The study, led by Chia-Te Chien of GEOMAR Helmholtz Centre for Ocean Research Kiel and National Taiwan University, together with Markus Pahlow, Christopher J. Somes, Markus Schartau, and Andreas Oschlies, is published in the journal Earth System Dynamics. Rather than running a single simulation with one best-guess set of model parameters, the team built an ensemble of 480 simulations of the Last Glacial Maximum, systematically varying both the environmental boundary conditions of the glacial world and the physiological parameters of the plankton ecosystem model embedded within an Earth system model. The design was aimed squarely at a blind spot that has haunted glacial carbon cycle modeling: most previous studies used a single parameter set calibrated against modern observations and simply assumed those values would hold in a radically different ocean.

The technical foundation of the work is the optimality-based plankton ecosystem model, or OPEM, coupled to the University of Victoria Earth system model of intermediate complexity. Unlike conventional models that fix the elemental composition of marine organic matter to the canonical Redfield ratio of 106 carbon to 16 nitrogen to 1 phosphorus, OPEM allows phytoplankton to flexibly adjust their carbon-to-nitrogen and carbon-to-phosphorus ratios in response to nutrient stress and light. This variable stoichiometry matters enormously for carbon sequestration, because if glacial plankton built organic matter with more carbon per unit of nutrient, the same nutrient supply could export more carbon into the deep ocean and pull down atmospheric carbon dioxide more efficiently.

The ensemble architecture was deliberately layered. The researchers first generated 600 parameter sets, each representing a unique combination of values for 19 parameters governing plankton physiology, detritus remineralization, and particle sinking. Each of the 600 was spun up for more than 10,000 years under pre-industrial conditions with a prescribed atmospheric carbon dioxide level of 284.3 parts per million. A likelihood-based cost function then scored each simulation against observations of phosphate, nitrate, and oxygen from the World Ocean Atlas across 17 biomes, penalizing mismatches in both mean values and spatial and temporal variability. The 20 parameter sets with the lowest costs, all within 1.3-fold of the best, were carried forward. Each was then run under 24 different combinations of glacial boundary conditions, producing the 480-simulation ensemble.

Those boundary conditions captured the major hypothesized differences between the glacial and pre-industrial ocean. On the physical side, the team applied Last Glacial Maximum orbital parameters, wind stress patterns drawn from PMIP3 models featuring a northward shift and intensification of Southern Ocean westerlies, and a halving of meridional moisture diffusivity over the Southern Ocean, which stratifies the surface and weakens deep water formation. On the biogeochemical side, the 120-meter drop in sea level exposed continental shelves, cutting benthic denitrification by roughly 40 percent and slashing sedimentary iron input by about 80 percent, while terrestrial erosion boosted the global phosphate inventory by 15 percent and glacial dust quadrupled atmospheric iron deposition to 6.1 gigamoles of iron per year.

The results were unambiguous about which lever mattered most. Iron supply, whether arriving as dust or released from sediments, exerted the most profound influence on marine biogeochemistry and atmospheric carbon dioxide. Quadrupling atmospheric iron deposition lowered carbon dioxide by about 26 parts per million on its own, while cutting sedimentary iron raised it by roughly 60 parts per million, a swing so large it dwarfed every other forcing tested. By contrast, increasing the phosphate inventory or reducing benthic denitrification, the two classic macronutrient hypotheses, produced only marginal effects. The reason lies in co-limitation: adding one major nutrient simply aggravates limitation by the others, and the model’s flexible stoichiometry partially offsets the gain by allowing plankton to incorporate relatively less carbon per unit of nutrient when nutrients become abundant.

The sedimentary iron finding carries a provocative implication. Because falling sea level stripped away shallow seafloor sources, the total dissolved iron inventory in the glacial ocean was actually about 1 percent lower than before the ice age, even as dust deposition soared. Lower temperatures also reduced productivity and therefore the organic carbon flux reaching the seafloor, further starving sedimentary iron release. Yet atmospheric carbon dioxide still fell, driven primarily by a spatial redistribution of surface dissolved inorganic carbon and altered air-sea fluxes rather than by a wholesale increase in oceanic iron fertilization. Earlier modeling studies that considered only enhanced dust deposition, without the sedimentary decline, may have substantially overestimated the glacial iron effect.

The ensemble also exposed a sobering uncertainty. Under full glacial conditions, atmospheric carbon dioxide decreased by 36 to 58 parts per million across the 20 parameter sets, a spread equal to roughly half the mean drawdown of 43.5 parts per million, even though all 20 sets reproduced pre-industrial biogeochemistry about equally well. Statistical analysis traced the variance to parameters governing phytoplankton nitrogen subsistence, temperature-dependent mortality, and zooplankton grazing, and revealed that strongly correlated parameter pairs could mask or amplify the carbon dioxide response depending on how their values co-varied. Notably, the drawdown achieved by each simulation was unrelated to its cost function score, meaning that a model fitting modern data well tells you almost nothing about how it will behave in a glacial ocean.

The full ensemble accounted for only about half of the observed 90 parts per million glacial drawdown, leaving room for processes outside the model’s scope, such as brine-induced stratification around Antarctica, air-sea carbon disequilibrium, and a larger terrestrial carbon transfer than the simulations produced. Still, the physical forcings alone, colder temperatures plus altered winds and moisture diffusivity, contributed about 35 parts per million, with the biological and stoichiometric effects adding the rest. The team estimates that flexible carbon-to-nutrient ratios contributed an additional 16 to 17 parts per million of drawdown when iron deposition alone is considered, consistent with earlier theoretical estimates.

Beyond solving an ice age puzzle, the findings speak directly to one of the most debated proposals for engineered carbon removal: ocean iron fertilization. The glacial ocean serves as a natural experiment in what happens when iron delivery to the sea surges, and the model’s behavior, including the nutrient-robbing effect in which fertilized regions deplete macronutrients downstream, mirrors the central criticism of iron fertilization as a climate intervention. Intriguingly, the variable stoichiometry feedback, which sustains productivity even as macronutrients decline, does not exist in fixed-ratio models and could partially buffer that limitation. The authors argue that the same parameter and boundary condition uncertainties that complicate glacial simulations apply equally to assessing carbon dioxide removal strategies, making ensemble approaches like this one essential for any credible accounting of what the ocean can, and cannot, be persuaded to store.

Subject of Research: Ensemble modeling of Last Glacial Maximum marine biogeochemistry and the role of the soft-tissue biological carbon pump in glacial atmospheric CO2 drawdown

Article Title: Ensemble simulation of the Last Glacial Maximum marine biogeochemistry and atmospheric pCO2 drawdown due to the soft-tissue biological carbon pump

Article References: Chien, C.-T., Pahlow, M., Somes, C. J., Schartau, M., & Oschlies, A. (2026). Ensemble simulation of the Last Glacial Maximum marine biogeochemistry and atmospheric p CO 2 drawdown due to the soft-tissue biological carbon pump. Earth System Dynamics, 17(5), 1277-1297. https://doi.org/10.5194/esd-17-1277-2026

Image Credits: AI Generated

DOI: 10.5194/esd-17-1277-2026

Keywords: Last Glacial Maximum, biological carbon pump, iron fertilization, marine biogeochemistry, atmospheric CO2, paleoceanography, Earth system modeling, variable stoichiometry, ocean circulation, nutrient limitation, carbon sequestration, ensemble simulation

The motivation comes from a growing catalogue of criticisms aimed at mainstream integrated assessment models, or IAMs. These tools, which feature heavily in IPCC reports and inform negotiations under the UNFCCC, have long been attacked for the extraordinary sensitivity of their results to how damage functions are specified. Estimates of the social cost of carbon can swing dramatically depending on those curves, yet their empirical foundations are thin and their inner workings opaque. Because the damage function is aggregated, researchers cannot trace how a hotter world disrupts specific macroeconomic mechanisms, test individual assumptions, or validate particular channels of harm. FRIDA’s designers set out to make every pathway visible.

The new module rests on four conceptual pillars drawn from economics rather than climate science alone. It is Schumpeterian, meaning that innovation and creative destruction drive growth from within the model rather than arriving as an exogenous trend. It operates in disequilibrium, abandoning the equilibrium assumptions that most IAMs inherit from neoclassical theory and that the authors argue are inconsistent with the deep uncertainty and market failures surrounding climate change. All growth components are endogenous, so productivity, investment and employment emerge from the model’s own feedback loops. And the financial architecture is stock-flow consistent, a bookkeeping discipline ensuring that every asset in the simulated economy is matched by a liability, so that money, debt and credit creation are tracked coherently through every transaction.

Structurally, the economy module contains 761 equations and 33 state variables organised into seven interconnected submodules: circular flow, finance, innovation, government, employment, GDP and inflation. The circular flow simulates production and consumption between households and firms, with income streams from labour and ownership tracked separately because workers and asset owners behave differently, saving and spending at different rates. Firms’ checking accounts receive consumption expenditure and pay out wages, rent and profits. On its own, this circuit would settle into a stable equilibrium; growth enters through bank lending and government spending, which the other submodules govern.

The finance submodule is where climate change first bites in a distinctly monetary way. A single aggregate banking sector extends credit, classifying new loans as performing, nonperforming or exploratory, the last category funding innovative ventures with transformative potential. Lending standards tighten as default rates rise, and the failure rate on new loans is dynamically adjusted for economic conditions and for the surface temperature anomaly itself: as global temperatures climb, extreme weather raises the probability that new investments fail. Sea level rise can sink even otherwise sound coastal businesses, innovation renders old assets obsolete and their loans with them, and interest rate hikes can push any class of loan into default. Each of these channels is explicit, parameterised against empirical literature, and traceable through the model’s stock and flow diagrams.

Innovation, in the Schumpeterian tradition, is a double-edged force. Exploratory loans and incumbent firms’ research and development jointly determine productivity growth, which expands the economy’s potential over the long run. But creative destruction displaces existing investments and workers in the short term, and firms become more inclined to innovate when their cash reserves stagnate, tying the pace of technological change to the business cycle. Realised labour productivity is then further eroded by heat: the model divides global work into high, low and no exposure classes across agriculture, industry and services, with outdoor strenuous labour suffering the most as temperatures rise, and the sectoral composition of the economy shifting with income as it does in observed development patterns.

The government submodule aggregates the world’s fiscal and monetary authorities. Tax revenue from wages, profits and rents funds public investment, consumption and transfers, with spending constrained by the debt-to-GDP ratio and deficits issuing public debt to the banking sector. Central banks adjust a policy rate to steer inflation toward two percent and unemployment toward five percent, while risk premia on private and government borrowing rise with perceived default risk. Transfers are countercyclical, expanding with unemployment and with the child and retiree cohorts tracked by FRIDA’s demographics module, and as climate losses mount, a growing share of public expenditure shifts from investment toward repairs and maintenance.

Calibration against historical data from 1980 to 2023, drawn from the World Bank, IMF, OECD, WID and ILO, shows the model producing macroeconomic dynamics broadly consistent with observations, including endogenous business-cycle fluctuations visible in GDP, unemployment and the labour share of income. The authors are careful about what these cycles mean: they arise from a single structural mechanism, the interaction of innovation and finance, and cannot reproduce specific historical downturns such as the 2009 banking crisis or the 2020 pandemic, which the model deliberately excludes as exogenous shocks. To confront uncertainty honestly, the team ran a 100,000-member ensemble using Sobol sequence sampling across parameter ranges, reporting medians and 67 and 95 percent confidence intervals through 2150 rather than seductive single-line forecasts.

The ensemble results carry a sobering message. Median real output continues to grow through the end of the century, but the growth decelerates as rising temperatures raise loan failure rates, tighten lending standards and widen the gap between safe and risky interest rates, directly raising the cost of financing investment. Unemployment climbs under the combined pressure of climate-driven defaults, productivity-driven displacement, demographic contraction and sticky wages, and the resulting transfer burdens push government debt upward, which in turn raises interest costs and squeezes discretionary spending. Within the 95 percent interval, some parameterisations consistent with historical data produce outright declines in real output, a risk that aggregate damage functions would struggle to surface.

The authors acknowledge the trade-offs of their approach: global aggregation masks local impacts, the disequilibrium framework forgoes formal policy optimisation, and calibration inevitably attributes historical volatility to the model’s single endogenous cycle mechanism. Yet they argue the payoff is a transparent, modifiable platform, freely available on GitHub, in which policy choices interact with evolving financial, fiscal and demographic pressures rather than being assessed in isolation. Whether a carbon tax lands in a recession or an expansion, how transition funding strategies compare, and how a disorderly transition might trigger system-wide financial distress are all questions the model is built to explore, and ones that the equilibrium, finance-free IAMs of the past generation could not even ask.

Subject of Research: A stock-flow consistent, disequilibrium economy module for the FRIDA v2.1 integrated assessment model of climate-economy interactions

Article Title: Schumpeterian disaggregation and integrated assessment: An endogenous, stock-flow consistent economy in disequilibrium for FRIDA v2.1

Article References: Schumpeterian disaggregation and integrated assessment: An endogenous, stock-flow consistent economy in disequilibrium for FRIDA v2.1. (n.d.). https://doi.org/10.5194/gmd-19-9411-2026

Image Credits: AI Generated

DOI: 10.5194/gmd-19-9411-2026

Keywords: integrated assessment models, climate economics, FRIDA v2.1, stock-flow consistency, Schumpeterian innovation, financial fragility, endogenous growth, business cycles, public debt, unemployment, climate damage functions, ensemble simulation

The research team, comprising Badriya Al Mawali, Issam Ali and Alaa Ibrahim from the Physics Department at Sultan Qaboos University, together with Martin Köhler and Maike Ahlgrimm of Deutscher Wetterdienst, the German national weather service, and Sultan Al Yahyai of Code Academy in Muscat, built their investigation around controlled sensitivity experiments. Al Mawali also holds an affiliation with Oman’s Directorate General of Meteorology, a connection that underscores the practical motivation behind the work: understanding how regional weather models respond to environmental conditions is directly relevant to forecasting and hazard assessment in a country that sits at the edge of the cyclone-prone Arabian Sea.

At the heart of the study is the ICON numerical weather prediction model, a state-of-the-art forecasting system run here as a regional ensemble at a spatial resolution of 6.5 kilometers. Ensemble simulations are a cornerstone of modern cyclone research because they generate many slightly perturbed versions of the same forecast, allowing scientists to distinguish robust signals from noise. Rather than trying to predict Shaheen as it happened, the team ran a control simulation of the storm and then systematically modified individual ingredients of the environment: the temperature of the sea surface, the height of Oman’s terrain, and the roughness of the ocean surface as felt by the wind.

The most dramatic response came from the ocean. When the researchers raised sea surface temperature by 2 degrees Celsius, the simulated cyclone intensified and rainfall along the coast increased. This result follows directly from the physics of tropical cyclones, which are heat engines fueled by the flux of energy and moisture from warm water into the atmosphere. A warmer sea surface raises the enthalpy available to the storm, encouraging stronger convection, lower surface pressure and higher wind speeds, and in this case the model translated that extra energy into heavier precipitation concentrated along the coastal zone where the storm interacted with land.

Terrain, by contrast, acted less as an energy source and more as a steering and organizing influence. When the team reduced the height of Oman’s mountains, the simulated cyclone’s track shifted, and the terrain-enhanced rainfall that typically develops when moist winds are forced up over topography weakened over the mountainous interior. The changes rippled through the storm’s vertical structure as well: the low-level flow, the convergence of moisture flux that feeds convective bands, the mid-level ascent that drives cloud formation, and the upper-level divergence that vents air out of the storm’s core were all reorganized when the mountains were lowered. In other words, topography did not simply block the storm; it reshaped the three-dimensional circulation that determines how a cyclone behaves as it approaches land.

The third variable, sea-surface roughness, produced a comparatively muted response. Roughness describes the frictional drag that the ocean surface exerts on the wind, and physically based changes to this parameter altered the simulated cyclone’s intensity and precipitation structure only modestly. This hierarchy of effects is one of the study’s clearest takeaways: of the surface properties tested, temperature dominated the storm’s energy budget and the magnitude of coastal rainfall, while terrain height dominated the storm’s path, its vertical circulation and the geographical distribution of where rain fell. Roughness, though physically important, sat in the background of this particular case.

For Oman and the surrounding region, the distinction between these mechanisms carries real weight. Coastal rainfall magnitude and storm intensity are the quantities that matter most for wind damage and flash flooding along the shore, and those responded most strongly to sea surface temperature. Track and rainfall distribution, which determine which valleys, cities and wadis are actually hit, responded most strongly to terrain. A forecasting system that misrepresents either ingredient could therefore err in different directions: too-warm sea surfaces in a model might inflate the amount of rain a storm delivers, while poorly resolved mountains might misplace the heaviest precipitation or shift the landfall zone. The study suggests that evaluating both of these surface conditions is essential for trustworthy regional cyclone simulation.

The work also speaks to a broader question that looms over the Arabian Sea: what happens as ocean waters warm? The researchers are careful on this point. The 2-degree increase was used as a controlled test of the model’s sensitivity, not as a prediction of how future cyclones will behave. Because the experiments involve a single storm and idealized modifications of the environment, the results should be read as a diagnostic of how the ICON model responds to changed conditions, not as a forecast of a warmer future. That caveat matters, because the relationship between sea surface temperature and cyclone behavior in the real climate involves many competing factors, from atmospheric stability to wind shear, that a single-storm sensitivity study cannot capture.

Even so, the methodology offers a template for future work. By isolating one environmental variable at a time within an ensemble framework, the researchers could attribute specific changes in the simulated storm to specific causes, something that is nearly impossible to do with observations alone. The approach can be extended to other cyclones, other regions and other model configurations, building a picture of which surface conditions regional prediction systems handle well and where they need improvement. The authors suggest the findings may support further evaluation and refinement of regional cyclone simulations and contribute to assessments of heavy-rainfall and flood hazards, capabilities that are increasingly urgent as coastal populations in the Middle East grow and extreme weather draws closer scrutiny.

Tropical Cyclone Shaheen itself remains a benchmark event for the region. Its direct impact on northern Oman in October 2021 was unusual enough that it provides a valuable, well-observed case for testing how models represent cyclones in a basin where such storms are historically infrequent but potentially devastating. By dissecting that storm in the laboratory of a 6.5-kilometer ensemble model, the Sultan Qaboos University team and their collaborators have shown that the ocean’s warmth and the mountains’ height are not interchangeable pieces of the forecast puzzle. One sets the storm’s power and the volume of rain it can deliver to the coast; the other sets its course and the map of where that rain will fall. For forecasters, modelers and hazard planners alike, keeping both in sharp focus may be the key to anticipating the next time warm Arabian Sea waters send a cyclone toward Oman’s shores.

Subject of Research: Sensitivity of regional ensemble simulations of Tropical Cyclone Shaheen to sea surface temperature and terrain-height changes in Oman

Article Title: Could warmer waters bring heavier coastal rain? Insights from Cyclone Shaheen simulations

Article References: Could warmer waters bring heavier coastal rain? Insights from Cyclone Shaheen simulations. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: Tropical Cyclone Shaheen, Oman, sea surface temperature, terrain height, ICON model, numerical weather prediction, ensemble simulation, coastal rainfall, Arabian Sea, Scientific Reports, Sultan Qaboos University, flood hazard