19 Aug 2022
19 Aug 2022
Status: this preprint is open for discussion.

Downscaling using Deep Convolutional Autoencoders, a case study for South East Asia

Oliver Douglas Levers1, Dorien Herremans1, Anurag Dipankar2, and Lucienne Blessing1 Oliver Douglas Levers et al.
  • 1Singapore University of Technology and Design (SUTD), 8 Somapah Rd, Singapore 487372
  • 2Institute for Atmospheric and Climate Science, ETH Zürich

Abstract. Inspired by recent advancements in the field of computer vision, specifically models for generating higher-resolution images from low-resolution images, we investigate the utility of a deep convolutional autoencoder for downscaling and bias correcting climate projections for South East Asia (SEA). Downscaled projections of 2 m surface temperature are generated, using autoencoders trained with data from the Coupled Model Intercomparison Project Phase 5 (CMIP5) and data from the fifth generation ECMWF atmospheric reanalysis (ERA5) project. Using CMIP5 projections as an input, three sets of downscaled data are generated using three methods of autoencoder training, which allow us to determine how autoencoder downscaling and bias correction modify temperature values. Where possible, the downscaled outputs are compared against the Southeast Asia Regional Climate Downscaling/Coordinated Regional Climate Downscaling Experiment–Southeast Asia (SEACLID/CORDEX–SEA) project and outputs from available CMIP6 experiments, to evaluate performance. The autoencoders are found to excel at the rapid generation of highly spatially-resolved climate projections for surface temperature. Realistic spatial features due to coastal and topographic variation are generated by the autoencoder, which are not present in the CMIP5 projections. Additionally, the autoencoders are capable of generating forecast data with regional temperature profiles exceeding that of those appearing in the training set (out-of-sample extrapolation). Seasonal temperature cycles are retained after downscaling throughout the region, despite the absence of temporal information provided to the model. However, autoencoders trained to carry out bias correction display a tendency to smooth daily average temperatures and reduce daily highs and lows beyond that which can be expected to be realistic. Without bias correction, downscaled outputs have a reduced improvement in spatial resolution but the daily temperature profiles of the CMIP5 input forecasts are maintained. Autoencoders rely on the presence of structural features in the datasets to carry out downscaling, and so performance over the oceans is reduced as strong temperature gradients are absent. For this reason, ocean warming is not well represented, an artefact which is not immediately clear in the downscaled outputs. This study demonstrates the importance of rigorous analysis of 'black-box' methods, which can generate non-obvious artefacts that could potentially create misleading results. Despite these limitations, Autoencoders are clearly capable of generating much needed high-resolution climate projections, and strategies to improve upon shortcomings are numerous and well established.

Oliver Douglas Levers et al.

Status: open (until 27 Oct 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2022-234', Anonymous Referee #1, 02 Sep 2022 reply
  • CEC1: 'Comment on egusphere-2022-234', Juan Antonio Añel, 20 Sep 2022 reply

Oliver Douglas Levers et al.

Model code and software

Trained models and code accompanying 'Downscaling using Deep Convolutional Autoencoders, a case study for South East Asia' Oliver Levers

Oliver Douglas Levers et al.


Total article views: 366 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
289 68 9 366 1 1
  • HTML: 289
  • PDF: 68
  • XML: 9
  • Total: 366
  • BibTeX: 1
  • EndNote: 1
Views and downloads (calculated since 19 Aug 2022)
Cumulative views and downloads (calculated since 19 Aug 2022)

Viewed (geographical distribution)

Total article views: 348 (including HTML, PDF, and XML) Thereof 348 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
Latest update: 28 Sep 2022
Short summary
Southeast Asia is a region which is very sensitive to climate change and has numerous islands and peninsulas which are not well resolved within many General Circulation Models (GCMs). Here, deep convolutional encoders are employed to increase the spatial resolution of climate model data (downscaling) and address systematic errors in model outputs (bias correction). Technique and region-specific issues are identified for surface temperature data and compared with other model outputs.