Fast and accurate learned multiresolution dynamical downscaling for precipitation

Abstract

Abstract. This study develops a neural network-based approach for emulating high-resolution modeled precipitation data with comparable statistical properties but at greatly reduced computational cost. The key idea is to use combination of low- and high- resolution simulations to train a neural network to map from the former to the latter. Specifically, we define two types of CNNs, one that stacks variables directly and one that encodes each variable before stacking, and we train each CNN type both with a conventional loss function, such as mean square error (MSE), and with a conditional generative adversarial network (CGAN), for a total of four CNN variants.We compare the four new CNN-derived high-resolution precipitation results with precipitation generated from original high resolution simulations, a bilinear interpolater and the state-of-the-art CNN-based super-resolution (SR) technique. Results show that the SR technique produces results similar to those of the bilinear interpolator with smoother spatial and temporal distributions and smaller data variabilities and extremes than the high resolution simulations. While the new CNNs trained by MSE generate better results over some regions than the interpolator and SR technique do, their predictions are still not as close as ground truth. The CNNs trained by CGAN generate more realistic and physically reasonable results, better capturing not only data variability in time and space but also extremes such as intense and long-lasting storms. The new proposed CNN-based downscaling approach can downscale precipitation from 50 km to 12 km in 14 min for 30 years once the network is trained (training takes 4 hours using 1 GPU), while the conventional dynamical downscaling would take 1 months using 600 CPU cores to generate simulations at the resolution of 12 km over contiguous United States.

Adjusting spatial dependence of climate model outputs with cycle-consistent adversarial networks

“…Due to the stochastic and high-dimensionality nature of many physical processes of the Earth system, GANs and conditional GANs are particularly appealing for atmospheric science problems. Recently, they have been used for various Earth-science related applications: for instance for statistical downscaling Wang et al 2021), temporal disaggregation of spatial rainfall fields (Scher and Peßenteiner 2020), sampling of extreme values (Bhatia et al 2020), modelling of chaotic dynamical systems (e.g., Xie et al 2018;Wu et al 2020), classification of snowflake images , weather forecasting (Bihlo 2020) and stochastic parameterization in geophysical models (Gagne II et al 2020).…”

Comment on gmd-2022-76 (response to reviewers)

“…Several authors have already demonstrated their potential in earth science related applications. Super resolution CNNs have been applied to radar data (Geiss and Hardin, 2020a), wind and solar modeling (Stengel et al, 2020), satellite remote sensing (Liebel and Körner, 2016;Lanaras et al, 2018;Müller et al, 2020), precipitation modeling (Wang et al, 2021), and climate modeling (Vandal et al, 2018;Baño Medina et al, 2020).…”

Using AI to build a hydrobiogeochemical soil model

“…AI has several methods that are well suited to this challenge. Artificial neural networks have proven successful at embedding multiple sources of data [30,31] and mapping complex non-linear input-outputs relationships. In particular, techniques such as generative adversarial networks have been successful at enhancing the resolution of outputs and will be used to obtain the most detailed representation of the quantities of interest [32].…”