Image: iLexx/Stock/Getty Images Plus
ECMWF has appointed three scientists as new Fellows to collaborate on research projects undertaken at the Centre. They add to five scientists whose Fellowships were renewed or started in 2023.
The new Fellows are Michaela Hegglin, a Director at the Institute of Climate and Energy Systems of the Forschungszentrum Jülich (Germany), Professor of Atmospheric Chemistry at the University of Reading (UK), and Professor of Atmospheric Physics at the University of Wuppertal (Germany); Ilias Pechlivanidis, an expert in hydrological forecasting and climate services at the Swedish Meteorological and Hydrological Institute (SMHI) and a Visiting Scientist at Uppsala University; and Anthony Weaver, a senior researcher at the French Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique (Cerfacs).
Our Fellowship programme fosters links between ECMWF and individual scientists carrying out pioneering work in areas relevant to our strategic goals. Fellows have access to ECMWF computing and database facilities. Areas of joint work have included data assimilation, novel numerical methods, the atmospheric water cycle, predictability, and the calibration and post-processing of forecasts.
The outcomes of this work have been shared through joint publications as well as at ECMWF seminars and workshops. Here, the three new Fellows set out their plans for their Fellowships:
Prof. Dr. Michaela Hegglin, Forschungszentrum Jülich (Germany), University of Reading (UK), and University of Wuppertal (Germany)
I am excited to have been offered this opportunity to become a fellow of ECMWF. My main research revolves around understanding the processes involved in chemistry–climate interactions through the synergistic use of models and observations treated as independent sources of information. I have been strongly involved in activities of the World Climate Research Programme (WCRP) project Atmospheric Processes And their Role in Climate (APARC) as well as European Space Agency (ESA) mission advisory groups and scientific studies, including the ESA Climate Change Initiative. I look forward to bringing this expertise to bear on the problem of predictability.
Together with the Earth System Predictability section at ECMWF, we will explore the role of composition (especially upper tropospheric and stratospheric ozone, water vapour, and aerosol) in enhancing the predictability of ECMWF weather forecasts on medium-range to sub-seasonal and seasonal timescales. Composition provides predictability through external forcings (e.g., halogen-induced ozone loss or the increase in stratospheric water vapour following the Hunga Tonga volcanic eruption), and it modifies the internal variability of the atmosphere through radiative effects. The effects of composition on predictability can be difficult to isolate statistically because of many confounding factors. I therefore plan to develop a process-oriented approach that brings together physical reasoning and parameter estimation within a causal framework.
Dr Ilias Pechlivanidis, Swedish Meteorological and Hydrological Institute (SMHI) and Uppsala University (Sweden)
My ECMWF Fellowship will focus on advancing the application of artificial intelligence (AI) to enhance the prediction of hydrological extremes, such as floods and droughts, across spatial and temporal scales. A central objective is to further develop and extend a hybrid AI framework that demonstrably improves forecast skill, particularly in data-scarce or ungauged regions. These methodologies are being co-developed within EU-funded Horizon Europe projects and are designed for compatibility with existing forecasting architectures, ensuring relevance for operational services.
A key aspect of the Fellowship will be to assess how these AI-enhanced methods could complement and add value to pan-European operational systems, including components of the EU’s Copernicus Emergency Management Service (CEMS), which are run computationally at ECMWF.
In addition, the Fellowship will investigate AI techniques that connect hydro-meteorological hazard indicators with observed societal impacts, thereby improving the skill and relevance of impact-based forecasting on sub-seasonal to seasonal (S2S) timescales. This effort will help to translate traditional physical indicators of extremes into more actionable, context-aware information for risk communication and decision-making. Overall, the Fellowship will contribute to both scientific innovation in hybrid AI modelling and the enhancement of user-relevant, risk-informed forecasting capabilities aligned with ECMWF’s mission.
Dr Anthony Weaver, Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique (Cerfacs) (France)
I have been collaborating with ECMWF for many years, with a primary focus on the development of the NEMOVAR ocean data assimilation software. NEMOVAR, a joint initiative between Cerfacs, ECMWF, the UK Met Office, and Inria (French National Institute for Research in Digital Science and Technology), has served as the foundation of ECMWF’s ocean reanalysis and operational forecasting suites for over a decade. It also plays a central role in ECMWF’s coupled data assimilation system, currently under active development.
My research has spanned multiple aspects of variational data assimilation, including 4D-Var, covariance modelling, balance constraints, and minimisation algorithms. While my main focus has been ocean data assimilation, several of the methodological advances I have contributed to, particularly in diffusion-based covariance modelling, have relevance and applicability across other components of the coupled system.
The ECMWF Fellowship presents an opportunity to strengthen ongoing collaborations in ocean and coupled data assimilation. Current and planned work, undertaken as part of the ERGO project funded by the EU’s Copernicus Climate Change Service (C3S), focuses on developing methods for accounting for spatially correlated errors in altimeter and sea-surface temperature observations, evaluating multi-scale, ensemble-based formulations of the background error covariance matrix, and advancing a weak-constraint formulation of the NEMOVAR data assimilation system to estimate and correct for model error.