This WP addresses RA2 and contributes to answering HTE RQs 1, 2 (‘What are the key model improvements needed to improve the reliability of extreme event prediction on climate timescales?’) and 3. Although highly valuable, the modern (post-1979) observational period considered in WP1 does not sample the full range of plausible/potential extreme events, which may lead to an underestimation of current and future risk. If the real-world climate system had evolved differently from observed, it could have exhibited different phases of multidecadal variability and different sampling of weather events.

WPs 2 and 3 will exploit a new generation of initial condition LE climate simulations, which represent different realisations of the modern observational era with identical external forcings and therefore provide more comprehensive estimates of plausible historical (post-1850) extreme events and their local and remote drivers. To quantify impacts on ice shelves, a physically sophisticated melt lake model will be added to the JULES (Joint UK Land Environment Simulator) ice surface module (WP2.4).

A statistical RCM emulator will be used to downscale LE climate simulations to produce output at local impact-relevant scales to drive the melt lake model (WP2.3). In terms of addressing HTE RQ2, we will identify required model improvements as follows:

• (1) by evaluating biases in the LE models and the attribution models from WP1.2 and identifying aspects contributing to biases in simulated extremes; and

• (2) with regard to modelling impacts, the planned development of a melt lake model will involve calibration and assessment against observations, which will help to inform required model improvements.

Tasks in WP2

WP2.1 Simulation of large-scale drivers and Antarctic extreme events in GCMs

• Specific objective 2.1: Evaluate performance of the global climate models (GCMs) used for the LE datasets.

WP2.2 Causes of multi-decadal trends and variability in Antarctic extreme events

• Specific objective 2.2: Assess causes of historical trends and variability in LE datasets.

WP2.3 Downscaling of LE to examine historical extreme meteorological impacts

• Specific Objective 2.3: Build statistical emulators for the high-resolution RCM hindcasts to produce downscaled output from the LE subset to examine historical local-scale extreme impacts.

WP2.4 Impacts of extreme events on ice shelf stability in present-day climate

• Specific Objective 2.4: Bring a surface energy balance and thermodynamic model with explicit treatment of densification and melt lakes into JULES. Explore implications of extreme atmospheric events for ice shelves (observations and modelling).