Global modulation of ENSO teleconnections by Pacific Decadal Variability

The El Niño Southern Oscillation (ENSO) is the leading mode of global climate variability on interannual timescales. On longer time-scales, decadal variability in the Pacific is responsible for modulating the rate of global warming (Meehl et al 2013, Maher et al 2014 Henley & King 2017). Whether Pacific Decadal Variability (PDV) modulates ENSO teleconnections is an important research question that has largely been investigated using the observational record. PDV is shown to modulate Australian rainfall (Power et al 1999, Arblaster et al 2002, Verdon et al, 2004, King et al 2013), which has impacts for flood frequency (Franks and Kuczera, 2002, Kiem et al, 2003, Pui et al 2011). PDV has also been shown to modulate ENSO precipitation teleconnections over Africa (Dong & Dai 2015), Texas (Khedun et al 2014), and Europe (Zanchettin et al 2008) as well as ENSO temperature teleconnections over New Zealand (Salinger et al 2001). While these observationally based studies suggest connections between ENSO teleconnections and PDV, the short observational record contains only two PDV phase changes. In addition, calculating PDV using a lowpass filter on the region that contains ENSO could also cause statistical artefacts in the results (Power at al, 2006, Westra et al 2015). These limitations can be addressed using climate models. Arblaster et al (2002) use atmosphere only simulations and find similar results to observational studies over Australia. Dong and Dai (2015) further investigate global modulation using 4 ensemble members of a single model. These modelling studies are limited in their use of single models and while they include a larger dataset than the observational record, previous work has only used small ensemble sizes. In this study, we address both the issue of small datasets and the dependence on results on the model used by utilising four single model initial-condition large ensembles.  Each model ensemble has a minimum of 20 members enabling investigation of multiple realizations of PDV and ENSO covariability. Over the historical period, using one ensemble member results in a record that is indeed too short to accurately quantify the influence of PDV on ENSO teleconnections. We then composite events for different phases of the PDV and ENSO using all ensemble members. Initial results show that PDV strongly influences ENSO temperature teleconnections over North America.  We find that stronger teleconnections occur when an El Niño occurs during a positive phase of PDV or a La Niña occurs in a negative phase of the PDV. Similarly, PDV phase affects precipitation over Australia, where co-occurring El Niño and positive PDV phases and La Niña and negative PDV phases have larger precipitation anomalies. Finally we investigate whether this modulation of ENSO teleconnections by the PDV is projected to change under strong anthropogenic forcing. We find greater inter-model agreement for precipitation teleconnections than for temperature teleconnections.  Ongoing work will assess the underlying physical mechanisms behind these results.