Scientist - Seasonal Forecasting

at  ECMWF

THE TEAM

The Long-range Forecasting Team, part of the Earth System Predictability Section, is responsible for the design and scientific underpinnings of the ECMWF seasonal prediction configuration, which is expected to cover the forecast ranges from one month to two years ahead. The team conducts predictability research to improve our understanding and representation of sources of seasonal predictability. The team also provides scientific support to the C3S multi-system seasonal forecast component.
The Earth System Predictability Section is part of ECMWF’s Research Department. The Section explores relevant directions to improve the skill of the ECMWF forecasting system across timescales. This involves both exploring the Earth-system predictability horizon and identifying elements limiting present-day forecast skill.

EDUCATION

• A very good university degree and doctorate degree in climate science, mathematics, physics or a related field

KNOWLEDGE, SKILLS AND EXPERIENCE

• Sound knowledge of meteorology, climate dynamics, climate variability and change
• Knowledge of concepts of predictability of weather and climate
• Experience with statistical techniques, including evaluation of ensemble simulations
• Ability to conduct numerical experimentation with GCMs in HPC environments
• Strong programming and scripting skills (python, bash, ideally also fortran)
• Experience working with large datasets, familiarity with ECMWF data formats
  You must be able to work effectively in English and interviews will be conducted in English

THE ROLE

We are looking for a motivated scientist to help us drive forward improvements in multi-system seasonal forecasts for climate services.
You will join a team at ECMWF and collaborate with colleagues from Member States, other partners and the wider scientific community to help us make further progress in the accuracy and reliability of seasonal forecast systems. Design and development of ensemble seasonal forecasting systems is complex and involves the scientific and computational balance of multiple components, including initialization, suitability of the forecast model, boundary conditions, ensemble generation and verification techniques. The very limited number of past cases available for model assessment, and the non-stationarity of the climate are serious challenges when it comes to evaluation and system design. The work is becoming ever more important as the impacts of climate change grow and past experience becomes an increasingly poor guide as to what weather to expect for the coming season.
Your work will focus on research to guide the design and evolution of future seasonal forecasting systems. This includes exploring hypotheses using numerical experimentation, diagnosing the performance of individual forecasting systems and developing additional metrics for system performance. Further research will include methods to extract predictable signals from the multi-system ensemble, and study of the impact of climate non-stationarity on climate predictability and methods to account for this.
This position is funded as part of the Copernicus Climate Change Service (C3S), and your work will support both the ECMWF contribution to the multi-system seasonal forecast ensemble and the wider development of C3S seasonal forecast activities.

MAIN DUTIES AND RESPONSIBILITIES

• To share in the scientific and technical activities of the long-range team, and to explore possible innovations to enhance the seasonal forecasting system
• To conduct numerical experimentation to advance the understanding of predictability at the seasonal time scale, and to test hypotheses related to improving forecast systems
• To develop and implement appropriate additional diagnostics to enhance performance metrics for model and forecast assessment
• To participate in research on the representation of climate non-stationarity in seasonal forecast systems, and methods to account for this in the creation of products
• To contribute to research on the predictable signals contained within model forecasts, and methods to extract these from the C3S multi-system ensemble
• To support users in the appropriate use and interpretation of seasonal model outputs and products