Searcharxiv⌕ Search

arXiv subjects

Niklas Boers

Publications and source records attributed to Niklas Boers.

At least 55 records · Page 3Linked to original sources
arXiv2024

Data-driven Global Ocean Modeling for Seasonal to Decadal Prediction

Accurate ocean dynamics modeling is crucial for enhancing understanding of ocean circulation, predicting climate variability, and tackling challenges posed by climate change. Despite improvements in traditional numerical models, predicting global ocean variability over multi-year scales remains challenging. Here, we propose ORCA-DL (Oceanic Reliable foreCAst via Deep Learning), the first data-driven 3D ocean model for seasonal to decadal prediction of global ocean circulation. ORCA-DL accurately simulates three-dimensional ocean dynamics and outperforms state-of-the-art dynamical models in capturing extreme events, including El Niño-Southern Oscillation and upper ocean heatwaves. This demonstrates the high potential of data-driven models for efficient and accurate global ocean forecasting. Moreover, ORCA-DL stably emulates ocean dynamics at decadal timescales, demonstrating its potential even for skillful decadal predictions and climate projections.

physics.ao-ph↗
arXiv2024

Internal noise interference to warnings of tipping points in generic multi-dimensional dynamical systems

A deterministic dynamical system that slowly passes through a generic fold-type (saddle-node) bifurcation can be reduced to one-dimensional dynamics close to the bifurcation because of the centre manifold theorem. It is often tacitly assumed that the same is true in the presence of stochasticity or noise so that, for example, critical slowing down (CSD) indicators can be applied as if the system were one-dimensional. In this work, we show that this is only true when given suitable system observables; specifically, we demonstrate that noise in other dimensions may interfere with indicators of CSD, also referred to as early warning signals (EWS). We point out a generic mechanism by which both variance and lag-1 autocorrelation [AC(1)], as well as other EWS, can fail to signal an approaching bifurcation. This can, in principle, occur whenever one noise source drives multiple system components simultaneously. Even under the favourable assumptions of uncoupled deterministic dynamics and stationary noise, some system observables can then exhibit false negative or false positive CSD indications. We isolate this phenomenon in an example that represents a generic two-dimensional fold-type bifurcation setting.

math.DS↗
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

Anticipating critical transitions in multidimensional systems driven by time- and state-dependent noise

Anticipating bifurcation-induced transitions in dynamical systems has gained relevance in various fields of the natural, social, and economic sciences. Before the annihilation of a system's equilibrium point by means of a bifurcation, the system's internal feedbacks that stabilize the initial state weaken and eventually vanish, a process referred to as critical slowing down (CSD). In one-dimensional systems, this motivates the use of variance and lag-1 autocorrelation as indicators of CSD. However, the applicability of variance is limited to time- and state-independent driving noise, strongly constraining the generality of this CSD indicator. In multidimensional systems, the use of these indicators is often preceded by a dimension reduction in order to obtain a one-dimensional time series. Many common techniques for such an extraction of a one-dimensional time series generally incur the risk of missing CSD in practice. Here, we propose a data-driven approach based on estimating a multidimensional Langevin equation to detect local stability changes and anticipate bifurcation-induced transitions in systems with generally time- and state-dependent noise. Our approach substantially generalizes the conditions under which CSD can reliably be detected, as demonstrated in a suite of examples. In contrast to existing approaches, changes in deterministic dynamics can be clearly discriminated from changes in the driving noise using our method. This substantially reduces the risk of false or missed alarms of conventional CSD indicators in settings with time-dependent or multiplicative noise. In multidimensional systems, our method can greatly advance the understanding of the coupling between system components and can avoid risks of missing CSD due to dimension reduction, which existing approaches suffer from.

math.DS↗
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

A Generative Machine Learning Approach for Improving Precipitation from Earth System Models

Quantifying the impacts of anthropogenic global warming requires accurate Earth system model (ESM) simulations. Statistical bias correction and downscaling can be applied to reduce errors and increase the resolution of ESMs. However, existing methods, such as quantile mapping, cannot effectively improve spatial patterns or temporal dynamics. We address this problem with a purely generative machine learning approach, combining unpaired domain translation with a super-resolution foundation model. Our results show realistic spatial patterns and temporal dynamics as well as reduced distributional biases in the processed ESM simulation.

physics.geo-ph↗
arXiv2024

Detection of Approaching Critical Transitions in Natural Systems Driven by Red Noise

Detection of critical slowing down (CSD) is the dominant avenue for anticipating critical transitions from noisy time-series data. Most commonly, changes in variance and lag-1 autocorrelation [AC(1)] are used as CSD indicators. However, these indicators will only produce reliable results if the noise driving the system is white and stationary. In the more realistic case of time-correlated red noise, increasing (decreasing) the correlation of the noise will lead to spurious (masked) alarms for both variance and AC(1). Here, we propose two new methods that can discriminate true CSD from possible changes in the driving noise characteristics. We focus on estimating changes in the linear restoring rate based on Langevin-type dynamics driven by either white or red noise. We assess the capacity of our new estimators to anticipate critical transitions and show that they perform significantly better than other existing methods both for continuous-time and discrete-time models. In addition to conceptual models, we apply our methods to climate model simulations of the termination of the African Humid Period. The estimations rule out spurious signals stemming from nonstationary noise characteristics and reveal a destabilization of the African climate system as the dynamical mechanism underlying this archetype of abrupt climate change in the past.

math.DS↗
arXiv2024

Intraseasonal synchronization of extreme rainfalls between North India and the Sahel

The Indian Summer Monsoon (ISM) and the West African Monsoon (WAM) are dominant drivers of boreal summer precipitation variability in tropical and subtropical regions. Although the regional precipitation dynamics in these two regions have been extensively studied, the intraseasonal interactions between the ISM and WAM remain poorly understood. Here, we employ a climate network approach based on extreme rainfall events to uncover synchronously occurring extreme rainfall patterns across the two monsoon systems. We reveal strong synchronization of extreme rainfall events during the peak monsoon period in July and August, linking heavy rainfall over North India to that over the Sahel with a lag of around 12 days. We find that La Niña-like conditions in combination with the Boreal Summer Intraseasonal Oscillation and an enhanced Tropical Easterly Jet (TEJ) foster the synchronization between the ISM and the WAM. Convective clouds are transported by an intensified TEJ from southwestern Asia toward North Africa, supporting anomalous deep convection over the Sahel region.

physics.ao-ph↗
arXiv2024

Stabilized Neural Differential Equations for Learning Dynamics with Explicit Constraints

Many successful methods to learn dynamical systems from data have recently been introduced. However, ensuring that the inferred dynamics preserve known constraints, such as conservation laws or restrictions on the allowed system states, remains challenging. We propose stabilized neural differential equations (SNDEs), a method to enforce arbitrary manifold constraints for neural differential equations. Our approach is based on a stabilization term that, when added to the original dynamics, renders the constraint manifold provably asymptotically stable. Due to its simplicity, our method is compatible with all common neural differential equation (NDE) models and broadly applicable. In extensive empirical evaluations, we demonstrate that SNDEs outperform existing methods while broadening the types of constraints that can be incorporated into NDE training.

cs.LG↗
arXiv2024

Glacial abrupt climate change as a multi-scale phenomenon resulting from monostable excitable dynamics

Paleoclimate proxies reveal abrupt transitions of the North Atlantic climate during past glacial intervals known as Dansgaard--Oeschger (DO) events. A central feature of DO events is a sudden warming of about 10$^{\circ}$C in Greenland marking the beginning of relatively mild phases termed interstadials. These exhibit gradual cooling over several hundred to a few thousand years until a final abrupt decline brings the temperatures back to cold stadial levels. As of now, the exact mechanism behind this millennia-scale variability remains inconclusive. Here, we propose an excitable model to explain Dansgaard-Oeschger cycles, where interstadials occur as noise-induced state space excursions. Our model comprises the mutual multi-scale interactions between four dynamical variables representing Arctic atmospheric temperatures, Nordic Seas' temperatures and sea ice cover, and the Atlantic Meridional Overturning Circulation. The model's atmosphere-ocean heat flux is moderated by the sea ice, which in turn is subject to large perturbations dynamically generated by fast-evolving intermittent noise. If supercritical, perturbations trigger interstadial-like state space excursions during which all four model variables undergo qualitative changes that consistently resemble the signature of interstadials in corresponding proxy records. As a physical intermittent process generating the noise we propose convective events in the ocean or atmospheric blocking events. Our model accurately reproduces the DO cycle shape, return times and the dependence of the interstadial and stadial durations on the background conditions. In contrast to the prevailing understanding that DO variability is based on bistability in the underlying dynamics, we show that multi-scale, monostable excitable dynamics provide a promising alternative to explain millennial-scale climate variability associated with DO events.

physics.ao-ph↗
arXiv2023

ResoNet: Robust and Explainable ENSO Forecasts with Hybrid Convolution and Transformer Networks

Recent studies have shown that deep learning (DL) models can skillfully predict the El Niño-Southern Oscillation (ENSO) forecasts over 1.5 years ahead. However, concerns regarding the reliability of predictions made by DL methods persist, including potential overfitting issues and lack of interpretability. Here, we propose ResoNet, a DL model that combines convolutional neural network (CNN) and Transformer architectures. This hybrid architecture design enables our model to adequately capture local SSTA as well as long-range inter-basin interactions across oceans. We show that ResoNet can robustly predict ESNO at lead times between 19 and 26 months, thus outperforming existing approaches in terms of the forecast horizon. According to an explainability method applied to ResoNet predictions of El Niño and La Niña events from 1- to 18-month lead, we find that it predicts the Niño3.4 index based on multiple physically reasonable mechanisms, such as the Recharge Oscillator concept, Seasonal Footprint Mechanism, and Indian Ocean capacitor effect. Moreover, we demonstrate that for the first time, the asymmetry between El Niño and La Niña development can be captured by ResoNet. Our results could help alleviate skepticism about applying DL models for ENSO prediction and encourage more attempts to discover and predict climate phenomena using AI methods.

physics.geo-ph↗
arXiv2023

Spatial correlation increase in single-sensor satellite data reveals loss of Amazon rainforest resilience

The Amazon rainforest (ARF) is threatened by deforestation and climate change, which could trigger a regime shift to a savanna-like state. Previous work suggesting declining resilience in recent decades was based only on local resilience indicators. Moreover, previous results are potentially biased by the employed multi-sensor and optical satellite data and undetected anthropogenic land-use change. Here, we show that the spatial correlation provides a more robust resilience indicator than local estimators and employ it to measure resilience changes in the ARF, based on single-sensor Vegetation Optical Depth data under conservative exclusion of human activity. Our results show an overall loss of resilience until around 2019, which is especially pronounced in the southwestern and northern Amazon for the time period from 2002 to 2011. The demonstrated reliability of spatial correlation in coupled systems suggests that in particular the southwest of the ARF has experienced pronounced resilience loss over the last two decades.

physics.geo-ph↗
arXiv2023

Measuring tropical rainforest resilience under non-Gaussian disturbances

The Amazon rainforest is considered one of the Earth's tipping elements and may lose stability under ongoing climate change. Recently a decrease in tropical rainforest resilience has been identified globally from remotely sensed vegetation data. However, the underlying theory assumes a Gaussian distribution of forest disturbances, which is different from most observed forest stressors such as fires, deforestation, or windthrow. Those stressors often occur in power-law-like distributions and can be approximated by $α$-stable Lévy noise. Here, we show that classical critical slowing down indicators to measure changes in forest resilience are robust under such power-law disturbances. To assess the robustness of critical slowing down indicators, we simulate pulse-like perturbations in an adapted and conceptual model of a tropical rainforest. We find few missed early warnings and few false alarms are achievable simultaneously if the following steps are carried out carefully: First, the model must be known to resolve the timescales of the perturbation. Second, perturbations need to be filtered according to their absolute temporal autocorrelation. Third, critical slowing down has to be assessed using the non-parametric Kendall-$τ$ slope. These prerequisites allow for an increase in the sensitivity of early warning signals. Hence, our findings imply improved reliability of the interpretation of empirically estimated rainforest resilience through critical slowing down indicators.

physics.ao-ph↗
arXiv2023

Deep learning for bias-correcting CMIP6-class Earth system models

The accurate representation of precipitation in Earth system models (ESMs) is crucial for reliable projections of the ecological and socioeconomic impacts in response to anthropogenic global warming. The complex cross-scale interactions of processes that produce precipitation are challenging to model, however, inducing potentially strong biases in ESM fields, especially regarding extremes. State-of-the-art bias correction methods only address errors in the simulated frequency distributions locally at every individual grid cell. Improving unrealistic spatial patterns of the ESM output, which would require spatial context, has not been possible so far. Here, we show that a post-processing method based on physically constrained generative adversarial networks (cGANs) can correct biases of a state-of-the-art, CMIP6-class ESM both in local frequency distributions and in the spatial patterns at once. While our method improves local frequency distributions equally well as gold-standard bias-adjustment frameworks, it strongly outperforms any existing methods in the correction of spatial patterns, especially in terms of the characteristic spatial intermittency of precipitation extremes.

physics.ao-ph↗
arXiv2023

Exploring Geometric Deep Learning For Precipitation Nowcasting

Precipitation nowcasting (up to a few hours) remains a challenge due to the highly complex local interactions that need to be captured accurately. Convolutional Neural Networks rely on convolutional kernels convolving with grid data and the extracted features are trapped by limited receptive field, typically expressed in excessively smooth output compared to ground truth. Thus they lack the capacity to model complex spatial relationships among the grids. Geometric deep learning aims to generalize neural network models to non-Euclidean domains. Such models are more flexible in defining nodes and edges and can effectively capture dynamic spatial relationship among geographical grids. Motivated by this, we explore a geometric deep learning-based temporal Graph Convolutional Network (GCN) for precipitation nowcasting. The adjacency matrix that simulates the interactions among grid cells is learned automatically by minimizing the L1 loss between prediction and ground truth pixel value during the training procedure. Then, the spatial relationship is refined by GCN layers while the temporal information is extracted by 1D convolution with various kernel lengths. The neighboring information is fed as auxiliary input layers to improve the final result. We test the model on sequences of radar reflectivity maps over the Trento/Italy area. The results show that GCNs improves the effectiveness of modeling the local details of the cloud profile as well as the prediction accuracy by achieving decreased error measures.

cs.LG↗
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↗
arXiv2023

Extreme rainfall propagation within Boreal Summer Intraseasonal Oscillation modulated by Pacific sea surface temperature

Intraseasonal variations of the South Asian Summer Monsoon (SASM) contain alternating extreme rainfall (active) and low rainfall (break) phases impacting agriculture and economies. Their timing and spatial location are dominated by the Boreal Summer Intraseasonal Oscillation (BSISO), a quasi-periodic movement of convective precipitation from the equatorial Indian Ocean to the Western Pacific. However, observed deviations from the BSISO's canonical north-eastward propagation are poorly understood. Utilizing climate networks to characterize how active phases propagate within the SASM domain and using clustering analysis, we reveal three distinct modes of BSISO propagation: north-eastward, eastward-blocked, and stationary. We further show that Pacific sea surface temperatures modulate the modes - with El Niño- (La Niña-) like conditions favoring the stationary (eastward-blocked) - by changing local zonal and meridional overturning circulations and the BSISO Kelvin wave component. Using these insights, we demonstrate the potential for early warning signals of extreme rainfall until four weeks in advance.

physics.ao-ph↗