing a closure that is genuinely suited to the flow regime, rather than merely convenient, is part of why the authors can dispense with near-wake corrections altogether.

The payoff is continuity. Where earlier parabolic models produced a patched-together picture, accurate far downstream but distorted near the rotor, FM-RANS yields a single smooth solution from the induction zone, through the near wake, into the far wake. That matters more than it might sound. The induction zone governs how much wind each turbine effectively intercepts and how it blocks its neighbors’ inflow in tightly packed arrays, while the far wake determines power losses hundreds of meters downwind. A model that captures both without seams lets engineers optimize layout, yaw strategies, and control schemes with consistent physics throughout the farm.

Validation focused on the toughest benchmark the authors could reasonably use as ground truth: full RANS simulations themselves. For an isolated turbine, FM-RANS closely reproduced the reference solutions for velocity deficit and turbulence across the entire wake. The test was then made harder with a column of four turbines, a configuration in which each rotor sits inside the compounded wakes of those upstream, and where small errors accumulate row by row into large discrepancies in the predicted power of the last machine. Again, the fast model tracked the full RANS results with very good agreement, suggesting that the analytical pressure closure retains the essential physics even in the interactive, multi-rotor environment of a real farm.

The performance figures are the headline grabber. Reducing per-core computational cost by more than three orders of magnitude means calculations that would monopolize a processor for hours on a RANS solver finish in seconds or minutes with FM-RANS. That changes the character of what wind farm analysis can be. Layout optimization, which might require evaluating tens of thousands of candidate turbine placements, or control studies exploring thousands of wake-steering yaw combinations, becomes tractable with a model that is physics-based rather than purely empirical. The authors are explicit that this is a proof of concept, presented as a brief communication, but the demonstrated combination of accuracy and speed is the kind of result that tends to reshape a subfield’s default toolkit.

The timing is propitious. Offshore wind is expanding into gigantic arrays where wake losses can reach double-digit percentages of annual energy production, and operators increasingly rely on dynamic wake steering, deliberately angling turbines to deflect their wakes away from downstream machines. Every such strategy depends on knowing where the wake goes and how it recovers, predictions that must be fast enough to run in real time and trustworthy enough to act on. A forward-marching RANS model with sound pressure physics could serve as the computational engine for such applications, sitting in the sweet spot between millisecond engineering formulas and hour-long high-fidelity solvers.

As with any preprint under open review, the findings will now face scrutiny from the community, with the discussion period open and the paper under consideration at Wind Energy Science. The authors note that one co-author sits on the journal’s editorial board, a disclosure handled under the journal’s standard transparency rules. Whether FM-RANS ultimately matures into an industry-standard tool will depend on extensions to three-dimensional terrain, atmospheric stability effects, and full farm scales. But the core demonstration stands: by pairing an elegant analytical description of rotor-induced pressure with a turbulence closure built for wakes, the model shows that the exhaustive and the instantaneous need not be enemies. The air through a wind farm, from first touch to final fade, may soon be simulated in one continuous, lightning-fast sweep.

Subject of Research: Fast parabolic RANS modeling of wind turbine induction zones and wakes in wind farm flows

Article Title: Brief communication: A fast parabolic RANS numerical model with analytical pressure closure for continuous induction-to-far-wake prediction in wind farm flows

Article References: Blondel, F., & Bastankhah, M. (2026). Brief communication: A fast parabolic RANS numerical model with analytical pressure closure for continuous induction-to-far-wake prediction in wind farm flows. https://doi.org/10.5194/wes-2026-172

Image Credits: AI Generated

DOI: 10.5194/wes-2026-172

Keywords: wind energy, wind farm wakes, RANS simulation, FM-RANS, turbulence closure, actuator disk, induction zone, parabolic model, wake modeling, computational fluid dynamics, k-epsilon model, offshore wind

The cores were drilled in July 2020 from the geotechnical drilling vessel MV Normand Flower, operated by Fugro, over just two days of fieldwork. In total, the team retrieved approximately 62 metres of sediment, with the deepest hole reaching 54.3 metres below the lowest astronomical tide. The drilling sites sit within the Hollandse Kust Zuid (HKZ) offshore wind farm zone, an area of roughly 225 square kilometres located 18 to 36 kilometres west of the Dutch coast, where water depths reach about 28 metres. That location was no accident. The team deliberately targeted the wind farm because it lies south of the maximum limits reached by Middle and Late Pleistocene ice sheets, which minimised the chance that glacial erosion had destroyed the sedimentary record, and because the site had already been imaged by an extraordinary wealth of high-resolution seismic survey data collected for wind farm development.

That seismic dataset proved central to the entire endeavour. Between 2016 and 2017, the survey company Fugro conducted comprehensive geophysical investigations on behalf of the Netherlands Enterprise Agency, imaging the shallow subsurface with a subbottom profiler, a single-channel sparker and a high-resolution multi-channel sparker. The single-channel sparker data, collected with a GSO 100-tip source operating at 200 to 300 joules and a peak frequency of 825 hertz, allowed the researchers to map two distinct high-amplitude reflector surfaces in the top 50 metres of sediment. Both reflectors correspond to peat deposits, which have distinctive acoustic properties because organic material attenuates and reflects sound differently from mineral sediments. The upper peat surface is continuous across most of the wind farm site, while the lower peat appears scattered but is attributed to the same stratigraphic level. By targeting boreholes where these peat reflections were present, the team maximised their chances of recovering sediments spanning one or more glacial-interglacial cycles.

Drilling in one of the busiest industrialised shallow seas in the world required careful negotiation of competing interests. The seabed of the HKZ zone is criss-crossed by cables and pipelines, dotted with platforms, and littered with shipwrecks, crashed aircraft and magnetic anomalies, some of which may be unexploded ordnance dating from the two World Wars. The researchers had to map all of these hazards before selecting safe drilling locations. Timing also mattered: final site investigations for the wind farm were underway in 2020, leaving only a narrow window before turbine construction began in July 2021. The wind farm, which now contains 139 turbines with a capacity of 1.5 gigawatts, enough to power 1.5 million Dutch households, was inaugurated in September 2023 and has been fully operational since December of that year. The scientific cores therefore represent a rare collaboration between academic research and the offshore energy industry, extracted from ground that is now literally the foundation of Europe’s renewable energy transition.

Because of pandemic-era operational restrictions, no scientists were aboard the drilling vessel. Instead, the team relied on remote communication while Fugro’s crew handled positioning and drilling. Surface positioning used two StarPack global navigation satellite system receivers, with underwater positioning provided by a Kongsberg Simrad HiPAP 500 ultra-short baseline system. Water depths were measured by multiple independent methods, including echosounders, the drill string itself, and conductivity-temperature-depth probes, and all depths were reduced to the lowest astronomical tide. Open-hole drilling removed the well-documented Holocene sands of the upper seabed, after which semi-continuous liner sampling began at 10 to 12 metres below the seafloor using thin-walled Shelby tubes 7.62 centimetres in diameter. The recovered cores were then transported to the British Ocean Sediment Core Research Facility at the National Oceanography Centre in Southampton, where they were split, imaged with X-ray radiography and laminography, and archived. Since 2021 they have been held in cold storage at the University of Leeds.

The sedimentary sequence inside the cores tells a striking story. Each of the five cores contains an upper and a lower peat unit, separated by semi-consolidated clays, silts, sands and gravels. The lowermost peat is primarily in situ woody peat, but its thickness varies dramatically between sites: in core RISeR_S1 it is less than 0.3 metres thick, an order of magnitude thinner than the roughly 3 metre thick peat deposit in RISeR_S7. In several cores the contact between the lower peat and overlying sediments is abrupt and erosional, while in others it grades gradually into clays, suggesting different histories of flooding and burial. Below the lower peat, the team identified a palaeosol, an ancient soil horizon, recognised by rooting features visible in X-ray laminography, a decalcified upper zone underlain by levels enriched with millimetre-scale calcium carbonate nodules, and distinctive steel-blue hues. These soils formed on land, in the open air, and their presence deep in the seabed is unambiguous evidence of a terrestrial environment.

Integrating the core descriptions with the seismic reflection data revealed how far these units extend laterally. The surface of the upper peat can be traced in the seismic profiles between cores RISeR_S1 and RISeR_S7, and between RISeR_S4 and RISeR_S6, even where a small channel form interrupts the continuity. However, a large channel incision separates the eastern part of the wind farm, where cores RISeR_S5 and RISeR_S6 were drilled, from the western part, so the team cannot yet confirm by seismic tracing alone that the peats on either side represent the same surface. The lower peat is laterally discontinuous between all the core sites, meaning correlations must rely on facies associations with the overlying and underlying units. The researchers caution that the vertical resolution of the multi-channel sparker data is limited to 50 to 100 centimetres, so thinner layers may be invisible, and sound-attenuating peat can mask the dominant lithology of adjacent units. These correlations, they stress, must ultimately be tested with biostratigraphical and geochronological laboratory analyses.

The interpretation of the wider sequence is where the findings become genuinely unexpected. While the in situ peats and their underlying palaeosols are unquestionably terrestrial, the intervening sands and clays could reflect fluvial deposits from the Rhine-Meuse system, including floodplains and channel-fills, or intertidal to shallow marine conditions in lower delta plains and lagoons. Notably, the team found no clear evidence of marine inundation within the cored interval, and the general lack of shells suggests open marine environments are not recorded there. The stacking of terrestrial soil-overprinted units without intercalated marine levels, yet preserved 40 to 50 metres below present sea level, requires a depositional model involving fluvio-deltaic cycles, slow subsidence rates and considerable elapsed time. The researchers tentatively assign the stratigraphy to a temporal window between 1.5 and 0.5 million years ago, after the southern North Sea basin was largely infilled by river deltas but before marine inundation during the sea-level highstands of the Middle and Late Quaternary interglacials. This is a marked contrast to the Late Pleistocene, Eemian age of around 129,000 to 116,000 years ago assigned in the initial wind farm site reports.

Establishing a firmer chronology is the team’s next challenge, and it is a formidable one. The sediments are far too old for radiocarbon dating, which reaches only about 50,000 years, and the traditional chronological framework for the Quaternary North Sea rests on pollen zonation schemes from the Netherlands and Britain that are complicated by variable pollen concentrations and the reworking of older taxa. The researchers are now pursuing detailed palynological analysis to tie the cores into the Dutch pollen chronostratigraphy, alongside experimental luminescence dating and palaeomagnetic investigation. Palaeomagnetic work is particularly promising because it could place the deposits before or after the Brunhes-Matuyama reversal at 774,000 years ago, cleanly separating Early from Middle Pleistocene. Together these parallel approaches should cross-validate the ages and the correlations between the five cores.

The stakes extend well beyond academic curiosity. The heterogeneous Quaternary sediments of the southern North Sea form the physical foundation for a rapidly expanding offshore wind industry, and understanding their architecture is critical input for ground models and geohazard assessment. More broadly, new chronologies for the RISeR cores will allow the vast archive of high-resolution seismic data gathered by the wind industry over the past decade to be transformed into modern geological subsurface models, enabling landscape-scale reconstructions of palaeoenvironmental change for chosen interglacials. During glacial lowstands, the exposed land south of the ice margins also served as a corridor for human migration into northwestern Europe, so refining the history of these drowned landscapes has implications for archaeology as well as climate science. In recovering a lost world from beneath the turbine foundations, the RISeR project demonstrates how industry and science can drill toward the same seabed and both come away with something valuable.

Subject of Research: Middle Pleistocene terrestrial stratigraphy recovered from sediment cores in the southern North Sea

Article Title: The RISeR cores: unexpected and extensive Middle Pleistocene terrestrial stratigraphy in the southern North Sea

Article References: McGuire, A. M., Cartelle, V., Rush, G., Busschers, F. S., Cohen, K. M., Hodgson, D. M., & Barlow, N. L. M. (2025). The RISeR cores: unexpected and extensive Middle Pleistocene terrestrial stratigraphy in the southern North Sea. Scientific Drilling, 34(1/2), 29-39. https://doi.org/10.5194/sd-34-29-2025

Image Credits: AI Generated

DOI: 10.5194/sd-34-29-2025

Keywords: North Sea, RISeR cores, Middle Pleistocene, palaeosols, peat deposits, seismic reflection, offshore wind, sea-level change, Quaternary stratigraphy, scientific drilling, Rhine-Meuse, palaeolandscape

The original work, by Simão Ferreira and colleagues, proposed a ceiling on the capacity factor of offshore wind farms, the ratio of actual electricity produced to the maximum a fleet of turbines could generate if running flat out at rated power. The authors built their limit from simplified analytical models: a Weibull distribution describing the wind speed climate at a site, an idealized turbine power curve, and a classical representation of how an infinitely large wind farm interacts with the atmospheric boundary layer, drawing on formulations by Frandsen from 1992 and Sørensen and Larsen from 2021. To make the infinite-farm model usable for real, finite wind farms, they applied a correction that accounts for turbines at the edges of a farm that still see undisturbed freestream wind. The model was then validated against the measured net capacity factors of 72 offshore wind farms across the Baltic, North, and Irish seas.

The stakes became political when the original authors calculated that Dutch national policy for planned offshore wind farms assumes capacity factors significantly in excess of their proposed limit, reportedly by as much as 49 percent when assuming ten percent additional losses. That finding suggested the Netherlands was planning more electricity from its crowded patch of the North Sea than physics would allow, and it fed directly into parliamentary debate. Van der Laan and Watson do not dispute that wake losses, the slowing of wind as turbines extract energy and leave turbulent, depleted flow behind them, are a genuine and growing problem for densely packed seas. What they dispute is whether the proposed ceiling deserves the label “theoretical.”

Their central argument is that the limit rests on heuristic model assumptions and unknown parameters rather than on a fundamental physical constraint. They contrast it with the Betz limit, the celebrated 59.3 percent cap on the power coefficient of a single wind turbine, which follows rigorously from momentum theory. The proposed wind farm limit, by comparison, is an analytical expression for normalized gross annual energy production multiplied by an ad hoc loss factor, and, crucially, different choices of its embedded parameters would yield distinctly different limits. The comment authors also point out that the plotted “theoretical limit” in the original paper actually represents normalized gross annual energy production, a subtlety they say is not clear from the original text and has been frequently misunderstood by readers.

The most damaging critique concerns reproducibility. To apply the infinite-farm model to real wind farms, the original authors corrected it by manually counting the number of freestream turbines at each of the 72 sites, taking into account the farm layout, the prevailing wind rose, and the wakes of neighboring wind farms. This manual procedure was only briefly described, and no scientific method was provided, meaning the results cannot be independently reproduced. Van der Laan and Watson built their own automated version of the procedure, fitting a polygon to the farm boundary, computing outward normal vectors for edge turbines, flagging inflow-edge turbines for each of twelve wind direction sectors, and filtering out turbines shadowed by upstream farms using a ray-casting method. For the Amrumbank West wind farm, their automated approach yielded 11.3 freestream turbines averaged over the wind rose, while the original manual method had reported 15.

The discrepancies run deeper than a single site. In community comments published during the open review process, the original authors themselves attempted an updated automated version of their correction and failed to reproduce their own manual results, with differences exceeding 30 percent for twenty wind farms, including Anholt, DanTysk, London Array, East Anglia One, and Borkum Riffgrund I and II. Van der Laan and Watson show that the weight given to the freestream-turbine count has an enormous influence on the predicted capacity factor, which can slide anywhere between the no-wake-loss gross value and the deep depletion of an infinite farm. When they replot the validation data using alternative correction methods, the quality of the fit degrades substantially, and the high correlation coefficient of 0.87 reported in the original work could not be replicated. A data point can even cross the normalized gross energy line, which the comment authors describe as a serious flaw in the validation.

There is also a measurement problem. The validation relied on net capacity factors, which bundle together wake losses with grid faults, curtailment, and turbine downtime. That mixture makes it impossible to isolate the wake losses the model is supposed to capture. The original authors attempted to correct for this with an additional loss factor of 0.9, but van der Laan and Watson find no compelling evidence for that particular value, noting that the choice directly shifts the apparent limit.

To expose how little of the model’s behavior is truly fixed by physics, the comment authors derived a striking simplification. Although the infinite-farm wake loss equation is implicit, requiring numerical solution of a geostrophic drag law, the 72 wind farms in question share such similar latitudes, hub heights, roughness lengths, and thrust characteristics that the entire expression collapses to a simple explicit formula depending on a single variable: the turbine spacing normalized by rotor diameter. The approximation matches the full implicit calculation to within 2.3 percent for all 72 farms. Likewise, the so-called wind farm wind factor, which condenses wake effects into one number, turns out to depend mainly on just two quantities, the freestream turbine fraction and the turbine spacing. That fragility, the authors argue, undercuts any claim that the model pins down a unique physical ceiling. They add that wake losses are strongly dependent on atmospheric turbulence intensity and stability, so collapsing them into a single variable may not even be possible in principle.

When these uncertainties are propagated to the Dutch policy case, the dramatic 49 percent overestimate dissolves. Van der Laan and Watson note that the references cited by the original study itself give a range of planned capacity densities between 4 and 10.5 megawatts per square kilometer and expected full-load hours between 3700 and 5100, corresponding to capacity factors between 0.42 and 0.58. While 0.58 is indeed optimistic, the lower value of 0.42 sits comfortably within the range of measured capacity factors of the 72 existing wind farms. Moreover, treating a planned 10-gigawatt Dutch build-out as one gigantic uniform wind farm is unrealistic; the more plausible scenario is a cluster of separate 1000 to 2000 megawatt farms built over time with gaps between them, a configuration the analytical framework can actually accommodate with far milder losses.

The comment’s conclusion is blunt: the proposed ceiling should be regarded as a model-derived estimate resting on a non-reproducible finite-farm correction, not a theoretical limit, and given the model’s sensitivity it cannot be used to judge the feasibility of national offshore wind policies. The episode is a cautionary tale for the energy transition, illustrating how a simple spreadsheet-friendly model, precisely because of its accessibility to policymakers, can acquire political force before its assumptions have been scrutinized. Robust tools for estimating large-scale wake losses do exist, from engineering wake models to mesoscale simulations, but they come with complexity and computational cost. As European seas fill with turbines and farms begin to shadow one another, the debate over how much energy the wind can truly deliver is far from settled, and the researchers argue it must be settled with reproducible science rather than ad hoc corrections.

Subject of Research: Critique of a proposed theoretical upper limit for offshore wind farm energy extraction and its reproducibility

Article Title: Comment on “A theoretical upper limit for offshore wind energy extraction” by Simão Ferreira et al. (2026)

Article References: van der Laan, M. P., & Watson, S. (2026). Comment on “A theoretical upper limit for offshore wind energy extraction” by Simão Ferreira et al. (2026). Wind Energy Science, 11(10), 3763-3774. https://doi.org/10.5194/wes-11-3763-2026

Image Credits: AI Generated

DOI: 10.5194/wes-11-3763-2026

Keywords: offshore wind, capacity factor, wake losses, wind farm modeling, atmospheric boundary layer, reproducibility, Betz limit, Dutch energy policy, North Sea, wind energy science, geostrophic drag law, peer review

The stakes are far from abstract. France’s offshore wind sector is projected to reach a cumulative capacity of 3.6 gigawatts by the end of 2027, driven by the commissioning of seven major projects. These turbines and their foundations are engineered to survive more than two decades in one of the most hostile environments on Earth, and their designs rest on a deceptively simple statistical assumption: that the statistics of the ocean, including the mean, the variance, and above all the extremes of wave height, remain stationary over time. Climate change breaks that assumption. Rising sea levels, migrating storm tracks, and shifting seasonal patterns mean that the wave climate a turbine faces in 2080 may be fundamentally different from the one measured during the historical record used to certify it.

At the heart of the study lies a class of statistical tools known as generalized extreme value, or GEV, models. These models describe the distribution of extreme values, such as the maximum wave height recorded in a given block of time, using three parameters: a location parameter that sets the central tendency of the extremes, a scale parameter that controls their spread, and a shape parameter that governs the behavior of the distribution’s tail, which determines how catastrophic the rarest events can become. Traditionally, engineers fit a GEV distribution to annual maxima, extracting the so-called 100-year return level, the wave height expected to be exceeded on average once per century. But annual maxima throw away most of the data, leaving only one point per year and producing wide, often impractically large, uncertainty bands.

Raillard and his team took a different route. Instead of annual maxima, they modeled monthly maxima, which multiplies the available data by a factor of twelve. The catch is that monthly maxima are not identically distributed, because wave climate is strongly seasonal. To handle this, the researchers turned to Generalized Additive Models, allowing the location and scale parameters of the GEV distribution to vary smoothly across the months of the year using spline functions, flexible curves fitted to the data. The shape parameter, notoriously difficult to estimate, was held constant. From the fitted monthly distributions, the team reconstructed the distribution of annual maxima by multiplying together the twelve monthly cumulative distribution functions, an operation justified by statistical tests showing no remaining time dependence in the model residuals once seasonality was accounted for.

The framework then went a step further into genuinely non-stationary territory. The researchers extended their model so that the seasonal cycles themselves evolve over time, using spline surfaces defined over the joint space of month and year. Cyclic P-splines captured the annual seasonality, while thin plate regression splines captured its long-term drift. Fitted independently to each of eight global climate models from the CMIP6 ensemble, under both a low-emission scenario (SSP1-2.6) and a high-emission scenario (SSP5-8.5), the model could shrink gracefully to a stationary form wherever the data showed no evidence of change, while flexibly tracking complex, non-linear evolutions elsewhere. Model selection was performed using the Akaike Information Criterion, and uncertainty was quantified through a Monte Carlo procedure simulating model parameters one thousand times.

The wave data themselves came from a numerical wave model, WaveWatch III, forced by the eight CMIP6 general circulation models, covering a historical period from 1985 to 2014 and a future window from 2071 to 2100 at three-hour temporal and half-degree spatial resolution. Because global climate models are too coarse to represent local conditions accurately, the team applied a statistical bias correction known as the CDF-t method, which maps the cumulative distribution of the global model onto that of a high-resolution local reanalysis, HYWAT for the English Channel and Atlantic and MED-WAV for the Mediterranean. The correction was applied month by month at six representative offshore sites chosen to span the diversity of the French seafront, and it markedly tightened the scatter of the climate models around the reanalysis reference.

The results paint a vivid picture of a changing wave climate. At the representative site in the Eastern English Channel, historical significant wave heights, the average height of the highest third of waves, ranged roughly between 3 and 7 meters. Under the high-emission scenario, that range widens to about 2 to 8 meters, with winter extremes pushing toward 8 meters and summer minima dipping near 2 meters. Perhaps most striking is the temporal redistribution: in most models the calm summer period grows longer and peaks later in the year, with the lowest wave heights shifting from June toward July, and January extremes exceeding December values by 0.2 to 0.5 meters. The seasonal cycle itself appears to be migrating toward the end of the year, hinting that conventional definitions of the seasons may need revision in a warming ocean.

When the researchers translated these shifts into design quantities, the ensemble average of the 100-year return level rose from approximately 6.15 meters in the historical period to 6.75 meters under SSP1-2.6 and 7 meters under SSP5-8.5 at the Eastern Channel site. Across all seafronts, the equivalent lifetime design level, the wave height that should not be exceeded over the operational life of a wind farm, increased almost everywhere: from about 7.4 meters historically to 8.2 meters under the high-emission scenario in the English Channel, and from 11 to 13 meters to 13 to 15 meters along the Atlantic coast. The North Atlantic site showed a clear increase of roughly 2.5 meters. The Mediterranean told a messier story, with weaker seasonal signals, much larger inter-model spread, and some physically implausible outliers, reaching simulated wave heights of up to 20 meters with uncertainties of 30 meters, likely reflecting coarse model resolution and poorly resolved coastal dynamics in that enclosed basin.

Methodologically, the payoff of using monthly maxima was substantial. Because the model is fitted to twelve times more data, the confidence intervals around the return levels narrowed dramatically compared with the classical annual-maxima approach, which the authors note can yield spurious results when data are scarce. Compared with an earlier study reporting an uncertainty of 13.15 meters for the 100-year return level under a comparable scenario, the new method achieved a maximum uncertainty of about 10 meters, a reduction the authors attribute to the larger effective sample size. The team also introduced a new definition of the design condition itself: rather than a fixed 100-year return level, they computed the quantile of the maximum wave height over the entire lifetime of the structure, calibrated so that the cumulative probability of failure over, say, thirty years matches the conventional stationary target.

The implications ripple outward from engineering offices to energy markets. Under-designed structures face reduced operational lifespans, escalating maintenance costs, and, in the worst case, catastrophic failure, while over-conservative designs inflate capital expenditure unnecessarily. By providing a statistically rigorous, uncertainty-aware way to translate climate projections into design loads, the framework offers engineers, designers, and stakeholders a tool for building offshore wind farms that remain both safe and economically viable across decades of environmental change. The authors see their work as a starting point rather than an endpoint, pointing toward refined wave modeling along the French coastline, parametric treatments of evolving seasonality, and joint analyses of wave height and peak period under future climates. As the ocean’s moods continue to shift, the mathematics used to tame them is shifting with it.

Subject of Research: Non-stationary extreme value analysis of projected changes in significant wave height for offshore wind farm design along the French coasts

Article Title: Non-stationary GEV models for estimating design sea-states in a changing climate – applications to offshore wind farms along the French coasts

Article References: Raillard, N., Poppeschi, C., Chevallier, T., Kervella, Y., & Dubus, L. (2026). Non-stationary GEV models for estimating design sea-states in a changing climate – applications to offshore wind farms along the French coasts. Advances in Statistical Climatology, Meteorology and Oceanography, 12(1), 195-210. https://doi.org/10.5194/ascmo-12-195-2026

Image Credits: AI Generated

DOI: 10.5194/ascmo-12-195-2026

Keywords: offshore wind, significant wave height, GEV distribution, non-stationary extremes, CMIP6, climate change, return levels, French coasts, English Channel, Mediterranean Sea, extreme value analysis, wave climate

The core problem pyMOODS addresses is one that optimization researchers have long acknowledged but rarely solved. Multi-objective optimization does not yield one optimal design; it yields an ensemble of Pareto-optimal candidates, each representing a different compromise among objectives such as cost, generation capacity, and environmental impact. Choosing among them is a human judgment call, informed by policy goals and stakeholder priorities. Yet the tools available for this post-optimization stage are fragmented. Algorithmic toolkits like pymoo, PlatEMO, and Borg generate Pareto fronts but offer little for the decision-maker. Multi-criteria decision analysis libraries such as pymcdm and pyDecision rank alternatives once objectives are fixed, but expose no exploratory interface. Visualization tools like PAVED and Parasol support interactive exploration but stop short of preference-weighted ranking. What has been missing, the PNNL team argues, is a single interactive platform that connects the visualization layer to a complete decision-support workflow on fronts that have already been computed.

pyMOODS fills that gap with a decoupled software architecture that separates heavy computation from interactive exploration. The frontend is a React and MUI application rendering linked interactive plots built with React Plotly, while a Python backend handles the analytics. User actions issue asynchronous requests to a Flask REST API, which mediates all communication between the two. Because interaction state lives in the frontend while analysis runs in the backend, computation never blocks the interface. A data access module loads, validates, and caches datasets held as CSV and JSON files, and a data processing module performs transformations and filtering before invoking the analytical routines. Notably, the dashboard evolved from early Plotly Dash prototypes into the current React-based implementation, gaining dynamic rendering of parameters, plots, and objective functions, along with better responsiveness under high-dimensional workloads.

One of the framework’s most elegant design decisions is its declarative input schema. A problem formulation is supplied as a single JSON configuration file that links to the data files and assigns each column to one of five categories: input parameters, hyperparameters, decision variables, objective functions, and control inputs. This means a single deployment can serve formulations whose objectives, decision variables, and hyperparameters differ in number and meaning, without any code changes. That flexibility matters because the framework targets real infrastructure applications, such as offshore wind co-design, MTDC-ESS system design, and load-shedding-constrained network expansion, where problem structures vary dramatically across use cases. It is a practical requirement that benchmark-oriented tools, typically evaluated on synthetic test problems like ZDT and DTLZ, simply do not address.

The analytical heart of pyMOODS combines two very different classes of methods. For structuring the solution space, the framework turns to unsupervised machine learning: a Uniform Manifold Approximation and Projection (UMAP) embedding projects the high-dimensional objective and decision space into two dimensions, and density-based HDBSCAN clustering groups similar solutions within that latent space. This gives users a navigable map of the entire front, where clusters of comparable candidates are visible at a glance and outliers stand out immediately. Users can also plot any pair of objectives or decision variables directly, or restrict the working set to solutions sharing a chosen hyperparameter configuration, such as a particular technology option.

For trade-off interpretation, pyMOODS takes a deliberately different path: a deterministic, parameter-free rank analysis. Every solution is ranked on each objective, and a solution’s generalizability is defined as its worst rank across all objectives. A solution is said to specialize in an objective when its rank there is strictly better than that of every more generalizable solution. Because this characterization derives entirely from the rank matrix, it introduces no trained model, no tuned threshold, and no random seed; identical inputs always produce identical generalizer and specializer assignments. This answers a question that ranking alone cannot: not merely which solutions score well under a given weighting, but which specific compromise each one represents. Solutions can additionally be ordered by an aggregate score reflecting user-specified objective weights, giving stakeholders a familiar mechanism for expressing their priorities, and the ranking is computed from the stored front rather than by re-solving, so stakeholders with different priorities can interrogate the same results interactively.

The framework’s workflow comes alive in its demonstration on the MoCoDo dataset for offshore wind farm planning with battery energy storage. The problem involves six objectives, including battery cost, cable material cost, day-ahead revenue, real-time revenue, and two reserve revenue streams, along with two decision variables: cable capacity and battery rated power. A conventional approach would aggregate all revenue streams into one total and all costs into another, reducing the problem to a simple two-way trade-off, but that aggregation obscures how individual solutions perform across specific revenue mechanisms. pyMOODS instead guides the user through four linked views: a scatter plot offering a first overview of the solution space, a ranking table for selecting a candidate, a parallel coordinates plot distinguishing generalizers from specializers across all six objectives, and beeswarm plots showing the distribution of all candidates per objective and decision variable. Selection propagates across views, so a candidate can be examined from several perspectives without losing context.

Performance testing suggests the architecture delivers on its responsiveness promise. On a local deployment running on an Apple M2 Pro with 16 GB of RAM, the team profiled three representative user tasks across two built-in formulations, Cameo_datacenter and MoCoDo_v2, repeating each measurement three times. Formulation initialization took roughly five milliseconds in both cases, while full dashboard load sequences averaged 3.838 seconds for the larger formulation and 0.475 seconds for the smaller one, with filter and selection interactions showing nearly identical latencies. The release also includes automated frontend tests using Playwright, which verify that dashboard components load correctly, the solution table sorts properly, and the generalizer-specializer controls update results, using simulated backend responses to validate the interface without requiring the full backend to run.

A capability comparison against representative open-source tools underscores what makes pyMOODS distinctive. Front-generation frameworks and metaheuristic platforms are not designed to be driven by a decision-maker; multi-criteria decision analysis libraries rank alternatives but expose no exploratory interface; and interactive visualization tools support exploration but stop short of preference-weighted ranking. In the published comparison, pyMOODS is the only tool that simultaneously offers an interactive user interface, operates on precomputed fronts from any solver, supports user-weighted ranking, provides automatic trade-off characterization, and accepts schema-agnostic problem formulations. The authors are candid about the release’s limitations, however: performance has been characterized on only two formulations on a single machine, rank-and-weight aggregation cannot recover non-convex regions of a Pareto front, and no usability evaluation has yet been conducted. Planned features include additional MCDM methods for comparison and an AI chatbot assistant.

The implications reach well beyond software engineering. As power systems worldwide absorb rising penetrations of onsite energy sources and face intensifying extreme weather, the co-optimization of generation, transmission, and storage investments is becoming one of the defining computational challenges of the energy transition. Tools like pyMOODS aim to ensure that the enormous computational effort poured into optimizing these systems does not end in an unreadable list of trade-offs. Released under the BSD 3-Clause License and archived on GitHub and Zenodo, version v0.0.3 puts a complete post-optimization decision-support pipeline into the hands of the planners, engineers, and researchers who must ultimately decide which compromises our future infrastructure will embody. Whether it changes real-world decisions, the authors acknowledge, is a question that future work, and future users, will have to answer.

Subject of Research: An open-source visual analytics framework for multi-objective decision support in large-scale infrastructure planning

Article Title: pyMOODS: An open-source framework for multi-objective decision support in large-scale infrastructure planning

Article References: Mattoo, P., Pham, J. N., Jain, M., Arendt, D., Avila, P., Ramachandran, T., Adetola, V., Yun, J. Y., & Wenskovitch, J. (2026). pyMOODS: An open-source framework for multi-objective decision support in large-scale infrastructure planning. SoftwareX, 36, Article 103095. https://doi.org/10.1016/j.softx.2026.103095

Image Credits: AI Generated

DOI: 10.1016/j.softx.2026.103095

Keywords: pyMOODS, multi-objective optimization, Pareto front, visual analytics, infrastructure planning, offshore wind, UMAP, HDBSCAN, decision support systems, open-source software, energy systems, Pacific Northwest National Laboratory

Decarbonizing an energy system is, in physical terms, a materials problem. Fossil fuel plants extract energy by burning dense fuels and require comparatively modest amounts of metal per megawatt of capacity. Renewable generators invert that logic: they harvest diffuse energy flows from sun and wind, so they must spread large areas of collectors and converters across the landscape, each kilogram of steel, glass and semiconductor doing less work than a tonne of coal. The result is that a renewable-heavy grid is dramatically more material-intensive per unit of electricity than the fossil system it replaces. Copper for cabling and generators, zinc for corrosion-resistant galvanizing of steel structures, silicon for photovoltaic wafers and rare earth elements for permanent magnets all scale directly with installed capacity.

What has been missing, the authors argue, is a rigorous accounting of what the EU’s specific legal commitments imply for these materials. Previous assessments, including work by the European Commission’s Joint Research Centre and the International Energy Agency, offered broad global or sectoral estimates, but the continental mineral demands of legally binding targets remained poorly quantified. The new study closes that gap by constructing three scenario-based projections of solar, onshore wind and offshore wind deployment across the EU27 through 2050, and translating each capacity trajectory into demand for copper, zinc, silicon and rare earth elements using technology-specific material intensities.

The scenarios matter because the three technologies are not interchangeable in their mineral appetites. Solar photovoltaics are dominated by silicon, silver and aluminum, with relatively little exposure to rare earths. Onshore wind turbines use permanent magnets in many modern designs but can also employ geared or doubly-fed generator architectures that reduce or eliminate rare earth content. Offshore wind is different in kind: the harsh marine environment and the premium on reliability and weight favor direct-drive turbines built around permanent magnet synchronous generators, which depend on neodymium, praseodymium, dysprosium and related rare earth elements. Every additional gigawatt of offshore wind therefore locks in a specific and largely non-negotiable rare earth demand.

The study’s most striking finding concerns exactly this lock-in. The researchers show that pathways prioritizing offshore wind maintain rare earth element exposure near full-target levels, at 69.2 kilotonnes, even when solar and onshore wind deployment falls short of their own targets. In other words, because the rare earths are technologically embedded in direct-driven turbine generators rather than being optional add-ons, a strategy that leans on offshore wind cannot shed its rare earth burden by trimming other technologies. The mineral demand follows the turbine design, not the overall ambition level. This asymmetry means that offshore-heavy pathways concentrate supply-chain risk in a small number of elements whose mining and, crucially, midstream processing are geographically concentrated far outside Europe.

The geographic concentration is where the analysis turns from engineering to geopolitics. The paper’s accompanying figures map world reserves of chromium, copper, manganese, molybdenum, nickel, rare earths and zinc across major reserve-holding countries, underscoring that no single region can supply the full basket of transition metals. For rare earths in particular, the bottleneck is less the existence of deposits than the refining and separation capacity needed to turn ore into magnet-grade material, a midstream stage dominated by a handful of suppliers. The authors, who have previously written in Nature that minerals will shape the future geopolitical order, frame the EU’s position as one of structural exposure: the bloc can diversify its turbine suppliers, but not easily the mineral inputs those suppliers all draw upon.

Supply-chain exposure, the study finds, is highly dependent on the chosen technology pathway. A portfolio that balances solar, onshore and offshore wind spreads demand across a wider set of minerals, each with its own reserve distribution and market structure, reducing dependence on any single chokepoint. An offshore-centric pathway, by contrast, delivers high capacity factors and valuable grid services, as other research has shown offshore wind can reduce the total installed capacity needed in zero-emissions grids, but it does so by concentrating risk in rare earths. The scenarios thus present policymakers with an explicit trade-off between the system-level efficiency of offshore wind and the material resilience of a more diversified build-out.

The authors’ central reframing is conceptual as much as quantitative: the clean energy transition should be understood as a continental materials procurement challenge in which mineral security functions as a binding feasibility constraint, on par with financing, grid infrastructure and public acceptance. If the required minerals cannot be procured on schedule, the legally binding targets are not merely delayed but rendered infeasible, regardless of political will or available capital. This shifts the policy question from how much renewable capacity to build toward whether the material inputs for that capacity can be secured, processed and delivered on a 25-year timeline.

To make that feasibility concrete, the paper proposes three institutional responses. First, member states should institutionalize material stress tests in their national energy and climate plans, subjecting deployment targets to the same kind of supply-chain scrutiny that financial regulators apply to bank balance sheets. Second, the EU should establish strategic component reserves, stockpiling not just raw ores but finished components such as permanent magnets and power electronics, where the true vulnerability lies. Third, the bloc should actively secure international midstream processing partnerships, investing in refining capacity in partner countries and diversifying away from single-source dependence. The recommendations echo the direction of the EU’s Critical Raw Materials Act and Net-Zero Industry Act, but the authors argue these instruments must be tied directly to the legally binding deployment targets rather than treated as industrial policy in the abstract.

The timing of the analysis is significant. Global investment in clean energy has surged, with the IEA’s World Energy Investment reports documenting record capital flows into renewables, and the EU’s Fit for 55 package has converted political ambition into binding law. Yet recent modeling of net-zero pathways and warnings in Science that the energy transition needs entirely new materials both point in the same direction as the new study: the physical substrate of decarbonization is finite, geographically uneven and increasingly contested. As the United States and other powers strike bilateral mineral agreements, the competition for transition metals is becoming a defining feature of twenty-first-century statecraft. For the EU, the message of this research is that gigawatts are the visible currency of the energy transition, but kilotonnes of copper, zinc, silicon and rare earths are the hidden ledger on which its climate commitments will ultimately be judged, and that ledger must be balanced starting now.

Subject of Research: Critical mineral demand projections for EU solar and wind energy deployment to 2050

Article Title: From gigawatts to critical mineral demands for EU solar and wind

Article References: Kammen, D. M., Baranowski, M., Jabkowski, P., & Dutta, S. (2026). From gigawatts to critical mineral demands for EU solar and wind. Nature Energy. https://doi.org/10.1038/s41560-026-02147-x

Image Credits: AI Generated

DOI: 10.1038/s41560-026-02147-x

Keywords: critical minerals, rare earth elements, offshore wind, solar energy, European Union, supply chains, energy transition, copper, zinc, silicon, Nature Energy, mineral security

The research team, led by Adrian Sobral-Lores of the University of Santiago de Compostela together with colleagues in Spain and Chile, examined twenty operational offshore wind farms in the United Kingdom. Their approach wove together life cycle assessment, or LCA, the standard method for tracing environmental impacts from raw material extraction through manufacturing, installation, operation and end of life, with data envelopment analysis, or DEA, a mathematical benchmarking technique that compares each facility against the best performers in the sample. The twist that sets this work apart is a contribution-analysis step inserted between the two: before any benchmarking takes place, the LCA identifies precisely which material flows dominate the environmental profile, and only those materials are fed into the efficiency model.

The results of that hotspot analysis are striking in their clarity. Low-alloy steel, the workhorse metal of towers, foundations and hubs, accounted for 49.37 percent of the average environmental contribution across eighteen impact categories. Copper, the metal of choice for cables and electrical equipment, contributed a further 21.99 percent. Adding chromium steel, glass fibre and aluminium brought the five selected materials to 78.65 percent of the total average environmental contribution. In other words, less than a handful of substances effectively determine whether an offshore wind farm is environmentally lean or heavy, a finding that dramatically narrows where engineers and policymakers should focus their attention.

With these five materials established as the environmentally decisive inputs, the team built a Slack-Based Measure DEA model under variable returns to scale, using lifetime electricity generation as the single output. The choice of model matters: unlike simpler radial approaches, the slack-based formulation captures non-proportional excesses in individual inputs and remains valid when inputs are measured in different physical units. Because the sample contained only twenty facilities, the researchers also respected a classical adequacy rule from the DEA literature, which limits the number of input and output variables relative to the number of units being compared, and they tested the robustness of their model by repeating the analysis with four and then three inputs.

The benchmarking verdict was split. Nine of the twenty wind farms achieved a perfect efficiency score of 1.00, defining the empirical frontier against which their peers were measured. The remaining eleven fell short, with scores ranging from 0.584 for the least efficient facility to 0.795 for the closest of the laggards. Crucially, the model did not merely rank the farms; it produced material-specific reduction targets for each inefficient one, quantifying how much less low-alloy steel, copper, chromium steel, glass fibre or aluminium each facility would need to generate the same lifetime electricity. The largest single gap involved copper at one wind farm, where a reduction of 78.66 percent was projected, while the least efficient farm showed substantial excesses across all five materials simultaneously.

Here the study makes its most methodologically important move. Rather than assuming that material savings translate automatically into environmental gains, the researchers transferred the DEA-derived reduction targets back into the life cycle inventories and ran a second, full LCA on the adjusted inventories, holding electricity generation and all other inventory flows constant. This verification step proved essential, because the environmental consequences of material savings vary enormously depending on the impact category. Median improvements across the eleven inefficient farms ranged from 8.85 percent for fossil resource scarcity to 30.32 percent for terrestrial and freshwater ecotoxicity, and the maximum reduction reached 59.18 percent for both ecotoxicity indicators at a single facility.

The pattern of these gains tells a coherent story about cause and effect. Toxicity, ecotoxicity, acidification and land-use categories, all of which are strongly influenced by copper and other metal inputs, showed the largest improvements, while global warming potential, with a median reduction of 10.62 percent, responded more modestly because the dominant driver, low-alloy steel, was subject to comparatively smaller projected cuts. The sensitivity analysis reinforced confidence in the overall picture: when aluminium and then glass fibre were excluded from the model, all eleven inefficient farms remained inefficient, classification agreement between scenarios ranged from 95 to 100 percent, and the material-reduction targets for the shared inputs were unchanged to the precision reported by the software.

The study also explored what distinguishes efficient farms from inefficient ones, and the correlations it uncovered are intriguing, if exploratory. Efficiency showed a statistically significant positive association with distance from shore in both Pearson and Spearman analyses, and a significant monotonic relationship with the number of turbines. Positive but non-significant trends linked efficiency to installed capacity and estimated electricity generation, hinting that larger farms may achieve higher material eco-efficiency. By contrast, efficiency was essentially unrelated to individual turbine capacity, rolling load factor or year of commissioning. With only twenty observations, the authors are careful to frame these correlations as hypothesis-generating rather than causal, and they acknowledge that because site-specific conditions such as wind resource quality were not embedded as control variables, part of the measured efficiency variation may reflect location and design circumstances rather than pure resource management.

The authors are equally careful about what the reduction targets do and do not mean. The DEA projections represent comparative benchmarks derived from peer farms in the sample, not engineering prescriptions that could be applied directly to a real project. Structural, geotechnical and site-specific constraints, from water depth and seabed characteristics to fatigue and installation limits, determine how much steel a foundation genuinely requires, and the optimal quantity differs for every site. Supporting engineering literature cited in the study suggests real headroom exists: integrated structural optimisation has been shown to cut the mass of a turbine support structure by nearly twenty percent, and alternative foundation designs have delivered life-cycle environmental scores roughly eighteen percent below conventional monopiles. The benchmark targets, in this light, mark the territory where such optimisation efforts are most worth pursuing.

One caveat deserves emphasis for anyone tempted to extrapolate the numbers to tomorrow’s projects. The underlying inventory comes from an earlier generation of UK offshore wind farms, and the authors caution that the efficiency scores and material targets should not be directly applied to modern installations using turbines above fifteen megawatts or floating foundations, whose material requirements can differ substantially. Still, the framework itself, which links hotspot identification, efficiency benchmarking and environmental verification in a single loop, remains applicable wherever comparable inventories exist, and recent assessments of floating wind confirm that steel, cables and composite blades remain the dominant hotspots. As Europe pursues its target of 300 gigawatts of installed offshore wind capacity by 2050, the lesson of this study is that the greenest kilowatt-hour is not only the one generated without emissions, but the one generated with the least steel and copper behind it, and that knowing exactly which materials matter is the first step toward building a wind fleet that is clean from seabed to scrapyard.

Subject of Research: Material eco-efficiency assessment of offshore wind farms using integrated life cycle assessment and data envelopment analysis

Article Title: Assessing the eco-efficiency of offshore wind energy using a combined life cycle assessment and data envelopment analysis approach

Article References: Sobral-Lores, A., Rebolledo-Leiva, R., Feijoo, G., & Moreira, M. T. (2026). Assessing the eco-efficiency of offshore wind energy using a combined life cycle assessment and data envelopment analysis approach. Journal of Industrial Ecology. https://doi.org/10.1007/s44498-026-00191-0

Image Credits: AI Generated

DOI: 10.1007/s44498-026-00191-0

Keywords: offshore wind, life cycle assessment, data envelopment analysis, eco-efficiency, low-alloy steel, copper, United Kingdom, renewable energy, environmental impacts, materials, sustainability, benchmarking