% 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{Ma:903941,
      author       = {Ma, Yueling and Montzka, Carsten and Naz, Bibi and Kollet,
                      Stefan},
      title        = {{A} {N}ovel {ML}-{B}ased {M}ethodology for {E}stimating
                      {W}ater {T}able {D}epth {A}nomalies at the {E}uropean
                      {C}ontinent {S}cale},
      reportid     = {FZJ-2021-05549},
      year         = {2021},
      abstract     = {Effective and efficient groundwater monitoring at the
                      continental scale is still a challenge, mainly due to the
                      scarcity of water table depth (wtd) observations. In this
                      study, we proposed a novel methodology based on advanced
                      machine learning techniques, Long Short-Term Memory (LSTM)
                      networks and transfer learning, for estimating monthly wtd
                      anomalies (wtda) over Europe with monthly precipitation and
                      soil moisture anomalies (pra and θa) as input. In the
                      methodology, the LSTM networks were trained on simulation
                      results and then, without additional training, utilized to
                      estimate wtda based on pra and θa from observational
                      datasets, thereby transferring the simulated input-output
                      relationship to the observation-based estimation. The
                      obtained estimates were evaluated based on in-situ wtd
                      measurements at 2,604 wells distributed over different
                      European regions, achieving R from 0.39 to 0.79 and RMSE
                      from 0.37 to 1.1 for regional averaged values. This
                      constitutes of a >0.10 increase in R and a >0.13 decrease in
                      RMSE compared to the simulation results used for training.
                      In addition, compared with the LSTM networks directly
                      trained on observations, the proposed methodology showed
                      slightly worse test performance at the individual pixel
                      level, lending confidence to applications in areas without
                      wtd observations. The study provides a validated methodology
                      for producing reliable wtda estimates over the European
                      domain in the absence of observations, which can be used for
                      data reconstruction and online groundwater monitoring useful
                      in European groundwater management.},
      month         = {Dec},
      date          = {2021-12-13},
      organization  = {AGU Fall Meeting 2021, New Orleans, LA
                       and online (The US and online), 13 Dec
                       2021 - 17 Dec 2021},
      subtyp        = {After Call},
      cin          = {IBG-3},
      cid          = {I:(DE-Juel1)IBG-3-20101118},
      pnm          = {2173 - Agro-biogeosystems: controls, feedbacks and impact
                      (POF4-217) / ERA-PLANET - The European network for observing
                      our changing planet (689443)},
      pid          = {G:(DE-HGF)POF4-2173 / G:(EU-Grant)689443},
      typ          = {PUB:(DE-HGF)24},
      url          = {https://juser.fz-juelich.de/record/903941},
}