Search

An anomalous signal-to-noise ratio (also called the signal-to-noise paradox) present in climate models has been widely reported, affecting predictions and projections from seasonal to centennial timescales and encompassing prediction skill from internal processes and external climate forcing. An anomalous signal-to-noise ratio describes a situation where the mean of a forecast ensemble correlates better with the corresponding verification than with its individual ensemble members. This situation has severe implications for climate science, meaning that large ensembles might be required to extract prediction signals. Although a number of possible physical mechanisms for this paradox have been proposed, none has been universally accepted. From a statistical point of view, an anomalous signal-to-noise ratio indicates that forecast ensemble members are not statistically interchangeable with the verification, and an apparent paradox arises only if such an interchangeability is assumed. It will be demonstrated in this study that an anomalous signal-to-noise ratio is a consequence of the relative magnitudes of the variance of the observations, the ensemble mean, and the error of the ensemble mean. By analysing the geometric triangle formed by these three quantities, and given that for typical seasonal forecasting systems both the correlation and the forecast signal are relatively small, it is concluded that an anomalous signal-to-noise ratio should, in fact, be expected in such circumstances.

The Arctic has warmed substantially over the last decades and will continue to do so owing to global warming in conjunction with polar amplification. The changing mean state poses many challenges to the construction, evaluation and calibration of subseasonal-to-seasonal forecasting systems, because it puts into question the representativeness of the system's retrospective forecasts (reforecasts). Furthermore, any inconsistencies with observed trends degrade the forecast skill and point to deficiencies in either the physical modelling or the initialization methods. Here, we assess the consistency of boreal winter trends of surface air temperature (SAT) in the Eurasian Arctic between the ERA5 reanalysis and ECMWF sub-seasonal reforecasts initialised from ERA5, for the 35-year period 1986-2021. We present methods to quantify robustness and importance of the observed trends, and to quantify the consistency of reforecast trends with these observed trends. We find that, in large parts of the marine Arctic, the reforecasts clearly underestimate the reanalsyis warming trend of about 1 K per decade at lead times beyond two weeks. For longer lead times, the reforecast trend is less than half of the reanalysis trend, with very high statistical significance. We present a series of numerical experiments to investigate potential reasons for the trend underestimation. These concern the sea-ice thermodynamic coupling to the atmosphere, impact of sea surface temperatures, and possible remote atmospheric influences from the North Atlantic and the Tropics. The outcome of these experiments provides guidance for future improvements in the physical forecast model and data assimilation methods needed to faithfully represent and predict Arctic climate variability and change.

The Indian Ocean Dipole (IOD) is a major source of seasonal climate variability. This dipole has strong impacts on the Indian Ocean region and through teleconnections can influence the seasonal climate of remote regions. The strong IOD of 2019/2020 influenced the state and predictability of the Northern Hemisphere extratropics. Thus, a good understanding of the mechanism of this teleconnection is desirable. In this contribution we investigate the special teleconnection of the winter 2019/2020.In model experiments with the OpenIFS from ECMWF we show that the NAO in the winter 2019/2020 is influenced by the IOD. We use hindcast ensembles of the DJF season 2019/2020 to analyse this teleconnection. In model experiments, the Sea Surface Temperature (SST) boundary conditions are perturbed. With these experiments we identify the relative importance of individual ocean regions to the state of the NAO in the winter of 2019/2020. We contrast the experiments to the operational ECMWF System5 forecast and ERA5. Experiments with the 2019/2020 SSTs in the Indian Ocean (with other boundary conditions set to climatology) reproduce many of the observed atmospheric 2019/2020 features. In contrast, experiments with SSTs in the Pacific show very different patterns. We identify eddy-mean-flow interactions as a mechanism that connects and transports information from the Indian Ocean to the North Atlantic. With Hoskins E-Vectors we show that anomalous eddy activity during IOD events impacts the position and strength of the Northern Hemisphere extratropical jet. This interaction provides a teleconnection mechanism in addition to the Rossby-wavetrain discussed in other studies., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

Heatwaves constitute a major threat to human health and ecosystems. Projected increases in heatwave frequency and severity thus lead to the need for prediction to enhance preparedness and minimize adverse impacts. In this Review, we document current capabilities for heatwave prediction at daily to decadal timescales and outline projected changes under anthropogenic warming. Various local and remote drivers and feedbacks influence heatwave development. On daily timescales, extratropical atmospheric blocking and global land–atmosphere coupling are most pertinent, and on subseasonal to seasonal timescales, soil moisture and ocean surface anomalies contribute. Knowledge of these drivers allows heatwaves to be skilfully predicted at daily to weekly lead times. Predictions are challenging beyond timescales of a few weeks, but tendencies for above-average temperatures can be estimated. Further into the future, heatwaves are anticipated to become more frequent, persistent and intense in nearly all inhabited regions, with trends amplified by soil drying in some areas, especially the mid-latitudes. There is also an increased occurrence of humid heatwaves, especially in southern Asia. A better understanding of the relevant drivers and their model representation, including atmospheric dynamics, atmospheric and soil moisture, and surface cover should be prioritized to improve heatwave prediction and projection. ISSN:2662-138X

Extreme weather attribution, quantifying the role of human influence in specific weather events, is of interest to scientists, adaptation planners and the general public. However, the devastating 2021 Pacific Northwest heatwave challenged conventional statistical approaches to attribution due to the absence of similar events in the historical record, and model-based approaches due to poor representation of key causal processes in current climate models. Here we use state-of-the-art operational medium-range and seasonal weather prediction systems, applied for the first time to this kind of climate question and unequivocally able to simulate the detailed physics of the heatwave in question, to show that human influence on the climate made this event at least 8 [2--30] times more likely to occur. Quantifying the absolute probability of such an unprecedented event is more challenging, but the length of the observational record suggests at least a multi-decade return-time in the current climate, with the likelihood doubling every 17 [10--50] years at the current rate of global warming. Our forecast-based approach synthesises the storyline approach, which examines human influence on the physical drivers of an event in a deterministic manner, and the probabilistic approach, which assesses how the frequency of a class of events has been affected by human influence. If developed as a routine service in a number of forecasting centres, it could provide reliable estimates of the changing probabilities of all extreme events that can be represented in forecast models, which is critical to supporting effective adaptation planning.

The simulation and prediction of winter Northern Hemisphere atmospheric blocking in the seasonal prediction systems from the European Centre for Medium‐Range Weather Forecasts (ECMWF) is analysed. Blocking statistics from the operational November‐initialised seasonal hindcasts are evaluated in three generations of models: System3, System4, and System5 (SEAS5). Improvements in the climatological representation of blocking are observed in the most recent model configurations, with reduced bias over North Pacific and Greenland. Minor progress is seen over the European sector, where SEAS5 still underestimates the observed blocking frequency. SEAS5 blocking interannual variability is underestimated too and is proportional to the climatological frequency, highlighting that a negative bias in the blocking frequency implies an underestimation of the interannual variance. SEAS5 predictive skill and signal‐to‐noise ratio remain low, but interesting positive results are found over Western and Central Europe. Improved forecasts with reduced ensemble spread are obtained during El Niño years, especially at low latitudes. Complementary experiments show that the statistics of blocking are improved following atmospheric and oceanic resolution increase. Conversely, they remain largely insensitive to coupled model sea‐surface temperature (SST) errors. On the other hand, the implementation of stochastic parameterisations tends to displace blocking activity equatorward. Finally, by comparing seasonal hindcasts with climate runs using the same model, we highlight that the largest contributors to the chronic underestimation of blocking are persistent errors in the atmospheric model. It is also shown that SST errors have a larger impact on blocking bias in climate runs than in seasonal runs, and that increased ocean model resolution contributes to improved blocking more effectively in climate runs. Seasonal forecasts can thus be considered a suitable test‐bed for model development targeting blocking improvement in climate models.

The European Climate Prediction system (EUCP) project aimed to lay the foundation for a future regional climate prediction system for Europe. An important element of this is the role of narrative or storylines approaches in data production and scientific investigation, as well as a user product. We will present the results of our investigations which sought to understand the potential advantages and challenges in developing physically based climate storylines as part of a climate service by addressing the questionsWhat are climate storylines and where are they useful? How could storylines bring together various outputs of EUCP state-of-the-art climate science attempting to reduce uncertainty and complexity in climate projections, and seamlessly combine them with decadal predictions? What are the challenges of producing them as a service? The body of EUCP work included two case studies of co-producing storylines as a user product, revealing the potential usefulness for applications. These included the heritage and water supply management sectors which are at different stages of adaptation management. We also reflect on the potential of event-based future storylines: one using a convection-permitting model to provide attribution statements for the Copenhagen flooding event in July 2011 and another using large ensembles to construct storylines of the European summer 2018 drought under different pseudo-global warming. Novel scientific studies were also performed which could form the scientific building blocks of climate storylines including an algorithmic clustering approach and approaches to producing more realistic climate variability. Finally, we will present a tool for performing multiple lines of evidence assessments that could aid storylines development.

In order to explore temporal changes of predictability of El Niño Southern Oscillation (ENSO), a novel set of global biennial climate reforecasts for the historical period 1901-2010 has been generated using a modern initialized coupled forecasting system. We find distinct periods of enhanced long-range skill at the beginning and at the end of the twentieth century, and an extended multi-decadal epoch of reduced skill during the 1930s-1950s. Once the forecast skill extends beyond the first spring barrier, the predictability limit is much enhanced and our results provide support for the feasibility of skillful ENSO forecasts up to 18 months. Changes in the mean state, variability (amplitude), persistence, seasonal cycle and predictability suggest that multi-decadal variations in the dynamical characteristics of ENSO rather than the data coverage and quality of the observations have primarily driven the reported non-monotonic skill modulations.

There is an increasing demand from society and policy makers for reliable, robust and actionable climate information for the upcoming 40 years. However, providing such seamless information poses a challenge to the scientific community. Traditionally, the scientific community developed targeted forecasts for specific time scales, e.g. short-term, seasonal or decadal predictions. These model integrations have thus a limited forecast period and do not provide seamless information on time scales up to 40 years.This work discusses two alternative approaches to combine information from initialized decadal predictions (providing information up to ten years) with uninitialized climate projections (available until 2100 and beyond). The first is a novel framework, which is designed to implicitly make use of the (added) values from initialization by constraining uninitialized climate projections using decadal predictions. This approach is applied to near-surface temperatures over the North Atlantic Subpolar gyre region from CMIP5 models. Results suggest that such a constraining approach is able to provide more skillful, seamless climate information beyond decadal time-scales compared to using unconstrained climate projections. The second approach is based on the simple temporal concatenation of decadal predictions and climate projections. It is shown that this can introduce inconsistencies, which may impact the usability for potential end users. Two different methods to overcome these issues are discussed: the application of a simple calibration method and a weighting scheme based on model performance. Results for the calibration method are in general promising, whereas the impact of the model weighting scheme is smaller. The latter is mainly associated with the small size of the decadal prediction ensemble, which hinders the usual application of the weighting scheme as done in previous studies based on much larger ensembles.

A major problem in climate science is providing reliable regional climate projections. Existing model ensembles are imperfect so a number of different methods for constraining future climate projections using the observational record have been developed, with the aim of improving the reliability of the projections. Whilst these constrained projections displayed some overlap and common characteristics, there were also some distinct differences. However, how these different projections should be interpreted is not clear. For example, are some of the methods more accurate and reliable than others?To address this question, we performed a novel out-of-sample “perfect model” experiment to test the different constraining methods. In this experiment we took advantage of newly available CMIP6 archive of coupled climate simulations to act as (anonymised) pseudo-observations over only the observational period, to test the methods which were all previously used CMIP5-era model data; for this reason, we refer to this as an out-of-sample test. Five different groups used these pseudo-observations to constrain their CMIP5-era model projections. Detailed results from these projections, verified against the target pseudo-observations, will be presented. The constraining methods all improve the projections for regional summer temperatures but for precipitation and in the winter season the improvements are less clear. There are substantial differences across the constraining methods and reasons for these and the utility of these different methods will be discussed.

Our study shows the relevance of the interactions between atmospheric chemistry and physics to achieve better weather predictions. We provide evidence of the role that stratospheric ozone plays to improve weather forecasts on several timescales, and highlight the need for seamless models to include realistic prognostic ozone interactive with radiation.Representing realistic feedbacks between ozone, radiation, temperature and dynamics is essential to correctly simulate the behaviour of the stratosphere and its links with tropospheric weather and climate. We have implemented an alternative stratospheric ozone model (Monge-Sanz et al., 2011) in the ECMWF system interactively with radiation, and we have assessed its performance and feedbacks with meteorological fields for different timescales, from medium-range to seasonal. Here we will discuss results from the experiments and analyses conducted in our study (Monge-Sanz et al., 2021), showing the feasibility of this ozone model for a seamless numerical weather prediction approach.We will show that the stratospheric ozone distribution provided by this new prognostic ozone model compares very well with observations even for unusual meteorological conditions. On assessing impacts on meteorological variables, the new ozone model improves the representation of the stratosphere, clearly reducing temperature biases in this region. We will also show the benefits it brings to tropospheric meteorological fields, highlighting the potential of this new ozone description to exploit stratospheric sources of predictability and improve weather predictions over Europe on a range of time scales.Therefore, our results show the value of this prognostic stratospheric ozone model for seamless Earth System Models, as well as for global systems where atmospheric composition is coupled to weather forecasts, like the systems being run within the Copernicus Atmosphere Monitoring Service (CAMS). We will also discuss challenges and possible strategies for the inclusion of chemistry-dynamics feedbacks in seamless Earth System Models. References:Monge-Sanz BM, Chipperfield MP, Cariolle D, Feng W. Results from a new linear O3 scheme with embedded heterogeneous chemistry compared with the parent full-chemistry 3-D CTM. Atmos. Chem. Phys. 11, 1227-1242, 2011.Monge-Sanz, B. M., Bozzo, A., Byrne, N., Chipperfield, M. P., Diamantakis, M., Flemming, J., Gray, L. J., Hogan, R. J., Jones, L., Magnusson, L., Polichtchouk, I., Shepherd, T. G., Wedi, N., and Weisheimer, A.: A stratospheric prognostic ozone for seamless Earth System Models: performance, impacts and future, Atmos. Chem. Phys., accepted, https://doi.org/10.5194/acp-2020-1261, 2021.

Here we explore the use of “counterfactual” weather forecasts, using perturbed initial condition runs of a state-of-the-art high-resolution coupled ocean-atmosphere-sea-ice ensemble NWP system, for the attribution of extreme weather events to anthropogenic climate change. We use the “record-shattering” heatwave experienced by Western North America during summer 2021 as a case study - though our forecast-based approach is applicable to other events.Since we cannot make direct observations of a world without human influence on climate, all approaches to extreme event attribution involve some kind of modelling, either statistical or numerical. Both approaches struggle with the most extreme weather events, which are poorly represented in both observational records and the climate models normally used for attribution studies. Recognising the compromises involved, researchers have traditionally relied on comparing results from several different approaches to assess the robustness of conclusions. We argue that a better approach would be to use initialised numerical models that have demonstrated their ability to simulate the event in question through a successful forecast.This work represents a continuation of a previous EGU talk and published study (https://meetingorganizer.copernicus.org/EGU21/EGU21-5731.html & https://doi.org/10.1073/pnas.2112087118), in which we used demonstrably successful weather forecasts to estimate the direct impact of increased CO2 concentrations (one component, but not the entirety, of human influence) on the 2019 European winter heatwave. In the previous and current work we use the operational ECMWF ensemble prediction system. This state-of-the-art weather forecast system is run at a much higher resolution (Tco639 / 18km) than most climate model simulations - important as even small reductions in resolution often change the representation of extreme events in numerical models. Using a reliable forecast ensemble allows us to quantify the associated uncertainties in our attribution analyses.We have built on this work with the aim of providing a more complete estimate of the human influence on an isolated extreme event. In addition to the reduction of CO2 concentrations back to pre-industrial levels, we now also remove an estimate of the human influence on 3D ocean temperatures since the pre-industrial period from the initial state of the forecast model. These changes allow the model to provide a “counterfactual” picture of what an extreme event might have looked like if it had occurred before human influence on the climate.Using this perturbed initial condition approach, we produce counterfactual forecasts of the Pacific Northwest heatwave at the end of June 2021. This event broke records throughout Western North America, including a new Canadian high temperature record of 49.6°C, shattering the previous record by almost 5°C. The heatwave was driven by a combination of meteorological factors, including an omega block and water vapour transport at the synoptic scale, and high solar irradiation and subsidence at the meso-scale (research into the drivers is ongoing). Crucially, the event was well-predicted by weather forecast models over a week in advance.We estimate the human contribution to this exceptional heatwave by comparing our counterfactual forecasts to the operational forecasts that successfully predicted the event.

We have implemented a new stratospheric ozone model in the European Centre for Medium-Range Weather Forecasts (ECMWF) system and tested its performance for different timescales to assess the impact of stratospheric ozone on meteorological fields. We have used the new ozone model to provide prognostic ozone in medium-range and long-range (seasonal) experiments, showing the feasibility of this ozone scheme for a seamless numerical weather prediction (NWP) modelling approach. We find that the stratospheric ozone distribution provided by the new scheme in ECMWF forecast experiments is in very good agreement with observations, even for unusual meteorological conditions such as Arctic stratospheric sudden warmings (SSWs) and Antarctic polar vortex events like the vortex split of year 2002. To assess the impact it has on meteorological variables, we have performed experiments in which the prognostic ozone is interactive with radiation. The new scheme provides a realistic ozone field able to improve the description of the stratosphere in the ECMWF system, as we find clear reductions of biases in the stratospheric forecast temperature. The seasonality of the Southern Hemisphere polar vortex is also significantly improved when using the new ozone model. In medium-range simulations we also find improvements in high-latitude tropospheric winds during the SSW event considered in this study. In long-range simulations, the use of the new ozone model leads to an increase in the correlation of the winter North Atlantic Oscillation (NAO) index with respect to ERA-Interim and an increase in the signal-to-noise ratio over the North Atlantic sector. In our study we show that by improving the description of the stratospheric ozone in the ECMWF system, the stratosphere–troposphere coupling improves. This highlights the potential benefits of this new ozone model to exploit stratospheric sources of predictability and improve weather predictions over Europe on a range of timescales.

The East Asian summer jet (EASJ) is an important component of the East Asian summer monsoon system and its variability is correlated with precipitation and surface temperature variations over this region. Whilst many studies have considered the interannual variability of the EASJ, less is known about variations on a decadal time-scale. This study investigates the relationship between decadal EASJ variability and sea surface temperatures (SSTs) and thus the potential predictability that SSTs may provide. Given the relatively short observational record, we make use of the long pre-industrial control simulations in the Coupled Model Intercomparison Project phase 6 (CMIP6) in addition to a large ensemble of atmosphere-only experiments, forced with random SST patterns. We then create an SST-based reconstruction of the dominant modes of EASJ variability in the CMIP6 models, finding a median EASJ–reconstruction correlation for the dominant mode of 0.43. Much of the skill in the reconstruction arises from variations in Pacific SSTs, however the tropical Atlantic also makes a significant contribution. These findings suggest the potential for multi-year predictions of the EASJ, provided that skilful SST forecasts are available.

Between the 21st and 27th February 2019, climatologically exceptional warm temperature anomalies of 10-15 °C were experienced throughout Northern and Western Europe. In particular, the 25th - 27th February saw record-breaking temperatures measured at many weather stations over wide areas of Iberia, France, the British Isles, the Netherlands, Germany and Southern Sweden. This heatwave was well-predicted by the European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble prediction system. Their forecasts indicated "extreme" heat was possible at a lead time of around two weeks, and likely at a lead time of around ten days. The performance of these forecasts in predicting the surface heat is also reflected in their ability to predict the synoptic situation. We exploit this successful forecast to perform an attribution analysis of the heatwave that differs from "conventional" analyses in several key regards. Firstly, we are not only confident that the model used is able to simulate the event in question; but that we are unequivocally studying the specific winter heatwave that occurred in Europe during February 2019. This analysis is carried out using a state-of-the-art coupled high-resolution forecast model ensemble, as opposed to the prescribed-SST experiments within climate model ensembles traditionally used for attribution.A crucial distinction between the typical climate model simulations used for attribution and the forecasts used here is that the climate model simulations are usually allowed to spin out for a sufficient length of time such that they have no memory of their initial conditions; an ensemble constructed in this way will therefore be representative of the climatology of the model (possibly conditioned on any prescribed-SST patterns). Unlike these climatological simulations, the successful forecasts used here are clearly heavily dependent on the initial conditions used. Within these simulations, the level of dynamical conditioning can therefore be specified by altering the lead time from initialisation to the event in question. We explore the implications of this aspect of forecast-based attribution, attempting to integrate between the "conventional" climatological and "storyline" frameworks of attribution. To simplify the interpretation of our experiments, here we have decided to only change a single feature between our factual and counterfactual experiments. The analysis presented is therefore limited to attributing the impact of diabatic heating due to increased CO2 concentrations above pre-industrial levels just over the days between the model initialisation date and the event. We carry out simulations at four different lead times from the event, allowing us to investigate the balance between the level of conditioning of the ensemble and the relaxation of the ensemble toward a new equilibrium at the lowered CO2 concentrations.

Internal climate variability will play a major role in determining change on regional scales under global warming. In the extratropics, large-scale atmospheric circulation is responsible for much of observed regional climate variability, from seasonal to multidecadal timescales. However, the extratropical circulation variability on multidecadal timescales is systematically weaker in coupled climate models. Here we show that projections of future extratropical climate from coupled model simulations significantly underestimate the projected uncertainty range originating from large-scale atmospheric circulation variability. Using observational datasets and large ensembles of coupled climate models, we produce synthetic ensemble projections constrained to have variability consistent with the large-scale atmospheric circulation in observations. Compared to the raw model projections, the synthetic observationally-constrained projections exhibit an increased uncertainty in projected 21st century temperature and precipitation changes across much of the Northern extratropics. This increased uncertainty is also associated with an increase of the projected occurrence of future extreme seasons. Future projections of northern hemisphere extratropical climate based on climate model simulations substantially underestimate the uncertainty that originates from large-scale atmospheric circulation variability, suggest synthetic ensemble projections constrained with observations.

As the leading climate mode of wintertime climate variability over Europe, the North Atlantic Oscillation (NAO) has been extensively studied over the last decades. Recently, studies highlighted the state of the Eurasian cryosphere as a possible predictor for the wintertime NAO. However, missing correlation between snow cover and wintertime NAO in climate model experiments and strong non-stationarity of this link in reanalysis data is questioning the causality of this relationship. Here we use the large ensemble of Atmospheric Seasonal Forecasts of the 20th Century (ASF-20C) with the European Centre for Medium-Range Weather Forecasts model, focusing on the winter season. Besides the main 110-year ensemble of 51 members, we investigate a second, perturbed ensemble of 21 members where initial (November) land conditions over the Northern Hemisphere are swapped from neighboring years. The Eurasian snow/NAO linkage is examined in terms of a longitudinal snow depth dipole across Eurasia. Subsampling the perturbed forecast ensemble and contrasting members with high and low initial snow dipole conditions, we found that their composite difference indicates more negative NAO states in the following winter (DJF) after positive west to east snow cover gradients at the beginning of November. Surface and atmospheric forecast anomalies through the troposphere and stratosphere associated with the anomalous positive snow dipole consist of colder early winter surface temperatures over Eastern Eurasia, an enhanced Ural ridge and increased vertical energy fluxes into the stratosphere, with a subsequent negative NAO-like signature in the troposphere. We thus confirm the existence of a causal connection between autumn snow patterns and subsequent winter circulation in the ASF-20C forecasting system.

The circumglobal teleconnection (CGT) is an important mode of circulation variability, with an influence across many parts of the northern hemisphere. Here, we examine the excitation mechanisms of the CGT in the ECMWF seasonal forecast model, and the relationship between the Indian summer monsoon (ISM), the CGT and the extratropical northern hemisphere circulation. Results from relaxation experiments, in which the model is corrected to reanalysis in specific regions, suggest that errors over northwest Europe are more important in inhibiting the model skill at representing the CGT, in addition to northern hemisphere skill more widely, than west-central Asia and the ISM region, although the link between ISM precipitation and the extratropical circulation is weak in all experiments. Thermal forcing experiments in the ECMWF model, in which a heating is applied over India, suggest that the ISM does force an extratropical Rossby wave train, with upper tropospheric anticyclonic anomalies over east Asia, the North Pacific and North America associated with increased ISM heating. However, this eastward-propagating branch of the wave train does not project into Europe, and the response there occurs largely through westward-propagating Rossby waves. Results from barotropic model experiments show a response that is highly consistent with the seasonal forecast model, with similar eastward- and westward-propagating Rossby waves. This westward-propagating response is shown to be important in the downstream reinforcement of the wave train between Asia and North America.

Observations facilitate model evaluation and provide constraints that are relevant to future predictions and projections. Constraints for uninitialized projections are generally based on model performance in simulating climatology and climate change. For initialized predictions, skill scores over the hindcast period provide insight into the relative performance of models, and the value of initialization as compared to projections. Predictions and projections combined can, in principle, provide seamless decadal to multi-decadal climate information. For that, though, the role of observations in skill estimates and constraints needs to be understood in order to use both consistently across the prediction and projection time horizons. This paper discusses the challenges in doing so, illustrated by examples of state-of-the-art methods for predicting and projecting changes in European climate. It discusses constraints across prediction and projection methods, their interpretation, and the metrics that drive them such as process accuracy, accurate trends or high signal-to-noise ratio. We also discuss the potential to combine constraints to arrive at more reliable climate prediction systems from years to decades. To illustrate constraints on projections, we discuss their use in the UK's climate prediction system UKCP18, the case of model performance weights obtained from the Climate model Weighting by Independence and Performance (ClimWIP) method, and the estimated magnitude of the forced signal in observations from detection and attribution. For initialized predictions, skill scores are used to evaluate which models perform well, what might contribute to this performance, and how skill may vary over time. Skill estimates also vary with different phases of climate variability and climatic conditions, and are influenced by the presence of external forcing. This complicates the systematic use of observational constraints. Furthermore, we illustrate that sub-selecting simulations from large ensembles based on reproduction of the observed evolution of climate variations is a good testbed for combining projections and predictions. Finally, the methods described in this paper potentially add value to projections and predictions for users, but must be used with caution. All authors were supported by the EUCP project funded by the European Commission's Horizon 2020 programme, Grant Agreement number 776613. JM was also supported by the french ANR MOPGA project ARCHANGE and by the EU-H2020 Blue Action (GA 727852) and 4C projects (GA 821003). MGD also received funding by the Spanish Ministry for the Economy, Industry and Competitiveness grant reference RYC-2017-22964. Peer Reviewed "Article signat per 14 autors/es: Gabriele C. Hegerl, Andrew P. Ballinger, Ben B. B. Booth, Leonard F. Borchert, Lukas Brunner, Markus G. Donat, Francisco J. Doblas-Reyes, Glen R. Harris, Jason Lowe, Rashed Mahmood, Juliette Mignot, James M. Murphy, Didier Swingedouw and Antje Weisheimer"

Accurate forecasting of variations in Indian monsoon precipitation and progression on seasonal time scales remains a challenge for prediction centres. We examine prediction skill for the seasonal-mean Indian summer monsoon and its onset in the European Centre for Medium-Range Weather Forecasts (ECMWF) seasonal forecasting system 5 (SEAS5). We analyse summer hindcasts initialised on 1st of May, with 51 ensemble members, for the 36-year period of 1981–2016. We evaluate the hindcasts against the Global Precipitation Climatology Project (GPCP) precipitation observations and the ECMWF reanalysis 5 (ERA5). The model has significant skill at forecasting dynamical features of the large-scale monsoon and local-scale monsoon onset tercile category one month in advance. SEAS5 shows higher skill for monsoon features calculated using large-scale indices compared to those at smaller scales. Our results also highlight possible model deficiencies in forecasting the all India monsoon rainfall.

Skillful, reliable and seamless climate information for the next 1-40 years is crucial for policy- and other decision makers to develop suitable planning strategies. This poses a challenge for the scientific community, which is split up into the prediction community (developing initialized predictions up to multi-annual time scales, e.g. 10 years), and the climate projection community (providing long-term projections). As predictions are initialized with the observed climate state at the start of the integration, they are often more skillful for lead times of a few years (depending on variable and region) compared to uninitialized climate projections, which can provide information beyond 10 years. Thus, most useful climate information for the next 1-40 years would likely need to draw upon information from both sources. However, temporal merging from different sources is challenging, e.g., it can lead to discontinuities in the central estimates at the respective transition points, which pose problems for interpretation and communication alike.The aim of this study is to explore if skillful and seamless climate information can be provided by applying a model weighting scheme to initialized decadal predictions and projections. The model specific weights are based on the respective past model performance compared to observations. Whereas for climate projections each model is assigned a single weight, for initialized decadal predictions these weights are calculated for each forecast year separately. Here, we apply the weighting technique to CMIP6 decadal predictions and climate projections from 8 different models.

Atmospheric seasonal predictability in winter over the Euro-Atlantic region is studied with an emphasis on the signal-to-noise paradox of the North Atlantic Oscillation. Seasonal hindcasts of the ECMWF model for the recent period 1981-2009 show, in agreement with other studies, that correlation skill over Greenland and parts of the Arctic is higher than the signal-to-noise ratio implies. This leads to the paradoxical situation where the real world appears more predictable than the models suggest, with the forecast ensembles being overly dispersive (or underconfident). However, it is demonstrated that these conclusions are not supported by the diagnosed relationship between ensemble mean RMSE and ensemble spread which indicates a slight underdispersion (overconfidence). Furthermore, long atmospheric seasonal hindcasts suggest that over the 110-year period from 1900 to 2009 the ensemble system is well calibrated (neither over- nor underdispersive). The observed skill changed drastically in the middle of the 20th Century and paradoxical regions during more recent hindcast periods were strongly underdispersive during mid-Century decades. Due to non-stationarities of the climate system in the form of decadal variability, relatively short hindcasts are not sufficiently representative for longer-term behaviour. In addition, small hindcast sample size can lead to skill estimates, in particular of correlation measures, that are not robust. It is shown that the relative uncertainty due to small hindcast sample size is often larger for correlation-based than for RMSE-based diagnostics. Correlation-based measures like the RPC are shown to be highly sensitive to the strength of the predictable signal, implying that disentangling of physical deficiencies in the models on the one hand, and the effects of sampling uncertainty on the other hand, is difficult. Given the current lack of a causal physical mechanism to unravel the puzzle, our hypotheses of non-stationarity and sampling uncertainty provide simple yet plausible explanations for the paradox.

As the leading climate mode to explain wintertime climate variability over Europe, the North Atlantic Oscillation (NAO) has been extensively studied over the last decades. Recently, studies highlighted the state of the Northern Hemispheric cryosphere as possible predictor for the wintertime NAO (Cohen et al. 2014). Although several studies could find seasonal prediction skill in reanalysis data (Orsolini et al. 2016, Duville et al. 2017,Han & Sun 2018), experiments with ocean-atmosphere general circulation models (AOGCMs) still show conflicting results (Furtado et al. 2015, Handorf et al. 2015, Francis 2017, Gastineau et al. 2017). Here we use two kinds ECMWF seasonal prediction ensembles starting with November initial conditions taken from the long-term reanalysis ERA-20C and forecasting the following three winter months. Besides the 110-year ensemble of 50 members representing internal variability of the atmosphere, we investigate a second ensemble of 20 members where initial conditions are split between low and high snow cover years for the Northern Hemisphere. We compare two recently used Eurasian snow cover indices for their skill in predicting winter climate for the European continent. Analyzing the two forecast experiments, we found that prediction runs starting with high snow index values in November result in significantly more negative NAO states in the following winter (DJF), which in turn modulates near surface temperatures. We track the atmospheric anomalies triggered by the high snow index through the tropo- and stratosphere as well as for the individual winter months to provide a physical explanation for the formation of this particular climate state.

Atmospheric circulation regimes can be used to study links between regional weather and other climate processes, like sudden stratospheric warmings. For these studies it is important to know whether there is any background non-stationarity in the regimes themselves. To identify regime non-stationarity model ensemble data is needed to have sufficient data. However, models are noisy in their representation of circulation regimes making obtaining the signal difficult. We propose a new method, in the form of a constraint on the ensemble-member similarity in the clustering method, to identify the signal of the non-stationary regime dynamics.We use ECMWF SEAS5 hindcast data to identify six wintertime circulation regimes over the Euro-Atlantic sector (NAO+/-, Atlantic Ridge (AR) +/- and Scandinavian Blocking (SB) +/-), which has been found to be the optimal number of regimes in a previous study. Implementing the constraint leads to more robust regimes and the identification of a stronger inter-annual signal in the regime occurrence rates than without the constraint. The clearest signal on inter-annual timescales is found during strong El Niño years. During those years the NAO+ becomes less frequent, while the SB- occurs more often. The signal in the occurrence rate of the NAO- is weaker than for the NAO+. Without the implementation of the constraint this difference in the strength of the signal between the two phases of the NAO cannot be detected. Thus, the constraint on the ensemble-member similarity allows for identifying a non-stationary signal that otherwise is more difficult to obtain.

Variability of the East Asian summer jet stream (EAJ) has a significant impact on the climate of East Asia, primarily through its modulation of East Asian precipitation. In recent decades the impact of sea surface temperatures (SSTs) in the tropical Indian and Pacific oceans on the EAJ have been studied in considerable detail, however much less is known about the drivers of EAJ variability on decadal or multi-decadal timescales. Investigating this problem is made more challenging by the temporal limitations of reanalysis datasets.In order to establish whether SSTs can provide a source of skill in predicting decadal variations of the EAJ, we analyse long pre-industrial control runs of the CMIP6 models. One issue with studying coupled model runs is that it is often unclear whether particular SST anomalies are forcing the atmosphere, and thus can provide a meaningful source of skill, or whether they are merely responding to local atmospheric anomalies. We address this issue by combining SST and turbulent heat flux information to indicate the direction of the forcing.

Weather forecasts rely heavily on general circulation models of the atmosphere and other components of the Earth system. National meteorological and hydrological services and intergovernmental organizations, such as the European Centre for Medium-Range Weather Forecasts (ECMWF), provide routine operational forecasts on a range of spatio-temporal scales by running these models at high resolution on state-of-the-art high-performance computing systems. Such operational forecasts are very demanding in terms of computing resources. To facilitate the use of a weather forecast model for research and training purposes outside the operational environment, ECMWF provides a portable version of its numerical weather forecast model, OpenIFS, for use by universities and other research institutes on their own computing systems. In this paper, we describe a new project (OpenIFS@home) that combines OpenIFS with a citizen science approach to involve the general public in helping conduct scientific experiments. Volunteers from across the world can run OpenIFS@home on their computers at home, and the results of these simulations can be combined into large forecast ensembles. The infrastructure of such distributed computing experiments is based on our experience and expertise with the climateprediction.net (https://www.climateprediction.net/, last access: 1 June 2021) and weather@home systems. In order to validate this first use of OpenIFS in a volunteer computing framework, we present results from ensembles of forecast simulations of Tropical Cyclone Karl from September 2016 studied during the NAWDEX field campaign. This cyclone underwent extratropical transition and intensified in mid-latitudes to give rise to an intense jet streak near Scotland and heavy rainfall over Norway. For the validation we use a 2000-member ensemble of OpenIFS run on the OpenIFS@home volunteer framework and a smaller ensemble of the size of operational forecasts using ECMWF's forecast model in 2016 run on the ECMWF supercomputer with the same horizontal resolution as OpenIFS@home. We present ensemble statistics that illustrate the reliability and accuracy of the OpenIFS@home forecasts and discuss the use of large ensembles in the context of forecasting extreme events.

The role of model resolution in simulating geophysical vortices with the characteristics of realistic tropical cyclones (TCs) is well established. The push for increasing resolution continues, with general circulation models (GCMs) starting to use sub-10-km grid spacing. In the same context it has been suggested that the use of stochastic physics (SP) may act as a surrogate for high resolution, providing some of the benefits at a fraction of the cost. Either technique can reduce model uncertainty, and enhance reliability, by providing a more dynamic environment for initial synoptic disturbances to be spawned and to grow into TCs. We present results from a systematic comparison of the role of model resolution and SP in the simulation of TCs, using EC-Earth simulations from project Climate-SPHINX, in large ensemble mode, spanning five different resolutions. All tropical cyclonic systems, including TCs, were tracked explicitly. As in previous studies, the number of simulated TCs increases with the use of higher resolution, but SP further enhances TC frequencies by ~30%, in a strikingly similar way. The use of SP is beneficial for removing systematic climate biases, albeit not consistently so for interannual variability; conversely, the use of SP improves the simulation of the seasonal cycle of TC frequency. An investigation of the mechanisms behind this response indicates that SP generates both higher TC (and TC seed) genesis rates, and more suitable environmental conditions, enabling a more efficient transition of TC seeds into TCs. These results were confirmed by the use of equivalent simulations with the HadGEM3-GC31 GCM.

Weather forecasts rely heavily on general circulation models of the atmosphere and other components of the Earth system. National meteorological and hydrological services and intergovernmental organisations, such as the European Centre for Medium-Range Weather Forecasts (ECMWF), provide routine operational forecasts on a range of spatio-temporal scales, by running these models in high resolution on state-of-the-art high-performance computing systems. Such operational forecasts are very demanding in terms of computing resources. To facilitate the use of a weather forecast model for research and training purposes outside the operational environment, ECMWF provides a portable version of its numerical weather forecast model, OpenIFS, for use by universities and other research institutes on their own computing systems. In this paper, we describe a new project (OpenIFS@home) that combines OpenIFS with a citizen science approach to involve the general public in helping conduct scientific experiments. Volunteers from across the world can run OpenIFS@home on their computers at home and the results of these simulations can be combined into large forecast ensembles. The infrastructure of such distributed computing experiments is based on our experience and expertise with the climateprediction.net and weather@home systems. In order to validate this first use of OpenIFS in a volunteer computing framework, we present results from ensembles of forecast simulations of tropical cyclone Karl from September 2016, studied during the NAWDEX field campaign. This cyclone underwent extratropical transition and intensified in mid-latitudes to give rise to an intense jet-streak near Scotland and heavy rainfall over Norway. For the validation we use a two thousand member ensemble of OpenIFS run on the OpenIFS@home volunteer framework and a smaller ensemble of the size of operational forecasts using ECMWF’s forecast model in 2016 run on the ECMWF supercomputer with the same horizontal resolution as OpenIFS@home. We present ensemble statistics that illustrate the reliability and accuracy of the OpenIFS@home forecasts as well as discussing the use of large ensembles in the context of forecasting extreme events.

Recent studies have found evidence of multidecadal variability in northern hemisphere wintertime seasonal forecast skill. Here we assess the robustness of this finding by extending the analysis to analysing a diverse set of ensemble atmospheric model simulations. These simulations differ in either numerical model or type of initialisation and include atmospheric model experiments initialised with reanalysis data and free‐running atmospheric model ensembles. All ensembles are forced with observed SST and seaice boundary conditions. Analysis of large‐scale Northern Hemisphere circulation indicesover the Northern Hemisphere (namely the North Atlantic Oscillation, Pacific North American pattern and the Arctic Oscillation) reveals that in all ensembles there is larger correlation skill in the late century periods than during periods in the mid‐century. Similar multidecadal variability in skill is found in a measure of total skill integrated over the whole of the extratropics. Most of the differences in large‐scale circulation skill between the skillful late period (as well as early period) and the less skillful mid‐century period seem to be due to a reduction in skill over the North Pacific and a disappearance in skill over North America and the North Atlantic. The results are robust across different models and different types of initialisation, indicating that the multidecadal variability in Northern Hemisphere winter skill is a robust feature of 20th century climate variability. Multidecadal variability in skill therefore arises from the evolution of the observed SSTs, likely related to a weakened influence of ENSO on the predictable extratropical circulation signal during the middle of the 20th century, and is evident in the signal‐to‐noise ratio of the different ensembles, particularly the larger ensembles.

We quantify the value of sub-seasonal forecasts for a real-world prediction problem: the forecasting of French month-ahead energy demand. Using surface temperature as a predictor, we construct a trading strategy and assess the financial value of using meteorological forecasts, based on actual energy demand and price data. We show that forecasts with lead times greater than 2 weeks can have value for this application, both on their own and in conjunction with shorter range forecasts, especially during boreal winter. We consider a cost/loss framework based on this example, and show that while it captures the performance of the short range forecasts well, it misses the marginal value present in the longer range forecasts. We also contrast our assessment of forecast value to that given by traditional skill scores, which we show could be misleading if used in isolation. We emphasise the importance of basing assessment of forecast skill on variables actually used by end-users., Comment: 22 pages, 8 figures, revised submission to QJRMS

Motivated by the high skill in predicting ENSO on seasonal time scales with ECMWF’s seasonal forecasting system SEAS5 and by previous findings of multi-decadal variability in seasonal forecast skill of extratropical dynamics, we have carried out an extensive set of 24-month long coupled hindcasts from 1901 to 2010. The hindcasts were run with SEAS5 in reduced resolution and are initialised from, and verified against, reanalyses of the 20thCentury. They allow us to analyse ENSO forecast skill beyond the first year, to study how skill varies on decadal time scales and to test sensitivities to atmospheric wind forcings and the assimilation of ocean observations in the initial conditions.First results show a substantial amount of multi-decadal variability in both ENSO mean state and forecast skill. We find periods in the early-to-mid 20thCentury with much reduced levels of skill, in particular after the spring barrier in the first forecast year. Periods at the beginning and at the end of the Century show broadly similar good performances with substantial skill even after the first year spring barrier. Combined effects of the wind forcing and the assimilation of ocean data on the initial state seem to play a crucial role in understanding this behaviour.

There is an increasing demand for robust, reliable and actionable climate information for the near-term future (1-40 years). Seamless information on these time-scales can only be derived from uninitialized climate projections, which however are not aligned with the observed, internal state of the climate system. Another source of information are Initialized predictions for which the observed state is taken into account, but these are only available for the upcoming decade. In this study, we test in how far decadal predictions can be used to constrain uninitialized projections to obtain skillful predictions on time-scales beyond decades. This is done by selecting a sub-ensemble of uninitialized projections, which are chosen by their proximity to the decadal predictions ensemble mean. This framework is applied to surface air temperatures from CMIP5 simulations over the North Atlantic Gyre region, as decadal predictions show largest added value over projections for this region. Skill is measured using anomalous correlation coefficient (ACC) and root-mean-square-error (RMSE). Results show that ACC values for forecast years 10-15 for the constrained sub-ensemble are similar to those derived for the non-constrained uninitialized ensemble. However, RMSEs are significantly decreased for the constrained sub-ensemble, not only for the first 10 forecast years but also beyond. This is mainly due to the fact that the constrained sub-ensemble has a much higher ability to capture the observed warming trend during the end of the 20th century compared to the uninitialized ensemble mean. Further to these results, the limitations of this framework are discussed, including an assessment of the potential upper limit of added value and it’s dependency on the skill of the decadal forecast system. This easy-to-apply framework can be used to provide crucial climate information for mitigation and adaptation strategies by filling the gap between initialized decadal predictions and uninitialized projections.

There is increasing demand for robust, reliable and actionable climate information for the next 1 to 50 years. This is challenging for the scientific community as the longest initialized predictions are limited to 10 years (decadal predictions). Thus, to provide seamless information for the upcoming 50 years, information from decadal predictions and uninitialized projections need to be merged. In this study, the ability to obtain valuable climate information beyond decadal time-scales by constraining uninitialized projections using decadal predictions is assessed. The application of this framework to surface temperatures over the North Atlantic Subpolar Gyre region, shows that the constrained uninitialized sub-ensemble has higher skill compared to the overall projection ensemble also beyond ten years when information from decadal predictions is no longer available. Though showing the potential of such a constraining approach to obtain climate information for the near-term future, its utility depends on the added value of initialization.

New seasonal retrospective forecasts for 1901-2010 show that skill for predicting ENSO, NAO and PNA is reduced during mid-century periods compared to earlier and more recent high-skill decades. Forecasts of seasonal climate anomalies using physically-based global circulation models are routinely made at operational meteorological centers around the world. A crucial component of any seasonal forecast system is the set of retrospective forecasts, or hindcasts, from past years which are used to estimate skill and to calibrate the forecasts. Hindcasts are usually produced over a period of around 20-30 years. However, recent studies have demonstrated that seasonal forecast skill can undergo pronounced multi-decadal variations. These results imply that relatively short hindcasts are not adequate for reliably testing seasonal forecasts and that small hindcast sample sizes can potentially lead to skill estimates that are not robust. Here we present new and unprecedented 110-year-long coupled hindcasts of the next season over the period 1901 to 2010. Their performance for the recent period is in good agreement with those of operational forecast models. While skill for ENSO is very high during recent decades, it is markedly reduced during the 1930s to 1950s. Skill at the beginning of the 20th Century is, however, as high as for recent high-skill periods. Consistent with findings in atmosphere-only hindcasts, a mid-century drop in forecast skill is found for a range of atmospheric fields including large-scale indices such as the NAO and the PNA patterns. As with ENSO, skill scores for these indices recover in the early 20th Century suggesting that the mid-century drop in skill is not due to lack of good observational data. A public dissemination platform for our hindcast data is available and we invite the scientific community to explore them.

The North Atlantic Oscillation (NAO) is the leading mode of variability in the large scale circulation over the North Atlantic in winter, and strongly influences the weather and climate of Europe. On synoptic timescales, the negative phase of the NAO often corresponds to the occurrence of a blocking episode over Greenland. Hence, the dynamics and predictability of these blocking events is of interest for the prediction of the NAO and its related impacts over a wide region. Ensemble sensitivity analysis utilises the information contained in probabilistic forecast ensembles to calculate a statistical relationship between a forecast metric and some precursor condition. Here the method is applied to 15‐day forecasts of a set of 26 Greenland blocking events using the state‐of‐the‐art European Centre for Medium‐Range Weather Forecasts (ECMWF) forecasting system. The ensemble sensitivity analysis shows that Greenland blocking does not develop in isolation in these forecasts, but instead the blocking is sensitive to remote precursors, such as 500 hPa and 50 hPa geopotential height, particularly in the low‐frequency flow. In general, there are more significant sensitivities to anomalies in the tropics than in the polar regions. Stratospheric sensitivities tend to emerge at later lead times than tropospheric sensitivities. The strongest and most robust sensitivities correspond to a Rossby wave precursor reaching from the Pacific basin across North America.

Statistical downscaling methods are popular post-processing tools which are widely used in many sectors to adapt the coarse-resolution biased outputs from global climate simulations to the regional-to-local scale typically required by users. They range from simple and pragmatic Bias Correction (BC) methods, which directly adjust the model outputs of interest (e.g. precipitation) according to the available local observations, to more complex Perfect Prognosis (PP) ones, which indirectly derive local predictions (e.g. precipitation) from appropriate upper-air large-scale model variables (predictors). Statistical downscaling methods have been extensively used and critically assessed in climate change applications; however, their advantages and limitations in seasonal forecasting are not well understood yet. In particular, a key problem in this context is whether they serve to improve the forecast quality/skill of raw model outputs beyond the adjustment of their systematic biases. In this paper we analyze this issue by applying two state-of-the-art BC and two PP methods to downscale precipitation from a multimodel seasonal hindcast in a challenging tropical region, the Philippines. To properly assess the potential added value beyond the reduction of model biases, we consider two validation scores which are not sensitive to changes in the mean (correlation and reliability categories). Our results show that, whereas BC methods maintain or worsen the skill of the raw model forecasts, PP methods can yield significant skill improvement (worsening) in cases for which the large-scale predictor variables considered are better (worse) predicted by the model than precipitation. For instance, PP methods are found to increase (decrease) model reliability in nearly 40% of the stations considered in boreal summer (autumn). Therefore, the choice of a convenient downscaling approach (either BC or PP) depends on the region and the season., This study was partially supported by the SPECS and EUPORIAS projects, funded by the European Commission through the Seventh Framework Programme for Research under grant agreements 308378 and 308291, respectively. JMG acknowledges partial support from the project MULTI-SDM (CGL2015-66583-R, MINECO/FEDER).