Our training resources

In this lecture we will give you a brief history of ECMWF and present the main areas of NWP research that is currently being carried out in the centre. We then look at current research challenges and present some of the latest developments that will soon become operational.

By the end of the lecture you should be able to:

• List the main research areas at ECMWF and describe the latest model developments.

The aim of this lecture is to introduce the concept of the EnKF in the context of atmospheric data assimilation. Strengths and weaknesses of the algorithm will be discussed and results of the ECMWF implementation will be presented.

By the end of the lecture the participants should be able to:

•    Describe the basic EnKF algorithm and its connections with    the Kalman Filter;

•    Discuss some of the advantages and the limitations of EnKF algorithms with respect to more established variational algorithms;

•    Be aware of recent developments in hybrid variational-EnKF data assimilation

The Ensemble of Data Assimilations (EDA) technique is used for the estimation of the analysis and background errors of the ECMWF assimilation system. This lecture describes the EDA formulation and implementation, and how it interacts with ECMWF 4DVar analysis and ECMWF Ensemble Prediction System.

By the end of the lecture the participants should be able to:

• Describe the theoretical basis and practical implementation of the EDA
• Explain the use of the EDA in the ECMWF analysis and ensemble prediction systems

This lecture provides an overview of a typical ocean data assimilation system for initialization and re-analyses application. The lecture uses as an example the ECMWF ocean data assimilation system, which is based the NEMOVAR (3Dvar FGAT). This will be used to discuss design of the assimilation cycle, formulation of error covariances, observations assimilated and evaluation procedure, among others.

By the end of the lecture students should be able to:

• describe the different components involved in a an ocean data assimilation system
• list the commonalities and and differences between ocean and atmosphere data assimilation
• describe the basics of the physical ocean observing system
• explain the essential multivariate relationships between ocean variables
• identify the limitations of the existing systems.

The primary purpose of this lecture is explore the implications of the fact that satellites can only measure radiation at the top of the atmosphere and do not measure the geophysical variables we require for NWP (e.g. temperature, humidity and wind). The link between the atmospheric variables and the measured radiances is the radiative transfer equation - the key elements of which are discussed. It is shown how - with careful frequency selection - satellite measurements can be made for which the relationship to geophysical variables is greatly simplified. Despite these simplifications, it is shown that the extraction of detailed profile information from downward looking radiance measurements is a formally ill posed inverse problem.

Data assimilation is introduced as the solution to this inverse problem, where background information and satellite observations are combined to produce a best or optimal estimate of the atmospheric state. The main elements of the assimilation scheme (such as the chain of observation operators for radiances) and its key statistical inputs are examined. In particular it is shown that incorrect specification of observation errors (R) and background errors (B) can severely limit the successful exploitation of satellite data.

By the end of this lecture you will:

• understand exactly what a satellite actually measures (radiance)
• appreciate the complex relationship between what is measured and what we wish to know for NWP
• how information is extracted from satellite measurements in data assimilation

In this lecture, the impact of model error on variational data assimilation will be presented. This lecture will introduce weak-constraint 4D-Var as a way to account for model error in the data assimilation process. Several examples of results from simplified implementations in the IFS will be shown.

By the end of the lecture the participants should be able to:

• describe the impact of model error on the data assimilation process,
• explain the difficulties in properly accounting for model error in data assimilation.

The aim of this session is to present an overview of the current observing systems used in Numerical Weather Prediction. We will discuss our observational requirement, and how close the current observing system is to meeting our needs. We will also discuss areas where our requirements are evolving. We will learn about WMO's OSCAR database that describes the Global Observing System. We will learn how the large diversity of observations now available, are monitored to ensure only good observations are presented to an operational system.

By the end of the session you should be able to:

• be able to describe the main types of observations used in data assimilation for Numerical Weather Prediction;
• be aware of how large volumes of observations are exchanged, implemented and monitored in operational systems;
• be aware of WMO's OSCAR database, how to access it and what type of information it can provide.

The aim of these sessions is to understand the role of land surface data assimilation on medium range weather forecasts.

We will give an overview of the different approaches used to assimilate land surface data and to initialise model variables in NWP.  We will  present the current observing systems and describe the land data assimilation structure within ECMWF system.

By the end of the session you should be able to:

• identify the different observations used for snow and soil moisture data assimilation
• define land surface data assimilation approaches used for NWP
• describe the role of land surface data assimilation on medium-range weather forecasts

The goal of this lecture is to familiarise the student with the notion of tangent linear and adjoint models, and their use in variational data assimilation.  A general overview of the current use of tangent linear and adjoint models in the ECMWF system will also be provided. Theoretical definitions and practical examples of tangent liner and adjoint models will be given. The student will be invited to work some simple tangent linear and adjoint derivations together with the instructor. A brief introduction to automatic differentiation software will also be given./

By the end of the session you should be able to:

• define what tangent linear and adjoint models are
• derive tangent linear and adjoint equations for a simple nonlinear equation
• describe the use of tangent linear and adjoint codes within the ECMWF's 4D-VAR system.

At ECMWF atmospheric composition data are assimilated into the IFS as part of the MACC-II project. On a global scale, atmospheric composition represents the full state of the global atmosphere covering phenomena such as desert dust plumes, long-range transport of atmospheric pollutants or ash plumes from volcanic eruptions, but also variations and long-term changes in the background concentrations of greenhouse gases.

The aim of this lecture is to give an overview of the work that is carried out at ECMWF regarding the assimilation of atmospheric composition data, and to address why this is of interest and which special challenges are faced when assimilating atmospheric composition data.

By the end of the session you should:

• have some understanding of the work carried out at ECMWF to assimilate data of atmospheric composition

The aim of these sessions is to understand the role of land surface data assimilation on medium range weather forecasts.

We will give an overview of the different approaches used to assimilate land surface data and to initialise model variables in NWP.  We will  present the current observing systems and describe the land data assimilation structure within ECMWF system.

By the end of the session you should be able to:

• identify the different observations used for snow and soil moisture data assimilation
• define land surface data assimilation approaches used for NWP
• describe the role of land surface data assimilation on medium-range weather forecasts

The aim of this session is to understand how data assimilation can improve our knowledge of past weather over long time-scales. We will present recent advances that help capture changes over time in observing system networks, and project this variation in information content into uncertainty estimates of the reanalysis products. We will also discuss the applications of reanalysis, which generally put weather events into the climate context.

By the end of the session you should be able to:

• explain what are the goals of data assimilation in a reanalysis data assimilation system
• list the key aspects that require particular attention in reanalysis, as compared to numerical weather prediction
• describe the most common problems in reanalysis products

In this lecture the variational bias correction scheme (VarBC) as used at ECMWF is explained. VarBC replaced the tedious job of estimating observation bias off-line for each satellite instrument or in-situ network by an automatic self-adaptive system. This is achieved by making the bias estimation an integral part of the ECMWF variational data assimilation system, where now both the initial model state and observation bias estimates are updated simultaneously.

By the end of the session you should be able to realize that:

• many observations are biased, and that the characteristics of bias varies widely between types of instruments
• separation between model bias and observation bias is often difficult
• the success of an adaptive system implicitly relies on a redundancy in the underlying observing system.

This one-hour lecture will identify the challenges associated with the use of physical parametrizations in the context of four-dimensional variational data assimilation (4D-Var). The importance of the linearity constraint in 4D-Var and the methods to address it will be detailed. The set of linearized physical parametrizations used at ECMWF will be briefly presented. Examples of the use of physical parametrizations in variational data assimilation and its impact on forecast quality will be given.

By the end of the lecture, the students should be able:

• to tell why physical parametrizations are needed in data assimilation.
• to recognize the importance of the regularization of the linearized code

Abstract: The lectures introduce methods of ensemble verification. They cover the verification of discrete forecasts (e.g. dry/wet) and continuous scalar forecasts (e.g. temperature). Various scores such as the Brier score and the continuous ranked probability score are introduced.

After the lectures you should be able to

• explain what a reliable probabilistic forecast is and how to measure reliability

• understand why resolution and sharpness of a probabilistic forecast matter

• compute several of the standard verification metrics used for ensemble forecasts

In this session the generation of the perturbed initial condition of the ECMWF ensemble will be presented. We will discuss the ratio behind using singular vectors in the ensemble and how they are calculated. Then it will be explained how the singular vectors are combined with perturbations from the ensemble of data assimilations to construct the perturbations for the ensemble.

By the end of the session you should be able to:

• explain the idea behind using singular vectors in the ensemble

• describe how singular vectors are calculated

• describe the construction of the ensemble perturbations

Abstract: The lectures introduce methods of ensemble verification. They cover the verification of discrete forecasts (e.g. dry/wet) and continuous scalar forecasts (e.g. temperature). Various scores such as the Brier score and the continuous ranked probability score are introduced.

After the lectures you should be able to

• explain what a reliable probabilistic forecast is and how to measure reliability

• understand why resolution and sharpness of a probabilistic forecast matter

• compute several of the standard verification metrics used for ensemble forecasts

The aim of this session is to introduce the main sources of uncertainty that lead to forecast errors. The weather prediction problem will be discussed, and stated it in terms of an appropriate probability density function (PDF). The concept of ensemble prediction based on a finite number of integration will be introduced, and the reason why it is to be the only feasible method to predict the PDF beyond the range of linear growth will be illustrated.

By the end of the session you should be able to:

• explain which are the main sources of forecast error

• illustrate why numerical prediction should be stated in probabilistic terms

• describe the rationale behind ensemble prediction

In this lecture we will give you a brief history of ECMWF and present the main areas of NWP research that is currently being carried out in the centre. We then look at current research challenges and present some of the latest developments that will soon become operational.

By the end of the lecture you should be able to:

• List the main research areas at ECMWF and describe the latest model developments.

The aim of this session is to introduce the ECMWF ensemble of data assimilation (EDA). The rationale and methodology of the EDA will be illustrated, and its use in to simulate initial uncertainties in the ECMWF ensemble prediction system (ENS) will be presented.

By the end of the session you should be able to:

• know what is the ECMWF EDA

• illustrate how the EDA is used to simulate initial uncertainty in ensemble prediction

• understand the main differences between singular vectors and EDA-based perturbations

This lecture covers the essentials of building a numerical seasonal forecast system, as exemplified by the present prediction system at ECMWF.

  By the end of this lecture, you should be able to:

• explain the scientific basis of seasonal forecast systems
• describe in outline ECMWF System 4 and its forecast performance
• discuss the critical factors in further improving forecast systems

Sources_predictability2.pdfAfter this lecture, students will be able to:

• explain the physical and practical motivations for using stochastic physics in an ensemble forecast;

• describe the two stochastic parameterization schemes used in the IFS ensemble, and their respective purposes;

• be able to identify the improvement in forecasting skill from the inclusion of stochastic physics.

You get the opportunity to experiment yourself with an ensemble prediction system for a chaotic low-dimensional dynamical system introduced by Edward Lorenz in 1995. Experiments permit to study the role of the initial condition perturbations and the representation of model uncertainties. Various metrics introduced in the ensemble verification lectures will be applied in this session.

After the practice session, you will be able to use the toy model as an educational tool.

The latest medium, monthly and seasonal forecasts will be discussed in terms of out look and performance.

This is a combined event with the weekly weather discussion that ECMWF staff attend.

This lecture will give an overview of the general circulation of the atmosphere as a refresher. Starting from a radiative energy balance perspective the dominating atmospheric circulation systems will be described. The coupling between Hadley cell dynamics and surface friction will be covered as well as midlatitude jets and baroclinic instability. Orographic steering of large scale Rossby waves and flow regime behaviour will also be discussed.

By the end of the lecture you should be able to:

• Describe the overall energy balance of the atmosphere.
• Explain the dynamics of the Hadley cell circulation.
• Describe factors that determine midlatitude flow variability.

The aim of this session is to illustrate the key characteristic of the nine operational global, medium-range ensemble systems. These are the ensembles available also within the TIGGE (Thorpex Interactive Grand Global Ensemble) project data-base. Similarity and differences in the approaches followed to simulate the sources of forecast uncertainties will be discussed, and their relevance for forecast performance will be illustrated.

By the end of the session you should be able to:

• illustrate the main similarities and differences of the 9 TIGGE global ensembles

• link the performance differences of TIGGE ensemble to their design

• describe the main differences between singular vectors and EDA-based perturbations

You get the opportunity to experiment yourself with an ensemble prediction system for a chaotic low-dimensional dynamical system introduced by Edward Lorenz in 1995. Experiments permit to study the role of the initial condition perturbations and the representation of model uncertainties. Various metrics introduced in the ensemble verification lectures will be applied in this session.

After the practice session, you will be able to use the toy model as an educational tool.

Abstract: The lecture is a short introduction to operational hydrological ensemble prediction systems, with focus on flooding. The European Flood Awareness System (EFAS) is described. The lecture also contains a short interactive exercise in decision making under uncertainty using prbabilistic forecasts as an example.

By the end of the session you should be able to:

• Describe the components in hydrological ensemble prediction systems (HEPS).

• Describe the major sources of uncertainty in HEPS and how they can be reduced.

• Explain the difficulties in using probabilistic flood forecasts in decision making.

The aim of this session is to understand the ECMWF clustering products.

By the end of the session you should be able to:

• explain how the cluster analysis works
• use the ECMWF clustering products

This lecture gives an overview of ensemble and post-processing and calibration techniques. The presentation is made from the medium-range forecast perspective. The (relative) benefits of calibration and multi-model combination for medium-range forecasting are also discussed.

  By the end of this lecture, you should be able to:

• describe a wide range of possible calibration methods for ensemble systems
• explain which methods are suitable in which circumstances
• discuss the merits of calibration and multi-model combination

The aim of this session is to understand the role of sea ice for medium to extended range forecasts. We will discuss the mechanisms that give rise to sea ice predictability on seasonal to interannual timescales. We will present an overview of the current observing systems and the representation of sea ice within ECMWF system.

By the end of the session you should be able to:

• explain the impact of sea ice cover on atmospheric and oceanic circulation
• describe the sources of predictability for Arctic sea ice
• describe the representation of sea ice within ECMWF forecast systems.

The aim of this session is to provide a general overview of the application of probabilistic medium-to-long-range forecasts in meteorological drought applications. We will review the main drought concepts and how probabilistic forecast can be adapted to typical drought indicators.

By the end of the session you should be able to:

• explain the benefit of using drought indicators

• describe the forecast products that can be used for drought monitoring and forecasting

• understand the different sources of drought indicators predictability

This lecture looks at calibration and multi-model ensembles from a seasonal forecasting perspective. The theoretical basis is given, followed by research results that strongly motivated a multi-model approach for these timescales. The operational EUROSIP multi-model system is described.

  By the end of this lecture, you should be able to:

• explain the difference between a model distribution and a pdf
• discuss why a multi-model approach is powerful for seasonal forecasts in particular
• outline the methods used by the EUROSIP multi-model system