The jovian anticyclone BA

The jovian anticyclone BA

Icarus 203 (2009) 486–498 Contents lists available at ScienceDirect Icarus journal homepage: The jovian anticyclone ...

2MB Sizes 4 Downloads 71 Views

Icarus 203 (2009) 486–498

Contents lists available at ScienceDirect

Icarus journal homepage:

The jovian anticyclone BA I. Motions and interaction with the GRS from observations and non-linear simulations E. García-Melendo a,*, J. Legarreta b, A. Sánchez-Lavega c, R. Hueso c, S. Pérez-Hoyos c, J. González c, J.M. Gómez-Forrellad a, IOPW Team 1 a

Esteve Duran Observatory Foundation, 08553 Seva, Spain Departamento de Ingenieria de Sistemas y Automática, EUITI, Universidad del País Vasco, Bilbao, Spain c Departamento de Física Aplicada I, E.T.S. Ingenieros, Universidad del País Vasco, Alameda Urquijo s/n, 48013 Bilbao, Spain b

a r t i c l e

i n f o

Article history: Received 17 December 2008 Revised 28 April 2009 Accepted 3 May 2009 Available online 17 June 2009 Keywords: Jupiter, atmosphere Atmospheres, dynamics

a b s t r a c t A study of the dynamics of the second largest anticyclone in Jupiter, Oval BA, and its red colour change that occurred in late 2005 is presented in a three part study. The first part, this paper, deals with its longterm kinematical and dynamical behaviour monitored since its formation in 2000 to September 2008 using ground-based observations archived at the public International Outer Planet Watch (IOPW) database. The vortex changed its zonal drift velocity from 1.8 m s1 in the period 2000–2002 to 0.8 m s1 in 2002–2003, and to 2.5 m s1 since late 2003. It also migrated southwards by 1.0 ± 0.5° in latitude between 2000 and 2004, remaining afterwards at an almost fixed latitude position. During the period 2000–2007, the oval also changed its triangular-like shape to a more symmetrical one. No latitudinal change was found in the months before the development of a red annulus in its interior. The colour change took place in less than 5 months in 2005–2006 and no red colour feature was observed to have been present or entrained by BA months before the annulus development. After detailed examination of the four encounters between BA and GRS that took place during this 9 year period, we did not detect any noticeable change in its drift rate or in apparent structure associated with the encounters at cloud level. Also, the area of BA did not significantly change in this period. Additionally, we found that BA displays a long-term oscillation of 160 days in its longitude position with peak to peak amplitude of 1.2°. Numerical experiments using the global circulation model EPIC reproduce accurately the shape, connecting it to its latitude migration, and morphology of the oval and confirm that no strong interaction between BA and the GRS is possible at least in the current situation. Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction BA, presently the second largest anticyclonic oval on Jupiter after the Great Red Spot, has received a great deal of attention due to its origin after the merging of three large ovals and more recently due to a colour change detected by amateur astronomers by the end of 2005 (Naeye, 2006; Simon-Miller et al., 2006; Cheng et al., 2008). The formation of BA, after the merging of its predecessors BC and DE to form the new Oval BE in 1998 (Sánchez-Lavega et al., 1999), and the final merging of BE and FA in 2000 (SánchezLavega et al., 2001), took place a few weeks after their conjunction with the GRS. Numerical models by Morales-Juberías et al. (2003) showed that the merger is a natural consequence of the encounter * Corresponding author. E-mail address: [email protected] (E. García-Melendo). 1 URL: (IOPW Team). 0019-1035/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.icarus.2009.05.031

of two vortices that occupy the same channel domain between opposed jets, and occurred following a mutual pair orbiting. Simulations by Youssef and Marcus (2003) indicate that oval pairs BC and DE, and BE and FA were trapped in a weakly stable dynamical system where a cyclone between the vortex pair provides repulsion between them. In their simulations a second cyclone perturbed the system and triggered the merging process. This paper I is the first part of three papers that deal with the long-term motions and shape evolution of BA (this part), its internal dynamics at high resolution (part-II, Hueso et al., 2009) and its cloud structure and colour characterisation (part-III, Pérez-Hoyos et al., 2009). Here we focus on the dynamics of the oval since its formation in 2000 up to 2008 and its encounters with the GRS. A second goal of this work is to constrain the environmental behaviour of BA and the temporal scale related to its colour change. A third goal is to explore to what extent a weak or a strong interaction between the GRS and BA is possible. This task is accomplished by examining


E. García-Melendo et al. / Icarus 203 (2009) 486–498

the outcome of the different passages of BA and the GRS and by using the General Circulation Model EPIC (Dowling et al., 1998) in its isentropic coordinate version. We show that the model is able to reproduce the oval structure and morphology in the different observing periods, which are linked to the latitude of BA, and that no strong interaction between both vortices is possible in the current dynamical situation.

information on BA’s morphological parameters and they can serve as a consistency test for the ground-based measurements. In addition, HST images in 2007 were used to determine the zonal wind profile in the South Tropical latitudes where BA is located. Detailed analyses of these HST images are presented on accompanying papers II (Hueso et al., 2009), and III (Pérez-Hoyos et al., 2009) where they are used to infer the internal dynamics of the oval and its cloud structure. Table 2 displays the observing log of the archived HST Jupiter imagery used in this study.

2. Ground-based observations and image navigation 2.2. Image navigation and measurement errors 2.1. Observations The dynamical behaviour of BA was monitored since its formation in early 2000 by using a long-term observational imaging survey performed with ground-based telescopes and the available archived HST WFPC2 and ACS images. Ground-based observations were collected by the International Outer Planet Watch (IOPW) team at the Planetary Virtual Observatory & Laboratory database ( Detectors range from conventional CCD astronomical devices to webcam cameras and general purpose CCD scientific cameras. Image resolution of selected frames, which ranged from 0.6–1 arcsec in 2000 to 0.35–0.9 arcsec in 2008, varied depending on telescope’s focal length, CCD detector characteristics and seeing conditions, but improved as a result of continuous advances in observers’ equipment and imaging techniques. During the first years of the IOPW record, from 2000 to 2004, Jupiter images were single filter or RGB colour images taken with conventional astronomical CCD cameras. In the later years, especially since 2005, image stacking procedures, a technique known as ‘‘shift and add” initially devised for high resolution imaging of binary stars (Christou, 1991), has been widely used by the planetary astronomy community (see, for example, Dantowich et al., 2000; Baumgardner et al., 2000). With the introduction of webcam CCD cameras and general purpose high resolution scientific CCD devices, ‘‘shift and add” has extended to small telescopes and bright objects like the planets and the Moon (Wöhler et al., 2006). In the case of Jupiter, this technique provides images with a very high signal to noise ratio that allows a significant improvement in the capability to contrast the cloud features through image processing techniques. More than 5000 images are available in the database from 2000 to 2008 but only the best of them were selected for this study according to their quality, spatial resolution and visibility of BA close to the central meridian. In total, 589 Jupiter images were used. Table 1 lists the observing log for the selected ground-based images. As a test of the quality of the selected images, we measured the GRS position when the vortex was present, retrieving a clear signature of its 89-day oscillation, including the correct phase and amplitude (Trigo et al., 2000). We also analysed the archived HST WFPC2 and ACS Jupiter frames between 2000 and 2007. Although they only cover a small portion of the ground-based monitored time they give precise

Table 1 Log of IOPW CCD observations. Observation interval (month-day-year)

July-22-2000 to April-23-2001 August-16-2001 to April-21-2002 September-02-2002 to June-01-2003 October-20-2003 to July-24-2004 November-06-2004 to July-15-2005 January-13-2006 to September-18-2006 February-09-2007 to September-13-2008

Julian day interval

2451748.3–2452023.3 2452138.3–2452386.3 2452520.3–2452791.6 2452933.4–2453210.9 2453316.4–2453567.0 2453749.3–2453996.9 2454141.3–2454723.0

Number of measured images 35 72 73 95 89 145 110

We used the LAIA software (Cano, 1998) as our image processing and analysis tool. It has been used by our team in previous works (see Sánchez-lavega et al., 2000; García-Melendo et al., 2000; Legarreta and Sánchez-Lavega, 2005). LAIA allows to measure longitudes and latitudes over the visible disk of Jupiter disk in the image after fitting an ellipse to the planet limb. The ellipse takes into account the planet oblateness. Longitudes and latitudes are calculated considering the Earth’s planetocentric declination D. We provide here a brief procedure to estimate measurement errors on ground-based images. In ground-based images we see a projection of the planetary disk on the sky background. If XYZ is a reference system coincident with the revolution ellipsoid of the planet, where the XZ plane coincides with the planet’s equatorial plane, and the Y axis is parallel to Jupiter’s rotation axis with positive values in the north direction, Jupiter’s shape can be described by

X2 R2e


Z2 R2e


Y2 R2p

¼ 1;


where Re and Rp are the respective equatorial and polar radii (Re = 71,492 km and Rp = 66,854 km as given by Seidelmann et al., 2007). We then choose a coordinate system on the CCD image where the x and y axes, centred on the planet’s disk, are coincident with the respective sky projected equatorial and polar radii defining Jupiter’s apparent disk. To complete the coordinate triad, we choose the z axis directed towards the observer. Any point on the planet’s surface with coordinates (X, Y, Z) is related to the (x, y, z) reference system by

9 > = Y ¼ y cosðDÞ þ z sinðDÞ ; > ; Z ¼ y sinðDÞ þ z cosðDÞ X¼x


where the Earth’s planetocentric declination, D, is typically a small angle, D < 3°. A measurement of the (x, y) position on the CCD frame of any point on Jupiter’s disk can be readily transformed into its (X, Y, Z) coordinates by simultaneously solving Eqs. (1) and (2). Longitude (k) with respect to the central meridian and planetocentric latitude (/c) are then given as:

Z cosðkÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 X þ Z2


Y tanðuc Þ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : 2 X þ Z2


Table 2 Measured archived HST frames. Date (yearmonth-day)

Dataset (first four characters)



2000-09-02 2005-01-19 2006-04-08 2007-02-26 2007-02-27 2007-03-08 2007-03-26



F673 F330 F330 F410 F410 F410 F410

N, F953 N W, FR656 N, F435 W W, F435 W, FR656 N, F658 N, M M M M


E. García-Melendo et al. / Icarus 203 (2009) 486–498

Expressions (1) and (3) allow us to estimate measurement uncertainties. There are two main error sources: image navigation and feature pointing on the CCD frame. To estimate errors we will assume a value of D = 0° to simplify computations, and therefore y ffi Y and z ffi Z. If we use image pixels instead of km to measure positions then we have the respective uncertainties DRp ffi DRe, Dx, and Dy. The related Dk and D/c uncertainties when x and y are not close to the limb or the central meridian are

Dk ffi

1 x2 ðRe DRe  f 2 yDyÞ  xDxðR2e  f 2 y2 Þ ; tanðkÞ ðR2e  x2  f 2 y2 ÞðR2e  f 2 y2 Þ

Duc ffi sinðuc Þ cosðuc Þ

R2e DRe  yRe Dy yðR2e  f 2 y2 Þ



ð4Þ ð5Þ

with f2 = (Re/Rp)2 = 1.13 for Jupiter. In most of the PVOL images the planet image extends only 300 pixels from one pole to the other. For a typical image where BA (/c = 30°) is 30° away from the central meridian, Rp = 150 pixels, y = 75 pixels and x = 70 pixels and (4) and (5) apply. Assuming an error of one pixel in the limb’s polar or equatorial radius and in the x and y coordinates, and assuming that uncertainties contribute constructively (DRp ffi DRe = ±1, Dx = 1, and Dy = 1) the resulting errors are Dk  1.2° and D/c  0.7°. The above example indicates that latitude uncertainties will have a tendency to be slightly smaller than longitude errors, and that errors are inversely proportional to image resolution. These results are consistent with Fig. 1, which shows a statistical analysis of the standard deviations of longitude and latitude measurements of BA as measured in the PVOL database plotted as functions of the observing season. The improvement in the ground-based observational techniques is responsible for the diminishing scatter. During the last seasons, typical position errors for BA in latitude and longitude on the ground-based images due to both navigation and feature pointing are 0.6°.

able changes in the environment around BA or the vortex itself; the morphology around BA in the last images before the colour change and in the first images after the red colouring is very similar (see Fig. 2a and b). In the recorded images we did not detect any ‘‘red colour” feature close to BA that could have been absorbed by the vortex. Therefore we conclude that the colour change was intrinsic to BA (at cloud level) and not due to horizontal cloud mixing. The composed colour images in the PVOL database indicate that the colour change was concentrated in a red annulus that developed inside BA in a time-scale of 100 days or less. Unfortunately, it is not possible to know whether the colour change of BA was sudden or gradual, but any physical model must take into account this upper time limit. We cannot provide either a numerical quantification of the colour change by examining the record of IOPW images since these are created using a variety of cameras, filters and image processing techniques. In general, colour in the IOPW images is created from the composition of three monochrome images taken with primary additive RGB filters and an infrared blocking filter to avoid infrared leakage, but some observers use only the red and blue filter creating a synthetic green image from the composition of red and blue. The properties of the colour change from HST images have been presented by Simon-Miller et al. (2006) and Cheng et al. (2008) and are discussed in detail in part-III by Pérez-Hoyos et al. (2009). In general terms these studies agree that the change is concentrated in blue wavelengths, which points to changes in the optical properties of the cloud particles (single scattering albedo and imaginary refractive index) and not in their number density or mean size. Possible sources for the red chromophores are vertical transport from lower levels, chemical or photochemical reactions, phase changes, etc. The lack of significant changes in the oval’s internal dynamics (part-II, Hueso et al., 2009) and the time-scale of colour variation are consistent with a diffusive process. A more in depth discussion of these sources is presented in part-III by Pérez-Hoyos et al. (2009).

3. BA properties and time evolution 3.1. Formation of a red annulus inside BA The highest quality ground-based images show an off-white, high albedo Oval BA from its formation in 2000–2005. By the end of the Jupiter–Sun conjunction on December 2005, BA showed a reddish colour in its interior (Fig. 2). This happened without notice-

Fig. 1. Standard deviation for latitude (black), and longitude (grey) measurements from ground-based images as a function of the observing season. The longitude scatter has been found after subtracting the mean drift rate of the oval in each observing season. A general trend of decreasing scatter can be seen as a result of the improvement in image resolution. Latitude measurements show a systematic lower scatter than longitude data. During the last two Jupiter apparitions the overall scatter was close to 0.6°.

Fig. 2. The Oval BA and the development of a red ring between years 2005 and 2006 as recorded by images from the IOPW. The black arrow marks the position of BA: (a) image taken on April 29, 2005 by Damian Peach; (b) April 12, 2006 by Damian Peach; (c) May 26, 2007 by Damian Peach; and (d) May 08, 2008 by Anthony Wesley. North is down and East is to the right (telescopic view). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

E. García-Melendo et al. / Icarus 203 (2009) 486–498

3.2. BA mean motions Longitude measurements of BA with respect to System III (XIII = 870.563 day1, Davies et al., 1986) are represented in Fig. 3a. A change in the zonal drift velocity can be clearly seen in 2002–2003 between the Julian Days 2452500 (August 2002) and 2452800 (June 2003) defining three velocity regimes. This allows us to compute three distinct drift velocities (see Table 3) by fitting different linear functions to the three linear regimes. The slopes of the fits gave us the zonal drift (longitude degrees per day), which can then be transformed into zonal velocity once the average latitude is known. Linear fits show that BA and the GRS move with similar zonal velocities (for BA ranged from +0.8 to +2.5 m s1 (Table 3) and for the GRS ranged from 3.8 to 4.0 m s1), and therefore their relative longitude drift is small.


The small velocity difference between both ovals makes them encounter at the same longitude every 2 years. The successive fusion of BC and DE and BE and FA took place following GRS passages in the 1998 and 2000 encounters. The GRS may have favoured the mergers by expelling an intermediate cyclone placed between the ovals (Sánchez-Lavega et al., 1999, 2001) but the role of the GRS is still an open question since there are simulations of the merger that do not require the GRS relying instead on interactions with nearby cyclonic features (Youssef and Marcus, 2003). We studied the movements of BA from its formation in 2000–2008 exploring possible interactions with the GRS in each passage. Fig. 3b shows the evolution of BA’s longitudinal position after removing the general linear trend present in Fig. 3a and represented by the equation L(t) = 0.177  t + 434554.7, where t is the time in Julian days. Fig. 3c shows the planetographic latitude

Fig. 3. (Panel a) Cumulative longitude position of BA centre for an arbitrary origin measured in System III. The slowing down of BA’s drift velocity during the 2002–2003 season (JD 2452520.3–2452791.6, see Table 2) is responsible for the longer time interval between the first two GRS passages. (Panel b) Residual longitude position of BA after removing the global linear trend. (Panel c) Planetographic latitude of BA centre. Grey-filled circles are individual latitude measurements obtained from ground-based observations with their mean value for each season marked as an empty circle along with its error bar. Solid circles are individual HST latitude measurements. Vertical grey bars mark GRS passages.


E. García-Melendo et al. / Icarus 203 (2009) 486–498

Table 3 BA’s zonal drift velocity and mean latitude for different time intervals. Average drift rate (m s1)

Average season date (Julian day and date)

Average Pg (PVOL images)

Average Pg (HST images)

1.8 (22 July 2000 to 21 April 2002)

2451839 (October 2000) 2452280 (January 2002)

32.2 ± 1.2 32.7 ± 1.0

32.7 ± 0.3 (September 2000) –

0.8 (2 September 2002 to 1 June 2003)

2452670 (January 2003)

32.2 ± 0.9

2.5 (20 October 2003 to 13 September 2008)

2453100 2453450 2453900 2454204 2454626

33.2 ± 0.8 33.6 ± 0.8 33.7 ± 0.7 33.3 ± 0.7 33.4 ± 0.5

– 33.6 ± 0.3 (January 2005) 33.1 ± 0.4 (April 2006) 33.6 ± 0.3 (March 2007) –

(April 2004) (March 2005) (June 2006) (April 2007) (June 2008)

of BA measured on ground-based and HST images. Vertical grey bars indicate the time of GRS passages. No clear correlation can be seen between the close presence of the GRS and changes in drift velocity. A deceleration in the drift rate of BA occurred following the first BA–GRS encounter (JD2452325, 19 February 2002) but during the second encounter (JD2453195, 8 July 2004) BA was already accelerating again, keeping this trend during its third (JD2453932, 15 July 2006) and fourth (JD2454650, 2 July 2008) passages with no significant drift changes. There also appears to be no correlation between small latitude changes of BA and the BA–GRS encounters. Table 3 lists average BA latitudes from ground-based observations along with their standard deviations as represented in Fig. 3c. To further check that the passages did not change the latitude of BA we computed the average latitude of BA in the seasons before and after the GRS passages and we did not find any significant differences larger than measurement uncertainties. For the first, second, and third passages these differences were 0.0°, 0.4°, and 0.3°, respectively, which are lower than the measurement errors. We also computed the latitude differences in the days immediately before and after the GRS passage within the same observing season. For the first, third, and fourth passages we obtained approximate absolute latitude differences of 0.2° and a lower difference in the second passage. There is, therefore, no detectable effect on the long-term dynamics of BA due to GRS passages. Given the error bars in latitude measurements around the average value of 33°S (Fig. 3c, dashed line), the long-term measured latitude change could be spurious as well. We tried to reduce this uncertainty in the latitude of BA by measuring its position on HST images. However, even in these images it is not possible to reduce the latitudinal uncertainty of this large feature below a 0.3° level (see Table 3). In any case, there is a strong correlation between the change from slow to fast drift velocities (Julian day intervals between 2,451,700 and 2,452,900, and 2,452,900 and 2,454,7000, respectively) and a global latitude change of BA from PVOL and HST images of 1.0 ± 0.5° and not related to GRS passages. Fig. 4 shows the latitude migration of BA’s centre and its drift rate. Fig. 4a shows BA latitude measurements plotted on the 2000 (Porco et al., 2003), and 2007 wind profiles in the region around BA. The 2007 zonal wind profile was measured by using an automatic one-dimensional correlator on HST image pairs (filters F410M, F673N, and F953N) separated by a jovian rotation (10 h) with typical rms errors of 5 m s1. Details on the one-dimensional correlation technique and measurement methods can be found in García-Melendo and Sánchez-Lavega (2001). Fig. 4a shows that the zonal wind did not change significantly at the latitudes of interest in this 7-year period. Fig. 4b shows the correlation between the mean latitude of the oval and its zonal drift rate. The stable latitude and zonal drift measured in the 2004–2008 period since BA formation suggest that the oval has reached an equilibrium position in the ambient wind, agreeing with previous vortices drift measurements (Morales-Juberías and Sánchez-Lavega, 2002).

3.3. BA’s longitudinal oscillations Although we did not detect any clear influence of GRS passages on the long-term zonal motion and latitudinal position of BA, we found a long-term oscillation of its longitudinal position that is not related to the GRS proximity. We explored this aspect by running a frequency analysis of BA’s longitude position. For every data series associated to each apparition we subtracted its linear trend, which was supposed to be constant for short time periods spanning only 7 or 8 months during one Jupiter’s observing season. Residual data were analysed searching for periodicities on the unevenly spaced time series taking into account the pathologies associated with the data spacing, including aliasing effects due to the properties of the spectral window (Deeming, 1975), which in this case is dominated by the synodic period of Jupiter. For that purpose we computed the Lomb–Scargle periodogram (Scargle, 1982) of the time series data. The Lomb–Scargle periodogram shows that the data are dominated by a strong signal with a periodicity of 157 ± 10 days, apparently not related to Jupiter’s synodic period of 399 days nor to a GRSBA encounter periodicity. The periodogram is shown on Fig. 5a and displays other strong peaks at periods of 253 ± 15, 193 ± 10, and 123 ± 3 days, all with false alarm probabilities lower than 1%. A careful analysis shows that these secondary periodicities are aliases of the 157-day period combined with the observing window imposed by the 399-day synodic period. We compared these results with a numerical analysis using Period04 (Lenz and Breger, 2005), a software program designed for the Fourier analysis of time series which searches the best least squares solution that simultaneously fits the frequency, amplitude, and phase of several periodic signals. This procedure effectively removes all the interactions of the true periodicities with the sampling spectral window. Empirical results (Breger et al., 1993) and numerical simulations (Kuschnig et al., 1997) on time data series show that the S/N of peak amplitudes should be greater than 4.0 to be significant. Those significant frequencies are listed in Table 4 and compared with the basic analysis using the simple Lomb–Scargle periodogram. The highest peak in the power spectrum at the frequency F1 corresponds to a period of 159 days. The peak to peak amplitude of this oscillation is 1.2° in longitude. The rest of the frequencies are equivalent to the frequencies found by the Lomb–Scargle periodogram. F2 is very close to F1 minus half of Jupiter’s synodic frequency, F3 is close to the first harmonic of F1 and both F3 and F4 are barely significant. BA’s residual longitude data was folded over the 159 day period by first choosing, as an arbitrary phase origin, the first point of the time series, and then computing the number of elapsed cycles for each data point keeping the fractional part as the phase. Folded data (Fig. 5b) suggest a continuous oscillatory movement of BA with a period of 159 days during the 2000–2008 interval analogous to the GRS 89-day oscillation (Trigo et al., 2000).

E. García-Melendo et al. / Icarus 203 (2009) 486–498


Fig. 4. (Panel a) Zonal wind profiles measured during the Cassini flyby (black, Porco et al., 2003) and in HST images in 2007 (grey, this work), compared to BA’s average (latitude, velocity) measurements from 2000 to 2007 (empty points). Identification of BA’s latitudes for the different ground-based observing seasons is also given. (Panel b) Correlation between BA’s drift rate (right axis, black points) and the average latitude of the vortex as measured from the ground-based images (left axis, grey points). Error bars are also included for both datasets, but in the case of drift rates, represented error bars are smaller than point size.

3.4. Size and morphology evolution There is no evidence of significant size variations of BA during the 9-year observation period. From analysis of the PVOL images average BA dimensions in the 2000–2008 periods are 10.5 ± 1.5° in the longitudinal direction by 7.5 ± 0.5° in the meridional direction, which approximately represents an ellipse of 11,000 km by 8800 km and an axis aspect ratio of 1.2. The outer dimensions of the internal red ring was also measured after its formation, yielding an average size of 4.9 ± 0.5° by 3.7 ± 0.6° in the respective longitudinal and meridional directions, or 5100 by 4300 km with an associated axis aspect ratio of 1.2. We also measured the size of BA over the HST images resulting in sizes of 9.5 ± 0.5° and 7 ± 0.5° with only negligible changes in BA in its North–South and East–West dimensions. Due to the presence around BA of a high albedo collar, there is a tendency in low-resolution groundbased images to systematically show a slightly larger vortex in longitude than in the HST high-resolution observations, but both sets of observations consistently show a vortex of the same size over the 2000–2007 period. The same high-resolution observations also show that its shape seems to have evolved since its formation in 2000 from a triangular-like shape towards a more elliptical and symmetrical one. On

the 2000–2001 high resolution Cassini images, BA had an asymmetric triangular-like shape with a lower curvature in its southern edge. On the northern limit of the vortex, at the stagnation point, a cloud filament connected the peripheral vortex circulation with the high albedo South Tropical Zone. In the high-resolution observations acquired in 2007 during the flyby of the New Horizons spacecraft to Pluto the morphology of BA had changed to a more elliptical vortex shape which can also be seen in previous HST images. Fig. 6 compares the morphology of the oval from 2000 to 2008 as seen in high resolution Cassini, HST, and New Horizon (NH) images. As a summary of this section, a detailed analysis reveals no significant changes of the dynamical and morphological properties of BA since its formation. In particular, small alterations in the drift rate, latitude position, and shape, cannot be correlated with GRS passages or its colour change. Papers II and III will also show that other physical properties such as the internal velocity field or upper vertical haze structure do not show signs of change after BA’s hue change. We will show in the following sections that the evolution of the morphology of BA can be explained by its small latitudinal migration southwards and that the Great Red Spot and BA exist in different isolated domains within the zonal flow with no possibility of interaction.


E. García-Melendo et al. / Icarus 203 (2009) 486–498

Fig. 5. (Panel a) Lomb–Scargle periodogram of the residual longitude position of BA centre in terms of the time periods of the oscillations. Horizonal lines show different false alarm probabilities (fap). A false alarm probability lower than 1% is generally considered as statistically significant. (Panel b) Residual longitude of BA from deviations of the linear drifts from 2000 to 2008 folded on the 159 day period. The continuous line is the fit for the main period of 159 days.

Table 4 Frequency components for BA longitude long-term oscillations. Frequency (cycles per day)

Period (days)


Period (days) (Lomb–Scargle)

False alarm probability (%)

F1 = 0.0063 F2 = 0.0051 F3 = 0.0130 F4 = 0.0069

159 197 77 145

7.4 6.1 4.7 5.6

157 ± 10 193 ± 10 77 Extends to the 157 period

1 1 10 <1

4. Numerical simulations of BA and its interactions with the GRS We used EPIC, the Explicit Coordinate Atmospheric Model (Dowling et al., 1998), to simulate the observed properties of Oval BA, the BA–GRS encounters, and the surrounding morphology. A detailed account of previous applications of EPIC to the simulation of jovian vortices can be found in LeBeau and Dowling (1998), Morales-Juberías et al. (2003), García-Melendo et al. (2007), and Legarreta and Sánchez-Lavega (2008).

4.1. Atmosphere model, simulation domain and grid resolution The model free parameters are the vertical thermal structure of the atmosphere, the zonal wind profile, and the vertical wind shear. The used vertical thermal structure is the same as the one described in García-Melendo et al. (2005), an extrapolation below the visible cloud deck (700 mbar) of the thermal profile mea-

Fig. 6. Evolution of BA’s shape from 2000 to 2008. (a) Cassini image taken in December 2000 in the CB2 filter (750 nm). (b) HST/ACS image taken in April 2006 in the FR656 N filter (A. Simon-Miller, I. de Pater and M. Wong, HST Proposal 10783), (c) New Horizons/LORRI panchromatic image taken in February 2007 (NASA/Johns Hopkins University Applied Physics Laboratory/Southwest Research Institute), and (d) HST/WFPC2 image (F410 M filter) taken on July 8th, 2008 during the GRS passage. The difference in cloud contrast between panels a, b, c and d is due to the different filters used.

sured by the Voyager 1 and 2 radio occultation experiments (Lindal et al., 1981). For the zonal wind profile, we used the one measured from HST images from 1995 to 2000 (García-Melendo and Sánchez-Lavega, 2001). The zonal wind profile at the region where

E. García-Melendo et al. / Icarus 203 (2009) 486–498

BA sits has not experienced important changes since it was first measured by Limaye (1986) up to our measurements in 2007 (Fig. 4), so we used the well characterised measurements near 2000 which agree with Cassini observations (Porco et al., 2003). The wind profile model U(k, p) = uh(k)uv(p) follows a two-segment linear structure in ln(p), and is a separable function of latitude and pressure. Above p0 = 700 mbar (cloud top level) we used the linear approach for the wind amplitude

8 > < uv ðpÞ ¼ 0 if

> 1 : uv ðpÞ ¼ 1:0 þ 2:4


p < 2:4; p0   otherwise: ln pp0



For p > p0, we assume there is no vertical shear but a constant value uv(p) = 1. This vertical profile is adopted after the extensive simulations of jovian vortices made by Legarreta and Sánchez-Lavega (2008), who obtained a good reproduction of the physical properties of the three classical White Ovals at the latitudes of BA and the GRS when the vertical shear below the p0 level is small or null. Most of the simulations where carried out in channels spanning 80–160° in longitude by 40° in latitude centred around latitude 30°. The corresponding grids had 256  128 and 512  128 points resulting in a uniform horizontal and vertical grid resolution of 0.31°, which is equivalent to 335 km resolution. With this resolution, BA was covered by a 23  16 point grid (368 points), and the GRS by a 48  30 point grid (1400 points), both able to reproduce major features in both vortices. In all cases the time step Dt was 30 s, well bellow the Courant stability limit (see Press et al., 2007) of 95 s. The vertical domain extended from 10 mbar in the upper layer to 7 bar at the abyssal bottom level and was divided in eight layers. 4.2. Vortex models: BA and GRS parameterization In isentropic coordinates the Montgomery potential M = CpT + gz plays the role of hydrostatic pressure in geometricheight coordinates (Dowling et al., 1998).

D~ uh ¼ f ~ k ~ uh  rh M: Dt


In this expression D/Dt is the total derivative operator, f is the Coriolis parameter, and ~ k is the vertical unitary vector. In EPIC, vortices are introduced as an ellipsoidal Gaussian perturbation DM on the Montgomery potential (Stratman et al., 2001) as (  2  2 ! n  2 ) /  /s k  ks lnðpÞ  lnðps Þ DM ¼ cRP V T bs f exp  þ  : as bs cs ð8Þ

For the vortex location, f is the Coriolis parameter, Rp is the planet’s curvature length, /s is the longitude, ks the latitude and ps is the pressure level of the vortex coordinates centre, as and bs are the major and minor semi-axes, and cs is the vertical extension in scale heights. VT is the vortex maximum tangential velocity and c is a shape factor defined as

  1 exp 1  2n c¼   1 ; 1 ð12nÞ 2n 1  2n

1 @ðDMÞ fR @k

and Dv ¼ þ

where R and r are the respective meridional and longitudinal planetary curvature radii. Afterwards the vortex is let to evolve freely. For instance, assuming that we want to obtain the tangential velocity zonal profile for a given vortex at p = ps and / = /s by using the vortex model defined on Eq. (8) we obtain



2n # k  ks k  ks : Du ¼ V T 2nc exp  bs bs


To check how realistic the model is, we simultaneously fitted the parameters (n, /s, ks, as, bs) to the profiles of the tangential velocity VT across the semi-major and minor axes of the GRS and BA at the cloud deck level p = ps, reducing the error through a least squares technique. The reference profiles are those for the GRS measured on Galileo images (Choi et al., 2007) and for BA from Cassini images as reported in paper II. Fig. 7 shows the results for the GRS and BA. Table 5 gives the best parameter fits for both anticyclones. For the GRS, the model represents quite satisfactorily the structure of its internal velocity field. The respective values of 7.4° and 4.7° for the size parameters as and bs, the latitude of the centre (planetographic /s = 22.65), and the peak tangential velocity, match very well those average values found by Legarreta and Sánchez-Lavega (2005) also measured on Galileo imagery data. The inner structure of the GRS is particularly difficult to simulate. The meridional structure of the strong GRS anticyclonic collar is best modelled by n = 2 but n = 3 produces a better comparison of the zonal structure of the GRS with the simulated vortex. This variation in the value of n that best reproduces the zonal or meridional properties of the GRS is due to the fact that the anticyclonic collar of the GRS preserves its width along the perimeter of the vortex, but the semi-major and the semi-minor axes of the ellipse which represents the simulated vortex have different lengths (aspect ratio about 1.6). For our simulations of the GRS we chose n = 3 because it gives a better representation of the overall properties of the GRS. In the case of BA the value n = 1 is able to represent quite precisely all the profiles. The models also give accurate values for the dimensions and latitude locations of the vortex according to Legarreta and Sánchez-Lavega (2005). All the parameters included in Eq. (8) can be experimentally determined by the least squares fitting procedure described above except ps and cs. We assumed ps to be at the level of the ammonia cloud base 700 mbar. The vertical extension of jovian vortices given by the model parameter cs is unknown, but a minimum value can be determined from the ratio between the horizontal and vertical scales which should be of the order of f/2N (Conrath et al., 1981), where N is the Brunt-Väïsälä frequency. In the case of Jupiter’s atmosphere vortices are expected to have vertical extents of the order of 2.0–4.5 scale heights. An intermediate value of 3.0 scale heights gives a good reproduction of BA morphology. Since the model bottom layer is located at 7 bar the model is not sensitive to vortices extending deeper in the atmosphere and all the simulations presented in this paper assume a vertical extension of the oval of 3.0 scale heights. 4.3. Results: BA dynamical structure


where n is an integer number that controls the spatial distribution of vorticity in the vortex. At t = 0 it is assumed that the vortex is in geostrophic, hydrostatic balance and therefore velocities are adjusted according to

Du ¼ 


1 @ðDMÞ ; fr @/


One of the goals of our simulations was to reproduce the observed BA morphology in the EPIC potential vorticity (PV) maps. Numerical experiments show that the shape of the vortex depends on the latitude where we locate BA. Fig. 8 shows different simulations with BA located at planetographic latitudes of 32°, 33°, and 34°. When BA is located at planetographic latitudes between 32° and 33°, it maintains an asymmetrical quasi-triangular shape during all the simulation period. This shape is similar to that observed by the Cassini mission at the end of 2000. In contrast, in


E. García-Melendo et al. / Icarus 203 (2009) 486–498

Fig. 7. Tangential velocities of the GRS (left panels) and BA (right panels). Meridional (a) and zonal (b) tangential velocity profiles of the GRS obtained from Galileo data (Choi et al., 2007), and meridional (c) and zonal (d) tangential velocity profiles of BA obtained from Cassini ISS images and detailed on paper II (Hueso et al., 2009). Solid dots are the observational measurements and the continuous thin line is the fitted model in a least squares sense according to expressions (8) and (9) and considering for the GRS n = 3 (a) and n = 2 (b) and n = 1 for BA (c and d). For BA the models compare well with the shape and intensity of the motions but cannot reproduce the peaked values found in the measurements.

Table 5 Vortices least squares fit parameters. Vortex


Semiaxis (°)

Maximum tangential velocity* (m s1)

GRS (E–W) GRS (N–S) BA (E–W) BA (N–S)

3 2 1 1

as = 7.4 bs = 4.7 as = 3.9 bs = 2.9

82 128 70 78


The speeds in this table are the tangential velocity of the vortex as defined in Eqs. (8) and (11) which does not take into account the interaction with the environment. When the vortex is initialised in that way in a region with zonal winds the initial vortex evolves quickly towards a superimposition of the vortex and the zonal wind motions. This effectively produces circulation pattern around BA on the order of 100 m s1.

Section 3.4 it was shown that BA’s shape has evolved towards a more symmetrical structure. This elliptical shape is found in simulations of BA when the vortex is located at planetographic latitude 34°. As shown in Section 3.2 and Figs. 3c and 4, BA probably migrated in latitude to readjust to the ambient zonal flow from 32° to planetographic latitudes between 33° and 34°. The simulation results summarised in Fig. 8 indicate that as BA moves slightly towards the south approaching the measured latitudes between 2005 and 2008, its shape becomes more symmetrical, which is fully consistent with the observed evolution of BA morphology and its position in latitude. The evolution of BA’s shape and its migration towards the south is consistent with some simplified shallow water vortex models (Marcus, 1993). Depending on the flow model (finite or infinite Rossby deformation radius LD), and the background zonal wind profile, stable anticyclones may display a triangular-like asymmetric shape as the one displayed by BA during the Cassini flyby. This happens in sit-

uations where the vortex fills the last closed streamline, and its north or south extremes enter a region where ambient vorticity has a different sign from the potential vorticity of the vortex itself. In the case of BA, as it slightly migrates to the south, the portion of the vortex in adverse ambient vorticity may become smaller and shape the vortex towards a more symmetric morphology. Regardless of the physical reasons for BA’s colour changes, if we assume that chromophores are passive tracers which tint the clouds, their distribution can be represented by the potential vorticity field if the weather layer is approximately adiabatic and inviscid. Our numerical simulations of BA show a potential vorticity field with an annular structure concentric with the anticyclone centre mimicking the colour distribution in BA’s interior. An example is shown on Fig. 9 for an initial vortex placed at 33° planetographic latitude; in this case the ring corresponds to a local minimum of PV. According to the series in Fig. 8, the PV field inside BA homogenises as the vortex evolves with increasing simulation time due to numerical dissipation, but in all simulations, the PV field inside the vortex has a maximum at the vortex centre and the lines of constant potential vorticity form a concentric ringed structure (Fig. 9). This raises the question, if PV has a role in pigment distribution, of how PV gradients may prevent lateral mixing. If PV is responsible for a stable chromophore confinement in an annulus, there must be other processes not reflected in the numerical simulations which must maintain the PV field. Some other specific aspects of the surrounding cloud morphology are also well-reproduced by our numerical models. For example, simulations show the formation of a cyclonic re-circulation present on the east side of BA’s northern half and a high albedo filament cloud which goes over the north tip of the vortex at the stagnation point.

E. García-Melendo et al. / Icarus 203 (2009) 486–498


Fig. 8. Potential vorticity maps for BA when placed at the 32°, 33, and 34° planetographic latitudes after 30, 60, 120, and 150 simulation days. The vortex becomes more symmetric as it is placed more southwards. Part of the GRS is seen in some of the snapshots.

Fig. 9. Comparison of BA structure between simulations and observations. Potential vorticity map (units of 105 m2 K s1 kg1) at the 680 mbar level from EPIC numerical simulations compared with a false colour image obtained taken on 8 April 2006 by the ACS/HST (A. Simon-Miller, I. de Pater and M. Wong, HST Proposal 10783).

4.4. Results: BA and GRS interactions A second goal of the simulations was to explore the outcome of interactions between BA and the GRS during GRS passages. If the interaction is weak, it might be noted in the zonal velocities of BA and the GRS and in their PV fields at cloud level when comparing PV-maps of the vortices when they are well separated and when they are close. If the interaction is strong, it could result in full or partial merging of the smaller anticyclone, as it occurred, for example, with a large anticyclone partially merged when it encountered the GRS (Sánchez-Lavega et al., 1998). A simulation where BA and the GRS are introduced at their observed latitudes and are allowed to freely evolve is summarised on

Fig. 10. Fig. 10a shows the morphology of the vortices a few days before their mutual passage. Fig. 10b plots the velocity drift rate of both vortices as a function of time. Open symbols represent the drift rates of both vortices when they are introduced simultaneously and the vertical line marks the passage epoch. Solid symbols represent the drift rates when the vortices are introduced individually in the simulation channel, i.e. when BA and GRS are isolated vortices. These simulations do not show any effect on the drift rates of both vortices. This result is fully consistent with our analysis of the observations from which no interaction could be detected. A strong interaction between GRS and BA was speculated as possible following BA genesis but in view of the simulations above and the observations both vortices should be closer in latitude to experiment a strong interaction. The centres of the Great Red Spot and BA were at the respective planetographic latitudes of 23.0 ± 0.3° and 33.7 ± 0.3° during the 2006 conjunction. The southern limit of the GRS in 2006 was at 26.9 ± 0.2°, and the northern one of BA at 29.2 ± 0.2°. The latitude span between these two limits is larger than 2° and the ‘‘impact parameter” between the vortices centres is about 10° in latitude. Clearly, to start any physical interaction between both vortices there should be a migration in latitude of BA, the GRS or both. However any large latitudinal migration of the vortices is unlikely to occur due to their constrained meridional motions by the alternating opposed jets that bound the anticyclonic domains where they sit. For example, the GRS latitude is stationary, at least since 1952, when precise measurements could be done on photographs of sufficient resolution. The average latitude of the GRS centre was at 22.3 ± 0.3° between 1952 and 1990 (Rogers, 1995) and 22.6 ± 0.3 between


E. García-Melendo et al. / Icarus 203 (2009) 486–498

Fig. 10. (a) Potential vorticity map (units of 105 m2 K s1 kg1) at the 680 mbar level from the EPIC numerical simulation of the interaction between the GRS and BA. (b) Simulated drift rates of BA and the GRS. Solid symbols represent the velocities when BA and GRS are introduced in the model as isolated vortices. Empty symbols show the drift rates obtained when both vortices are introduced simultaneously and can interact. The GRS passage in the simulation is marked by the vertical line.

1996 and 2000 (Legarreta and Sánchez-Lavega, 2005). Although BA has migrated southwards since its formation, this migration is small and does not affect the separation distance between both vortices. We conducted EPIC experiments where Oval BA was placed at different latitudes to explore its behaviour at different points of the zonal wind profile. We run multiple simulations of a single vortex with the physical properties described in Section 4.2 introduced at the planetographic latitudes of 30°, 31°, 32°, and 33° in an 80  40° channel at the standard 0.31 deg pix1 resolution. If BA drifted to the north 2° or 3° from its natural position to approach the GRS, its centre would enter the cyclonic domain and it would quickly break apart as Fig. 11 shows when BA is introduced at 30°; the same happens when the initial latitude is 31°. In this scenario part of the vortex remains in its natural anticyclonic region but the rest migrates quickly to the north to form another vortex in the closest anticyclonic domain. These simulations indicate that no strong interaction between the GRS and BA is possible. Both vortices are confined at their latitudes with very small excursions allowed by the zonal wind system and a strong interaction between both vortices can be ruled out. This is the expected result in shallow water simulations of vortices. When a vortex of constant potential vorticity q is placed in a zonal flow u with adverse ambient vorticity r(y) = ou/oy, i.e. when qr < 0, the vortex is quickly destroyed (Marcus, 1993). If the ambient vorticity is sufficiently strong compared to q, vortices break apart in potential vorticity filaments and patches. Potential vorticity filaments numerically dissipate and patches migrate towards regions of prograde vorticity (qr > 0) where they settle as stable vortices.

5. Conclusions Our long-term monitoring of BA dynamical properties does not reveal any detectable interaction with the GRS during passages. BA has undergone two sudden drift velocity changes since its formation in 2000, but none of them can be associated to the GRS influence. Instead, there seems to be a correlation between small latitude changes and the changes in the drift velocity of BA. Additionally, we have detected a 160 day longitude oscillation in the centre of BA with peak to peak amplitude of 1.2°. This oscillation is similar the oscillatory behaviour detected for the GRS (Solberg, 1969; Trigo et al., 2000) and the vortices in Neptune (Ingersoll et al., 1995). Further monitoring during the next several years is recommended to check the long-term behaviour of this oscillation and its properties such as its amplitude and phase. Our numerical simulations reproduce reasonably well BA’s cloud morphology. The changes in the morphology of the oval seem associated to be associated with the small southward migration of the oval. The PV distribution in the inner part of BA of concentric rings mimics the annular structure seen when BA turned red, and suggests that the red chromophore distribution might be following approximate iso-potential vorticity trajectories. Red chromophores are probably dynamically confined in a PV-annulus region. The red colour change is not related to the absorption by BA of any red feature outside the vortex. Entrainment of red material from the GRS by BA across their peripheries during close passage can also be ruled out since we do not observe colour changes following any of the passages and numerical experiments support the observed lack of any type of interaction (weak or strong) between GRS and BA.

E. García-Melendo et al. / Icarus 203 (2009) 486–498


Acknowledgments We thank two anonymous referees for their useful suggestions, which helped to improve the final version of this paper. This work has been funded by Spanish MEC AYA2006-07735 with FEDER support and Grupos Gobierno Vasco IT-464-07. RH acknowledges a ‘‘Ramón y Cajal” contract from MEC. This research made use of the computing facilities at CESCA in Barcelona with the help of the Ministerio de Educación y Ciencia. References

Fig. 11. Temporal sequence of an EPIC simulation of vortex BA placed at a 30° latitude within the ambient wind profile in the contiguous cyclonic domain. The panels show potential vorticity maps (units of 105 m2 K s1 kg1). The simulation shows that the vortex breaks apart in two patches of potential vorticity which migrate to the nearest regions of anticyclonic vorticity where they settle as stable vortices.

Baumgardner, J., Mendillo, M., Wilson, J.K., 2000. A digital high-definition imaging system for spectral studies of extended planetary atmospheres. I. Initial results in white light showing features on the hemisphere of Mercury unimaged by Mariner 10. Astron. J. 119, 2458–2464. Breger, M., Stich, J., Garrido, R., Martin, B., Jiang, S., Li, Z.P., Hube, D.P., Ostermann, W., Paparo, M., Scheck, M., 1993. Nonradial pulsation of the Delta-Scuti Star BuCancri in the Praesepe cluster. Astron. Astrophys. 271, 482–486. Cano, J.A., 1998. L.A.I.A.: Laboratorio de Análisis de Imágenes Astronámicas.Grup d’Estudis Astronòmics, Barcelona. Cheng, A.F., and 14 colleagues, 2008. Changing characteristics of Jupiter’s little red spot. Astron. J. 135, 2446–2452. Choi, D.S., Banfield, D., Gierasch, P., Showman, A.P., 2007. Velocity and vorticity measurements of Jupiter’s Great Red Spot using automated cloud feature tracking. Icarus 188, 35–46. Christou, J.C., 1991. Image quality, tip-tilt correction, and shift-and-add infrared imaging. Publ. Astron. Soc. Pac. 103, 1040–1048. Conrath, B.J., Flasar, F.M., Pirraglia, J.A., Gierasch, P.J., Hunt, G.E., 1981. Termal structure and dynamics of the jovian atmosphere. 2. Visible cloud features. J. Geophys. Res. 86, 8769–8775. Dantowich, R.F., Scott, W.T., Kozubal, M.J., 2000. Ground-based high-resolution imaging of Mercury. Astron. J. 119, 2455–2457. Davies, M.E., Abalakin, V.K., Bursa, M., Lederle, T., Lieske, J.H., Rapp, R.H., Seidelman, P.K., Sinclair, A.T., Tejfel, V.G., Tjuflin, Y.S., 1986. Report of the IAU IAG COSPAR working group on cartographic coordinates and rotational elements of the planets and satellites: 1985. Celest. Mech. 39, 102–113. Deeming, T.J., 1975. Fourier analysis with unequally-spaced data. Astrophys. Space Sci. 36, 137–158. Dowling, T.E., Fisher, A.S., Gierasch, P.J., Harrington, J., LeBeau, R.P., Santori, C.M., 1998. The explicit planetary isentropic-coordinate (EPIC) atmospheric model. Icarus 132, 221–238. García-Melendo, E., Sánchez-Lavega, A., Gómez, J.M., Lecacheux, J., Colas, F., Miyazaki, I., Parker, D., 2000. Long-lived vortices and profile changes in the 23.7°N high-speed jovian jet. Icarus 146, 514–524. García-Melendo, E., Sánchez-Lavega, A., 2001. A study of the stability of jovian zonal winds from HST images: 1995–2000. Icarus 152, 316–330. García-Melendo, E., Sánchez-Lavega, A., Dowling, T.E., 2005. Jupiter’s 24°N highest speed jet: Vertical structure deduced from nonlinear simulations of a largeamplitude natural disturbance. Icarus 176, 272–282. García-Melendo, E., Sánchez-Lavega, A., Hueso, R., 2007. Nonlinear simulations of Saturn’s long-lived anticyclones. Icarus 191, 665–677. Hueso, R., Legarreta, J., García-Melendo, E., Pérez-Hoyos, S., Sánchez-Lavega, A., 2009. The jovian anticyclone BA: II. Circulation and models of its interaction with the zonal jets. Icarus. 203, 499-515. Ingersoll, A.P., Barnet, C.D., Beebe, R.F., Flasar, F.M., Hinson, D.P., Limaye, S.S., Sromovsky, L.A., Suomi, V.E., 1995. Dynamic meteorology of Neptune. In: Cruikshank, D.P. (Ed.), Neptune and Triton. University of Arizona Press, Tucson, pp. 613–682. Kuschnig, R., Weiss, W.W., Gruber, R., Bely, P.Y., Jenkner, H., 1997. Microvariability survey with the hubble space telescope fine guidance sensors. Exploring the instrumental properties. Astron. Astrophys. 328, 544–550. Legarreta, J., Sánchez-Lavega, A., 2005. Jupiter’s cyclones and anticyclones vorticity from Voyager and Galileo images. Icarus 174, 178–191. Legarreta, J., Sánchez-Lavega, A., 2008. Vertical structure of Jupiter’s troposphere from nonlinear simulations of long-lived vortices. Icarus 196, 184–201. LeBeau, R.P., Dowling, T.E., 1998. EPIC simulations of time-dependent, threedimensional vortices with application to Neptune’s Great Dark Spot. Icarus 132, 239–265. Lenz, P., Breger, M., 2005. Period04 user guide. CoAst 146, 53–136. Limaye, S.S., 1986. Jupiter: New estimates of the mean zonal flow at the cloud level. Icarus 65, 335–352. Lindal, G.F., and 11 colleagues, 1981. The atmosphere of Jupiter: An analysis of the Voyager radio occultation measurements. J. Geophys. Res. 86, 8721–8727. Marcus, Philip S., 1993. Jupiter’s Great Red Spot and other vortices. Annu. Rev. Astron. Astrophys. 31, 523–573. Morales-Juberías, R., Sánchez-Lavega, A., 2002. A comparative study of anticyclone properties from a six-year (1994–2000) survey. Icarus 157, 76–90. Morales-Juberías, R., Sánchez-Lavega, A., Dowling, T., 2003. EPIC simulations of the merger of Jupiter’s White Ovals BE and FA: Altitude dependent behavior. Icarus 166, 63–74. Naeye, R., 2006. A new red spot. Sky Telescope 111 (6), 18.


E. García-Melendo et al. / Icarus 203 (2009) 486–498

Pérez-Hoyos, S., Sánchez-Lavega, A., Hueso, R., García-Melendo, E., Legarreta, J., 2009. The jovian anticyclone BA: III. Colour change and aerosol properties. Icarus, 203, 516-530. Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P., 2007. Numerical Recipes: The Art of Scientific Computing. Cambridge University Press, Cambridge. Porco, C., and 23 colleagues, 2003. Cassini imaging of Jupiter’s atmosphere, satellites, and rings. Science 299, 1541–1547. Rogers, J.H., 1995. The Giant Planet Jupiter. Cambridge University Press, New York. Sánchez-Lavega, A., Hueso, R., Lecacheux, J., Colas, F., Rojas, J.F., Gomez, J.M., Miyazaki, I., Parker, D., 1998. Dynamics and interaction between a large-scale vortex and the Great Red Spot in Jupiter. Icarus 136, 14–26. Sánchez-Lavega, A., Rojas, J.F., Hueso, R., Lecacheux, J., Colas, F., Acarreta, J.R., Miyazaki, I., Parker, D., 1999. Interaction of jovian White Ovals BC and DE in 1998 from Earth-based observations in the visual range. Icarus 142, 116–124. Sánchez-Lavega, A., Rojas, J.F., Sada, P.V., 2000. Saturn’s zonal winds at cloud level. Icarus 147, 405–420. Sánchez-Lavega, and 12 colleagues, 2001. The merger of two giant anticyclones in the atmosphere of Jupiter. Icarus 149, 491–495. Scargle, J.F., 1982. Studies in astronomical time series analysis, II. Statistical aspects of spectral analysis of unevenly spaced data. Astrophys. J. 263, 835–853.

Seidelmann, P.K., and 14 colleagues, 2007. Report of the IAU/IAG working group on cartographic coordinates and rotational elements: 2006. Celestial Mech. Dyn. Astr. 98, 155–180. Simon-Miller, A.A., Chanover, N.J., Orton, G.S., Sussman, M., Tsavaris, I.G., Karkoschka, E., 2006. Jupiter’s White Oval turns red. Icarus 185, 558–562. Solberg, H.G., 1969. A 3-month oscillation in the longitude of Jupiter’s Red Spot. Planet. Space Sci. 17, 1573–1580. Stratman, P.W., Showman, A.P., Dowling, T.E., Sromovsky, L.A., 2001. EPIC simulations of bright companions to Neptune’s Great Dark Spots. Icarus 151, 275–285. Trigo, J.M., Sánchez-Lavega, A., Gómez, J.M., Lecacheux, J., Colas, F., Miyazaki, I., 2000. The 90 day oscillation of Jupiter’s Great Red Spot revisited. Planet. Space Sci. 48, 331–339. Wöhler, C., Lena, R., Lazzarotti, P., Phillip, J., Wirths, M., Pujic, Z.Geologic lunar research (GLR) group, 2006. A combined spectrophotometric and morphometric study of the lunar mare dome fields near Cauchy, Arago, Hortensius, and Milichius. Icarus 186, 237–264. Youssef, A., Marcus, P.S., 2003. The dynamics of jovian White Ovals from formation to merger. Icarus 162, 74–93.