The combined influence of the stratospheric polar vortex and ENSO on zonal asymmetries in the southern hemisphere upper tropospheric circulation during austral spring and summer

Abstract

The influence of El Niño Southern Oscillation (ENSO) and the Stratospheric Polar Vortex (SPV) on the zonal asymmetries in the Southern Hemisphere atmospheric circulation during spring and summer is examined. The main objective of the work is to explore if the SPV can modulate the ENSO teleconnections in the extratropics. We use a large ensemble of seasonal hindcasts from the European Centre for Medium-Range Weather Forecasts Integrated Forecast System to provide a much larger sample size than is possible from the observations alone. We find a small but statistically significant relationship between ENSO and the SPV, with El Niño events occurring with weak SPV and La Niña events occurring with strong SPV more often than expected by chance, in agreement with previous works. We show that the zonally asymmetric response to ENSO and SPV can be mainly explained by a linear combination of the response to both forcings, and that they can combine constructively or destructively. However the nature of this interference evolves through the spring and summer period, and is not aligned with the traditional seasons. From this perspective, we find that the tropospheric asymmetries in response to ENSO are more intense when El Niño events occur with weak SPV and La Niña events occur with strong SPV, at least from September through December. In the stratosphere, the ENSO teleconnections are mostly confounded by the SPV signal. The analysis of Rossby Wave Source and of wave activity shows that both are stronger when El Niño events occur together with weak SPV, and when La Niña events occur together with strong SPV.

Appendix

1.1 A Wave activity flux decomposition

Wave activity fluxes are used to investigate the propagation of waves. We use the horizontal component of the wave activity flux ${\varvec{W}}$ derived by Takaya and Nakamura (2001) which describes the propagation of wave disturbances on a zonally varying basic flow:

$$\begin{aligned} {\varvec{W}} = \frac{1}{2\left| {\overline{U}} \right| } \begin{bmatrix} {\overline{u}}(\psi ^{*2}_{x}-\psi ^{*}\psi ^{*}_{xx})+{\overline{v}}(\psi ^{*}_{x}\psi ^{*}_{y}-\psi ^{*}\psi ^{*}_{xy})\\ {\overline{u}}(\psi ^{*}_{x}\psi ^{*}_{y}-\psi ^{*}\psi ^{*}_{xy})+{\overline{v}}(\psi ^{*2}_{y}-\psi ^{*}\psi ^{*}_{yy}) \end{bmatrix} \end{aligned}$$

where ${\overline{U}}=({\overline{u}},{\overline{v}})$ is the basic wind field and $\psi ^{*}$ is the streamfunction obtained from the zonal asymmetries $Z^{*}$. We decompose the ensemble mean of ${\varvec{W}}$ (EM) into its linear term ($EM_{LIN}$) and non linear term ($EM_{NL}$). The decomposition follows Fletcher and Kushner (2011) and is illustrated here for the first term in the x-direction, but similar arguments apply to the rest of the terms. For each realization in the ensemble, we have

$$\begin{aligned} \psi ^{*}= & {} \left\langle \psi ^{*} \right\rangle + \psi ^{*'}, \quad \psi ^{*}_{x} = \left\langle \psi ^{*}_{x} \right\rangle + \psi ^{*'}_{x},\\ \psi ^{*}_{xx} = & {} \left\langle \psi ^{*}_{xx} \right\rangle + \psi ^{*'}_{xx} \end{aligned}$$

where the angle brackets denote an ensemble mean and the prime a departure from the ensemble mean. The mean response, $\varDelta \left\{ \ldots \right\}$, for the first term in the x-direction can be decomposed as

$$ \begin{aligned} \varDelta \left\{ \psi ^{*2}_{x} - \psi ^{*}\psi ^{*}_{xx} \right\}&=\varDelta \left\{ \left\langle \psi ^{*}_{x} \right\rangle ^2 - \left\langle \psi ^{*}\right\rangle \left\langle \psi ^{*}_{xx}\right\rangle \right\} + \varDelta \left\{ \left\langle \psi ^{*'2}_{x} \right\rangle \right. \left. - \left\langle \psi ^{*'} \psi ^{*'}_{xx} \right\rangle \right\} . \end{aligned}$$

The first term of the right-hand-side of the equation is the response associated with the ensemble mean eddy response (EM), while the second term is the response associated with the departures from the ensemble mean (FL). The EM term can be further decomposed if we separate the ensemble mean as follows:

$$\begin{aligned} \left\langle \psi ^{*}\right\rangle = \psi ^{*}_c + \varDelta \left\langle \psi ^{*} \right\rangle , \quad \left\langle \psi ^{*}_{x} \right\rangle = \psi ^{*}_{xc} + \varDelta \left\langle \psi ^{*}_{x} \right\rangle, \quad \left\langle \psi ^{*}_{xx} \right\rangle = \psi ^{*}_{xxc} + \varDelta \left\langle \psi ^{*}_{xx} \right\rangle \end{aligned}$$

where the subscript c refers to the climatology. Then,

$$\begin{aligned} EM = EM_{LIN} + EM_{NL} \end{aligned}$$

where $EM_{LIN}$ and $EM_{NL}$ in the mentioned example turn out to be

$$\begin{aligned} EM_{LIN} = \left\{ 2 \psi ^{*}_{xc} \varDelta \left\langle \psi ^{*}_{x} \right\rangle - \psi ^{*}_c \varDelta \left\langle \psi {*}_{xx} \right\rangle - \psi ^{*}_{xxc} \varDelta \left\langle \psi {*} \right\rangle \right\} \end{aligned}$$

and

$$\begin{aligned} EM_{NL} = \left\{ ( \varDelta \left\langle \psi ^{*}_{x} \right\rangle )^2 - \varDelta \left\langle \psi ^{*}\right\rangle \varDelta \left\langle \psi ^{*}_{xx}\right\rangle \right\} . \end{aligned}$$

The $EM_{LIN}$ term represents the linear interference effect, which involves the phase difference between the wave response and the climatological wave. The $EM_{NL}$ term reflects the wave activity change intrinsic to the wave response itself.