Elena Saggioro (University of Reading, UK)
The coupling between stratospheric and tropospheric dynamics is currently a topic of major interest [1,2]. In the context of the Southern Hemisphere (SH) summertime, reanalysis data seem to suggested that the stratospheric polar vortex acts as an organizing influence on the variability of the tropospheric zonal-mean circulation [3, 4, 5].
In order to test this potentially causal influence, which is currently quantified in terms of lagged cross-correlations, a novel inference technique based on Time-Series Bayesian Causal Networks [6,7,8] is adopted. Comparisons of a causal measure with the standard lagged cross-correlation analysis shows how the latter leads to highly biased results, if interpreted as a diagnostic for influence. In fact, cross-correlation is strongly affected by autocorrelation of individual time-series [9], which are particularly relevant for the stratospheric index.
The causal framework nonetheless supports the conclusion of a downward stratospheric influence during the spring-to-summer transition, which can be interpreted as a source of statistical predictability relevant to inform optimal model set up for seasonal forecasting.
References:
[1] J. Kidston et al., Nature Geoscience 8.6 (May 2015), pp. 433–440.
[2] O. P. Tripathi et al. , Quarterly Journal of the Royal Meteorological Society 141.689 (Oct. 2014), pp. 987–1003.
[3] R. X. Black et al. , Journal of the Atmospheric Sciences 64.8 (Aug. 2007), pp. 2968–2974.
[4] N. J. Byrne et al. , Journal of Climate 30.18 (Sept. 2017), pp. 7125–7139.
[5] N. J. Byrne et al. , Journal of Climate 31.9 (May 2018), pp. 3467–3483.
[6] I. Ebert-Uphoff et al. , Journal of Climate 25.17 (Sept. 2012), pp. 5648–5665.
[7] J. Runge et al, Physical Review Letters 108.25 (June 2012).
[8] J. Runge., Chaos 28.7 (July 2018), p. 075310.
[9] J. Runge et al., Journal of Climate 27.2 (Jan. 2014), pp. 720–739.
Invited by Sebastian Reich
