Category: PublicationsRA2

Framework for an Ocean-Connected Supermodel of the Earth System

Counillon, F., Keenlyside, N., Wang, S., Devilliers, M., Gupta, A., Koseki, S., Shen, M.-L. 2023: Framework for an Ocean-Connected Supermodel of the Earth System. JAMES. https://doi.org/10.1029/2022MS003310

Summary: Observed and future winter Arctic sea ice loss is strongest in the Barents Sea. However, the anthropogenic signal of the sea ice decline is superimposed by pronounced internal variability that represents a large source of uncertainty in future climate projections. A notable manifestation of internal variability is rapid ice change events (RICEs) that greatly exceed the anthropogenic trend. These RICEs are associated with large displacements of the sea ice edge which could potentially have both local and remote impacts on the climate system. In this study we present the first investigation of the frequency and drivers of RICEs in the future Barents Sea, using multi-member ensemble simulations from CMIP5 and CMIP6. A majority of RICEs are triggered by trends in ocean heat transport or surface heat fluxes. Ice loss events are associated with increasing trends in ocean heat transport and decreasing trends in surface heat loss. RICEs are a common feature of the future Barents Sea until the region becomes close to ice-free. As their evolution over time is closely tied to the average sea ice conditions, rapid ice changes in the Barents Sea may serve as a precursor for future changes in adjacent seas.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Super-resolution data assimilation

Barthélémy, S., Brajard, J., Bertino, L., Counillon, F. 2022: Super-resolution data assimilation. Ocean Dyn. https://doi.org/10.1007/s10236-022-01523-x

Summary: Increasing model resolution can improve the performance of a data assimilation system because it reduces model error, the system can more optimally use high-resolution observations, and with an ensemble data assimilation method the forecast error covariances are improved. However, increasing the resolution scales with a cubical increase of the computational costs. A method that can more effectively improve performance is introduced here. The novel approach called “Super-resolution data assimilation” (SRDA) is inspired from super-resolution image processing techniques and brought to the data assimilation context. Starting from a low-resolution forecast, a neural network (NN) emulates the fields to high-resolution, assimilates high-resolution observations, and scales it back up to the original resolution for running the next model step. The SRDA is tested with a quasi-geostrophic model in an idealized twin experiment for configurations where the model resolution is twice and four times lower than the reference solution from which pseudo-observations are extracted. The assimilation is performed with an Ensemble Kalman Filter. We show that SRDA outperforms both the low-resolution data assimilation approach and a version of SRDA with cubic spline interpolation instead of NN. The NN’s ability to anticipate the systematic differences between low- and high-resolution model dynamics explains the enhanced performance, in particular by correcting the difference of propagation speed of eddies. With a 25-member ensemble at low resolution, the SRDA computational overhead is 55 percent and the errors reduce by 40 percent, making the performance very close to that of the high-resolution system (52 percent of error reduction) that increases the cost by 800 percent. The reliability of the ensemble system is not degraded by SRDA.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Training a supermodel with noisy and sparse observations: a case study with CPT and the synch rule on SPEEDO – v.1

Schevenhoven, F., Carrassi, A. 2022: Training a supermodel with noisy and sparse observations: a case study with CPT and the synch rule on SPEEDO – v.1. Geosci. Model Dev. https://doi.org/10.5194/gmd-15-3831-2022

Summary: As an alternative to using the standard multi-model ensemble (MME) approach to combine the output of different models to improve prediction skill, models can also be combined dynamically to form a so-called supermodel. The supermodel approach enables a quicker correction of the model errors. In this study we connect different versions of SPEEDO, a global atmosphere-ocean-land model of intermediate complexity, into a supermodel. We focus on a weighted supermodel, in which the supermodel state is a weighted superposition of different imperfect model states. The estimation, “the training”, of the optimal weights of this combination is a critical aspect in the construction of a supermodel. In our previous works two algorithms were developed: (i) cross pollination in time (CPT)-based technique and (ii) a synchronization-based learning rule (synch rule). Those algorithms have so far been applied under the assumption of complete and noise-free observations. Here we go beyond and consider the more realistic case of noisy data that do not cover the full system’s state and are not taken at each model’s computational time step. We revise the training methods to cope with this observational scenario, while still being able to estimate accurate weights. In the synch rule an additional term is introduced to maintain physical balances, while in CPT nudging terms are added to let the models stay closer to the observations during training. Furthermore, we propose a novel formulation of the CPT method allowing the weights to be negative. This makes it possible for CPT to deal with cases in which the individual model biases have the same sign, a situation that hampers constructing a skillfully weighted supermodel based on positive weights. With these developments, both CPT and the synch rule have been made suitable to train a supermodel consisting of state of the art weather and climate models.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Estimation of Ocean Biogeochemical Parameters in an Earth System Model Using the Dual One Step Ahead Smoother: A Twin Experiment

Singh, T., Counillon, F., Tjiputra, J., Wang Y., El Gharamti, M. 2022: Estimation of Ocean Biogeochemical Parameters in an Earth System Model Using the Dual One Step Ahead Smoother: A Twin Experiment. Front. Mar. Sci. https://doi.org/10.3389/fmars.2022.775394 .

For an easy-to-understand overview of this publication, produced in collaboration with the TRIATLAS project, we recommend starting with this neat article written by Henrike Wilborn, at NERSC: “Making climate models more accurate by improving their tuning“.

Summary: Ocean biogeochemical (BGC) models utilise a large number of poorly-constrained global parameters to mimic unresolved processes and reproduce the observed complex spatio-temporal patterns. Large model errors stem primarily from inaccuracies in these parameters whose optimal values can vary both in space and time. This study aims to demonstrate the ability of ensemble data assimilation (DA) methods to provide high-quality and improved BGC parameters within an Earth system model in an idealized perfect twin experiment framework. We use the Norwegian Climate Prediction Model (NorCPM), which combines the Norwegian Earth System Model with the Dual-One-Step ahead smoothing-based Ensemble Kalman Filter (DOSA-EnKF). We aim to estimate five spatially varying BGC parameters by assimilating salinity and temperature profiles and surface BGC (Phytoplankton, Nitrate, Phosphate, Silicate, and Oxygen) observations in a strongly coupled DA framework—i.e., jointly updating ocean and BGC state-parameters during the assimilation. We show how BGC observations can effectively constrain error in the ocean physics and vice versa. The method converges quickly (less than a year) and largely reduces the errors in the BGC parameters. Some parameter error remains, but the resulting state variable error using the estimated parameters for a free ensemble run and for a reanalysis performs nearly as well as with true parameter values. Optimal parameter values can also be recovered by assimilating climatological BGC observations or sparse observational networks. The findings of this study demonstrate the applicability of the DA approach for tuning the system in a real framework.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Basin Interactions and Predictability. In: Interacting Climates of Ocean Basins: Observations, Mechanisms, Predictability, and Impacts

Keenlyside, N., Y. Kosaka, N. Vigaud, A. Robertson, Y. Wang, D. Dommenget, J.-J. Luo, and D. Matei. 2020: Basin Interactions and Predictability, In: Mechoso (Ed.). Interacting Climates of Ocean Basins Observations, Mechanisms, Predictability, and Impacts. Cambridge University Press, 2020, 258-292 .
Summary: The general public is familiar with weather forecasts and their utility, and the field of weather forecasting is well-established. Even the theoretical limit of the weather forecasting – two weeks – is known. In contrast, familiarity with climate prediction is low outside of the research field, the theoretical basis is not fully established, and we do not know the extent to which climate can be predicted. Variations in climate, however, can have large societal and economic consequences, as they can lead to droughts and floods, and spells of extreme hot and cold weather. Thus, improving our capabilities to predict climate is important and urgent, as it can enhance climate services and thereby contribute to the sustainable development of humans in this era of climate change.

Link to chapter. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Relating model bias and prediction skill in the equatorial Atlantic

Counillon, F., Keenlyside, N., Toniazzo, T., Koseki, K., Demissie, T., Bethke, I., Wang, Y. 2021: Relating model bias and prediction skill in the equatorial Atlantic. Climate Dynamics. https://doi.org/10.1007/s00382-020-05605-8

For a nice overview of the article, check out this news piece by our partner NERSC, also involved in our collaborative projects TRIATLAS and STERCP.

Summary: We investigate the impact of large climatological biases in the tropical Atlantic on reanalysis and seasonal prediction performance using the Norwegian Climate Prediction Model (NorCPM) in a standard and an anomaly coupled configuration. Anomaly coupling corrects the climatological surface wind and sea surface temperature (SST) fields exchanged between oceanic and atmospheric models, and thereby significantly reduces the climatological model biases of precipitation and SST. NorCPM combines the Norwegian Earth system model with the ensemble Kalman filter and assimilates SST and hydrographic profiles. We perform a reanalysis for the period 1980–2010 and a set of seasonal predictions for the period 1985–2010 with both model configurations. Anomaly coupling improves the accuracy and the reliability of the reanalysis in the tropical Atlantic, because the corrected model enables a dynamical reconstruction that satisfies better the observations and their uncertainty. Anomaly coupling also enhances seasonal prediction skill in the equatorial Atlantic to the level of the best models of the North American multi-model ensemble, while the standard model is among the worst. However, anomaly coupling slightly damps the amplitude of Atlantic Niño and Niña events. The skill enhancements achieved by anomaly coupling are largest for forecast started from August and February. There is strong spring predictability barrier, with little skill in predicting conditions in June. The anomaly coupled system show some skill in predicting the secondary Atlantic Niño-II SST variability that peaks in November–December from August 1st.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

North Atlantic climate far more predictable than models imply

Smith, D.M., Scaife, A.A., Eade, R. et al. 2020: North Atlantic climate far more predictable than models imply. Nature. https://doi.org/10.1038/s41586-020-2525-0 .

Summary: Quantifying signals and uncertainties in climate models is essential for the detection, attribution, prediction and projection of climate change1,2,3. Although inter-model agreement is high for large-scale temperature signals, dynamical changes in atmospheric circulation are very uncertain4. This leads to low confidence in regional projections, especially for precipitation, over the coming decades5,6. The chaotic nature of the climate system7,8,9 may also mean that signal uncertainties are largely irreducible. However, climate projections are difficult to verify until further observations become available. Here we assess retrospective climate model predictions of the past six decades and show that decadal variations in North Atlantic winter climate are highly predictable, despite a lack of agreement between individual model simulations and the poor predictive ability of raw model outputs. Crucially, current models underestimate the predictable signal (the predictable fraction of the total variability) of the North Atlantic Oscillation (the leading mode of variability in North Atlantic atmospheric circulation) by an order of magnitude. Consequently, compared to perfect models, 100 times as many ensemble members are needed in current models to extract this signal, and its effects on the climate are underestimated relative to other factors. To address these limitations, we implement a two-stage post-processing technique. We first adjust the variance of the ensemble-mean North Atlantic Oscillation forecast to match the observed variance of the predictable signal. We then select and use only the ensemble members with a North Atlantic Oscillation sufficiently close to the variance-adjusted ensemble-mean forecast North Atlantic Oscillation. This approach greatly improves decadal predictions of winter climate for Europe and eastern North America. Predictions of Atlantic multidecadal variability are also improved, suggesting that the North Atlantic Oscillation is not driven solely by Atlantic multidecadal variability. Our results highlight the need to understand why the signal-to-noise ratio is too small in current climate models10, and the extent to which correcting this model error would reduce uncertainties in regional climate change projections on timescales beyond a decade.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Combining data assimilation and machine learning to emulate a dynamical model from sparse and noisy observations: A case study with the Lorenz 96 model.

Brajard, J., Carrassi, A., Bocquet, M., Bertino, L. 2020: Combining data assimilation and machine learning to emulate a dynamical model from sparse and noisy observations: A case study with the Lorenz 96 model. Geoscientific Model Development. https://doi.org/10.1016/j.jocs.2020.101171 .

Summary: A novel method, based on the combination of data assimilation and machine learning is introduced. The new hybrid approach is designed for a two-fold scope: (i) emulating hidden, possibly chaotic, dynamics and (ii) predicting their future states. The method consists in applying iteratively a data assimilation step, here an ensemble Kalman filter, and a neural network. Data assimilation is used to optimally combine a surrogate model with sparse noisy data. The output analysis is spatially complete and is used as a training set by the neural network to update the surrogate model. The two steps are then repeated iteratively. Numerical experiments have been carried out using the chaotic 40-variables Lorenz 96 model, proving both convergence and statistical skill of the proposed hybrid approach. The surrogate model shows short-term forecast skill up to two Lyapunov times, the retrieval of positive Lyapunov exponents as well as the more energetic frequencies of the power density spectrum. The sensitivity of the method to critical setup parameters is also presented: the forecast skill decreases smoothly with increased observational noise but drops abruptly if less than half of the model domain is observed. The successful synergy between data assimilation and machine learning, proven here with a low-dimensional system, encourages further investigation of such hybrids with more sophisticated dynamics.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Ocean Biogeochemical Predictions—Initialization and Limits of Predictability

Fransner, F., Counillon, F., Bethke, I., Tjiputra, J., Samuelsen, A., Nummelin, A., Olsen, A. 2020: Ocean Biogeochemical Predictions—Initialization and Limits of Predictability. Front Mar Sci. https://doi.org/10.3389/fmars.2020.00386 .

Summary: Predictions of ocean biogeochemistry, such as primary productivity and CO2 uptake, would help to understand the changing marine environment and the global climate. There is an emerging number of studies where initialization of ocean physics has led to successful predictions of ocean biogeochemistry. It is, however, unclear how much these predictions could be improved by also assimilating biogeochemical data to reduce uncertainties of the initial conditions. Further, the mechanisms that lead to biogeochemical predictability are poorly understood. Here we perform a suite of idealized twin experiments with an Earth System Model (ESM) with the aim to (i) investigate the role of biogeochemical tracers’ initial conditions on their predictability, and (ii) understand the physical processes that give rise to, or limit, predictability of ocean carbon uptake and export production. Our results suggest that initialization of the biogeochemical state does not significantly improve interannual-to-decadal predictions, which we relate to the strong control ocean physics exerts on the biogeochemical variability on these time scales. The predictability of ocean carbon uptake generally agrees well with the predictability of the mixed layer depth (MLD), suggesting that the predictable signal comes from the exchange of dissolved inorganic carbon (DIC) with deep-waters. The longest predictability is found in winter in at high latitudes, as for sea surface temperature and salinity, but the predictability of the MLD and carbon exchange is lower as it is more directly influenced by the atmospheric variability, e.g., the wind. The predictability of the annual mean export production is, on the contrary, nearly non-existing at high latitudes, despite the strong predictive skill for annual mean nutrient concentrations in these regions. This is related to the low predictability of the physical state of the summer surface ocean. Due to the shallow mixed layer it is decoupled from the ocean below and therefore strongly influenced by the chaotic atmosphere. Our results show that future studies need to target the predictability of the mixed layer to get a better understanding of the real-world predictability of ocean biogeochemistry.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.

Seasonal to decadal predictions of regional Arctic sea ice by assimilating sea surface temperature in the Norwegian Climate Prediction Model

Dai, P., Gao, Y., Counillon, F., Wang, Y., Kimmritz, M., Langehaug, H.R. 2020: Seasonal to decadal predictions of regional Arctic sea ice by assimilating sea surface temperature in the Norwegian Climate Prediction Model. Clim Dyn 54, 3863–3878. https://doi.org/10.1007/s00382-020-05196-4 .

Summary: The version of the Norwegian Climate Prediction Model (NorCPM) that only assimilates sea surface temperature (SST) with the Ensemble Kalman Filter has been used to investigate the seasonal to decadal prediction skill of regional Arctic sea ice extent (SIE). Based on a suite of NorCPM retrospective forecasts, we show that seasonal prediction of pan-Arctic SIE is skillful at lead times up to 12 months, which outperforms the anomaly persistence forecast. The SIE skill varies seasonally and regionally. Among the five Arctic marginal seas, the Barents Sea has the highest SIE prediction skill, which is up to 10–11 lead months for winter target months. In the Barents Sea, the skill during summer is largely controlled by the variability of solar heat flux and the skill during winter is mostly constrained by the upper ocean heat content/SST and also related to the heat transport through the Barents Sea Opening. Compared with several state-of-the-art dynamical prediction systems, NorCPM has comparable regional SIE skill in winter due to the improved upper ocean heat content. The relatively low skill of summer SIE in NorCPM suggests that SST anomalies are not sufficient to constrain summer SIE variability and further assimilation of sea ice thickness or atmospheric data is expected to increase the skill.

Link to publication. You are most welcome to contact us or the corresponding author(s) directly, if you have questions.