Evaluation of seasonal teleconnections to remote drivers of Australian rainfall in CMIP5 and CMIP6 models

Abstract

This study describes how coupled climate models participating in the sixth phase of the Coupled Model Intercomparison Project (CMIP6) simulate the primary climate drivers that affect Australian climate, and their seasonal relationship to Australian rainfall, namely the El Niño–Southern Oscillation (ENSO), the Indian Ocean Dipole (IOD) and the Southern Annular Mode (SAM). As results from the earlier generation of models (CMIP5) are still in use, the CMIP6 multi-model mean teleconnections between climate drivers and seasonal Australian rainfall are compared to CMIP5. Collectively, an improvement is found in CMIP6 relative to CMIP5 in the representation of the relationship between ENSO and IOD events and Australia’s springtime rainfall. Overall, CMIP6 models are also able to reproduce the asymmetric relationship between ENSO and eastern Australian rainfall, which exhibits a more robust signal during La Niña than during El Niño years. Both CMIP5 and CMIP6 models are also generally able to capture the stronger relationship between Central Pacific La Niñas, compared to Eastern Pacific La Niñas. However, the large spread in model-to-model behaviour, and among ensemble members, remains a source of uncertainty. Although CMIP6 models have improved in their representation of SAM variability, the simulated relationship between SAM and Australian rainfall has not materially improved. Additionally, this study is accompanied by an extensive Appendix in which each model’s ENSO, IOD and SAM seasonal teleconnection patterns to rainfall are presented and ranked.

A Statistical Forecasting Model for Extremes of the Fire Behaviour Index in Australia

“…For instance, where climate drivers may have a less pronounced or less clear influence on fire danger [20], a linear model may prove less effective. This also holds for regions and fire seasons where multicollinearity and teleconnections between drivers are different or have different strengths [83,112,113]. In these cases, there may be benefits in considering other non-linear forms of forecasting, such as more developed machine-learning models, or the AI-based methods alluded to above.…”

Bayesian Structure Learning for Climate Model Evaluation

“…Having said that, the typical approach by which model biases and performance is assessed is via systematic comparison of individual metrics, usually focused on the mean climate and the large scale climate modes based on SST differences between CMIP model and reanalysis products such as ENSO, IOD, interdecadal Pacific oscillation (IPO) and the Atlantic multidecadal oscillation (AMO) (Rashid et al., 2013; Stoner et al., 2009). Alternate common approaches include estimating the influence of temporal biases in given climate modes on specific variables, for example, precipitation and temperature (Chung et al., 2023). Due to the maturity of the data and range of available intercomparisons, we have chosen to focus on a subset of CMIP5 models however, in common with CMIP3 (Stoner et al., 2009) and CMIP5 (Rashid et al., 2013), the most recent phase 6 of CMIP (Rashid et al., 2022) reveals that, whereas the spatial structures of the large scale oceanic climate modes (ENSO, IOD, IPO, and AMO) compare favorably with the structures of their observed counterparts, there remain major and systematic differences in the simulated temporal variability.…”

Bayesian structure learning for climate model evaluation

“…(Stoner et al, 2009;Rashid et al, 2013). Alternate common approaches include estimating the influence of temporal biases in given climate modes on specific variables e.e., precipitation and temperature (Chung et al, 2023). Due to the maturity of the data and range of available intercomparisons, we have chosen to focus on a subset of CMIP5 models however, in common with CMIP3 (Stoner et al, 2009) and CMIP5 (Rashid et al, 2013), the most recent phase 6 of CMIP (Rashid et al, 2022) reveals that, whereas the spatial structures of the large scale oceanic climate modes (ENSO, IOD, IPO and AMO) compare favourably with the structures of their observed counterparts, there remain major and systematic differences in the simulated temporal variability.…”