Technical note: Deep learning for creating surrogate models of precipitation in Earth system models
Open Access
- 26 February 2020
- journal article
- research article
- Published by Copernicus GmbH in Atmospheric Chemistry and Physics
- Vol. 20 (4) , 2303-2317
Abstract
We investigate techniques for using deep neural networks to produce surrogate models for short-term climate forecasts. A convolutional neural network is trained on 97 years of monthly precipitation output from the 1pctCO2 run (the CO2 concentration increases by 1 % per year) simulated by the second-generation Canadian Earth System Model (CanESM2). The neural network clearly outperforms a persistence forecast and does not show substantially degraded performance even when the forecast length is extended to 120 months. The model is prone to underpredicting precipitation in areas characterized by intense precipitation events. Scheduled sampling (forcing the model to gradually use its own past predictions rather than ground truth) is essential for avoiding amplification of early forecasting errors. However, the use of scheduled sampling also necessitates preforecasting (generating forecasts prior to the first forecast date) to obtain adequate performance for the first few prediction time steps. We document the training procedures and hyperparameter optimization process for researchers who wish to extend the use of neural networks in developing surrogate models.Keywords
All Related Versions
This publication has 42 references indexed in Scilit:
- Retrospective prediction of the global warming slowdown in the past decadeNature Climate Change, 2013
- A verification framework for interannual-to-decadal predictions experimentsClimate Dynamics, 2012
- Systematic Estimates of Initial-Value Decadal Predictability for Six AOGCMsJournal of Climate, 2012
- The representative concentration pathways: an overviewClimatic Change, 2011
- Carbon emission limits required to satisfy future representative concentration pathways of greenhouse gasesGeophysical Research Letters, 2011
- Uncertainties in the 20th century carbon budget associated with land use changeGlobal Change Biology, 2010
- How will Earth's surface temperature change in future decades?Geophysical Research Letters, 2009
- Cassini detection of Enceladus' cold water‐group plume ionosphereGeophysical Research Letters, 2009
- On Early Stopping in Gradient Descent LearningConstructive Approximation, 2007
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998