arXiv subjects

Sebastian Bathiany

Publications and source records attributed to Sebastian Bathiany.

18 recordsLinked to original sources

arXiv2026

Flexible generation of daily Earth system model projections across radiative forcing scenarios

Earth system model (ESM) projections of the climate system's response to anthropogenic forcing are central to assess the impacts of climate change and inform adaptation and mitigation policies. However, given their high computational cost, projections are only made for a limited set of standardized forcing scenarios with limited temporal extent, such as the Shared Socioeconomic Pathways (SSPs), the spatiotemporal resolution remains too low for direct impact assessments, and uncertainties cannot be comprehensively quantified. Recent data-driven models offer efficient and accurate high-resolution simulations for weather prediction, but cannot extrapolate to future greenhouse gas concentrations because they cannot capture the responses to unprecedented forcing, limiting their value for climate change projections. Here, we combine response theory with a tailored generative machine learning framework to address this challenge. Our approach extracts the physical forced response to radiative forcing from monthly low-resolution ESM fields, and uses this response to guide a generative model to infer consistent daily global high-resolution temperature and precipitation projections. Our probabilistic approach generalizes across ESMs and provides long-term, bias-corrected responses to radiative forcing at high spatiotemporal resolution. It efficiently generates large ensembles needed for uncertainty quantification, effectively fills the gaps between existing SSPs, and readily extends climate projections to 2300 and beyond. Our framework hence complements ESM projections by providing efficient, stable, and high spatiotemporal resolution long-term climate projection ensembles across emission scenarios, enabling detailed impact assessment and exploration of long-term climate commitment.

physics.ao-ph↗

arXiv2026

Extrapolation from historical data cannot reliably predict the time of a potential AMOC collapse

Ditlevsen and Ditlevsen [Nature Communications, 2023] (DD23 hereafter) propose a statistical framework to estimate the timing of a potential collapse of the Atlantic Meridional Overturning Circulation (AMOC) based on extrapolating information from observed sea-surface temperature (SST) variability. By fitting a stochastic one-dimensional fold-bifurcation model to an SST-based fingerprint of the AMOC using Maximum Likelihood Estimation (MLE), they conclude that a collapse is most likely to occur in the middle of the 21st century, with a reported 95% confidence interval covering the time span from 2037 to 2109. Given the profound implications of such a claim for both climate and society, it is essential to thoroughly test the robustness of this result, to critically assess the underlying assumptions and uncertainties, and to estimate the extent to which the reported confidence interval reflects the true limits of current knowledge. Here we examine the sensitivity of DD23's results and argue that four types of uncertainty are insufficiently explored in their analysis: (i) structural uncertainty associated with the assumed low-order bifurcation model, (ii) statistical uncertainty in their model fit, (iii) uncertainty in the representativeness of SST-based fingerprints as proxies for the high-dimensional AMOC dynamics, and (iv) uncertainty in the underlying data, arising from non-stationary observational coverage and dataset preprocessing. Using synthetic experiments and a systematic analysis of alternative fingerprints and observational products, we show that the tipping times estimated by DD23 are highly sensitive to the uncertainties listed above, and extend several millennia into the future when these uncertainties are thoroughly propagated.

physics.geo-ph↗

arXiv2026

Statistical warning indicators for abrupt transitions in dynamical systems with slow periodic forcing

There is growing interest in anticipating critical transitions in natural systems, often pursued through statistical detection of early warning signals associated with dynamical bifurcations. In stochastic dynamical systems, such signals commonly rely on manifestations of critical slowing down. However, we still need additional development for the underlying theory for critical transitions in non-autonomous systems. This extension is relevant for natural systems, whose behaviour often emerges from seasonal periodic forcing. In this study, we systematically investigate the feasibility of anticipating the termination of oscillatory behavior in a bistable system with slow periodic forcing. In this setting, existing approaches of estimating linear characteristics of the return map fail in practical scenarios due to the unfavourable time-scale separation. Instead, we propose two statistical indicators for the anticipation of critical transitions in the periodic behaviour: (i) conventional early warning indicators, such as increasing variance and autocorrelation, evaluated across system cycles, and (ii) indicators derived from the phase of the seasonal forcing. By statistically comparing their predictive performance, we find that phase-based indicators provide the strongest early warning capability. Our results offer guidance for the detection of critical transitions in periodically forced systems and, more broadly, systematically extend early-warning signs towards non-autonomous dynamical systems.

math.DS↗

arXiv2026

Generative deep learning improves reconstruction of global historical climate records

Accurate assessment of anthropogenic climate change relies on historical instrumental data, yet observations from the early 20th century are sparse, fragmented, and uncertain. Conventional reconstructions rely on disparate statistical interpolation, which tends to smooth local features and create unphysical artifacts, often leading to an underestimation of intrinsic variability and extremes. While recent machine learning approaches have improved reconstruction accuracy, they remain confined to purely spatial inpainting of coarse-resolution fields. Here, we present a unified, probabilistic generative deep learning framework that overcomes these limitations and reveals previously unresolved historical climate variability back to 1850. Leveraging a learned generative prior of Earth system dynamics, our model performs probabilistic inference to estimate spatiotemporally consistent historical temperature and precipitation fields from sparse observations. Our approach preserves the higher-order statistics of climate dynamics, transforming reconstruction into a robust uncertainty-aware assessment. We demonstrate that our reconstruction mitigates the smoothing effects inherent in widely used historical reference products, including those underlying IPCC assessments, especially regarding extreme weather events. Notably, we uncover higher early 20th-century global warming levels compared to existing reconstructions, primarily driven by more pronounced polar warming, with mean Arctic warming trends exceeding established benchmarks by 0.15--0.29C per decade for 1900--1980. Conversely, for the modern era, our reconstruction indicates that the broad Arctic warming trend is likely overestimated in recent assessments, yet explicitly resolves previously unrecognized intense, localized hotspots in the Barents Sea and Northeastern Greenland.

physics.geo-ph↗

arXiv2025

NeuralCrop: Combining physics and machine learning for improved crop yield projections

Global gridded crop models (GGCMs) are crucial to project the impacts of climate change on agricultural productivity and assess associated risks for food security. Despite decades of development, state-of-the-art GGCMs retain substantial uncertainties stemming from process representations. Recently, machine learning approaches trained on observational data provide alternatives in crop yield projections. However, these models have not demonstrated improved performance over traditional GGCMs and are not suitable for projecting crop yields under a changing climate due to their poor out-of-distribution generalization. Here we introduce NeuralCrop, a differentiable hybrid GGCM that combines the strengths of an advanced process-based GGCM, resolving important processes explicitly, with data-driven machine learning components. NeuralCrop is first trained to emulate a competitive GGCM before it is fine-tuned on observational data. We show that NeuralCrop produces projections with accuracy comparable to state-of-the-art GGCMs across site-level and large-scale crop simulations. NeuralCrop can accurately project the interannual yield variability in European wheat regions and the US Corn Belt. Capturing yield anomalies is essential for developing adaptation strategies in the context of climate change. NeuralCrop can more accurately reproduce yield anomalies across various climatic conditions, with particularly notable improvements under drought extremes. For large-scale, long-term simulations, our approach is orders of magnitude more computationally efficient. Our results show that end-to-end hybrid crop modelling offers more reliable yield projections that are essential for food risk assessments under climate change and intensifying extreme weather events.

cs.LG↗

arXiv2025

Anticipating AMOC transitions via deep learning

Key components of the Earth system can undergo abrupt and potentially irreversible transitions when the magnitude or rate of external forcing exceeds critical thresholds. In this study, we use the example of the Atlantic Meridional Overturning Circulation (AMOC) to demonstrate the challenges associated with anticipating such transitions when the system is susceptible to bifurcation-induced, rate-induced, and noise-induced tipping. Using a calibrated AMOC box model, we conduct large ensemble simulations and show that transition behavior is inherently probabilistic: under identical freshwater forcing scenarios, some ensemble members exhibit transitions while others do not. In this stochastic regime, traditional early warning indicators based on critical slowing down are unreliable in predicting impending transitions. To address this limitation, we develop a convolutional neural network (CNN)-based approach that identifies higher-order statistical differences between transitioning and non-transitioning trajectories within the ensemble realizations. This method enables the real-time prediction of transition probabilities for individual trajectories prior to the onset of tipping. Our results show that the CNN-based indicator provides effective early warnings in a system where transitions can be induced by bifurcations, critical forcing rates, and noise. These findings underscore the potential in identifying safe operating spaces and early warning indicators for abrupt transitions of Earth system components under uncertainty.

cs.CE↗

arXiv2025

Decadal sink-source shifts of forest aboveground carbon since 1988

Forest ecosystems are vital to the global carbon cycle, yet their long-term aboveground carbon (AGC) dynamics remain uncertain. Here, we integrate multi-source satellite observations with probabilistic deep learning models to reconstruct a harmonized, uncertainty-aware global forest AGC record from 1988 to 2021 at 0.25-deg. We find that, although global forests sequestered 6.2 PgC, moist tropical and boreal forests have progressively transitioned toward carbon sources since the early 2000s. This shift coincides with a strengthening negative correlation between tropical AGC variability and atmospheric CO2 growth rates (r = -0.63 in 2011-2021), suggesting tropical forests increasingly modulate the global carbon cycle. Notably, in the Brazilian Amazon, the contribution of intact forests to the year-to-year variations in AGC losses increased from 33% in the 1990s to 76% in the 2010s, surpassing that of deforested areas (from 60% to 13%). Our findings highlight the vulnerability of carbon stocks in key biomes and provide a benchmark to track emerging sink-source shifts under anthropogenic climate change.

physics.geo-ph↗

arXiv2025

The influence of data gaps and outliers on resilience indicators

The resilience, or stability, of major Earth system components is increasingly threatened by anthropogenic pressures, demanding reliable early warning signals for abrupt and irreversible regime shifts. Widely used data-driven resilience indicators based on variance and autocorrelation detect `critical slowing down', a signature of decreasing stability. However, the interpretation of these indicators is hampered by poorly understood interdependencies and their susceptibility to common data issues such as missing values and outliers. Here, we establish a rigorous mathematical analysis of the statistical dependency between variance- and autocorrelation-based resilience indicators, revealing that their agreement is fundamentally driven by the time series' initial data point. Using synthetic and empirical data, we demonstrate that missing values substantially weaken indicator agreement, while outliers introduce systematic biases that lead to overestimation of resilience based on temporal autocorrelation. Our results provide a necessary and rigorous foundation for preprocessing strategies and accuracy assessments across the growing number of disciplines that use real-world data to infer changes in system resilience.

nlin.AO↗

arXiv2025

Generating realistic global precipitation fields from modelled atmospheric circulation

Improving the representation of precipitation in Earth system models (ESMs) is critical for assessing the impacts of climate change and especially of extreme events like floods and droughts. In existing ESMs, precipitation is not resolved explicitly, but represented by parameterizations. These typically rely on resolving approximated but computationally expensive column-based physics, not accounting for interactions between locations. They struggle to capture fine-scale precipitation processes and introduce significant biases. We present a novel approach, based on generative machine learning, which integrates a conditional diffusion model with a UNet architecture to generate accurate, high-resolution (0.25{\deg}) global daily precipitation fields from a small set of prognostic atmospheric variables. Unlike traditional parameterizations, our framework efficiently produces ensemble predictions, capturing uncertainties in precipitation, and does not require fine-tuning by hand. We train our model on the ERA5 reanalysis and present a method that allows us to apply it to unseen ESM data, enabling fast generation of probabilistic forecasts and climate scenarios. By leveraging interactions between global prognostic variables, our approach provides an alternative parameterization scheme that mitigates biases present in the ESM precipitation while maintaining consistency with its large-scale (annual) trends. This work demonstrates that complex precipitation patterns can be learned directly from large-scale atmospheric variables, offering a computationally efficient method to obtain high-resolution precipitation without the cost of running the dynamical model at such high resolution.

cs.LG↗

arXiv2025

Reconciled warning signals in observations and models imply approaching AMOC tipping point

Paleoclimate proxy records and models suggest that the Atlantic Meridional Overturning Circulation (AMOC) can transition abruptly between a strong and a weak state. Empirical warning signals in observational fingerprints indeed suggest a stability decline and raise concerns that the system may be approaching a tipping point. However, state-of-the-art Earth System Models (ESMs) do not consistently show such a stability loss, hence inconclusive in their projections of AMOC collapse under global warming. It remains unclear whether warning signals of AMOC tipping are overlooked in ESMs or overinterpreted in observations, calling for further investigations of AMOC stability. Here, based on the concept of critical slowing down, AMOC stability decline that can be interpreted as a warning signal of AMOC tipping is identified in the eastern SPNA in both observations and ESMs. This warning signal is in accordance with a physical indicator of AMOC stability, AMOC-induced freshwater convergence into the Atlantic basin. The observed signal can be reconciled with the modeled one only under warming exceeding the Paris Agreement goal, suggesting that AMOC stability is overestimated in ESMs. Our results suggest that the observed AMOC may indeed be losing stability and could thus reach a tipping point much earlier than state-of-the-art ESMs suggest.

physics.ao-ph↗

arXiv2024

Assessment of Abrupt Shifts in CMIP6 Models using Edge Detection

Past research has shown that multiple climate subsystems might undergo abrupt shifts, such as the Arctic Winter sea ice or the Amazon rainforest, but there are large uncertainties regarding their timing and spatial extent. In this study we investigated when and where abrupt shifts occur in the latest generation of earth system models (CMIP6) under a scenario of 1\% annual increase in CO$_2$. We considered 82 ocean, atmosphere, and land variables across 57 models. We used a Canny edge detection method to identify abrupt shifts occurring on yearly to decadal timescales, and performed a connected component analysis to quantify the spatial extent of these shifts. The systems analyzed include the North Atlantic subpolar gyre, Tibetan Plateau, land permafrost, Amazon rainforest, Antarctic sea ice, monsoon systems, Arctic summer sea ice, Arctic winter sea ice, and Barents sea ice. Except for the monsoon systems, we found abrupt shifts in all of these across multiple models. Despite large inter-model variations, higher levels of global warming consistently increase the risk of abrupt shifts in CMIP6 models. At a global warming of 1.5 $^\circ \text{C}$, six out of ten studied climate subsystems already show large-scale abrupt shifts across multiple models.

physics.ao-ph↗

arXiv2024

Deep Learning for predicting rate-induced tipping

Nonlinear dynamical systems exposed to changing forcing can exhibit catastrophic transitions between alternative and often markedly different states. The phenomenon of critical slowing down (CSD) can be used to anticipate such transitions if caused by a bifurcation and if the change in forcing is slow compared to the internal time scale of the system. However, in many real-world situations, these assumptions are not met and transitions can be triggered because the forcing exceeds a critical rate. For example, given the pace of anthropogenic climate change in comparison to the internal time scales of key Earth system components, such as the polar ice sheets or the Atlantic Meridional Overturning Circulation, such rate-induced tipping poses a severe risk. Moreover, depending on the realisation of random perturbations, some trajectories may transition across an unstable boundary, while others do not, even under the same forcing. CSD-based indicators generally cannot distinguish these cases of noise-induced tipping versus no tipping. This severely limits our ability to assess the risks of tipping, and to predict individual trajectories. To address this, we make a first attempt to develop a deep learning framework to predict transition probabilities of dynamical systems ahead of rate-induced transitions. Our method issues early warnings, as demonstrated on three prototypical systems for rate-induced tipping, subjected to time-varying equilibrium drift and noise perturbations. Exploiting explainable artificial intelligence methods, our framework captures the fingerprints necessary for early detection of rate-induced tipping, even in cases of long lead times. Our findings demonstrate the predictability of rate-induced and noise-induced tipping, advancing our ability to determine safe operating spaces for a broader class of dynamical systems than possible so far.

cs.LG↗

arXiv2024

Targeted Calibration to Adjust Stability Biases in Complex Dynamical System Models

Models of complex dynamical systems like the Earth's climate often involve large numbers of uncertain parameters. Comprehensive exploration of the parameter space is typically prohibitive due to excessive computational costs. Systematic gradient-based parameter optimization is not feasible because such models are typically not differentiable. This is especially problematic in cases where the models describe highly nonlinear and possibly abrupt dynamics, where sensitivity to parameter changes is high. Components of Earth's climate system, such as the North Atlantic Overturning Circulation or the polar ice sheets, are at risk of undergoing critical transitions in response to anthropogenic climate change. Concerns have been raised that these Earth system components are too stable in state-of-the-art models. In my presentation, we will see examples how new scenario simulations allow studying the possibility and the consequences of tipping events in Earth system models. Also, we will discuss a method for efficient, systematic, and objective calibration of dynamical complex system models, targeted at adjusting system stability. Given a number of physical or observational constraints, the method can efficiently adjust stability biases in a range of complex system models and help reveal potentially hidden instabilities, with important implications for Earth system modelling.

physics.ao-ph↗

arXiv2024

Uncertainties too large to predict tipping times of major Earth system components from historical data

One way to warn of forthcoming critical transitions in Earth system components is using observations to detect declining system stability. It has also been suggested to extrapolate such stability changes into the future and predict tipping times. Here, we argue that the involved uncertainties are too high to robustly predict tipping times. We raise concerns regarding (i) the modeling assumptions underlying any extrapolation of historical results into the future, (ii) the representativeness of individual Earth system component time series, and (iii) the impact of uncertainties and preprocessing of used observational datasets, with focus on nonstationary observational coverage and gap filling. We explore these uncertainties in general and specifically for the example of the Atlantic Meridional Overturning Circulation. We argue that even under the assumption that a given Earth system component has an approaching tipping point, the uncertainties are too large to reliably estimate tipping times by extrapolating historical information.

physics.geo-ph↗

arXiv2024

Conditional diffusion models for downscaling and bias correction of Earth system model precipitation

Climate change exacerbates extreme weather events like heavy rainfall and flooding. As these events cause severe socioeconomic damage, accurate high-resolution simulation of precipitation is imperative. However, existing Earth System Models (ESMs) struggle to resolve small-scale dynamics and suffer from biases. Traditional statistical bias correction and downscaling methods fall short in improving spatial structure, while recent deep learning methods lack controllability and suffer from unstable training. Here, we propose a machine learning framework for simultaneous bias correction and downscaling. We first map observational and ESM data to a shared embedding space, where both are unbiased towards each other, and then train a conditional diffusion model to reverse the mapping. Only observational data is used for the training, so that the diffusion model can be employed to correct and downscale any ESM field without need for retraining. Our approach ensures statistical fidelity and preserves spatial patterns larger than a chosen spatial correction scale. We demonstrate that our approach outperforms existing statistical and deep learning methods especially regarding extreme events.

physics.geo-ph↗

arXiv2023

Differentiable Programming for Earth System Modeling

Earth System Models (ESMs) are the primary tools for investigating future Earth system states at time scales from decades to centuries, especially in response to anthropogenic greenhouse gas release. State-of-the-art ESMs can reproduce the observational global mean temperature anomalies of the last 150 years. Nevertheless, ESMs need further improvements, most importantly regarding (i) the large spread in their estimates of climate sensitivity, i.e., the temperature response to increases in atmospheric greenhouse gases, (ii) the modeled spatial patterns of key variables such as temperature and precipitation, (iii) their representation of extreme weather events, and (iv) their representation of multistable Earth system components and their ability to predict associated abrupt transitions. Here, we argue that making ESMs automatically differentiable has huge potential to advance ESMs, especially with respect to these key shortcomings. First, automatic differentiability would allow objective calibration of ESMs, i.e., the selection of optimal values with respect to a cost function for a large number of free parameters, which are currently tuned mostly manually. Second, recent advances in Machine Learning (ML) and in the amount, accuracy, and resolution of observational data promise to be helpful with at least some of the above aspects because ML may be used to incorporate additional information from observations into ESMs. Automatic differentiability is an essential ingredient in the construction of such hybrid models, combining process-based ESMs with ML components. We document recent work showcasing the potential of automatic differentiation for a new generation of substantially improved, data-informed ESMs.

cs.LG↗

arXiv2023

Uncertainties in critical slowing down indicators of observation-based fingerprints of the Atlantic Overturning Circulation

Observations are increasingly used to detect critical slowing down (CSD) to measure stability changes in key Earth system components. However, most datasets have non-stationary missing-data distributions, biases and uncertainties. Here we show that, together with the pre-processing steps used to deal with them, these can bias the CSD analysis. We present an uncertainty quantification method to address such issues. We show how to propagate uncertainties provided with the datasets to the CSD analysis and develop conservative, surrogate-based significance tests on the CSD indicators. We apply our method to three observational sea-surface temperature and salinity datasets and to fingerprints of the Atlantic Meridional Overturning Circulation derived from them. We find that the properties of these datasets and especially the specific gap filling procedures can in some cases indeed cause false indication of CSD. However, CSD indicators in the North Atlantic are still present and significant when accounting for dataset uncertainties and non-stationary observational coverage.

physics.ao-ph↗

arXiv2016

Trends in sea-ice variability on the way to an ice-free Arctic

It has been widely debated whether Arctic sea-ice loss can reach a tipping point beyond which a large sea-ice area disappears abruptly. The theory of dynamical systems predicts a slowing down when a system destabilises towards a tipping point. In simple stochastic systems this can result in increasing variance and autocorrelation, potentially yielding an early warning of an abrupt change. Here we aim to establish whether the loss of Arctic sea ice would follow these conceptual predictions, and which trends in sea ice variability can be expected from pre-industrial conditions toward an Arctic that is ice-free during the whole year. To this end, we apply a model hierarchy consisting of two box models and one comprehensive Earth system model. We find a consistent and robust decrease of the ice volume's annual relaxation time before summer ice is lost because thinner ice can adjust more quickly to perturbations. Thereafter, the relaxation time increases, mainly because the system becomes dominated by the ocean water's large heat capacity when the ice-free season becomes longer. Both trends carry over to the autocorrelation of sea ice thickness in time series. These changes are robust to the nature and origin of climate variability in the models and hardly depend on the balance of feedbacks. Therefore, characteristic trends can be expected in the future. As these trends are not specific to the existence of abrupt ice loss, the prospects for early warnings seem very limited. This result also has implications for statistical indicators in other systems whose effective mass changes over time, affecting the trend of their relaxation time. However, the robust relation between state and variability would allow an estimate of sea-ice variability from only short observations. This could help one to estimate the likelihood and persistence of extreme events in the future.

physics.ao-ph↗