Global warming has led to an increase in La Nia events
Inter-model consensus on the change of triple La Nia under high-emission warming: Evidence from the pre-industrial control and inter-model democracy approaches
Further, our analysis indicates that there is no inter-model consensus on the change of triple La Niña events from 1900–1999 to 2000–2099, even under a high-emission warming scenario SSP585. Only a total of 8 out of 20 (40%) models show an increase in triple La Niña events with an MME increase of 12.2 ± 28.1%. Triple La Niña is relatively rare in the historical record and its mechanism is still uncertain. A preliminary analysis indicates that there is no systematic change in the meridional structure of the second-year La Niña between 1900–1999 and 2000–2099, which may provide a possible explanation for why there is no inter-model consensus on the change in triple La Niña frequency under greenhouse warming.
Further, our sensitivity test indicates that, even if data quality was not an issue, detecting observed change may depend on the time period in which the frequency of multi-year La Niña is diagnosed. For example, when comparing 1945–1979 with 1980–2014, there is no change in the frequency of multi-year La Niña events in observations (Extended Data Fig. There is still a strengthened east-west SST in the Pacific. The La Nia change is subject to uncertainty and may be influenced by natural variability.
As in previous studies37,50,51, we use one experiment only from each model (the ‘model democracy’ approach) to avoid dominance by models with which many experiments are carried out such that each model is represented equally in the assessment of inter-model consensus and the ensemble mean change. Projected changes in ENSO might be subject to internal variability46,60. The longer the time window is used to diagnose ENSO variability change, the lower the noise level of unforced natural variability is compared with a warming-induced ENSO SST variability change signal42. The influence from internal variability can be mitigated using two 100-year periods, but details of the time-varying features within each period may not be captured. We use a multi-century pre-industrial control (piControl) simulation of each model to examine the influence of internal variability in the ENSO cycle and the results show distinct increase in the frequency of multi-year La Niña emerging from background natural variability in the twenty-first century (Fig. 2b). We also use all available models under different emission scenarios (SSP126, SSP245 and SSP370; Supplementary Table 1) to test the sensitivity of our results.
in which LH denotes latent heat flux, u is the 10-m zonal wind and W is the total wind speed. The mean state has a big impact on the intensity of WE Sp. A more intense WES feedback can come from a stronger mean zonal wind and a warmer background.
