We design a new non-linear model to construct an accurate relationship between auroral oval boundaries and 18 space physical parameters, and explore the influence of every single space physical parameter on auroral oval boundary in this paper. As a result, we found the combination of some space physical parameters can strengthen each other's influence on aurora oval boundary prediction, and this model can achieve the best performance when only partial space physical parameters are used as input.

Negentropy anomalies of borehole strain associated with the Wenchuan earthquake are analysed. The cumulative anomalies are studied on both large and small scales. Through a comparison discussion, we compare the cumulative anomalies of different time periods and stations with those at the Guza station during the study period and preliminarily exclude meteorological factors. We suspect negentropy anomalies at the Guza station to have recorded abnormal changes related to the Wenchuan earthquake.

This study investigates non-Gaussian statistics of the data from a 10240-member ensemble Kalman filter. The large ensemble size can resolve the detailed structures of the probability density functions (PDFs) and indicates that the non-Gaussian PDF is caused by multimodality and outliers. While the outliers appear randomly, large multimodality corresponds well with large analysis error, mainly in the tropical regions and storm track regions where highly nonlinear processes appear frequently.

Baroclinic and barotropic instabilities are well known as the processes responsible for the production of the most important energy-containing eddies in the atmospheres and oceans of Earth and other planets. Linear and nonlinear instability theories provide insights into when such instabilities may occur, grow to a large amplitude and saturate, with examples from the laboratory, simplified numerical models and planetary atmospheres. We conclude with a number of open issues for future research.

Seasonal prediction of the of the North Atlantic Oscillation (NAO) has been improved in recent years by improving dynamical models and ensemble predictions. One step therein was the so-called sub-sampling, which combines statistical and dynamical predictions. This study generalises this approach and makes it much more accessible. Furthermore, it presents a new verification approach for such predictions.

Statistical post-processing aims to increase the predictive skill of probabilistic ensemble weather forecasts by learning the statistical relation between historical pairs of observations and ensemble forecasts within a given training data set. This study compares four different training schemes and shows that including multiple years of data in the training set typically yields a more stable post-processing while it loses the ability to quickly adjust to temporal changes in the underlying data.

The chaotic-oscillation inertia weight back propagation (COPSO-BP) algorithm is proposed for transient electromagnetic inversion. The BP's initial weight and threshold parameters were trained by COPSO, overcoming the BP falling into a local optimum. Inversion of the layered geoelectric model showed that the COPSO-BP method is accurate and stable and needs less training time. It can be used in 1-D direct current sounding, 1-D magnetotelluric sounding, seismic-wave impedance and source detection.

Atmospheric stably stratified boundary layers display transitions between regimes of sustained and intermittent turbulence. These transitions are not well represented in numerical weather prediction and climate models. A prototype explicitly stochastic turbulence parameterization simulating regime dynamics is presented and tested in an idealized model. Results demonstrate that the approach can improve the regime representation in models.

I provide a personal view on the role of bifurcation analysis of climate models in the development of a theory of variability in the climate system. By outlining the state of the art of the methodology and by discussing what has been done and what has been learned from a hierarchy of models, I will argue that there are low-order phenomena of climate variability, such as El Niño and the Atlantic Multidecadal Oscillation.

Computational models involving adaptive meshes can both evolve dynamically and be remeshed. Remeshing means that the state vector dimension changes in time and across ensemble members, making the ensemble Kalman filter (EnKF) unsuitable for assimilation of observational data. We develop a modification in which analysis is performed on a fixed uniform grid onto which the ensemble is mapped, with resolution relating to the remeshing criteria. The approach is successfully tested on two 1-D models.

An adaptive PSO optimization method is proposed based on the average population velocity in order to solve for location parameters of the seismic source in a location model. Combined with the actual need to solve for seismic source parameters, the model constraints of inertia weight, accelerating constants, the maximum flight velocity of particles, and other parameters are discussed in order to improve the optimization capacity of the PSO algorithm and avoid being trapped in a local optimum.

This paper describes an innovative way to use data assimilation to infer the dynamics of a physical system from its observation only. The method can operate with noisy and partial observation of the physical system. It acts as a deep learning technique specialised to dynamical models without the need for machine learning tools. The method is successfully tested on chaotic dynamical systems: the Lorenz-63, Lorenz-96, and Kuramoto–Sivashinski models and a two-scale Lorenz model.

Neural networks are a technique that is widely used to predict the time evolution of physical systems. For this the past evolution of the system is shown to the neural network – it is trained – and then can be used to predict the evolution in the future. We show some limitations in this approach for certain systems that are important to consider when using neural networks for climate- and weather-related applications.

ll-posedness of the inverse problem and sparse noisy data are two major challenges in the modeling of high-dimensional spatiotemporal processes. We present a Bayesian inference method with a strongly regularized posterior to overcome these challenges, enabling joint state-parameter estimation and quantifying uncertainty in the estimation. We demonstrate the method on a physically motivated nonlinear stochastic partial differential equation arising from paleoclimate construction.

This paper aims to analyse and simulate correlations between two fields in a scale-invariant framework. It starts by theoretically assessing and numerically confirming the behaviour of renormalized multiplicative power law combinations of two fields with known scale-invariant properties. Then a new indicator of correlation is suggested and tested on rainfall data to study the correlation between the common rain rate and drop size distribution features.

Current space weather problems are discussed for young researchers. We have discussed some of the major problems that need to be solved for space weather forecasting to become a reality.

This study presents a novel method for reduced-rank data assimilation of multiscale highly nonlinear systems. Time-varying dynamical properties are used to determine the rank and projection of the system onto a reduced subspace. The variable reduced-rank method is shown to succeed over other fixed-rank methods. This work provides implications for performing strongly coupled data assimilation with a limited number of ensemble members on high-dimensional coupled climate models.

We present a new way to adaptively improve weather forecasts by incorporating last-minute observation information. The recently measured error, together with a statistical model, gives us an indication of the expected future error of wind speed forecasts, which are then adjusted accordingly. This new technique can be especially beneficial for customers in the wind energy industry, where it is important to have reliable short-term forecasts, as well as providers of extreme weather warnings.

We study a model containing the characteristic of self-similarity (invariance under scale) which allows for scaling between lab experiments and geographical-scale seismicity. Particular to this model is the dependency of the earthquake rates on the magnitude difference between events that are causally connected. We present results of a statistical analysis of magnitude correlations for the model along with its implications for the ongoing efforts in earthquake forecasting.

The conspicuous feature of the atmosphere–ocean system is the large difference in the masses of the two media. In this respect there is a strong analogy to Brownian motion, with light and fast molecules colliding with heavy and slow Brownian particles. I apply the tools of non-equilibrium statistical mechanics for studying Brownian motion to air–sea interaction.

A methodology is developed to simplify the mathematical description of activity in a time-dependent driven system. The method describes the response in the system that is most-closely related to the driver. This reduced description has advantages: low noise, high prediction efficiency, linearity in the described system response to the driver, and compactness. The analysis of the Earth’s magnetospheric system is demonstrated.

Forecasting the height of new snow is crucial for avalanche hazard, road viability, ski resorts and tourism. The numerical models suffer from systematic and significant errors which are misleading for the final users. Here, we applied for the first time a state-of-the-art statistical method to correct ensemble numerical forecasts of the height of new snow from their statistical link with measurements in French Alps and Pyrenees. Thus the realism of automatic forecasts can be quickly improved.

A popular variational ensemble smoother for data assimilation and history matching is simplified. An exact relationship between ensemble linearizations (linear regression) and adjoints (analytic derivatives) is established.

In this paper, we highlight a specific example of large-scale flows. We discuss a graph-theory-based Lagrangian technique for identifying regions of strong mixing (in the sense of diffusion) in the flow and compare it to previous Lagrangian approaches used in this context.

We utilized N-S, C-V and S-V fractal models to delineate different grades of mineralization in the Pulang deposit. The high-grade mineralization determined by S–V has a better correlation with potassic alteration zones resulting from the 3-D geological model than from N-S and C-V models. The highly and moderately mineralized zones obtained from the fractal models are both situated in the southern and central parts of the Pulang deposit and coincide with potassic and phyllic alteration zones.

We examined the spatial diversity of Indian rainfall teleconnection at different timescales, first by identifying homogeneous communities and later by computing non-linear linkages between the identified communities (spatial regions) and dominant climatic patterns, represented by climatic indices such as El Nino–Southern Oscillation, Indian Ocean Dipole, North Atlantic Oscillation, Pacific Decadal Oscillation and Atlantic Multi-Decadal Oscillation.

Statistical hypothesis tests in wavelet analysis are used to asses the likelihood that time series features are noise. The choice of test will determine which features emerge as a signal. Tests based on area do poorly at distinguishing abrupt fluctuations from periodic behavior, unlike tests based on arclength that do better. The application of the tests suggests that there are features in Indian rainfall time series that emerge from background noise.

We explore the nature of instabilities in a well-known meteorological toy model, the Lorenz 96, to unravel key mechanisms of interaction between scales of different resolutions and time scales. To do so, we use a mathematical machinery known as Lyapunov analysis, allowing us to capture the degrees of chaoticity associated with fundamental directions of instability. We find a non-trivial group of such directions projecting significantly on slow variables, associated with long term dynamics.

This paper is based on a deep learning neural network to invert the Rayleigh surface wave velocity of the crustal thickness, which is a new geophysical inversion solution that proved to be effective and practical. Through comparative experiments, we found that deep learning neural networks can more accurately reveal the non-linear relationship between phase velocity and crustal thickness than traditional shallow networks. Deep learning neural networks are more efficient than Monte Carlo methods.