Spectral signatures of the tropical Pacific dynamics from model and altimetry: a focus on the meso-/submesoscale range

Abstract

Abstract. The processes that contribute to the flat sea surface height (SSH) wavenumber spectral slopes observed in the tropics by satellite altimetry are examined in the tropical Pacific. The tropical dynamics are first investigated with a 1∕12∘ global model. The equatorial region from 10∘ N to 10∘ S is dominated by tropical instability waves with a peak of energy at 1000 km wavelength, strong anisotropy, and a cascade of energy from 600 km down to smaller scales. The off-equatorial regions from 10 to 20∘ latitude are characterized by a narrower mesoscale range, typical of midlatitudes. In the tropics, the spectral taper window and segment lengths need to be adjusted to include these larger energetic scales. The equatorial and off-equatorial regions of the 1∕12∘ model have surface kinetic energy spectra consistent with quasi-geostrophic turbulence. The balanced component of the dynamics slightly flattens the EKE spectra, but modeled SSH wavenumber spectra maintain a steep slope that does not match the observed altimetric spectra. A second analysis is based on 1∕36∘ high-frequency regional simulations in the western tropical Pacific, with and without explicit tides, where we find a strong signature of internal waves and internal tides that act to increase the smaller-scale SSH spectral energy power and flatten the SSH wavenumber spectra, in agreement with the altimetric spectra. The coherent M2 baroclinic tide is the dominant signal at ∼140 km wavelength. At short scales, wavenumber SSH spectra are dominated by incoherent internal tides and internal waves which extend up to 200 km in wavelength. These incoherent internal waves impact space scales observed by today's along-track altimetric SSH, and also on the future Surface Water Ocean Topography (SWOT) mission 2-D swath observations, raising the question of altimetric observability of the shorter mesoscale structures in the tropics.

Revised Global Wave Number Spectra From Recent Altimeter Observations

“…So, even though the variable wavelengths resolved in this region are above the local eddy length scale and the local noise level (cf. Figure ), we are still fitting a curve to a statistical combination of the signals of balanced and unbalanced motions, since the latter can extend out to wavelengths greater than 200 km in the tropics (Tchilibou et al, ). In contrast, as we approach regions with more energetic mesoscale structures (high latitudes and western boundary current systems), we observe spectral slopes closer to QG predictions ( k ‐4 to k ‐5 ).…”

“…The global distribution of SSH spectral slope has revealed a strong geographical dependence (Dufau et al, 2016;Xu & Fu, 2012;Zhou et al, 2015), with low values (k -1.5k -2 ) in the intertropical band (20°S-20°N) and in low-energy eastern boundary regions, and progressively steeper slopes toward higher latitudes, reaching k -11/3 in the highly energetic regions. These low SSH spectral slopes have been attributed to the presence of internal gravity waves (IGWs) from atmospheric or tidal forcing that can dominate the SSH in regions of low eddy energy, as shown in simulations (Richman et al, 2012;Tchilibou et al, 2018) and in altimetric observations (Dufau et al, 2016;. Indeed, the unresolved internal tides lead to peaks in spectral energy around 100-200 km wavelength band, depending on the region, that can even lead to positive spectral slope estimates in low eddy energy regions (Dufau et al, 2016).…”

“…The along-track measurements are subsampled inside a 15°× 15°box. Segments of constant length (1,500 km) are selected, to ensure adequate coverage in the tropics following Tchilibou et al (2018). These segment lengths are longer than those used by Dufau et al (2016) of 560 km and Xu and Fu (2012) of 1,000 km.…”

“…This may be well adapted to the midlatitudes but poorly frames the steepest slopes in the tropics and high latitudes (Dufau et al, 2016). Spectral processing techniques also play a role in modifying the spectral slope, and recently, Tchilibou et al (2018) highlighted that short segment lengths and certain windowing choices for the spectral estimates can contribute to reducing the SSH wave number spectral slopes in the tropics.…”

“…(2) We will also use a geographically variable "mesoscale" wavelength range to compute the spectral slopes, taking into account the geographical dependence of the characteristic eddy length scales based on their local Rossby radius, following Eden (2007). (3) We choose spectral preprocessing techniques that minimize the spectral slope modification, with particular care taken in the tropics, following from the work of Tchilibou et al (2018). (4) We will test the impact of the phaselocked internal tide on our estimates of SSH wave number spectral slopes by applying a recently developed phase-locked internal tide correction (Zaron & Ray, 2017) that reduces the main internal tide SSH power peaks.…”