A spatial hybrid approach for downscaling of extreme precipitation fields

Abstract

For a few decades, climate models are used to provide future scenarios of precipitation with increasingly higher spatial resolution. However, this resolution is not yet sufficient to describe efficiently what happens at local scale. Dynamical and statistical methods of downscaling have been developed and allow us to make the link between two levels of resolution and enable us to get values at a local scale based on large-scale information from global or regional climate models. Nevertheless, both the extreme behavior and the spatial structures are not well described by these downscaling methods. We propose a two-step methodology, called spatial hybrid downscaling (SHD), to solve this problem. The first step consists in applying a univariate (i.e., one-dimensional) statistical downscaling to link the high-and low-resolution variables at some given locations. Once this 1d-link is performed, a conditional simulation algorithm of max-stable processes adapted to the extremal t process enables us to get conditional distributions of extreme precipitation at any point of the region. An application is performed on precipitation data in the south of France where extreme (Cevenol) events have major impacts (e.g., floods). Different versions of the SHD approach are tested. Most of them show particularly good results regarding univariate and multivariate criteria and overcome classical downscaling techniques tested in comparison. Furthermore, these conclusions are robust to the choice of the 1d-link functions tested and to the choice of the conditioning points to drive the conditional local-scale simulations performed by the SHD approach.

Assessment of the Simulation of Gravity Waves Generation by a Tropical Cyclone in the High‐Resolution WACCM and the WRF

“…Downscaling is widely used to bridge the gap between GCMs and RCMs (e.g., Bao et al, ; Bechler et al, ; He et al, ; Xu & Yang, ). In this paper, a similar approach is presented to investigate the ability of high‐resolution WACCM to resolve TGWs, by conducting WRF simulations using initial and boundary conditions obtained from the high‐resolution WACCM in a regional domain.…”

“…Dynamical downscaling simulations have proven to be a valuable approach to study fine‐scale features of climate simulations, and to evaluate climate model parameterizations (e.g., precipitation; Bao et al, ; Bechler et al, ; He et al, ; Xu & Yang, ). Besides, as the first GCM to be able to resolve mesoscale TGWs, there is currently little or no local data available for comparison and evaluation (Liu et al, ).…”

A new discrete multiplicative random cascade model for downscaling intermittent rainfall fields

“…An important advantage of the microcanonical model is that the distribution of the cascade weights can be studied directly from the data through the calculation of empirical breakdown coefficients (Cârsteanu and Foufoula-Georgiou, 2016;Licznar et al, 2015). The latter are estimated by successively aggregating grid cells in the input field to larger spatial scales and by studying how the rainfall volumes in aggregated grid cells split up as a function of area and rainfall intensity.…”

Statistical downscaling for future extreme wave heights in the North Sea

“…Since most multivariate extreme value models [Coles and Tawn (1994), de Haan and de Ronde (1998)] and all max-stable processes [Bechler, Vrac and Bel (2015), Davison, Padoan and Ribatet (2012)] can model asymptotic dependence only, we shall use the conditional extreme value model of Heffernan and Tawn (2004), which includes both asymptotic dependence and asymptotic independence. Heffernan and Tawn (2004) and Keef, Papastathopoulos and Tawn (2013) found that for a wide range of copulas there exist parameters −1 ≤ α ≤ 1 and −∞ < β < 1 and a nondegenerate distribution function Q(z) such that for w > 0 and z ∈ R, (2.12) lim…”

“…First, it models the joint distribution of (H, W, θ H , θ W ) at each location, both in the past and in the future, gaining insight over methods that only downscale H . Second, it uses state-of-the-art univariate and multivariate extreme value theory including more efficient threshold methods, and a broad class of asymptotically justified dependence models, which encompass methods used previously by Bechler, Vrac and Bel (2015) as a special sub-class. Finally, our approach uses a novel form of distributional downscaling that overcomes some major weaknesses with existing methods.…”