Multivariate analysis of Kelvin wave seasonal variability in ECMWF L91 analyses

Abstract

<p><strong>Abstract.</strong> The paper presents the seasonal variability of Kelvin waves (KWs) in 2007–2013 ECMWF analyses on 91 model levels. The waves are filtered using the normal-mode function decomposition which simultaneously analyses wind and mass field based on their relationships from linear wave theory. Both spectral as well as spatiotemporal features of the KWs are examined in terms of their seasonal variability in comparison with background wind and stability. Furthermore, a differentiation is made using spectral bandpass filtering between the slow horizontal barotropic KW response and the fast vertical projection response observed as vertically-propagating KWs. <br><br> Results show a clear seasonal cycle in KW activity which is predominantly at the largest zonal scales (wavenumber 1–2) where up to 50 % more energy is observed during the solstice seasons in comparison with spring and autumn. The spatiotemporal structure of the KW reveals the slow response as a robust <q>Gill-type</q> structure with its position determined by the location of the dominant convective outflow winds throughout the seasons. Its maximum strength occurs during northern summer when easterlies in the Eastern Hemisphere are strongest. The fast response in the form of free traveling KWs occur throughout the year with seasonal variability mostly found in the wave amplitudes being dependent on background easterly winds.</p>

Waves and coherent flows in the tropical atmosphere: New opportunities, old challenges

“…The Kelvin and MRG wave contributions are much smaller when viewed in the context of the total subseasonal variability which is dominated by variability in Rossby modes (Žagar et al, 2020b), but are relevant for the formulation of simplified tropical models as in Section 2.5. Furthermore, the most energetic scales of the two special wave solutions are characterised by significant interseasonal dynamics, with the largest variance in boreal summer and winter for the Kelvin waves and MRG waves, respectively (not shown), in agreement with previous studies (Yang and Hoskins, 2017;Blaauw and Žagar, 2018).…”

“…Spatial filtering using the wave eigenstructures is a vastly different approach; in addition to the Hough harmonics decomposition on the sphere (e.g. Žagar et al, 2009;Castanheira and Marques, 2015;Blaauw and Žagar, 2018), the parabolic cylinder functions can be employed on the equatorial β −plane (e.g. Yang et al, 2003Yang et al, , 2007Tindall et al, 2006).…”

“…Filtering relies on the Fourier series expansion in longitude and time, with the dispersion relations for linear waves on the plane overlaid on the wavenumber–frequency variance diagrams to connect the observed power spectra with the Kelvin, MRG and other equatorially trapped waves (e.g., Wheeler and Kiladis, 1999; Kiladis et al ., 2009). Spatial filtering using the wave eigenstructures is a vastly different approach; in addition to the Hough harmonics decomposition on the sphere (e.g., Žagar et al ., 2009; Castanheira and Marques, 2015; Blaauw and Žagar, 2018), the parabolic cylinder functions can be employed on the equatorial plane (e.g., Yang et al ., 2003, 2007, Tindall et al ., 2006). The procedure can be multivariate, meaning that the geopotential height and winds are analyzed simultaneously (as in ()).…”

Estimating Subseasonal Variability and Trends in Global Atmosphere Using Reanalysis Data

“…The prevalence of the EIG‐related variance on large scales is most likely related to the role of the tropical Kelvin waves in subseasonal variability, whereas the subsynoptic prevalence of the WIG variance is likely due to IG waves associated with eastward‐moving baroclinic waves in the extratropics. The EIG component also shows a seasonal cycle (Figure c) which is associated with a strong seasonality in vertically propagating equatorial waves, primarily the Kelvin waves (Blaauw & Žagar, )…”

“…For the scale‐dependent analysis, the circulation is decomposed by horizontal and vertical scales using the normal‐mode function (NMF) expansion (Kasahara & Puri, ). The NMF decomposition has been extensively applied for the initialization of weather prediction models (Daley, ) but also for studies of atmospheric energetics and variability (e.g., Blaauw & Žagar, ; Castanheira et al, ; Tanaka & Kung, ; Žagar et al, ). It applies the Hough harmonics for the multivariate representation of the horizontal winds and geopotential height on terrain‐following levels after the vertical decomposition.…”