Fidelity of Precipitation Extremes in High Resolution Global Climate Simulations

Abstract

Precipitation extremes have tangible societal impacts. Here, we assess if current state of the art global climate model simulations at high spatial resolutions (0.35 • x0.35 • ) capture the observed behavior of precipitation extremes in the past few decades over the continental US. We design a correlation-based regionalization framework to quantify precipitation extremes, where samples of extreme events for a grid box may also be drawn from neighboring grid boxes with statistically equal means and statistically significant temporal correlations. We model precipitation extremes with the Generalized Extreme Value (GEV) distribution fits to time series of annual maximum precipitation. Non-stationarity of extremes is captured by including a timedependent parameter in the GEV distribution. Our analysis reveals that the high-resolution model substantially improves the simulation of stationary precipitation extreme statistics particularly over the Northwest Pacific coastal region and the Southeast US. Observational data exhibits significant non-stationary behavior of extremes only over some parts of the Western US, with declining trends in the extremes. While the high resolution simulations improve upon the low resolution model in simulating this non-stationary behavior, the trends are statistically significant only over some of those regions.

Reconstructing High Resolution ESM Data Through a Novel Fast Super Resolution Convolutional Neural Network (FSRCNN)

“…The resolution of finer scales in the high resolution model and scale-agnostic nature of current sub-grid scale physical parameterizations used in climate models imply that a low-resolution model simulation is not statistically equivalent to coarsened data from a high-resolution model. For example, the simulation of precipitation extremes is found to be stronger in high resolution simulations than the low resolution simulations, even after conservative mapping of high-resolution simulation data to the low-resolution grid (Mahajan et al, 2015;Wehner et al, 2010Wehner et al, , 2014. In order to apply FSRCNN-ESM directly to low-resolution model output to generate high resolution images at the skill of a high resolution simulation, we would thus require a bias-correction step.…”

Model Resolution Sensitivity of the Simulation of North Atlantic Oscillation Teleconnections to Precipitation Extremes

“…GEV models are fitted to the monthly maximum daily total precipitation during the winter months from November to February at each grid point using the R package ismev (Heffernan & Stephenson, ). Such a covariate‐based block maxima or peak over threshold approach has been widely used to represent nonstationary climate extremes (Brown et al, ; Evans et al, ; Kharin & Zwiers, ; Kharin et al, ; Mahajan et al, ; Wehner et al, ), as well as to link climate variability modes (e.g., Brown et al, ) and large‐scale meteorological patterns (e.g., Grotjahn et al, ; Whan & Zwiers, ; X. Zhang et al, ) to climate extremes.…”

“…High‐resolution models (< 0.5°) also tend to simulate stronger extremes on daily and subdaily time scales generally improving upon the negative bias noted in their counterpart low‐resolution models (>1°; e.g., Johnson et al, ; J. Li et al, ; Mahajan et al, ; Terai et al, ; Wehner et al, ; Yu et al, ) and produce lower drizzle (e.g., Rauscher et al, ; Wehner et al, ; L. Zhang et al, ). The improvements are particularly noted over regions with varied topography (e.g., Johnson et al, ; J. Li et al, ; Mahajan et al, ; Wehner et al, ; Yu et al, ; L. Zhang et al, ) as finer‐scale physical and dynamical processes such as orographic lifting are explicitly resolved around complex terrain.…”

“…High‐resolution models (< 0.5°) also tend to simulate stronger extremes on daily and subdaily time scales generally improving upon the negative bias noted in their counterpart low‐resolution models (>1°; e.g., Johnson et al, ; J. Li et al, ; Mahajan et al, ; Terai et al, ; Wehner et al, ; Yu et al, ) and produce lower drizzle (e.g., Rauscher et al, ; Wehner et al, ; L. Zhang et al, ). The improvements are particularly noted over regions with varied topography (e.g., Johnson et al, ; J. Li et al, ; Mahajan et al, ; Wehner et al, ; Yu et al, ; L. Zhang et al, ) as finer‐scale physical and dynamical processes such as orographic lifting are explicitly resolved around complex terrain. These improvements are largely associated with an increase in large‐scale (stratiform) precipitation resulting from resolved scales of motion, which increasingly contribute more to total precipitation with increasing resolution (e.g., Boyle & Klein, ; Demory et al, ; Kooperman et al, ; F. Li et al, ; O'Brien et al, ; Rauscher et al, ; Yang et al, ).…”

The Infrastructure Trolley Problem: Positioning Safe‐to‐fail Infrastructure for Climate Change Adaptation

“…Ongoing debate in the scientific community on whether to use stationary or nonstationary statistical distributions to predict the frequency and/or intensity of future climate extremes highlights the need for infrastructure development practices that transcend the risk-based approach to future weather extremes (Warren et al, 2018). High-resolution climate simulations can improve the representation of extreme weather in certain regions (Mahajan et al, 2015). Also, there have been advancements in modeling convective storms, which may help design infrastructure for future flash floods (Prein et al, 2017).…”