Is time a variable like the others in multivariate statistical downscaling and bias correction?

Yoann Robin,Mathieu Vrac

doi:10.5194/esd-12-1253-2021

Abstract

Abstract. Bias correction and statistical downscaling are now regularly applied to climate simulations to make then more usable for impact models and studies. Over the last few years, various methods were developed to account for multivariate – inter-site or inter-variable – properties in addition to more usual univariate ones. Among such methods, temporal properties are either neglected or specifically accounted for, i.e. differently from the other properties. In this study, we propose a new multivariate approach called “time-shifted multivariate bias correction” (TSMBC), which aims to correct the temporal dependency in addition to the other marginal and multivariate aspects. TSMBC relies on considering the initial variables at various times (i.e. lags) as additional variables to be corrected. Hence, temporal dependencies (e.g. auto-correlations) to be corrected are viewed as inter-variable dependencies to be adjusted and an existing multivariate bias correction (MBC) method can then be used to answer this need. This approach is first applied and evaluated on synthetic data from a vector auto-regressive (VAR) process. In a second evaluation, we work in a “perfect model” context where a regional climate model (RCM) plays the role of the (pseudo-)observations, and where its forcing global climate model (GCM) is the model to be downscaled or bias corrected. For both evaluations, the results show a large reduction of the biases in the temporal properties, while inter-variable and spatial dependence structures are still correctly adjusted. However, increasing the number of lags too much does not necessarily improve the temporal properties, and an overly strong increase in the number of dimensions of the dataset to be corrected can even imply some potential instability in the adjusted and/or downscaled results, calling for a reasoned use of this approach for large datasets.

Highlights

Climate and Earth system models (ESMs) and their simulations are the main physical tools to investigate the potential future evolutions of the climate system (e.g. Flato et al, 2013; Kirtman et al, 2013)
They are clearly indispensable for testing how different scenarios of greenhouse gas emission trajectories might induce climate changes and, for trying to anticipate potential impacts of those changes (e.g. IPCC, 2019). Such elaborate models contain many relevant and complex processes characterizing the climate properties and their dependencies, the numerical simulations they generate are often tainted with biases and disagreements with respect to observations. Those can stem (i) from the spatial resolution of the simulations, usually too low compared to needs of impact models that may require very local or high-resolution input climate data, e.g. kilometres, hundreds of metres or below, down to the weather stations (e.g. Chen et al, 2011; Maraun and Widmann, 2018), and/or (ii) from inherent biases in the model simulations, due to parameter
We apply the time-shifted multivariate bias correction (TSMBC) method with the underlying dynamical optimal transport correction (dOTC) method to the bias correction and downscaling of the Institut Pierre Simon Laplace (IPSL) global climate model (GCM) simulations with respect to the regional climate model (RCM) simulations taken as references

Summary

Introduction

Climate and Earth system models (ESMs) and their simulations are the main physical tools to investigate the potential future evolutions of the climate system (e.g. Flato et al, 2013; Kirtman et al, 2013). Flato et al, 2013; Kirtman et al, 2013) They are clearly indispensable for testing how different scenarios of greenhouse gas emission trajectories might induce climate changes and, for trying to anticipate potential impacts of those changes (e.g. IPCC, 2019). Such elaborate models contain many relevant and complex processes characterizing the climate properties and their dependencies, the numerical simulations they generate are often tainted with biases and disagreements with respect to observations. Vrac: Is time a variable like the others in multivariate statistical downscaling and bias correction?

Objectives

Results

Conclusion