The Madden–Julian oscillation strengthens its reach

Abstract

Predicting weather beyond about two weeks eludes the most adept forecasters, but certain climate system phenomena can give them an edge. An example is the Madden–Julian oscillation (MJO), which begins as a pocket of deep convection in the equatorial Indian Ocean and crawls east along the equator, circling the globe. As it travels, the MJO has a ripple effect much like El Niño, impacting the extratropical circulation and weather in regions far away. These impacts are called teleconnections, and Northern Hemisphere hotspots include the North Pacific and North Atlantic, affecting North American and European climate downstream. Understanding these teleconnections can boost regional prediction skill into the subseasonal-to-seasonal (S2S) timescale of weeks to months. Assessing how they may evolve with climate warming is similarly valuable, and this is the question that Savini Samarasinghe and colleagues address in an article in Geophysical Research Letters. With a focus on the North Atlantic region, the author team found MJO teleconnections strengthened, and attributed this to a changing tropospheric pathway. This grew out of a historical MJO teleconnection study (J. Geophys. Res. Atmos. 124, 9356–9371; 2019). Causality techniques showed that the MJO impacts the North Atlantic Oscillation (NAO) — a regional mode of North Atlantic climate variability — via two pathways. In a tropospheric pathway, the MJO drives large-scale atmospheric disturbances, called Rossby waves, that travel from the Pacific Ocean to directly modulate North Atlantic climate. An alternative stratospheric pathway occurs in two parts: MJO-driven Rossby waves first disrupt the stratospheric vortex encircling the polar region, and then atmospheric waves travel back down to affect the North Atlantic. “After that paper, we started thinking about MJO teleconnections under climate warming”, explains Elizabeth Barnes, a co-author and principal investigator of the current study, and lead author of the previous work. “We wanted to see what the causality methods said about this.” The researchers addressed this question using historical and future climate simulations during 1850–2094 from the Community Earth System Model version 2 (CESM2) Whole Atmosphere Community Climate Model (WACCM). CESM2-WACCM’s atmosphere reaches up to about 130 km, higher than the default version at around 40 km, making it particularly equipped to simulate stratosphere-dependent dynamics. The authors used the model output to produce indices of MJO, NAO and stratospheric vortex strength over time. To understand whether MJO teleconnection strength changes with climate warming, the researchers used a statistical approach to distil MJO’s remote impacts across its lifecycle, finding a robust increase in teleconnection strength (Fig. 1). Next, they used graphical causal models to determine whether the tropospheric or stratospheric pathway was responsible. This technique measures the strength and causality of pairwise connections and was applied to the MJO, NAO and stratospheric vortex indices, revealing that the stronger teleconnections were a result of changes to the tropospheric pathway alone. These findings challenged the researchers’ expectations. Recent work suggests that with warming, the tropical atmosphere becomes more stable and propagates MJO signals less efficiently, implying the MJO’s extratropical impacts could weaken (see J. Adv. Model. Earth Sy. 9, 307–331; 2017 and Nat. Clim. Change 9, 26–33; 2019). “The fact that we saw the North Atlantic teleconnections strengthen — or at least happen more often/ robustly — was a surprise”, notes Barnes. They hypothesized this was a result of changes in the mid-latitude circulation rather than in the MJO itself. Another study found MJO teleconnections along the US west coast may be amplified with warming, due to a speedup and eastward extension of the Pacific jet stream winds that steer mid-latitude storms (Nat. Clim. Change 10, 654–660; 2020). “This jet-stream extension may act as an even better waveguide to the North Atlantic”, Barnes suggests. This study has important takeaways. First, the methods can be applied widely. “I’m excited about graphical causal models as a method to separate teleconnection pathways”, says Barnes. “They’re powerful and alleviate at least some of the issues we face with separating pathways and timescales from data alone, when it’s not possible to run carefully designed model experiments to test causality. We’re not the only group doing this, but I think more climate scientists could benefit from this approach.” Second, if MJO teleconnections grow stronger with climate change then S2S predictability may benefit, but this needs to be tested. “While our analysis suggests a strengthening of the MJO–North Atlantic teleconnection, we have not quantified what this means for actual predictability — so to me, that is the next big step.” Another highlight of the work is the composition of the research team. “The first four authors are women across a range of career stages”, Barnes explains. “This field of data science and climate could have more gender diversity at the moment, so this was a particular point of pride for me.”