NP2.1

Recently, there has been much interest in issuing subseasonal to seasonal (S2S) forecasts, although their skill is often debated. In addition to large systematic errors, ensemble systems are often overconfident, i.e. have incorrect information about the uncertainty of a particular forecast. Stochastic parameterization schemes are used routinely to remedy the problem of overconfidence, but also have the potential to reduce systematic model errors. 

Here, we study the impact of adding a stochastic parameterization scheme in coupled simulations with the climate model CESM.  Physical processes associated with S2S-predictability, like the Madden-Julian  Oscillation (MJO) and Northern Hemispheric blocking are analyzed. In the simulations with a stochastic parameterization scheme, the northward propagation of the MJO is captured better, leading to an improved MJO lifecycle. The impact on other atmospheric fields like precipitation and winds will be discussed. 

How to cite: Berner, J.: Benefits of stochastic parameterizations in subseasonal to seasonal (S2S) forecasts, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16321, https://doi.org/10.5194/egusphere-egu21-16321, 2021.

Stochastic subgrid-scale parametrizations aim to incorporate effects of unresolved processes in an effective model by sampling from a distribution usually described in terms of resolved modes. This is an active research area in climate, weather and ocean science where processes evolved in a wide range of spatial and temporal scales. In this study, we evaluate the performance of conditional generative adversarial network (GAN) in parametrizing subgrid-scale effects in a finite-difference discretization of stochastically forced Burgers equation. We define resolved modes as local spatial averages and deviations from these averages are the unresolved degrees of freedom. We train Wesserstein GAN (WGAN) conditioned on the resolved variables to learn the distribution of subgrid flux tendencies for resolved modes and, thus, represent the effect of unresolved scales. Resulting WGAN is then used in an effective model to reproduce the statistical features of resolved modes. We demonstrate that various stationary statistical quantities such as spectrum, moments, autocorrelation, etc. are well approximated by this effective model.

How to cite: Timofeyev, I. and Alcala, J.: Subgrid-scale parametrization of unresolved scales in forced Burgers equation using Generative Adversarial Networks (GAN), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6910, https://doi.org/10.5194/egusphere-egu21-6910, 2021.

Intrinsic ocean variability is essential for climate prediction because it is less sensitive to stochastic process, but it is very difficult to be identified due to internal climate variability. Here we use regional interactive ensemble applied on ocean-atmosphere interface (RIE-OA) to suppress atmosphere stochastic variability and to reveal intrinsic variability as well as to understand climate dynamic across multiple timescales. Five atmosphere general circulation models (AGCM) are coupled to an ocean general circulation model (OGCM) over the North Atlantic basin (20^oN to Denmark Strait and Greenland-Scotland ridge). The OGCM interacts with fluxes from a selected AGCM globally except over the North Atlantic basin where the OGCM interacts with the ensemble averaged fluxes from the five AGCMs. The five AGCMs, on the other hand, feel the same ocean states. Hence, the atmosphere stochastic variability impacting the ocean is one-fifth weaker than stand-alone configuration (control case). This leads to reduction of the local climate variability, such as Atlantic Multidecadal Variability, but should not reduce intrinsic variability. Comparing control cases and RIE-OA case, we found the intrinsic ocean variability, a narrow-banded low-frequency (about 8 to 20 years) signal over the North Atlantic Subtropical Gyre, is not influenced by the weakened stochastic variability. More details will be discussed.

How to cite: Shen, M.-L., Keenlyside, N., and Chiu, P.-G.: Understanding intrinsic ocean variability by suppressing regional stochastic variability, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14346, https://doi.org/10.5194/egusphere-egu21-14346, 2021.

I will present a comprehensive inter-comparison of linear regression, stochastic and deep-learning-based models for reduced-order statistical modelling of the simplified ocean circulation. The reference dataset is provided by the top 150 empirical orthogonal functions (EOFs) and principal components (PCs) of an idealized, eddy-resolving, double-gyre ocean model. Our goal is to have a systematic and comprehensive assessment of the skills, costs and complexities of all the models considered.

The model based on linear regression is considered as a baseline. Additionally, we investigate stochastic models (linear regression plus additive-noise and a multi-level approach), deep-learning models (a feed-forward Artificial Neural Network (ANN), a Long Short Term Memory (LSTM)), and deep-learning augmented linear regression models (also called hybrid models). We also explored stochastically improved deep learning methods by adding spatially correlated white noise in the deep learning models to account for the residuals and left out variance in the discarded PCs. The assessment metrics considered are climatology, variance, RMSE, instantaneous correlation coefficients, frequency map, prediction horizon, and computational costs for training and predictions.  

Until now, we found that the hybrid LSTM models perform the best, followed by the multi-level linear stochastic model and multiplicative white noise model. Additionally, hybrid models found to perform better when augmented by spatially correlated white noise.  This suggests that an amalgam of physics, memory effects, and stochasticity provides the best strategy for low-order representation of oceanic process. However, LSTM was also found to be most expensive to train and forecast amongst all. Skills of simple stochastic models are similar to those of the linear regression model but superior to those of the pure deep learning models, as evidenced by relatively better frequency maps, infinite prediction horizon, and low running cost.

Overall, our analysis promotes multi-level stochastic methods, with memory effects, and stochastic hybrid methods for low-dimensional ocean models as a more practical option when compared to pure deep-learning solutions as they are more accurate, stable, and low-cost. Furthermore, this is an ongoing research project and more updated results will be discussed at the time of presentation.

How to cite: Agarwal, N., Kondrashov, D., Dueben, P., Ryzhov, E., and Berloff, P.: A comparison of data-driven approaches to build low-dimensional ocean models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10194, https://doi.org/10.5194/egusphere-egu21-10194, 2021.

Moist equatorial waves, responsible for a large fraction of synoptic and intraseasonal tropical variability, are visible in satellite observations of cloud top temperature and outgoing longwave radiation, with familiar dispersion relations appearing in space time spectra once a background red-like spectrum is removed (Kiladis et al., 2009).

Studies have suggested that the large-scale planetary waves in the equatorial region can be excited by smaller scale gravity waves (Yang & Ingersoll, 2013), baroclinic waves from the extratropics (Wedi & Smolarkiewicz, 2010), or localized synoptic scale heating (Gill 1980, 1982). In this study we examine the possibility that a continuous forcing of anomalies at the mesoscale can, through a turbulent upscale cascade, excites these waves, thus explaining both: the peaks along dispersion relations and the background red spectrum.

Our underlying assumption is that within the tropics (excluding wave forcing from the extratropics), a prerequisite to form coherent heating of ≈1000 km zonal length on an aqua planet is self-aggregation. Over the last two decades, self-aggregation has been studied over a wide range of scenarios up to the atmospheric mesoscale. In this study we examine the aggregation from the mesoscale up to the planetary scale, by applying mesoscale stochastic forcing in idealized spherical shallow water model. In particular, we examine the dependence of the large scale spectra on the field being stochastically forced, and on the existence of moisture.

We find that indeed, continuous stochastic forcing at the mesoscale can excite two-dimensional turbulence and linear tropical wave modes. When vorticity or moisture are forced in the simulations at wavenumber 100, a classical -5/3 slope of the eddy kinetic energy spectrum forms in an upscale energy cascade up to the planetary scale. Furthermore, equatorial waves emerge and Wheeler-Kiladis plots reveal a rich temporal and spatial structure of Rossby, Kelvin, Yanai, and Inertial Gravity waves. On the other hand, stochastic forcing of the divergence, or height fields only leads to a turbulent field when applied at planetary scales. Some explanations for this strong dependence on the type of forcing, and the role of moisture, will be discussed. 

How to cite: Schröttle, J., Suhas, D., Harnik, N., and Sukhatme, J.: Coexistence of equatorial waves & turbulence in moist shallow water simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7722, https://doi.org/10.5194/egusphere-egu21-7722, 2021.

Atmosphere and ocean dynamics display many complex features and are characterized by a wide variety of processes and couplings across different timescales. Here we use Multivariate Empirical Mode Decomposition (MEMD; Rehman and Mandic, 2010) to investigate the multivariate and multiscale properties of a low-order model of the ocean-atmosphere coupled dynamics (Vannitsem, 2017). The MEMD allows us to decompose the original data into a series of oscillating patterns with time-dependent amplitude and phase by exploiting the local features of data and without any a priori assumptions on the decomposition basis. Moreover, each oscillating pattern, usually named Multivariate Intrinsic Mode Function (MIMF), can be used as a source of local (in terms of scale) fluctuations and information. This information allows us to derive multiscale measures when looking at the behavior of the generalized fractal dimensions at different scales (Hentschel and Procaccia, 1983) that can be seen as a sort of multivariate and multiscale generalized fractal dimensions (Alberti et al., 2020). With these two approaches, we demonstrate that the coupled ocean-atmosphere dynamics presents a rich variety of common features, although with a different nature of the fractal properties between the ocean and the atmosphere at different timescales. The MEMD results allow us to capture the main dynamics of the phase-space trajectory that can be used for reconstructing the skeleton of the phase-space dynamics, while the evaluation of the fractal dimensions at different timescales characterize the intrinsic complexity of oscillating patterns that can be related to the attractor properties. Our results support the interpretation of the coupled ocean-atmosphere dynamics as well as the investigation of general deterministic-chaotic dissipative dynamical systems in terms of invariant manifolds, bifurcations, as well as (strange) attractors in their phase-space, whose geometric and topological properties are a reflection of the dynamical regimes of the system at different scales. We compare the results obtained for the low-order dynamical model with those derived from the reanalysis data and demonstrate that a similar scale-dependent behavior is found, thus also confirming the suitability of the proposed system to model the ocean-atmosphere dynamics at different timescales and to describe topological and geometrical features of its phase-space.

References

Alberti, T., Consolini, G., Ditlevsen, P. D., Donner, R. V., Quattrociocchi, V. (2020). Multiscale measures of phase-space trajectories. Chaos 30, 123116.

Alberti, T., Donner, R. V., and Vannitsem, S. (2021). Multiscale fractal dimension analysis of a reduced order model of coupled ocean-atmosphere dynamics. Earth Syst. Dynam. Discuss. [preprint], https://doi.org/10.5194/esd-2020-96, in review.

Hentschel, H. G. E., Procaccia, I. (1983). The infinite number of generalized dimensions of fractals and strange attractors. Physica D 8, 435–444.

Rehman, N., Mandic, D. P. (2010). Multivariate empirical mode decomposition. Proceedings of the Royal Society A, 466, 1291–1302.

Vannitsem S., Predictability of large-scale atmospheric motions: Lyapunov exponents and error dynamics, Chaos, 27, 032101, 2017. 

How to cite: Alberti, T., Donner, R., and Vannitsem, S.: On the multiscale fractal features of a low-order coupled ocean-atmosphere model in comparison with reanalysis data, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-135, https://doi.org/10.5194/egusphere-egu21-135, 2021.

It has been well recognized that, for most climatic records, their current states are influenced by both past conditions and current dynamical excitations. However, how to properly use this idea to improve the climate predictive skills, is still an open question. In this study, we evaluated the decadal hindcast experiments of 11 models (participating in phase 5 of the Coupled Model Intercomparison Project, CMIP5) in simulating the effects of past conditions (memory part, M(t)) and the current dynamical excitations (non-memory part, ε(t)). Poor skills in simulating the memory part of surface air temperatures (SAT) are found in all the considered models. Over most regions of China, the CMIP5 models significantly overestimated the long-term memory (LTM) of SAT. While in the southwest, the LTM was significantly underestimated. After removing the biased memory part from the simulations using fractional integral statistical model (FISM), the remaining non-memory part, however, was found reasonably simulated in the multi-model means. On annual scale, there were high correlations between the simulated and the observed ε(t) over most regions of the country, and for most cases they had the same sign. These findings indicated that the current errors of dynamical models may be partly due to the unrealistic simulations of the impacts from the past. To improve predictive skills, a new strategy was thus suggested. As FISM is capable of extracting M(t) quantitatively, by combining FISM with dynamical models (which may produce reasonable estimations of ε(t)), improved climate predictions with the effects of past conditions properly considered may become possible.

How to cite: Xiong, F., Yuan, N., Ma, X., Lu, Z., and Gao, J.: On memory and non-memory parts of surface air temperatures over China: can they be simulated by decadal hindcast experiments in CMIP5?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3628, https://doi.org/10.5194/egusphere-egu21-3628, 2021.

One of the most used metrics to gauge the effects of climate change is the equilibrium climate sensitivity, defined as the long-term (equilibrium) temperature increase resulting from instantaneous doubling of atmospheric CO2. Since global climate models cannot be fully equilibrated in practice, extrapolation techniques are used to estimate the equilibrium state from transient warming simulations. Because of the abundance of climate feedbacks – spanning a wide range of temporal scales – it is hard to extract long-term behaviour from short-time series; predominantly used techniques are only capable of detecting the single most dominant eigenmode, thus hampering their ability to give accurate long-term estimates. Here, we present an extension to those methods by incorporating data from multiple observables in a multi-component linear regression model. This way, not only the dominant but also the next-dominant eigenmodes of the climate system are captured, leading to better long-term estimates from short, non-equilibrated time series.

How to cite: Bastiaansen, R., Dijkstra, H., and von der Heydt, A.: Multivariate Estimations of Equilibrium Climate Sensitivity from Short Transient Warming Simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-187, https://doi.org/10.5194/egusphere-egu21-187, 2021.

How to cite: Lembo, V., Lucarini, V., and Ragone, F.: Predicting Climate Change through Response Operators in a Coupled General Circulation Model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2221, https://doi.org/10.5194/egusphere-egu21-2221, 2021.

Nonlinear atmospheric dynamics produce rare events that are hard to predict and attribute due to many interacting degrees of freedom. A sudden stratospheric warming is a spectacular example in which the winter polar vortex in the stratosphere rapidly breaks down, inducing a shift in midlatitude tropospheric weather patterns that persist for up to 2-3 months.  In principle, lengthy numerical simulations can be used to predict and understand these rare transitions.  For complex models, however, the cost of the direct numerical simulation approach is often prohibitive.  We describe an alternative approach which in principle only requires relatively short duration computer simulations of the system.  Applying this methodology to a classical idealized stratospheric model with stochastic forcing, we compute optimal forecasts of sudden warming events and quantify the limits of predictability.  Statistical analysis relates these optimal forecasts to a small number of easy-to-interpret physical variables.Remarkably, we are able to estimate these quantities using a data set of simulations much shorter than the return time of the warming event.

How to cite: Finkel, J., Webber, R. J., Gerber, E. P., Abbot, D. S., and Weare, J.: Forecasting rare stratospheric transitions using short simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16323, https://doi.org/10.5194/egusphere-egu21-16323, 2021.

In this work, we consider turbulence closures of LES (Large Eddy Simulation) type for classical decaying 2D turbulence in a priori and a posteriori experiments using explicit filtering approach. According to Germano 1986 decomposition, full subfilter stress for Gaussian filter is decomposed into three Galilean-invariant parts: Leonard, Cross and Reynolds stresses. By analysing spectral transfer of energy and enstrophy, we show that Leonard stress redistributes resolved energy toward large scales and dissipates substantial part of enstrophy, while Cross stress provides an additional enstrophy dissipation at subfilter scales and Reynolds stress predominantly injects energy into middle scales (i.e., Kinetic Energy Backscatter). Substantial part of enstrophy dissipation is located on subfilter scales, and it should be accounted for by choosing base filter wide enough compared to mesh step of LES model. Otherwise, significant fraction of enstrophy dissipation will correspond to subgrid scale stress, which is less universal and harder to approximate. As a result of a priori analysis, we propose LES closure consisting of three parts: SSM (Scale Similarity Model), which is equivalent to Leonard stress, biharmonic Smagorinsky damping as a Cross stress counterpart and ADM (Approximate Deconvolution Model) approximation for Reynolds stress ("backscatter"). The proposed model have two free parameters: Smagorinsky constant and amplitude of the backscatter. These parameters are estimated in a posteriori experiments utilizing dynamic approach and energy-enstrophy balance equation, correspondingly. The proposed model have the following distinctive features: it reproduces energy and enstrophy transfer spectra in accordance to the individual components of the subfilter forces, "reproduces" base filter and reproduces energy growth in accordance to the filtered DNS (Direct Numerical Simulation) solution.

The work was supported by the Russian Foundation for Basic Research (projects 19-35-90023, 18-05-60184) and Moscow Center for Fundamental and Applied Mathematics (agreement with the Ministry of Education and Science of the Russian Federation No. 075-15-2019-1624).

How to cite: Perezhogin, P. and Glazunov, A.: A priori and a posteriori analysis in Large eddy simulation of the two-dimensional decaying turbulence using Explicit filtering approach, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2382, https://doi.org/10.5194/egusphere-egu21-2382, 2021.

Wetlands, affected by the hydro-climatic condition and human activities, are key elements in providing valuable ecosystem services for ecology, environment, and human. Wetlands can exist in various states (e.g., area, volume, depth, etc.) driven by both natural and human forcing, and are often distributed in a wetlandscape. In these specific landscapes, wetlands (node) and dispersal path (link) of inhabiting species organize ecological networks. Here, we generated the three ecological networks with three dispersal models (threshold distance, exponential kernel, and heavy-tailed dispersal model) and analyzed network characteristics (degree, efficiency and clustering coefficient) associated with the seasonal change of hydro-climatic condition on wetland hydrology. To identify the role of small wetlands, we analyzed two different scenarios in which the sum of wetland areas are similar but their area distributions are distinct. In the first scenario, most of the small wetlands are hydrologically disappeared while the second scenario maintains the small wetlands with a shrunk area of large wetlands. When the area of large wetlands was reduced, a slight decrease in the values of network metrics was observed due to an increase in distances between wetlands. On the other hand, when a number of small wetlands were hydrologically disappeared, all the metric values were significantly decreased compared to the network in which all wetlands were hydrologically maintained. Especially, when the disappeared wetlands were not recovered even after rainfall, possibly due to long-term dehydration of supporting soil, the network characteristics also did not recover even if the total area of wetlands were recovered. However, when the dried small wetlands were hydrologically recovered, the network characteristics also recovered rapidly. Based on our observation, we confirmed that the small wetlands, despite their extremely low areal portion in the entire wetlandscape, play a key role in maintaining the ecological network resilience. Our findings can be used for a decision-making process for wetland conservation and restoration by reflecting the functional importance of small wetlands with physical characteristics requirements such as wetland areas.

How to cite: Kim, B. and Park, J.: The role of small wetlands for resilient ecological networks in a wetlandscape, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13712, https://doi.org/10.5194/egusphere-egu21-13712, 2021.