Evaluating Tropical Cyclone Representation in ERA5‐Driven NARCliM2.0 WRF Simulations Over CORDEX‐Australasia Using Multiple Trackers

Abstract

We evaluate tropical cyclone (TC) representation in CORDEX ERA5‐driven simulations from the New South Wales and Australian Regional Climate Modelling Project (NARCliM2.0) over the Australian region at ∼20 km resolution for the period 1980–2020. The simulations are performed using the Weather Research and Forecasting (WRF) Model with two configurations that differ only in their planetary boundary layer (PBL) physics: R3 (MYNN2.5) and R5 (ACM2). TCs are identified using two independent tracking algorithms to assess the robustness of the results across detection methods (a) local minimum sea level pressure combined with a warm‐core criterion (UZ; physics‐based), and (b) Okubo‐Weiss‐Zeta parameter (OWZ; dynamics‐based). Both trackers were applied to ERA5 reanalysis and ERA5‐driven simulations, with simulated TC characteristics evaluated against Bureau of Meteorology observations. Using both TC trackers, the ERA5‐driven simulations reasonably reproduce the broad observed spatial patterns of TC genesis and track density, including their modulation by the El Niño–Southern Oscillation. The regional differences are largely driven by PBL physics rather than by the choice of trackers. R3 outperforms R5 over northwestern Australia in both trackers, where moisture deficits and less favorable dynamical conditions in R5 suppress TC genesis during both El Niño and La Niña. Although both configurations result in too many Category‐1 TCs over the Solomon Sea region, R3 reduces this bias, particularly with the UZ tracker. Both configurations underestimate TC intensity, similar to ERA5 reanalysis. Overall, the ERA5‐driven simulations provide a reasonable representation of Australian TC climatology, supporting their potential use as a benchmark for CMIP6‐driven regional TC projections.