Figure 11. Median time of emergence (TOE) for TXx and TNn based on SNR thresholds under SSP5-8.5 and RCP8.5. (a) TOE for TXx under SSP5-8.5 when SNR > 1; (b) TOE for TXx under SSP5-8.5 when SNR > 2; (c, d) same as (a) and (b), but for TNn; (e-h) same as (a-d), but for RCP8.5.
Over some central and tropical parts of Australia, the multi-model median TOE in TXx for SNR > 1 can occur as early as the second decade of this century (2010-2020). Generally, the signal emerges earlier over northwestern region than the southeast for both thresholds (Fig. 11a, b, e and f), in which the signal emerges in 2020s for SNR > 1 and 2040s for SNR > 2, as there indicate relative smaller noise and larger signal (Figs. S17 and S21). Over the southeast regions, the TOE occurs within 2030-2050 for SNR > 1. In contrast, for TNn, the signal emerges from the noise in 2020s over Australia (SNR > 1; Fig. 11c and g); while for SNR > 2, the TOE is within the fifth decade (2040-2050) over western and central regions (Fig. 11d and h). Compared to RCP8.5, the multi-model medians of TOE for TXx and TNn in CMIP6 show earlier TOE over more regions based on the threshold SNR > 2, implying the larger median SNR in the middle of this century as shown in Figs. 8 and 9. However, the uncertainty surrounding these TOE estimates remains large (Figs. 8 and 9). For example, for SNR = 2, the range (inter-model spread) of TOE for TXx over AUS can be from 2010s to 2060s (Fig. 8a). For lower scenarios, the multi-model medians of TOE commonly occur later and over smaller regions than the stronger pathways. For example, TOE (SNR > 1) for TNn under SSP2-4.5&RCP4.5 usually can be 10 years later over some southeast regions (Fig. S28g and k) than that shown in SSP5-8.5&RCP4.5 (Fig. 11c and g), as the signal is lower compared to that in higher pathways.
The analysis on SNR and TOE has useful implications for Australia. Under the highest-emission scenarios SSP5-8.5&RCP8.5, in which the medians of the signal for TXx are comparable, the early emergence over northwest Australia suggests that there is less time for stakeholders and policy makers to implement effective measures, compared to southeast Australia. In contrast, if under lower scenarios, the TOE for TXx can be postponed, especially for southeast regions which exhibit larger variability for TXx in the extratropics. However, the adaptation policy may change for different extremes, even under same future pathways. For TNn, the TOE (SNR > 1) can occur over most regions even under lower-emission scenarios; while the “unfamiliar” climate (SNR > 2) can be largely postponed if taking a more sustainable pathway (lower emission). It is also noted that the large uncertainty in the estimate of SNR and TOE highlights further challenges for stakeholders and policy makers.
3.3 Large Ensembles in CMIP6
Previous research has demonstrated the model uncertainty in estimating the effects of internal variability on the TXx and TNn trends, shown in LEs during 1950-2014 over Australian regions (Deng et al., 2021). Therefore, how internal variability influences the projected changes and TOE/SNR (including signal and the noise) needs further investigation. In Fig. 12, which represents the boxplots of projected changes in TXx and TNn for CanESM5-LE and MIROC6-LE over Australian regions under SSP5-8.5 and SSP1-2.6, model uncertainty for representing internal variability still exists, and the relative magnitudes of the spreads for projected changes resemble the results in Figs. 12 and 13 in Deng et al. (2021). The projected changes in TXx for MIROC6-LE span larger ranges than CanESM5-LE by a factor of ~3 or more over the regions, which can be larger than that in Fig. 3a. Moreover, there exhibit larger ranges of the projected changes for TXx over SEA, MEA, and SSA, and relatively narrower spreads over TA for CanESM5-LE and SWA for MIROC6-LE. For TNn, the relative magnitude for the two LEs are comparable over the regions. The different effects of internal variability for different LEs and regions complicate the assessment of the uncertainty on projected changes.