Abstract
In a recent paper, Baño-Medina et al. (Configuration and Intercomparison of deep learning neural models for statistical downscaling. preprint, 2019) assessed the suitability of deep convolutional neural networks (CNNs) for downscaling of temperature and precipitation over Europe using large-scale ‘perfect’ reanalysis predictors. They compared the results provided by CNNs with those obtained from a set of standard methods which have been traditionally used for downscaling purposes (linear and generalized linear models), concluding that CNNs are well suited for continental-wide applications. That analysis is extended here by assessing the suitability of CNNs for downscaling future climate change projections using Global Climate Model (GCM) outputs as predictors. This is particularly relevant for this type of “black-box” models, whose results cannot be easily explained based on physical reasons and could potentially lead to implausible downscaled projections due to uncontrolled extrapolation artifacts. Based on this premise, we analyze in this work the two key assumptions that are made in perfect prognosis downscaling: (1) the predictors chosen to build the statistical model should be well reproduced by GCMs and (2) the statistical model should be able to reliably extrapolate out of sample (climate change) conditions. As a first step to test the suitability of these models, the latter assumption is assessed here by analyzing how the CNNs affect the raw GCM climate change signal (defined as the difference, or delta, between future and historical climate). Our results show that, as compared to well-established generalized linear models (GLMs), CNNs yield smaller departures from the raw GCM outputs for the end of century, resulting in more plausible downscaling results for climate change applications. Moreover, as a consequence of the automatic treatment of spatial features, CNNs are also found to provide more spatially homogeneous downscaled patterns than GLMs.
Highlights
As a result of several decades of intense research, a variety of statistical downscaling approaches and techniques are nowadays available to fill the gap between the coarse resolution outputs provided by Global Climate Models (GCMs) and the local or regional information required for impact1 3 Vol.:(0123456789)J
Baño-Medina et al (2019) assessed the suitability of deep convolutional neural networks (CNNs) for the downscaling of temperature and precipitation over Europe using large-scale ‘perfect’ reanalysis predictors. They used the experimental framework defined in VALUE (Experiment 1) to compare the results provided by CNNs with those obtained from a set of other more classical, standard techniques i.e., generalized linear models, concluding that CNNs are well suited for continental-wide applications
bias adjustment (BA) consists in adjusting the EC-Earth monthly means towards the corresponding reanalysis values, gridbox by gridbox (Eqs. 1, where j refers to a particular variable, i = 1, 2, ..., 12 to the month of the year, and h and f denote historical and RCP periods, respectively)
Summary
As a result of several decades of intense research, a variety of statistical downscaling approaches and techniques are nowadays available to fill the gap between the coarse resolution outputs provided by Global Climate Models (GCMs) and the local or regional information required for impact. Baño-Medina et al (2019) assessed the suitability of deep convolutional neural networks (CNNs) for the downscaling of temperature and precipitation over Europe using large-scale ‘perfect’ reanalysis predictors To do so, they used the experimental framework defined in VALUE (Experiment 1) to compare the results provided by CNNs with those obtained from a set of other more classical, standard techniques i.e., generalized linear models, concluding that CNNs are well suited for continental-wide applications. Only a few studies have focused on the impact that the two key assumptions of perfect prognosis downscaling may have for climate change applications: (1) the predictors should be well reproduced by GCMs and (2) the statistical model should be able to generalize and extrapolate out-of-sample (e.g. climate change) conditions This is crucial to assess the credibility of future climate information and avoid misadaptation (Pryor and Schoof 2020).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
