Changes over time in the 100-year return value of climate model variables

TL;DR

This study evaluates how the 100-year return values of wind, solar irradiance, and temperature variables from climate models change over time due to climate forcing, using advanced statistical models and multiple climate scenarios.

Contribution

It introduces non-stationary extreme value and Gaussian regression models to quantify changes in climate extremes and means over a century, considering multiple GCM outputs and scenarios.

Findings

Strong evidence for changes in 100-year return values of solar irradiance and temperature.

Less evidence for wind variables' changes over time and scenarios.

Quantified the distribution of future changes in climate extremes and means.

Abstract

We assess evidence for changes in tail characteristics of wind, solar irradiance and temperature variables output from CMIP6 global climate models (GCMs) due to climate forcing. We estimate global and climate zone annual maximum and annual means for period (2015, 2100) from daily output of seven GCMs for daily wind speed, maximum wind speed, solar irradiance and near-surface temperature. We calculate corresponding annualised data for individual locations within neighbourhoods of the North Atlantic and Celtic Sea region. We consider output for three climate scenarios and multiple climate ensembles. We estimate non-stationary extreme value models for annual extremes, and non-homogeneous Gaussian regressions for annual means, using Bayesian inference. We use estimated statistical models to quantify the distribution of (i) the change in 100-year return value for annual extremes, and (2) the change in annual mean, over the period (2025, 2125). To summarise results, we estimate linear mixed effects models for observed variation of (i) and (ii). Evidence for changes in the 100-year return value for annual maxima of solar irradiance and temperature is much stronger than for wind variables over time and with climate scenario.