Category: PublicationsRA2

Impact of initialization methods on the predictive skill in NorCPM: an Arctic–Atlantic case study

Passos, L., Langehaug, HR., Årthun, M., Eldevik, T., Bethke, I., Kimmritz, M. 2022: Impact of initialization methods on the predictive skill in NorCPM: an Arctic–Atlantic case study. Clim Dyn. https://doi.org/10.1007/s00382-022-06437-4

Summary: The skilful prediction of climatic conditions on a forecast horizon of months to decades into the future remains a main scientific challenge of large societal benefit. Here we assess the hindcast skill of the Norwegian Climate Prediction Model (NorCPM) for sea surface temperature (SST) and sea surface salinity (SSS) in the Arctic–Atlantic region focusing on the impact of different initialization methods. We find the skill to be distinctly larger for the Subpolar North Atlantic than for the Norwegian Sea, and generally for all lead years analyzed. For the Subpolar North Atlantic, there is furthermore consistent benefit in increasing the amount of data assimilated, and also in updating the sea ice based on SST with strongly coupled data assimilation. The predictive skill is furthermore significant for at least two model versions up to 8–10 lead years with the exception for SSS at the longer lead years. For the Norwegian Sea, significant predictive skill is more rare; there is relatively higher skill with respect to SSS than for SST. A systematic benefit from more complex data assimilation approach can not be identified for this region. Somewhat surprisingly, skill deteriorates quite consistently for both the Subpolar North Atlantic and the Norwegian Sea when going from CMIP5 to corresponding CMIP6 versions. We find this to relate to change in the regional performance of the underlying physical model that dominates the benefit from initialization.

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

Editorial: Past Reconstruction of the Physical and Biogeochemical Ocean State

Masina, S., Counillon, F., Grégoire, M., Storto, A., Tsujino, H. 2022: Editorial: Past Reconstruction of the Physical and Biogeochemical Ocean State. Front Earth Sci. https://doi.org/10.3389/feart.2022.890370

Summary: Knowledge of the ocean’s physical, biogeochemical and ecosystem state and variability is crucial for understanding the evolution of our climate system and better predicting its future. However, the sparseness and inhomogeneous distribution of observations hinder the creation of sound 4-dimensional reconstructions of the past (for an overview of ocean observing systems see the Research Topic Oceanobs’19: An Ocean of Opportunity). Instead, we must rely on a combination of ocean modeling and data analysis to infer past changes. Over the last decade the quality of ocean reanalyses has improved mainly thanks to advances in data assimilation methods and more quality-controlled observation data sets. Reanalyses provide the best-possible state estimate by assimilating observations into a dynamical model (Balmaseda et al., 2015; Masina and Storto, 2017; Storto et al., 2019). In addition, advanced statistical mapping methods (e.g., objective or variational analysis) provide observation-based gridded fields whose resolution depends on the amount of available data (among many Cheng et al., 2017, Ishii et al., 2017; Boyer et al., 2018). For many variables, particularly biogeochemical, the lack of observations more strongly limits the spatial and temporal resolution of these gridded products (Fennel et al., 2019).

The Research Topic gathers contributions aiming at reconstructing the past physical, sea ice and biogeochemical state of the ocean using models in combination with data. Ocean reanalyses and observation-mapping are proposed to further our knowledge, to demonstrate their use in supporting various applications, and to increase confidence in these reconstructions within the scientific community. The products and applications described in this topic provide a foundation for their use in ecosystem-based management, policy advice to support mitigation and adaptation strategies, and in the identification of pathways towards a sustainable ocean.

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.

WMO Global Annual to Decadal Climate Update: A Prediction for 2021–25

Hermanson, L., Smith, D., Seabrook, M., Bilbao, R., Doblas-Reyes, F., Tourigny, E., Lapin, V., Kharin, V.V., Merryfield, W.J., Sospedra-Alfonso, R., Athanasiadis, P., Nicoli, D., Gualdi, S., Dunstone, N., Eade, R., Scaife, A., Collier, M., O’Kane, T., Kitsios, V., Sandery, P., Pankatz, K., Früh, B., Pohlmann, H., Müller, W., Kataoka, T., Tatebe, H., Ishii M., Imada, Y., Kruschke, T., Koenigk, T., Pasha Karami, M., Yang, S., Tian, T., Zhang, L., Delworth, T., Yang, X., Zeng, F., Wang, Y., Counillon, F., Keenlyside, N.S., Bethke, I., Lean, J., Luterbacher, J., Kumar Kolli, R., Kumar, A. 2022: WMO Global Annual to Decadal Climate Update: A Prediction for 2021–25. BAMS https://doi.org/10.1175/BAMS-D-20-0311.1 .

Summary: As climate change accelerates, societies and climate-sensitive socioeconomic sectors cannot continue to rely on the past as a guide to possible future climate hazards. Operational decadal predictions offer the potential to inform current adaptation and increase resilience by filling the important gap between seasonal forecasts and climate projections. The World Meteorological Organization (WMO) has recognized this and in 2017 established the WMO Lead Centre for Annual to Decadal Climate Predictions (shortened to “Lead Centre” below), which annually provides a large multimodel ensemble of predictions covering the next 5 years. This international collaboration produces a prediction that is more skillful and useful than any single center can achieve. One of the main outputs of the Lead Centre is the Global Annual to Decadal Climate Update (GADCU), a consensus forecast based on these predictions. This update includes maps showing key variables, discussion on forecast skill, and predictions of climate indices such as the global mean near-surface temperature and Atlantic multidecadal variability. it also estimates the probability of the global mean temperature exceeding 1.5°C above preindustrial levels for at least 1 year in the next 5 years, which helps policy-makers understand how closely the world is approaching this goal of the Paris Agreement. This paper, written by the authors of the GADCU, introduces the GADCU, presents its key outputs, and briefly discusses its role in providing vital climate information for society now and in the future..

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.

NorCPM1 and its contribution to CMIP6 DCPP

Bethke, I., Wang, Y., Counillon, F., Keenlyside, N., Kimmritz, M., Fransner, F., Samuelsen, A., Langehaug, H., Svendsen, L., Chiu, P.-G., Passos, L., Bentsen, M., Guo, C., Gupta, A., Tjiputra, J., Kirkevåg, A., Olivié, D., Seland, Ø., Solsvik Vågane, J., Fan, Y., Eldevik, T. 2021: NorCPM1 and its contribution to CMIP6 DCPP. Geosci Model Dev. https://doi.org/10.5194/gmd-14-7073-2021 .

For an easy-to-understand overview, we recommend starting with this neat article written by the Climate Futures team, a project connected to BCPU: “New Study: Decadal Climate Forecasts From The Norwegian Climate Prediction Model” (les heller på norsk).

Summary: The Norwegian Climate Prediction Model version 1 (NorCPM1) is a new research tool for performing climate reanalyses and seasonal-to-decadal climate predictions. It combines the Norwegian Earth System Model version 1 (NorESM1) – which features interactive aerosol–cloud schemes and an isopycnic-coordinate ocean component with biogeochemistry – with anomaly assimilation of sea surface temperature (SST) and -profile observations using the ensemble Kalman filter (EnKF).

We describe the Earth system component and the data assimilation (DA) scheme, highlighting implementation of new forcings, bug fixes, retuning and DA innovations. Notably, NorCPM1 uses two anomaly assimilation variants to assess the impact of sea ice initialization and climatological reference period: the first (i1) uses a 1980–2010 reference climatology for computing anomalies and the DA only updates the physical ocean state; the second (i2) uses a 1950–2010 reference climatology and additionally updates the sea ice state via strongly coupled DA of ocean observations.

We assess the baseline, reanalysis and prediction performance with output contributed to the Decadal Climate Prediction Project (DCPP) as part of the sixth Coupled Model Intercomparison Project (CMIP6). The NorESM1 simulations exhibit a moderate historical global surface temperature evolution and tropical climate variability characteristics that compare favourably with observations. The climate biases of NorESM1 using CMIP6 external forcings are comparable to, or slightly larger than those of, the original NorESM1 CMIP5 model, with positive biases in Atlantic meridional overturning circulation (AMOC) strength and Arctic sea ice thickness, too-cold subtropical oceans and northern continents, and a too-warm North Atlantic and Southern Ocean. The biases in the assimilation experiments are mostly unchanged, except for a reduced sea ice thickness bias in i2 caused by the assimilation update of sea ice, generally confirming that the anomaly assimilation synchronizes variability without changing the climatology. The i1 and i2 reanalysis/hindcast products overall show comparable performance. The benefits of DA-assisted initialization are seen globally in the first year of the prediction over a range of variables, also in the atmosphere and over land. External forcings are the primary source of multiyear skills, while added benefit from initialization is demonstrated for the subpolar North Atlantic (SPNA) and its extension to the Arctic, and also for temperature over land if the forced signal is removed. Both products show limited success in constraining and predicting unforced surface ocean biogeochemistry variability. However, observational uncertainties and short temporal coverage make biogeochemistry evaluation uncertain, and potential predictability is found to be high. For physical climate prediction, i2 performs marginally better than i1 for a range of variables, especially in the SPNA and in the vicinity of sea ice, with notably improved sea level variability of the Southern Ocean. Despite similar skills, i1 and i2 feature very different drift behaviours, mainly due to their use of different climatologies in DA; i2 exhibits an anomalously strong AMOC that leads to forecast drift with unrealistic warming in the SPNA, whereas i1 exhibits a weaker AMOC that leads to unrealistic cooling. In polar regions, the reduction in climatological ice thickness in i2 causes additional forecast drift as the ice grows back. Posteriori lead-dependent drift correction removes most hindcast differences; applications should therefore benefit from combining the two products.

The results confirm that the large-scale ocean circulation exerts strong control on North Atlantic temperature variability, implying predictive potential from better synchronization of circulation variability. Future development will therefore focus on improving the representation of mean state and variability of AMOC and its initialization, in addition to upgrades of the atmospheric component. Other efforts will be directed to refining the anomaly assimilation scheme – to better separate internal and forced signals, to include land and atmosphere initialization and new observational types – and improving biogeochemistry prediction capability. Combined with other systems, NorCPM1 may already contribute to skilful multiyear climate prediction that benefits society.

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

Training of supermodels in the context of weather and climate forecasting (PhD thesis)

Schevenhoven, Francine (2021-02-08). Training of supermodels in the context of weather and climate forecasting (PhD thesis, University of Bergen, Bergen, Norway). https://bora.uib.no/bora-xmlui/handle/11250/2727454 .

Summary: Given a set of imperfect weather or climate models, predictions can be improved by combining the models dynamically into a so called `supermodel’. The models are optimally combined to compensate their individual errors. This is different from the standard multi-model ensemble approach (MME), where the model output is statistically combined after the simulations. Instead, the supermodel can create a trajectory closer to observations than any of the imperfect models. By intervening during the forecast, errors can be reduced at an early stage and the ensemble can exhibit different dynamical behavior than any of the individual models. In this way, common errors between the models can be removed and new, physically correct behavior can appear.
In our simplified context of models sharing the same evolution function and phase space, we can define either a connected or a weighted supermodel. A connected supermodel uses nudging to bring the models closer together, while in a weighted supermodel all model states are replaced at regular time intervals (i.e., restarted) by the weighted average of the individual model states. To obtain optimal connection coefficients or weights, we need to train the supermodel on the basis of historical observations. A standard training approach such as minimization of a cost function requires many model simulations, which is computationally very expensive. This thesis has focused on developing two new methods to efficiently train supermodels. The first method is based on an idea called cross pollination in time, where models exchange states during the training. The second method is a synchronization-based learning rule, originally developed for parameter estimation.
The techniques are developed on low-order systems, such as Lorenz63, and later applied to different versions of the intermediate-complexity global coupled atmosphere-ocean-land model SPEEDO. Here the observations are from the same models, but with different parameters. The applicability of the method to real observations is tested using sensitivity to noisy and incomplete data. The characteristics the individual models should have in order to be combined together into a supermodel are identified, as well as which physical variables should be connected in a supermodel, and which ones should not. Both training methods result in supermodels that outperform both the individual models and the MME, for short term predictions as well as long term simulations. Furthermore, we show that the novel use of negative weights can improve predictions in cases where model errors do not cancel (for instance, all models are too warm with respect to the truth). A crucial advantage of the proposed training schemes is that in the present context relatively short training periods suffice to find good solutions. Although the validity of our conclusions in the context of real observations and model scenarios has yet to be proved, our results are very encouraging. In principle, the methods are suitable to train supermodels constructed using 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.

Relating model bias and prediction skill in the equatorial Atlantic

Counillon, F., Keenlyside, N., Toniazzo, T., Koseki, S., 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.

Mechanisms and pathways of ocean heat anomalies in the Arctic-Atlantic region (PhD thesis)

Asbjørnsen, Helene (2020-12-10). Mechanisms and pathways of ocean heat anomalies in the Arctic-Atlantic region (PhD thesis, University of Bergen, Bergen, Norway). https://bora.uib.no/bora-xmlui/handle/11250/2712025 .

Summary: Along the Atlantic water pathway, from the Gulf Stream in the south to the Arctic Ocean in the north, variability in ocean heat content is pronounced on interannual to decadal time scales. Ocean heat anomalies in this Arctic-Atlantic sector are known to affect Arctic sea ice extent, marine ecosystems, and continental climate. However, there is at present neither consensus nor any complete understanding of the mechanisms causing such heat anomalies. This dissertation obtains a more robust understanding of regional ocean heat content variability by assessing the mechanisms and pathways of ocean heat anomalies in the Arctic-Atlantic region. The results are presented in three papers.

The first paper investigates the link between a variable Nordic Seas inflow and large- scale ocean circulation changes upstream. Using a global, eddy-permitting ocean hind- cast together with a Lagrangian analysis tool, numerical particles are seeded at the Iceland-Scotland Ridge and tracked backward in time. Water from the subtropics sup- plied by the North Atlantic Current (NAC) is found to be the main component of the Nordic Seas inflow (64%), while 26% of the inflow has a subpolar or Arctic origin. Different atmospheric patterns are seen to affect the circulation strength along the advective pathways, as well as the supply of subtropical and Arctic-origin water to the ridge through shifts in the NAC and the subpolar front. A robust link between a high transport of Arctic-origin water and a cold and fresh inflow is furthermore established, while a high transport of subtropical water leads to higher inflow salinities. The second paper investigates the mechanisms of interannual heat content variability in the Norwegian Sea downstream of the Iceland-Scotland Ridge, using a state-of-the-art ocean state estimate and closed heat budget diagnostics. Ocean advection is found to be the primary contributor to heat content variability in the Atlantic domain of the Norwegian Sea, although local surface fluxes also play an active role. Anomalous heat advection furthermore depends on the strength of the Atlantic water inflow and the conditions upstream of the ridge. Combined, the two papers demonstrate the importance of gyre dynamics and large-scale wind forcing in causing variability at the ridge, while high- lighting the impacts on Norwegian Sea heat content downstream.

For the third paper, warming trends in the Barents Sea and Fram Strait are explored, and, thus, the mechanisms underlying recent Atlantification of the Arctic Ocean. The Barents Sea is seen to transition to a warmer state, with reduced sea ice concentrations and Atlantic water extending further poleward. The mechanisms driving the warming are, however, found to be regionally dependent and not stationary in time. In the ice- free region, ocean advection is found to be a major driver of the warming trend due to increasing inflow temperatures in the late 1990s and early 2000s, while reduced ocean heat loss is contributing to the warming trend from the mid-2000s and onward. A considerable upper-ocean warming and a weakened stratification is seen in the ice- covered northwestern Barents Sea. However, in contrast to what has been previously hypothesized, the results do not point to increased upward heat fluxes from the Atlantic water layer to the Arctic surface layer as the source of the upper-ocean warming.

The supply of Atlantic heat to the Nordic Seas and the Arctic Ocean has been scrutinized using both Lagrangian methods and heat budget diagnostics. Combined, the three papers demonstrate the important role of ocean heat transport in causing regional heat content variability and change in the Arctic-Atlantic region. A better understanding of interannual to decadal ocean heat content variability has implications for future prediction efforts, and for how we understand the ocean’s role in ongoing and future climate change.

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.