ClimAlign: Unsupervised statistical downscaling of climate variables via normalizing flows

Abstract

Downscaling is a common task in climate science and meteorology in which the goal is to use coarse scale, spatio-temporal data to infer values at finer scales. Statistical downscaling aims to approximate this task using statistical patterns gleaned from an existing dataset of downscaled values, often obtained from observations or physical models. In this work, we investigate the application of domain alignment to the task of statistical downscaling. We present ClimAlign, a novel method for unsupervised, generative downscaling using adaptations of recent work in normalizing flows for variational inference. We evaluate the viability of our method using several different metrics on two datasets consisting of daily temperature and precipitation values gridded at low (1 • latitude/longitude) and high ( 1 4• and 1 8 • ) resolutions. We show that our method achieves comparable predictive performance to existing supervised statistical downscaling methods while simultaneously allowing for both conditional and unconditional sampling from the joint distribution over high and low resolution spatial fields. To the best of our knowledge, this is the first proposed method for unsupervised statistical downscaling, and one of very few proposed methods that allows for efficient sampling of synthetic data. We provide publicly accessible implementations of our method, as well as the baselines used for comparison, on GitHub 3 . I. MOTIVATIONGlobal Climate Models (GCMs) provide valuable information about Earths past, present, and future climate. The predictions made by these models have a wide range of possible applications in the Earth sciences, as well as in engineering and risk analysis. However, climate models typically operate at very coarse spatial scales, often summarizing over one hundred square kilometers or more. Predictions at this scale are useful for global analyses but provide very limited local or regional information.