Preprints
https://doi.org/10.5194/hess-2022-334
https://doi.org/10.5194/hess-2022-334
 
20 Sep 2022
20 Sep 2022
Status: this preprint is currently under review for the journal HESS.

Hybrid forecasting: using statistics and machine learning to integrate predictions from dynamical models

Louise Slater1, Louise Arnal2, Marie-Amélie Boucher3, Annie Y.-Y. Chang4,5, Simon Moulds1, Conor Murphy6, Grey Nearing7, Guy Shalev8, Chaopeng Shen9, Linda Speight1, Gabriele Villarini10, Robert L. Wilby11, Andrew Wood12, and Massimiliano Zappa4 Louise Slater et al.
  • 1School of Geography and the Environment, University of Oxford, Oxford, UK
  • 2University of Saskatchewan, Centre for Hydrology, Canmore, Canada
  • 3Université de Sherbrooke, Canada
  • 4Swiss Federal Research Institute WSL, Birmensdorf, Switzerland
  • 5ETH, Zurich, Switzerland
  • 6Irish Climate Analysis and Research Units, Department of Geography, Maynooth University, Kildare, Ireland
  • 7Google Research, Mountain View, CA, USA
  • 8Google Research, Tel Aviv, Israel
  • 9Civil and Environmental Engineering, The Pennsylvania State University, State College, PA 16801, USA
  • 10IIHR–Hydroscience and Engineering, University of Iowa, Iowa, USA
  • 11Geography and Environment, Loughborough University, Loughborough, UK
  • 12National Center for Atmospheric Research, Climate and Global Dynamics, Boulder, CO, USA

Abstract. Hybrid hydroclimatic forecasting systems employ data-driven (statistical or machine learning) methods to harness and integrate a broad variety of predictions from dynamical, physics-based models – such as numerical weather prediction, climate, land, hydrology and Earth system models – into a final prediction product. They are recognised as a promising way of enhancing prediction skill of meteorological and hydroclimatic variables and events, including rainfall, temperature, streamflow, floods, droughts, tropical cyclones, or atmospheric rivers. Hybrid forecasting methods are now receiving growing attention due to advances in weather and climate prediction systems at sub-seasonal to decadal scales, a better appreciation of the strengths of machine learning, plus expanding access to computational resources and methods. Such systems are attractive because they may avoid the need to run a computationally-expensive offline land model, can minimize the effect of biases that exist within dynamical outputs without explicit bias correction and downscaling, benefit from the strengths of machine learning models, and can learn from large datasets, while combining different sources of predictability with varying time-horizons. Here we review recent developments in hybrid hydroclimatic forecasting and outline key challenges and opportunities. These include obtaining physically-explainable results, assimilating human influences from novel data sources, integrating new ensemble techniques to improve predictive skill, creating seamless prediction schemes that merge short to long lead times, incorporating modelled initial land surface and ocean/ice conditions, acknowledging spatial variability in landscape and atmospheric forcing, and increasing the operational uptake of hybrid prediction schemes.

Louise Slater et al.

Status: open (until 15 Nov 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Louise Slater et al.

Louise Slater et al.

Viewed

Total article views: 892 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
652 234 6 892 4 5
  • HTML: 652
  • PDF: 234
  • XML: 6
  • Total: 892
  • BibTeX: 4
  • EndNote: 5
Views and downloads (calculated since 20 Sep 2022)
Cumulative views and downloads (calculated since 20 Sep 2022)

Viewed (geographical distribution)

Total article views: 857 (including HTML, PDF, and XML) Thereof 857 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 

Discussed

Latest update: 29 Sep 2022
Download
Short summary
Hybrid forecasting systems employ data-driven methods to integrate predictions from dynamical physics-based models into a final prediction product. They are recognised as a promising way of enhancing prediction skill of meteorological and hydroclimatic variables and events – including rainfall, temperature, streamflow, floods, droughts, tropical cyclones, or atmospheric rivers. Here we review recent developments in hybrid forecasting and outline key challenges and opportunities.