NPG - recent papers

Conditional updates of neural network weights for increased out of training performance

2026-10-02T21:07:19+02:00

Conditional updates of neural network weights for increased out of training performance
Jan Saynisch-Wagner and Saran Rajendran Sari
Nonlin. Processes Geophys., 33, 503–520, https://doi.org/10.5194/npg-33-503-2026, 2026
Neural networks are limited in situations that differ from the learned conditions. We propose a 3-step solution to this out of distribution problem: (1) Derive the anomalies of a trained neural networks' internal parameters by retraining on subsets of its training data. (2) Relate the ensuing network-parameter anomalies to differences in the training data subsets that caused them. (3) Extrapolate the found relations to generate a network that performs better outside the org. training distribution.

Data assimilation of coherent structures in parametric and gridded spaces: an idealized ocean eddy study

2026-09-21T21:07:19+02:00

Data assimilation of coherent structures in parametric and gridded spaces: an idealized ocean eddy study
Solène Dealbera, Stéphane Raynaud, Carlos Granero Belinchon, Brahim Boussidi, Clément Le Goff, and Pierre Tandeo
Nonlin. Processes Geophys., 33, 489–501, https://doi.org/10.5194/npg-33-489-2026, 2026
The goal of data assimilation in geoscience is to merge information from models and observations. In the presence of structures such as ocean eddies, classic assimilation schemes fail to merge those two sources, resulting in position and shape errors. Here, we propose a new scheme which assimilates only the parameters describing a simplified eddy structure. Comparison with classic schemes shows that our approach preserves the eddy parameters and may reduce the computational cost.

Ensemble Kalman-guided model predictive path integral control for spatially localized suppression of extremes in chaotic geophysical flows

2026-09-04T21:07:19+02:00

Ensemble Kalman-guided model predictive path integral control for spatially localized suppression of extremes in chaotic geophysical flows
Haru Kuroki, Kazumune Hashimoto, Yuki Uehara, Yohei Sawada, Duc Le, and Masashi Minamide
Nonlin. Processes Geophys., 33, 473–487, https://doi.org/10.5194/npg-33-473-2026, 2026
We developed a method to plan tiny, local nudges that reduce spikes in weather-like simulations. It was motivated by the lack of reliable ways to design small interventions in unpredictable systems. The method uses many parallel forecasts to suggest where a nudge matters, then tests and refines it with repeated full-model simulations. In two models, it reduced extremes and peak regional wind speeds with similar or smaller input, enabling safer study of intervention ideas.

Covariations between persistent synoptic features and record low Antarctic sea ice events via unsupervised regression learning

2026-08-27T21:07:19+02:00

Covariations between persistent synoptic features and record low Antarctic sea ice events via unsupervised regression learning
Andrew R. Axelsen, Terence J. O'Kane, Courtney R. Quinn, and Andrew P. Bassom
Nonlin. Processes Geophys., 33, 455–472, https://doi.org/10.5194/npg-33-455-2026, 2026
Recent increases in the variability of Antarctic sea ice have elicited much interest and research on these changes. Here, we examine observations taken from three specific years (2016, 2021, and 2023) which either contain or precede a period of anomalously low sea ice extent. To understand the combined influence of weather systems, surface temperatures, and atmospheric pressure on sea ice formation and decay, we apply novel methods from machine learning and dynamical systems.

Bayesian data selection to quantify the value of data for landslide runout calibration

2026-08-25T21:07:19+02:00

Bayesian data selection to quantify the value of data for landslide runout calibration
V. Mithlesh Kumar, Anil Yildiz, and Julia Kowalski
Nonlin. Processes Geophys., 33, 425–453, https://doi.org/10.5194/npg-33-425-2026, 2026
The reliability of Bayesian calibration depends on the quality and availability of observational data. But are we choosing the right data? We address this question by measuring the information gained during calibration to quantify how data selection influences the Bayesian calibration of physics-based landslide runout models. We find that more data does not always yield better results – observations that capture the dynamics governed by a parameter are more effective for its calibration.

Elucidating the performance of data assimilation neural networks for chaotic dynamics

2026-08-11T21:07:19+02:00

Elucidating the performance of data assimilation neural networks for chaotic dynamics
Marc Bocquet, Tobias Sebastian Finn, Sibo Cheng, and Alban Farchi
Nonlin. Processes Geophys., 33, 401–424, https://doi.org/10.5194/npg-33-401-2026, 2026
Deep learning has been used to discover new data assimilation methods to track chaotic dynamical systems. Strikingly, these data assimilation networks can match the accuracy of ensemble-based methods using only a single state forecast. This paper first investigates the reasons for this efficacy, and then shows that their performance in more nonlinear regimes matches that of the most accurate scalable data assimilation methods, with greater efficiency and robustness.

A simple dynamical system for representing climate tipping points with hysteresis

2026-08-06T21:07:19+02:00

A simple dynamical system for representing climate tipping points with hysteresis
Chris Huntingford, Paul D. L. Ritchie, and Joseph Clarke
Nonlin. Processes Geophys., 33, 385–399, https://doi.org/10.5194/npg-33-385-2026, 2026
The risk of climate change triggering tipping points is a concern. While Earth System Models (ESMs) predict these points, their timing often varies. Equations of nonlinear dynamical systems can model tipping, hysteresis, and inertia, enabling comparison with ESMs. We propose a method to map tipping timing estimates onto a dynamical system that simulates a full hysteresis loop. We focus on solutions to the equation under overshoot forcing in global temperature and different inertia levels.

An improved noise model for representing westerly wind bursts in the recharge oscillator model of ENSO

2026-07-20T21:07:19+02:00

An improved noise model for representing westerly wind bursts in the recharge oscillator model of ENSO
Georg A. Gottwald, Eli Tziperman, and Alexey Fedorov
Nonlin. Processes Geophys., 33, 373–383, https://doi.org/10.5194/npg-33-373-2026, 2026
The recharge oscillator model has served as a simple yet powerful toy model for modelling the El Niño–Southern Oscillation. Whereas the addition of Gaussian noise is sufficient to capture the observed overall statistical features, it fails to resolve dynamical signatures associated with major El Niño events. Here we propose to employ a conditional non-Gaussian noise model that better captures the effect of large westerly wind bursts and their effect on major warming events.

Formulation of parametric uncertainty forecasts towards operational wildfire smoke assimilation

2026-07-14T21:07:19+02:00

Formulation of parametric uncertainty forecasts towards operational wildfire smoke assimilation
Annika Vogel, Richard Ménard, James Abu, and Jack Chen
Nonlin. Processes Geophys., 33, 347–371, https://doi.org/10.5194/npg-33-347-2026, 2026
Recent wildfire activity in Canada has been rising the public demand for fast, yet accurate operational air quality forecasts along with related forecast uncertainties. This study explores the potential of an efficient process-based approach to estimate forecast uncertainties of operational air quality models. Our results show its potential for understanding how uncertainties of operational forecast evolve over time and an improved use of sparse observation signals at remote wildfire regions.

Noise-scaled accuracy of the ensemble Kalman filter with an instability-based minimum ensemble size

2026-07-06T21:07:19+02:00

Noise-scaled accuracy of the ensemble Kalman filter with an instability-based minimum ensemble size
Kota Takeda and Takemasa Miyoshi
Nonlin. Processes Geophys., 33, 335–346, https://doi.org/10.5194/npg-33-335-2026, 2026
This study examines the minimum ensemble size for accurate geophysical forecasting using a method called the ensemble Kalman filter. We reformulate accuracy via observation noise-dependency to classify filter performance qualitatively. Through numerical experiments with a chaotic model, we link the minimum ensemble size for the accuracy to system's instability and propose an effective ensemble downsizing method that ensures both stability and accuracy.

Quantitative comparison of causal inference methods for climate tipping points

2026-06-19T21:07:19+02:00

Quantitative comparison of causal inference methods for climate tipping points
Niki Lohmann, David Strahl, Annika Högner, Willem Huiskamp, Matthias Boehm, and Nico Wunderling
Nonlin. Processes Geophys., 33, 313–334, https://doi.org/10.5194/npg-33-313-2026, 2026
Causal inference methods could be used to study the interaction of climate tipping elements, which may degrade abruptly due to climate change. We compare three of these methods to determine their reliability and apply two of them to the Arctic summer sea ice and the Atlantic Meridional Overturning Circulation (AMOC). Our results imply that a weaker AMOC would stabilize Arctic summer sea ice, and that a loss of Arctic summer sea ice would likely stabilize the AMOC in the short term.

Nonlinear quantitative relationship between the duration and occurrence frequency of droughts

2026-06-08T21:07:19+02:00

Nonlinear quantitative relationship between the duration and occurrence frequency of droughts
Pengcheng Yan, Guolin Feng, Cailing Zhao, Ping Yang, Hao Wu, and Dongdong Zuo
Nonlin. Processes Geophys., 33, 303–312, https://doi.org/10.5194/npg-33-303-2026, 2026
In this study, we examine the relationship between drought duration and frequency across China using daily data. We find a clear double-logarithmic relationship between the duration and the frequency. We also show that droughts in dry northwestern areas tend to last for months, while those in wet southeastern regions are shorter but more frequent. This pattern holds across all drought intensities. Overall, our findings offer a simple tool for drought risk assessment and water management.

Structural joint modeling of magnetotelluric data and Rayleigh wave dispersion curves using Pareto-based particle swarm optimization: an example to delineate the crustal structure of the southeastern part of the Biga Peninsula in western Anatolia

2026-06-02T21:07:19+02:00

Structural joint modeling of magnetotelluric data and Rayleigh wave dispersion curves using Pareto-based particle swarm optimization: an example to delineate the crustal structure of the southeastern part of the Biga Peninsula in western Anatolia
Ersin Büyük, Ekrem Zor, and Mustafa Cengiz Tapırdamaz
Nonlin. Processes Geophys., 33, 267–302, https://doi.org/10.5194/npg-33-267-2026, 2026
We introduce a Pareto-based multi-objective particle swarm optimization framework for joint modeling of magnetotelluric and Rayleigh wave dispersion data from the southeastern Biga Peninsula. The approach uses a shared structural parameterization without enforcing a fixed petrophysical link between resistivity and velocity. The study shows that magnetotelluric data are more affected by model trade-offs, whereas Rayleigh wave dispersion is more sensitive in data space.

Boosting ensembles for statistics of tails at conditionally optimal advance split times

2026-05-28T21:07:19+02:00

Boosting ensembles for statistics of tails at conditionally optimal advance split times
Justin Finkel and Paul A. O'Gorman
Nonlin. Processes Geophys., 33, 233–265, https://doi.org/10.5194/npg-33-233-2026, 2026
Estimating small probabilities of high-impact extreme weather events is a persistent computational challenge, motivating techniques such as rare event sampling and ensemble boosting: lightly perturbing simulated moderate events into more extreme ones. We formulate a new, flexible sampling strategy and characterizes a critical parameter – the advance split time, dictating when to perturb – in a simple atmospheric turbulence model, with generalizable entropy-based criteria.

Sandy beaches' chaos: shoreline-sandbar coupling inferred from observational time series

2026-04-21T21:07:19+02:00

Sandy beaches' chaos: shoreline-sandbar coupling inferred from observational time series
Marius Aparicio, Sylvain Mangiarotti, Salomé Frugier, Laurent Lacaze, Marcan Graffin, and Rafael Almar
Nonlin. Processes Geophys., 33, 197–231, https://doi.org/10.5194/npg-33-197-2026, 2026
We studied how sandy beaches evolve by tracking the shoreline and offshore sandbars from satellites over many years. By rebuilding beach behavior directly from observations, we show that beaches follow organized but chaotic motion shaped by internal feedbacks. Beyond the seasonal rhythm imposed by waves, shorelines and sandbars exchange energy through the surf zone, producing repeated erosion and recovery cycles with limited predictability, explaining why beaches remain difficult to forecast.

Bayesian inference based on algorithms: MH, HMC, MALA and Lip-MALA for prestack seismic inversion

2026-04-20T21:07:19+02:00

Bayesian inference based on algorithms: MH, HMC, MALA and Lip-MALA for prestack seismic inversion
Richard Perez-Roa, Saba Infante, Gabriel Barragan, and Raul Manzanilla
Nonlin. Processes Geophys., 33, 173–195, https://doi.org/10.5194/npg-33-173-2026, 2026
We explored four methods to improve how underground rock properties are estimated from seismic data. By comparing these methods on both simulated and real-world oilfield data, we found that techniques using gradient information give better accuracy but require more computing time. Our results help guide the choice of method depending on whether speed or precision is more important in subsurface exploration.

Spatiotemporal variation in rainfall predictability in Serbia under a changing climate

2026-03-24T21:07:19+01:00

Spatiotemporal variation in rainfall predictability in Serbia under a changing climate
Tatijana Stosic, Ivana Tošić, Antonio Samuel Alves da Silva, Vladimir Djurdjević, and Borko Stosic
Nonlin. Processes Geophys., 33, 157–172, https://doi.org/10.5194/npg-33-157-2026, 2026
In this work we address the change in rainfall predictability in Serbia due to climate change, using a novel entropy-based method that highlights both small and large fluctuations. The study is performed on data from 14 stations from 1961–2020. While rainfall average remains rather stable between two subperiods, the predictability of large and small fluctuations has changed, suggesting that climate change has affected rainfall dynamics in ways not observable by standard statistical methods.

Beyond static forecasts: a dynamic stress gradient framework for high-resolution aftershock prediction and mitigation

2026-03-19T21:07:19+01:00

Beyond static forecasts: a dynamic stress gradient framework for high-resolution aftershock prediction and mitigation
Boi-Yee Liao
Nonlin. Processes Geophys., 33, 123–155, https://doi.org/10.5194/npg-33-123-2026, 2026
After major earthquakes, smaller shocks often follow, yet predicting where they will occur remains difficult. This study introduces a new method for tracking changes in underground stress after a large earthquake. Using the 2018 Hualien earthquake in Taiwan as a case study, we found that areas with strong stress differences provide clearer signals of future aftershocks than stress magnitude alone. This approach can improve short-term earthquake risk assessment and disaster response planning.

Inferring the role of Interdecadal Pacific Oscillation phase on tropical-extratropical teleconnection dependencies

2026-03-04T21:07:19+01:00

Inferring the role of Interdecadal Pacific Oscillation phase on tropical-extratropical teleconnection dependencies
Mark A. Collier, Dylan Harries, and Terence J. O'Kane
Nonlin. Processes Geophys., 33, 103–122, https://doi.org/10.5194/npg-33-103-2026, 2026
Here we apply Bayesian methods to reconstructed and simulated climate model data over past decades to determine the role of long timescale phase dependencies, and extratropical teleconnections, on the major drivers of tropical climate variability.

MESMER-RCM: a probabilistic climate emulator for regional warming projections

2026-02-12T21:07:19+01:00

MESMER-RCM: a probabilistic climate emulator for regional warming projections
Hao Pan, Lukas Gudmundsson, Mathias Hauser, Jonas Schwaab, Yann Quilcaille, and Sonia I. Seneviratne
Nonlin. Processes Geophys., 33, 73–83, https://doi.org/10.5194/npg-33-73-2026, 2026
Existing regional climate model (RCM) emulators mainly provide deterministic emulations, while internal RCM variability is typically not represented. We develop MESMER-RCM, a probabilistic RCM emulator for annual 2-m temperature, using a simple and physically interpretable approach. We demonstrate its ability to emulate both RCM trends and internal variability in a high-dimensional spatial setting, where existing approaches typically struggle.