Report

Thermal Structure and Dynamics of Saturn’s Northern Springtime Disturbance

See allHide authors and affiliations

Science  17 Jun 2011:
Vol. 332, Issue 6036, pp. 1413-1417
DOI: 10.1126/science.1204774

Abstract

Saturn’s slow seasonal evolution was disrupted in 2010–2011 by the eruption of a bright storm in its northern spring hemisphere. Thermal infrared spectroscopy showed that within a month, the resulting planetary-scale disturbance had generated intense perturbations of atmospheric temperatures, winds, and composition between 20° and 50°N over an entire hemisphere (140,000 kilometers). The tropospheric storm cell produced effects that penetrated hundreds of kilometers into Saturn’s stratosphere (to the 1-millibar region). Stratospheric subsidence at the edges of the disturbance produced “beacons” of infrared emission and longitudinal temperature contrasts of 16 kelvin. The disturbance substantially altered atmospheric circulation, transporting material vertically over great distances, modifying stratospheric zonal jets, exciting wave activity and turbulence, and generating a new cold anticyclonic oval in the center of the disturbance at 41°N.

In contrast to Jupiter’s frequent large-scale storms and vortices (1), Saturn typically appears less active, with short-lived white spots occurring infrequently at mid-latitudes (25). Saturn’s temperatures, winds, and circulation patterns above the visible cloud decks evolve slowly in response to the changing levels of sunlight (6), and eruptions of planetary-scale disturbances are extremely rare (7, 8), even though convective processes and eddy mixing are thought to play a key role in transporting Saturn’s internal heat and maintaining the zonal jets (2). This slow evolution was disrupted on 5 December 2010 by a convective plume of bright cloud material near 40°N (planetographic latitude) in Saturn’s northern springtime hemisphere, which created an expanding system of white cloud material that was redistributed east and west by the prevailing zonal winds. Historical records show that Saturn’s atmosphere exhibits such planetary-scale disturbances approximately once per saturnian year (29.4 Earth years), usually after the summer solstice [heliocentric longitude Ls in the range from 110° to 167° for previous storms (7, 8), measured from the spring equinox at Ls = 0°]. The 2010–2011 storm occurred earlier in the seasonal cycle than usual (Ls ~ 16°), and is an exception to the 30-year cadence [the previous equatorial eruption occurred in September 1990 (812) at Ls = 120°]. It is the first to occur at this latitude in over a century [the last eruption at 36°N was in 1903 (7) at Ls ~ 134°], and only the sixth disturbance to be recorded since 1876. However, although previous disturbances have been intensively studied in reflected sunlight (712), their effect on the atmospheric thermal structure, chemistry, and circulation patterns has never been measured.

We combined orbital spectroscopy (5 to 200 μm) from the Cassini spacecraft with high-resolution Earth-based thermal imaging from the European Southern Observatory (ESO) Very Large Telescope (VLT) in Chile (7 to 20 μm) to reveal the vertically coupled dynamics of the storm system from the upper troposphere (at pressures ranging from 70 mbar to 3 bar) to the stratosphere (0.5 to 20 mbar). The VLT images taken on 19 January 2011 (Fig. 1) (13) reveal large-amplitude perturbations of the usually quiescent thermal field between 20° and 50°N and extending from east to west over an entire hemisphere (140,000 km at 40°N) 45 days after the onset of the disturbance. Cooler temperatures were due to vertical upwelling and adiabatic cooling of air parcels (typical of Saturn’s zones in the upper troposphere), whereas warmer temperatures were due to subsidence. The VLT images are centered on a compact cold oval (79 ± 2 K at 100 mbar, 4000 × 5500 km in size ± 800 km) near 41°N, 314°W, which appeared bluish-white in visible imaging and was surrounded by a collar of white clouds (Fig. 1, B and C). Thermal contrasts in Fig. 1D are consistent with a newly formed anticyclonic vortex in geostrophic balance whose peripheral velocity decreases with altitude. The central latitude of the disturbance, near the peak of a weak westward jet (39°N), had been previously identified as a region of potential dynamic instability and eddy activity derived from the mean zonal flow (14) and was the site of the highest concentration of discrete vortices before Cassini’s arrival (15). The “storm head” of visibly bright clouds extended 60° west of the cold vortex (toward 15°W on January 19). The vortex marked a distinct boundary between warmer temperatures over the western storm head and cooler temperatures to the east (an east-west contrast of 5 ± 2 K at 70 to 300 mbar, Fig. 1D), and was a point of bifurcation into two cool eastward branches of the disturbance: a northern branch between 42° and 48°N and a southern branch between 31° and 38°N, separated by a faint warm sector between 38° and 42°N. The two cool branches, characterized by undulating white cloud structures in visible observations, extended eastward with the prevailing zonal flow (15) at 32°N and 47°N (Fig. 1C). Conversely, the compact vortex at the center of the disturbance propagated westward with the retrograde jet at 39°N.

Fig. 1

(A) Thermal images acquired with VISIR (23) on the VLT on 19 January 2011, sensitive to Saturn’s tropospheric temperatures (10.7- and 18.7-μm, 200 to 500 mbar) and stratospheric temperatures (8.6- and 12.3-μm emission from methane and ethane, respectively, sensitive to the 1- to 10-mbar level), compared to a visible-light RGB image from T. Barry [left, from the International Outer Planets Watch Planetary Visual Observatory & Laboratory (13) database]. The storm clouds appear white in Saturn’s spring hemisphere (visible image), but the disturbance is dark and cold in the thermal images and flanked in the stratosphere by regions of bright, warm emission. (B) Cool tropospheric brightness temperatures at 18.7 μm are correlated with (C) albedo features in the visible. White vertical arrows highlight the central vortex (41°N, 314°W) and the undulating features of the eastern branches. TB is the brightness temperature in kelvin. Black horizontal arrows in (C) show the prevailing zonal wind (u) direction at each latitude. (D) Saturn’s longitudinal temperature structure near 40°N retrieved from the eight VLT images (7 to 20 μm). The dark vortex core at 314°W (B) can be seen as a cold region near 100 mbar in the longitudinal temperature cross-section (D).

We used the VLT images to place Cassini spectroscopic observations into their global context. The Composite Infrared Spectrometer (CIRS) (16) obtained two nadir maps of the northern hemisphere between 7 and 200 μm at 2.5 cm−1 resolution; the first on 22 October 2010 before the storm onset, and the second at the same time as the VLT observations on 19 January 2011, during the mature phase of the disturbance (13). We derived tropospheric (70 to 400 mbar) and stratospheric (0.5 to 5.0 mbar) temperatures from each CIRS map using an optimal estimation retrieval algorithm (Fig. 2) (13). Assuming that seasonal changes were minimal over the 89 days separating the observations (6), differences between the retrieved temperatures demonstrate substantial thermal perturbations generated by the disturbance (Fig. 2). In particular, the upper tropospheric cooling served to increase the lapse rate by 0.2 to 0.3 K/km in the region from 70 to 300 mbar [compared to a dry adiabatic lapse rate of 1 K/km in the same range (13)], destabilizing the troposphere to convection and reinforcing the growing disturbance.

Fig. 2

(A) Differences in retrieved north-south temperatures (ΔT, measured in kelvin) averaged over the disturbance longitudes before and after the onset of the storm from Cassini/CIRS spectroscopy (13) (fig. S3). (B) Longitudinal thermal anomalies (ΔT, measured in kelvin) generated by the storm from CIRS spectroscopy at 40°N, covering the same longitude range as Fig. 1D. (C) Storm-induced perturbations to the zonal jets (velocity difference Δu, measured in meters per second) calculated via the thermal windshear equation. (D) Vertical displacements in the stratosphere (Δz, measured in kilometers) calculated by assuming adiabatic motions over the disturbance longitudes (13).

The CIRS spectra, along with VLT filters sensitive to stratospheric emission from methane and ethane (7.9, 8.6, and 12.3 μm), demonstrate that the tropospheric disturbance had a substantial effect on the stably stratified upper atmosphere at pressures as low as 0.5 mbar, the upper limit to CIRS sensitivity in nadir spectra (16). Before the storm, the northern stratosphere (10° to 50°N) had warmed considerably (6) since the start of the Cassini mission (Ls = 292°, northern winter), but the upwelling associated with the disturbance reversed this warming trend (260° to 320°W had cooled 7 to 9 K by 19 January, Fig. 2B). Although some degree of convective overshooting to the 60-mbar level (15 to 20 km above the tropopause) was expected from moist convection models (5, 17), such large perturbations at 1 to 5 mbar (Fig. 2, A and B) cannot be attributed to direct convection. Instead, they are a response to the mechanical forcing from below, possibly as a result of wave activity (e.g., Rossby waves) initiated by the strong tropospheric upwelling and transporting momentum vertically into the stratosphere (18). Indeed, the cool stratosphere over the disturbance was flanked on all sides by warm regions consistent with subsidence at the periphery of the central upwelling, which produced warm “stratospheric beacons” of emission near 225° to 255°W and 25° to 330°W (over the storm “head,” Fig. 1). The 16 ± 3 K contrast between the beacons and the cool region above the central disturbance (Fig. 2B) is the largest-amplitude zonal perturbation detected in Saturn’s stratosphere. The beacons persisted throughout January to March 2011 in Cassini and ground-based observations (13) and remained nearly stationary with respect to the tropospheric disturbance.

We used the thermal perturbations to estimate the vertical displacements of air parcels associated with the stratospheric anomalies (Fig. 2D) (13) and found that air over the central disturbance at 5 to 10 mbar was displaced upward by 20 ± 5 km (half a scale height) over the 310° to 270°W range, whereas air subsided by 10 ± 4 km to produce the warm stratospheric beacons, consistent with an atmospheric response to elevated pressure from below. Such displacements are consistent with the 20-km amplitude of convective overshooting predicted by numerical modeling (17). If this stratospheric displacement in response to the convective outburst was continuous over the 45 days of the disturbance, then we found minimum estimates of the vertical velocities of 2 to 5 mm/s. Given that horizontal mixing will accompany these displacements and that the nature of the waves excited by the disturbance cannot be unambiguously identified, these estimates should be considered as lower limits and are considerably smaller than the expected velocities within the tropospheric updrafts (17).

The latitudinal thermal anomalies of 6 ± 2 K (Fig. 2A) modify the vertical shear on the zonal winds above the disturbance (Fig. 2C) (13), causing the acceleration of the eastward jet near 30°N and a westward jet near 40°N. The jet perturbations were strongest in the stratosphere near the 1-mbar region, suggesting the vertical transport of momentum by waves as modeled for the 1990 disturbance (18). The absence of traceable cloud features at stratospheric altitudes prevents direct confirmation of these velocity changes, and the perturbation at the tropospheric altitudes of the visible clouds is predicted to be small (less than 20 m/s). Finally, the thermal wind equation implies that the longitudinal thermal gradients (Fig. 2B) are associated with strong meridional velocities, consistent with the generation of closed vortices (such as the compact oval in Fig. 1) or zonal jets that meander in latitude as on Earth. The thermal contrasts resulting from the planetary-scale disturbance have a considerable effect on atmospheric motion.

Because of its long photochemical lifetime, stratospheric acetylene (C2H2) can be used as a passive tracer of atmospheric motion (19). We retrieved the C2H2 spatial distribution (Fig. 3A) from the 2.5 cm−1 spectral maps, finding it to be zonally homogenous [0.3 parts per million (ppm) at 1 mbar] on 22 October 2010 but significantly depleted (0.1 ppm near 287°W, Fig. 3A) over the disturbance on 19 January 2011, consistent with upward advection of C2H2-depleted air. Assuming no chemical sources or sinks over the disturbance, evaluation of the continuity equation (13) yields estimates of the vertical displacement of 40 to 45 km (approximately a scale height) in the stratosphere (a velocity of 10 mm/s if the depletion was continuous over 45 days). In addition, motion in the upper troposphere (400 to 700 mbar) can be traced using the disequilibrium species phosphine (PH3) (20) retrieved from far-infrared CIRS spectra at the northern edge of the disturbance on 2 January 2011 (21). PH3 was enhanced over both the cold region east of the central oval and the western storm head by 20 ± 5% as compared to its concentration at quiescent longitudes east of 240°W (Fig. 3B), which implies vertical mixing over the whole longitude range of the disturbance (10° to 240°W). Assuming that perturbations to PH3 are solely caused by dynamics [the lifetime of PH3 may be multiple years at the pressure levels of interest (21)], vertical displacements over the northern edge of the storm must be approximately 2 to 10 km greater than those at the quiescent longitudes to produce the 20% enhancement. The vertical displacements estimated from the gaseous composition are thus broadly consistent with those from the temperature anomalies (10 to 20 km, Fig. 2D).

Fig. 3

(A) The longitudinal distribution of acetylene at 40°N (C2H2, retrieved from 750 to 900 cm–1 spectroscopy) (13) before and after the disturbance. Mole fractions are given at 1 mbar. The black star indicates the approximate longitude of the central vortex on each date. (B) Fractional enhancement of tropospheric PH3 (pressure > 550 mbar) near the northern edge of the storm (45° to 55°N) on 2 January 2011 (13). The horizontal dashed line shows the 20% enhancement over the disturbance, from the storm head to the easterly branches, 240° to 360°W.

East of the central vortex, undulations of tropospheric temperatures and cloud albedo are reminiscent of planetary wave activity with a 15° longitudinal wavelength (11,700 km), causing vertical oscillations of air parcels, condensation of ices at the peaks, and sublimation of ices at the troughs. Similar horizontal wave phenomena were reported and modeled for the 1990 equatorial disturbance (10, 18). At least three of these undulations are well resolved in the southern branch in Fig. 1, B and C. The Cassini Visual and Infrared Mapping Spectrometer (VIMS) (22) obtained nightside 4.6- to 5.1-μm spectroscopy of cloud opacity associated with these undulations near 30°N (Fig. 4). Retrieved tropospheric properties within this southern branch confirm (i) elevated cloud opacity associated with dark features silhouetted against Saturn’s 5-μm emission (both in a “haze” extending above 1.4 bar and in deeper clouds condensing near 2.8 bar); (ii) elevated PH3 in the 1- to 2-bar range, particularly associated with the largest dark feature between 190° and 195°W; and (iii) depleted NH3 vapor, suggesting removal by condensation in the cooler updrafts. These VIMS and CIRS retrievals suggest wave-induced vertical oscillations throughout the 0.1- to 3-bar range and enhanced condensation at the cool peaks of the wave, extending east from the central vortex.

Fig. 4

(A) Cassini/VIMS 5-μm nightside emission over a narrow longitude east of the main disturbance on 9 January 2011, showing clouds as dark silhouettes illuminated by Saturn’s internal heat. We retrieved cloud opacity and composition (13) for two latitude circles, one within the undulating cloud structures (30°N, black circles) and another at a quiescent latitude (22°N, open diamonds). (B) Optical depth (τ2) of a 1.4-bar cloud deck. (C) Optical depth (τ1) of the deeper 2.8-bar cloud deck. (D) Longitudinal PH3 distribution in the 1- to 2-bar region associated with the undulating clouds. (E) NH3 vapor in the 1- to 3-bar region. Error bars on derived quantities are larger in the dark cloud regions because of the smaller signal.

Thermal imaging of this intense northern springtime disturbance revealed strong coupling between the tropospheric outburst [possibly moist convection initiated within water clouds at 9 to 12 bar (18)] and the atmospheric structure between 1 mbar and 3 bar (vertically separated by 350 km) over an enormous area of Saturn’s northern mid-latitudes. The increased insolation after the spring equinox (August 2009, Ls = 0°) may permit convective outbursts to penetrate to higher altitudes than during other seasons, where they become accessible to infrared remote sensing; but the initial instability presumably occurred at deep levels where solar insolation has no influence. The large amplitude of the stratospheric perturbations, which continue to dominate Saturn’s infrared emission at the time of writing, suggests a strong atmospheric response to the storm cells below and implies vertical coupling over hundreds of kilometers between the convective deep atmosphere and the radiatively cooled upper atmosphere. Such large perturbations substantially altered atmospheric circulation, transporting energy and material vertically over great distances, perturbing stratospheric zonal jets, exciting wave activity, and disrupting the slow seasonal evolution of Saturn’s atmosphere. The newly formed oval and the aftermath of this planetary disturbance could affect Saturn’s northern hemisphere for years to come.

Supporting Online Material

www.sciencemag.org/cgi/content/full/science.1204774/DC1

Materials and Methods

Figs. S1 to S5

Table S1

References (2428)

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. Acknowledgments: L.N.F. is supported by a Glasstone Science Fellowship at the University of Oxford. We thank members of both the Cassini CIRS and VIMS teams for their support in the analysis of Cassini data, and the director and staff of the ESO VLT for their assistance with the execution of these observations. This investigation was partially based on observations acquired at the Paranal UT3/Melipal Observatory under ID 386.C-0096. We are extremely grateful to all those observers contributing to the International Outer Planets Watch Planetary Visual Observatory & Laboratory (IOPW-PVOL), including T. Barry (Australia) for his contribution to Fig. 1. The UK authors acknowledge the support of the Science and Technology Facilities Council. A.S.L., R.H., and T.d.R. are supported by the Spanish MICIIN (Spanish Ministry of Science and Innovation) project AYA2009-10701 with FEDER and Grupos Gobierno Vasco IT-464-07. G.S.O. is supported by grants from NASA to the Jet Propulsion Laboratory, California Institute of Technology. L.N.F. performed the optimal estimation retrievals on both ground-based and space-based data sets and analyzed the results, and wrote the initial draft. B.E.H., G.L.B., and A.S.-M. designed the CIRS mapping observations; calibration assistance was provided by A.M. T.W.M. and K.H.B. provided the VIMS nightside spectra. G.S.O. supported ground-based VLT observations, and A.S.-L., R.H., and T.d.R.-G. provided IOPW-PVOL images. J.M.G. provided cylindrical mapping of IOPW-PVOL images. P.G.J.I. developed the optimal estimation retrieval codes used in this study. All authors discussed the results and commented on the manuscript.
View Abstract

Navigate This Article