Understanding changes in the tropical Pacific atmospheric circulation with global warming

The weakening of the Walker circulation in the tropical Pacific is a robust response to global warming in climate models. This can have a global impact on climate, because the convection in the ascending branch of the Walker circulation triggers planetary scale waves that radiate to higher latitudes. In a recent article in the Journal of Atmospheric Sciences (Wills et al. 2017), we study the physical mechanisms responsible for the Walker circulation weakening in an idealized model. Here, we discuss how this work applies to the real-world climate system and how Walker circulation changes are related to tropical Pacific sea-surface temperature changes.

The Walker circulation in the modern climate consists of ascending air in the warm pool, around Indonesia, and descending air in the eastern equatorial Pacific, off the coast of Peru. The westward surface winds that complete this circulation sustain an east-west gradient in sea-surface temperature (SST) across the Pacific, a factor that helps maintain the Walker circulation. In this way, variability of the Walker Circulation is coupled to variability of the underlying ocean, in what is called the El Niño-Southern Oscillation (ENSO).

The weakening of the Walker circulation with warming is associated with an El-Niño-like change in Pacific-ocean temperatures. However, La-Niña-like trends could occur in the near term as a result of ocean changes, either due to internal variability, or if the climate is changing too fast for the Pacific ocean circulation to stay in equilibrium (Kohyama et al. 2017).

Even in the simpler case where the ocean circulation is not allowed to change (an artificial construct only realizable in climate models), the precise mechanisms for the Walker circulation weakening remain unclear. The weakening of the Walker circulation with global warming has previously been explained as part of the global reduction in convective circulations in response to the moistening of the atmosphere with warming, a consequence of global energetic constraints that prevent global precipitation from increasing as fast as atmospheric moisture content (Betts 1998, Held and Soden 2006, Vecchi and Soden 2007, Schneider et al. 2010). However, this mechanism does not constrain the Walker circulation strength in particular, because there can be compensating changes in other convective circulations.

In Wills et al. 2017, we study the response of the Walker circulation to global warming in an idealized general circulation model (GCM), where we can simulate a wide range of mean climates, from 10°C colder than modern to more than 25°C warmer than modern. We find that the Walker circulation weakens consistently with warming over the full range of climates. Compared to comprehensive climate models, our model has simplified convection and radiation schemes and lacks ocean-circulation variability, the diurnal cycle, the seasonal cycle, the water-vapor feedback, clouds, and continents. By making these large simplifications, we can more easily analyze the dominant physical mechanisms setting the Walker circulation strength and use these insights to understand more complex models.

Figure 1: Schematic of Walker circulation changes with warming. In a warmer climate, deep convection reaches higher into the atmosphere, increasing the potential energy needed for each convecting parcel. This leads to a reduction in circulation strength (arrows) when energy input from the ocean is fixed. Increases in lower-tropospheric temperature and moisture content (blue shading) decrease the energy required per unit mass of deep convection, moderating the convection depth effect. Positive (negative) quantities are denoted by a + (-). Increasing (decreasing) quantities are denoted by a ↑ (↓). From Wills et al. 2017.

We develop a theory that determines the Walker circulation strength in terms of the local net energy input from the ocean and top-of-atmosphere radiative fluxes. We use the moist static energy (MSE) budget to account for the influence of latent heating on the energy balance within atmospheric columns. Our theory accounts for the weakening of the Walker circulation with warming, showing that it is a consequence of an increase in the gross moist stability (GMS) felt by tropical circulations. The GMS is a measure of the tropical stratification, which is closely related to the MSE difference Δh between the boundary layer and the tropopause. It measures the energy required per unit mass for deep convection. There are contributions to Δh from the temperature, moisture, and geopotential differences between the boundary layer and the tropopause (depicted schematically in Fig. 1). The largest term is the geopotential difference ΔΦ, which increases with climate change as the tropopause rises, increasing the potential energy required for a convecting parcel to reach its level of neutral buoyancy near the tropopause. This effect wins out over competing influences from the increasing boundary-layer temperature and moisture, which would both lead to a reduction in GMS with warming. The net result is an increase in GMS with warming, which results in a decrease in Walker circulation strength in the absence of large changes in the column energy budget.

In the real world, there could be changes in the column energy budget due to changes in ocean circulation, changes in cloud cover, or changes in atmospheric moisture content. However, these energy budget changes would have to be substantial and systematic in order to overcome the increased GMS and lead to a Walker circulation strengthening with warming. Further work, using comprehensive climate models, is needed to understand how these processes influence the Walker circulation changes predicted for the next century.

In general, GMS is an empirical quantity that can be useful for diagnosing the mechanisms of changes in tropical circulations. However, convective quasi-equilibrium theory can help to quantitatively constrain GMS by determining the vertical structure of perturbations in the tropical atmosphere, based on the assumption that temperature is determined by a saturated moist adiabat above the boundary layer (Arakawa and Schubert 1974; Emanuel et al. 1994). Our work shows that quasi-equilibrium theory accounts for GMS changes in the idealized GCM. In this case, MSE differences between the boundary layer and the tropopause can result only when atmospheric relative humidity is less than 100%, such that GMS is related to the saturation deficit of the atmosphere.

Changes in the Walker circulation are often associated with changes in the east-west SST difference across the Pacific. However, our work shows that there is a clearer mechanistic link between Walker circulation strength and the east-west difference in net energy input (i.e., east-west differences in ocean heat uptake when east-west differences in radiation are small). Regardless, the Walker circulation strength and east-west temperature difference have similar fractional changes across the range of climates simulated by the idealized GCM (Fig. 2). What causes Walker circulation strength and east-west SST difference to change together under global warming? Merlis and Schneider (2011) show that the east-west SST difference decreases with warming when the ocean energy input is fixed because evaporation becomes more efficient at reducing the temperature difference as the saturation deficit of the boundary layer increases with warming. When net energy input from the ocean is fixed, the reductions in Walker circulation strength and east-west temperature difference with warming both result from a common cause: the increase in atmospheric saturation deficit.

Figure 2: Walker circulation strength (y-axis) and east-west tropical SST difference (x-axis) across the range of climates simulated by our idealized GCM. From Wills et al. 2017.

In summary, we have found that the Walker circulation strength decreases with warming, because the increase in tropopause height increases the energy required for convection. This is true as long as the net energy input in the Walker circulation ascent/descent regions does not change very much. Cloud and ocean circulation changes in particular could be important for modifying the net energy input in the real world and should be investigated further. We have shown that the Walker circulation strength can be quantified in the idealized GCM by using convective quasi-equilibrium theory to calculate the gross moist stability of tropical circulations. Further work is needed on the extent to which this theory applies to models with more realistic convection schemes.