% IMPORTANT: The following is UTF-8 encoded.  This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.

@INPROCEEDINGS{Langguth:1035267,
      author       = {Langguth, Michael and Lessig, Christian and Schultz, Martin
                      and Luise, Ilaria},
      title        = {{D}ownscaling with the foundation model {A}tmo{R}ep},
      reportid     = {FZJ-2025-00343},
      year         = {2024},
      abstract     = {In recent years, deep neural networks (DNN) to enhance the
                      resolution of meteorological data, known as statistical
                      downscaling, have surpassed classical statistical methods
                      that have been developed previously with respect to several
                      validation metrics. The prevailing approach for DNN
                      downscaling is to train deep learning models in an
                      end-to-end manner. However, foundation models trained on
                      very large datasets in a self-supervised way have proven to
                      provide new SOTA results for various applications in natural
                      language processing and computer vision. To investigate the
                      benefit of foundation models in Earth Science applications,
                      we deploy the large-scale representation model for
                      atmospheric dynamics AtmoRep (Lessig et al., 2023) for
                      statistical downscaling of the 2m temperature over Central
                      Europe. AtmoRep has been trained on almost 40 years of ERA5
                      data from 1979 to 2017 and has shown promising skill in
                      several intrinsic and downstream applications. By extending
                      AtmoRep’s encoder-decoder with a tail network for
                      downscaling, we super-resolve the coarse-grained 2 m
                      temperature field from ERA5-data (Δx = 25 km) to attain the
                      high spatial resolution (Δx = 6 km) of the COSMO REA6
                      dataset. Different coupling approaches between the core and
                      tail network (e.g. with and without fine-tuning the core
                      model) are tested and analyzed in terms of accuracy and
                      computational efficiency. Preliminary results show that
                      downscaling with a task-specific extension of the foundation
                      model AtmoRep can improve the downscaled product in terms of
                      standard evaluation metrics such as the RMSE compared to a
                      task-specific deep learning model. However, deficiencies in
                      the spatial variability of the downscaled product are also
                      revealed, highlighting the need for future work to focus
                      especially on target data that inhibit a high degree of
                      spatial variability and intrinsic uncertainty such as
                      precipitation.},
      month         = {Apr},
      date          = {2024-04-14},
      organization  = {European Geosciences Union General
                       Assembly 2024, Vienna (Austria), 14 Apr
                       2024 - 19 Apr 2024},
      subtyp        = {After Call},
      cin          = {JSC},
      cid          = {I:(DE-Juel1)JSC-20090406},
      pnm          = {5111 - Domain-Specific Simulation $\&$ Data Life Cycle Labs
                      (SDLs) and Research Groups (POF4-511) / MAELSTROM - MAchinE
                      Learning for Scalable meTeoROlogy and cliMate (955513) /
                      Verbundprojekt: MAELSTROM - Skalierbarkeit von Anwendungen
                      des Maschinellen Lernens in den Bereichen Wetter und
                      Klimawissenschaften für das zukünftige Supercomputing
                      (16HPC029) / Earth System Data Exploration (ESDE)},
      pid          = {G:(DE-HGF)POF4-5111 / G:(EU-Grant)955513 / G:(BMBF)16HPC029
                      / G:(DE-Juel-1)ESDE},
      typ          = {PUB:(DE-HGF)24},
      doi          = {10.5194/egusphere-egu24-18331},
      url          = {https://juser.fz-juelich.de/record/1035267},
}