Origin of bright ring-shaped craters in radar images of Venus

The surface of Venus viewed in Arecibo radar images has a small population of bright ring-shaped features. These features are interpreted as the rough or blocky deposits surrounding craters of impact or volcanic origin. Population densities of these bright ring features are small compared with visually identified impact craters on the surface of the Moon and volcanic craters on Io. However, they are comparable to the short-lived radar-bright haloes associated with ejecta deposits of young craters on the Moon. This suggests that bright radar signatures of the deposits around Venusian craters are obliterated by an erosional or sedimentary process. We have evaluated the hypothesis that bright radar crater signatures were obliterated by a global mantle deposited after impacts of very large bolides. The mechanism accounts satisfactorily for the population of features with internal diameters greater than 64 km. The measured population of craters with internal diameters between 32 and 64 km is difficult to account for with the model but it may be underestimated because of poor radar resolution (5 to 20 km). Other possible mechanisms for the removal of radar bright crater signatures include in situ chemical weathering of rocks and mantling by young volcanic deposits. All three alternatives may be consistent with existing radar roughness and cross-section data and Venera 8, 9, and 10 data. However, imaging observations from a lander on the rolling plains or lowlands may verify or disprove the proposed global mantling. New high-resolution ground-based radar data can also contribute new information on the nature and origin of these radar bright ring features.     

Shallow radar (SHARAD) sounding observations of the Medusae Fossae Formation, Mars

The SHARAD (shallow radar) sounding radar on the Mars Reconnaissance Orbiter detects subsurface reflections in the eastern and western parts of the Medusae Fossae Formation (MFF). The radar waves penetrate up to 580 m of the MFF and detect clear subsurface interfaces in two locations: west MFF between 150 and 155° E and east MFF between 209 and 213° E. Analysis of SHARAD radargrams suggests that the real part of the permittivity is ∼3.0, which falls within the range of permittivity values inferred from MARSIS data for thicker parts of the MFF. The SHARAD data cannot uniquely determine the composition of the MFF material, but the low permittivity implies that the upper few hundred meters of the MFF material has a high porosity. One possibility is that the MFF is comprised of low-density welded or interlocked pyroclastic deposits that are capable of sustaining the steep-sided yardangs and ridges seen in imagery. The SHARAD surface echo power across the MFF is low relative to typical martian plains, and completely disappears in parts of the east MFF that correspond to the radar-dark Stealth region. These areas are extremely rough at centimeter to meter scales, and the lack of echo power is most likely due to a combination of surface roughness and a low near-surface permittivity that reduces the echo strength from any locally flat regions. There is also no radar evidence for internal layering in any of the SHARAD data for the MFF, despite the fact that tens-of-meters scale layering is apparent in infrared and visible wavelength images of nearby areas. These interfaces may not be detected in SHARAD data if their permittivity contrasts are low, or if the layers are discontinuous. The lack of closely spaced internal radar reflectors suggests that the MFF is not an equatorial analog to the current martian polar deposits, which show clear evidence of multiple internal layers in SHARAD data.     

Arecibo radar observations of Rhea, Dione, Tethys, and Enceladus

We have measured the bulk radar reflectance properties of the mid-size saturnian satellites Rhea, Dione, Tethys, and Enceladus with the Arecibo Observatory's 13 cm wavelength radar system during the 2004 through 2007 oppositions of the Saturn system. Comparing to the better studied icy Galilean satellites, we find that the total reflectivities of Rhea and Tethys are most similar to Ganymede while Dione is most similar to Callisto. Enceladus' reflectivity falls between those of Ganymede and Europa. The mean circular polarization ratios of the saturnian satellites range from ∼0.8 to 1.2, and are on average lower than those of the icy Galilean satellites at this wavelength although still larger than expected for single reflections off the surface. The ratio for the trailing hemisphere of Enceladus may be the exception with a value ?0.56. The 13 cm wavelength radar albedos and polarization ratios may be systematically lower than similar results from the Cassini orbiter's RADAR instrument at 2.2 cm wavelength [Ostro, S.J., and 19 colleagues, 2006. Icarus 183, 479-490]. Overall, these reflectivities and polarization properties, together with the shapes of the echo spectra, suggest subsurface multiple scattering to be the dominant reflection mechanism although operating less efficiently than on the large icy moons of Jupiter. All these saturnian moons and icy jovian moons are atmosphere-less, low temperature water ice surfaces, and any differences in radar properties may be indicative of differences in composition or the effects of various processes that modify the regolith structure. The degree of variation in radar properties with wavelength on each satellite may constrain the thickness and efficiency of the scattering layer.     

Solar radar astronomy with the low-frequency array

Initial studies of the Sun's corona using a solar radar were done in the 1960s and provided measurements of the Sun's radar cross-section at about 38 MHz. These initial measurements were done at a time when the large-scale phenomenon known as a coronal mass ejection was unknown; however, these data suggest that coronal mass ejections (CMEs) may have been detected but were unrecognized. That solar radar facility, which was located at El Campo, TX, no longer exists. New solar radar investigations are motivated by our modern understanding of CMEs and their effects on the Earth. A radar echo from an Earthward-directed coronal mass ejection may be expected to have a frequency shift proportional to velocity; thus providing a good estimate of arrival time at Earth and the possible occurrence of geomagnetic storms. Solar radar measurements may also provide new information on electron densities in the corona. The frequencies of interest for solar radars fall in the range of about 10–100 MHz, corresponding to the lower range planned for the low-frequency array. In combination with existing or new high-power transmitters, it is possible to use the low-frequency array to re-initiate radar studies of the Sun's corona. In this report, we review the basic requirements of solar radars, as developed in past studies and as proposed for future investigations.     

The surficial nature of lunar swirls as revealed by the Mini-RF instrument

Lunar swirls are optically bright, sinuous albedo features found on the Moon. The Mini-RF synthetic aperture radar on the Lunar Reconnaissance Orbiter has provided a comprehensive set of X- and S-Band radar images of these enigmatic features, including the first radar observations of swirls on the farside of the Moon. A few general remarks can be made about the nature of the lunar swirls from this data set. First, the average radar properties of lunar swirls are identical to nearby non-swirl regions, in both total radar backscatter and circular polarization ratio (CPR). This implies that average centimeter-scale roughness and composition within the high-albedo portions of the swirls do not differ appreciably from the surroundings, and that the high optical reflectance of the swirls is related to a very thin surface phenomenon (less than several decimeters thick) not observable with X- or S-Band radar. Secondly, bright swirl material appears to be stratigraphically younger than a newly discovered impact melt flow at Gerasimovich D. This observation indicates that the swirls are capable of forming over timescales less than the age of the crater. The Mini-RF data set also provides clues to the origin of the lunar swirls. In at least one case, the presence of an enhanced crustal magnetic field appears to be responsible for the preservation of a high-albedo ejecta blanket around an otherwise degraded crater, Descartes C. The degree of degradation of Descartes C suggests it should not be optically bright, yet it is. This implies that the enhanced albedo is related to its location within a magnetic anomaly, and hence supports an origin hypothesis that invokes interaction between the solar wind and the magnetic anomaly.     

Blocky craters: Implications about the lunar megaregolith

Radar, infrared, and photogeologic properties of lunar craters have been studied to determine whether there is a systematic difference in blocky craters between the maria and terrae and whether this difference may be due to a deep megaregolith of pulverized material forming the terra surface, as opposed to a layer of semi-coherent basalt flows forming the mare surface. Some 1310 craters from about 4 to 100 km diameter have been catalogued as radar and/or infrared anomalies. In addition, a study of Apollo Orbital Photography confirmed that the radar and infrared anomalies are correlated with blocky rubble around the crater.Analysis of the radar and infrared data indicated systematic terra—mare differences. Fresh terra craters smaller than 12 km were less likely to be infrared and radar anomalies than comparable mare craters: but terra and mare craters larger than 12 km had similar infrared and radar signatures. Also, there are many terra craters which are radar bright but not infrared anomalies.Our interpretation of these data is that while the maria are rock layers (basaltic flow units) where craters eject boulder fields, the terrae are covered by relatively pulverized megaregolith at least 2 km deep, where craters eject less rocky rubble. Blocky rubble, either in the form of actual rocks or partly consolidated blocks, contributes to the radar and infrared signatures of the crater. However, aging by impacts rapidly destroys these effects, possibly through burial by secondary debris or by disintegration of the blocks themselves, especially in terra regions.PSI Contribution No. 110.     

Finite difference time domain simulation of radar wave propagation through comet nuclei dielectric models

Abstract— The 90 MHz radar‐wave experiment, Comet Nucleus Sounding Experiment by Radiowave Transmission (CONSERT), on board the Rosetta mission (ESA, 2004) is expected to probe the nucleus of the comet 67P/Churyumov‐Gerasimenko (67P/C‐G) to reveal information on its physical properties, chemical composition, and internal structure. This investigation assesses the potential to recognize lithological structure in the comet nucleus with a radar experiment such as CONSERT. Radar simulations at 90 MHz were performed with a finite difference time domain (FDTD) method. The amplitude and losses of the transmitted and reflected electric field components of an incident radar pulse were evaluated as a function of time. Seven different dielectric models of sections of a hypothetical comet nucleus were used, representative of existing theories of comet nuclei. Values of dielectric constant assigned to these models are based on mixing laws for a porous mixture of ice and meteoritic dust, employing laboratory measured values of relative permittivity for mainly chondritic meteorites. Our results confirm that structural differences such as layers or inclusions are discernable from transmitted and reflected radar signals at 90 MHz, the central frequency of the CONSERT instrument. Such simulations help to constrain the ambiguities that might exist in future radar data associated with the nature of the comet nuclei, whether conglomerate or layered in nature.     

Arecibo radar observations of Phobos and Deimos

In November 2005, we observed the moons of Mars using the Arecibo 2380-MHz (13-cm) radar, obtaining a result for the OC radar albedo of Phobos (0.056±0.014) consistent with its previously reported radar albedo and implying an upper bound on its near-surface bulk density of . We detected Deimos by radar for the first time, finding its OC radar albedo to be 0.021±0.006, implying an upper bound on its near-surface density of , consistent with a high-porosity regolith. We briefly discuss reasons for these low radar albedos, Deimos' being possibly the lowest of any Solar System body yet observed by radar.     

THE METEOR FLUX: IT DEPENDS HOW YOU LOOK

In this paper, we use radar observations from a 50 MHz radar stationed near Salinas, Puerto Rico, to study the variability of specular as well as non-specular meteor trails in the E-region ionosphere. The observations were made from 18:00 to 08:00 h AST over various days in 1998 and 1999 during the Coqui II Campaign [Urbina et al., 2000, Geophys. Rev. Lett. 27, 2853–2856]. The radar system had two sub-arrays, both produced beams pointed to the north in the magnetic meridian plane, perpendicular to the magnetic field, at an elevation angle of approximately 41 degrees. The Coqui II radar is sensitive to at least two types of echoes from meteor trails: (1) Specular reflections from trails oriented perpendicular to the radar beam, and (2) scattering, or, non-specular reflections, from trails deposited with arbitrary orientations. We examine and compare the diurnal and seasonal variability of echoes from specular and non-specular returns observed with the Coqui II radar. We also compare these results with meteor head echo observations made with the Arecibo 430 MHz radar. We use common region observations of these three types of meteor echoes to show that the diurnal and seasonal variability of specular trails, non-specular trails, and head echoes are not equivalent. The implications of these results on global meteor mass flux estimates obtained from specular meteor observations remains to be examined.     

Mars surface properties observed by earth-based radar at 70-, 12.5-, and 3.8-cm wavelengths

A review of Mars radar data obtained through the 1973 opposition confirms that the surface of the planet has many diverse characteristics. Analysis of the quasi-specular echo component shows changes in apparent reflectivity of at least 5 to 1. If attributed entirely to variations in surface material, these correspond to dielectric constants between 1.6 and 4.0. Values of rms surface slope on 1- to 100-m scales range from as low as 0.5° in tablelands near Vlles Marineris to more than 3.0° (the upper limit for which these analysis techniques are appropriate) in certain other areas such as inside Coprates Chasma itself. There is weak correlation between the small-scale surface characteristics inferred from radar and those inferred from Mariner 9 images, geologic maps derived from those images, and other remote sensing data sets. Topography, a large-scale surface characteristic for which good correlation exists between radar and other data sets, was not considered in this study. A search for guidelines which would allow extrapolation of radar properties to new areas on the basis of those studied has been singularly unsuccessful. Data obtained during the 1973 opposition at Arecibo, Goldstone, and Haystack Observatories indicate that the scattering behavior of Mars varies little over the 70- to 3.8-cm wavelength range. Comparison of 1971 and 1973 Goldstone results shows no detectable variation with time.     

A comparison of the thermal and radar characteristics of Mars

There is a correlation between Martian thermal inertia and radar cross section data centered on +22° latitude. The correlation is strongest with 70-cm radar, except between longitudes 10 and 90° where there is a slight anticorrelation, and gets progressively weaker at 12.5- and 3.8-cm wavelengths, respectively. A correlation is expected because of the dependence of both properties on density, but an increase in the average particle size of the surface with increasing dielectric constant is also required in order to explain the data. This may take the form of an increased number of small rocks. The anticorrelation may result from either the effects of atmospheric dust on the surface temperature or from the effects on radar of local variations in large-scale roughness or scattering by rocks. The relative behavior between the wavelengths can be understood in terms of appropriately sized rocks which act as radar scatterers. The trend of the correlation agrees with the dichotomy of the planet into two types of terrain, as noted in other remote-sensing data, and is consistent with an erosional versus depositional surface nature. Variations in the surface dielectric constant, inferred from the 3.8-cm radar data, can explain discrepancies between 2.8-cm radio emission observations and a simple model based on the global distribution of thermal inertia and albedo.     

On the feasibility of RADAR detection of high-energy neutrino-induced showers in ice

In this article we try to answer the question whether the radar detection technique can be used for the detection of high-energy-neutrino induced particle cascades in ice. A high-energy neutrino interacting in ice will induce a particle cascade, also referred to as a particle shower, moving at approximately the speed of light. Passing through, the cascade will ionize the medium, leaving behind a plasma tube. The different properties of the plasma-tube, such as its lifetime, size and the charge-density will be used to obtain an estimate if it is possible to detect this tube by means of the radar detection technique. Next to the ionization electrons a second plasma due to mobile protons induced by the particle cascade is discussed. An energy threshold for the cascade inducing particle of 4 PeV for the electron plasma, and 20 PeV for the proton plasma is obtained. This allows the radar detection technique, if successful, to cover the energy-gap between several PeV and a few EeV in the currently operating neutrino detectors, where on the low side IceCube runs out of events, and on the high side the Askaryan radio detectors begin to have large effective volumes.     

Solar, interplanetary, planetary, and related extra-solar system science for LOFAR

The low frequency array (LOFAR) radiotelescope will be a powerful instrument for answering fundamental, unresolved scientific questions concerning solar system radio phenomena and related emissions from nearby stellar systems. This paper reviews the phenomena, emission mechanisms, open scientific questions, and LOFAR's capabilities. LOFAR will detect metric solar radio bursts in the corona and interplanetary medium, out to distances of order 10 solar radii, as well as Jovian radio emissions. Arguments are given that LOFAR may be sufficiently sensitive to detect stellar analoges of solar type II and III bursts, and may detect cyclotron-maser emissions from extra-solar planets. LOFAR may also aid space weather research, by passively detecting coronal mass ejections (CMEs) via scintillation and Faraday rotation effects, or by detecting radar signals bounced off CMEs and coronal density structures if a suitable solar radar is developed.     

Determination of the area and mass distribution of orbital debris fragments

An important factor in modeling the orbital debris environment is the loss rate of debris due to atmospheric drag and luni/solar perturbations. An accurate knowledge of the area-to-mass ratio of debris fragments is required for the calculation of the effect of atmospheric drag. In general, this factor is unknown and assumed values are used. However, this ratio can be calculated for fragments for which changes in the orbital elements due to atmospheric drag as a function of time are known. This is the inverse of the technique used to determine the atmospheric density from the decay of satellites with accurately known area-to-mass ratios. These kinds of propagation programs are routinely used in predicting the decay of an orbiting vehicle. In this work the area-to-mass ratio of about 2600 fragments arising from the breakup of 24 artificial satellites have been determined. An analysis of the data on about 200 objects (rocket bodies, scientific satellites, etc.) with known mass, size, and shape has also been made. The value of the radar cross-section (RCS), as measured by the Eglin radar operating at 70 cm wavelength, has been correlated to the effective area of these objects. The measurements of the area-to-mass ratio of these objects then provide a calibration of the actual to the calculated mass. It has been shown that the debris mean mass, m, is related to the mean effective area, A, by a power law relation, m = k A ^1.86. However, for a given effective area the mass distribution is very broad. Moreover, the cumulative mass distribution, N(>m), can be expressed as N(>m) = D(m + b), where D, b, and c are constants. The asymptotic slope, c, of low intensity explosions is on the average lower than the slope for high intensity explosions, but there is considerable spread of this slope in each class. Part of the flattening, as indicated by the finite value of the parameter, b, can be understood as arising out of the spread in the RCS values due to the tumbling motion of the fragments and effects related to the detectability of the fragment by the Eglin radar. It has been established that the mass in a given breakup calculated using this technique is in good agreement with the expected mass value. These results can be used in modeling the breakups of other artificial earth satellites and safety analysis.     

海洋潮汐模型在浅海区域的应用

海洋潮汐和大气、海洋、海冰之间存在复杂的相互作用,它对地球气候有复杂而深远的影响。海潮对流经大陆沿岸或大陆架的洋流有很强烈的作用。潮汐流产生混合湍动;潮汐耗散和内潮波效应对海洋环流的传输和循环也有一定的影响。1995年前后,使用TOPEX／POSEIDON测高卫星资料。建立了十多个海潮模型。研究表明,1994-1996年期间发展起来的正压波海潮模型在深海的精度为2—3cm,空间分辨率为50km量级,在浅海区域的精度显著下降。近年来运用更加成熟精细的流体动力学理论模型,在数据同化技术中使用时间跨度更长的测高资料,已经建立了一些改进的海潮模型。该文使用验潮站潮汐常数、测高资料以及交叉点资料,评估了6个海潮模型在浅海区域(包括中国海海域)的表现,以应用于今后对海平面的研究。初步分析表明,浅海区域的海平面高度的误差仍然相当显著。要发展海洋潮汐模型需要进一步减小潮汐混淆效应,提高长周期潮汐的精度,尤其在浅海区域。模型的改进必将增进对潮汐现象的认识,促进学科间进行相互融合和相互渗透的研究(例如潮汐摩擦引起的月球自转的长期缓慢减速、地球内部结构的物理学研究等)。     

Radar detection of near-Earth Asteroids 1915 Quetzalcoatl, 3199 Nefertiti, 3757 (1982 XB), and 4034 (1986 PA)

We describe Arecibo (2380 MHz, 12.6 cm) Doppler-only radar detections of near-Earth Asteroids 1915 Quetzalcoatl, 3199 Nefertiti, 3757 (1982 XB), and 4034 (1986 PA) obtained between 1981 and 1989. Estimates of the echo spectral bandwidths, radar cross-sections, and circular polarization ratios of these objects constrain their sizes, radar albedos, surface roughnesses, taxonomic classes, rotation periods, and pole directions. Our radar constraints on the diameters of Quetzalcoatl and Nefertiti are most consistent with sizes determined using thermal-radiometry and the Fast Rotation Model (FRM); this consistency may indicate that these asteroids have surfaces of high thermal inertia (i.e., little or no regolith). Constraints on Quetzalcoatl's radar albedo rule out a “metallic M” classification. The radar constraints for Nefertiti are inconsistent with a rotation pole published by Kaasalainen et al. (2004, Icarus 167, 178). Our estimates of 1982 XB's size are consistent with previously published estimates. The radar bandwidth of 1986 PA places an upper bound of about 24 h on its rotation period.     

A comparison of infrared,radar, and geologic mapping of lunar craters

Between 1000 and 2000 infrared (eclipse) and radar anomalies have been mapped on the nearside hemisphere of the Moon. A study of 52 of these anomalies indicates that most are related to impact craters and that the nature of the infrared and radar responses is compatible with a previously developed geologic model of crater aging processes. The youngest craters are pronounced thermal and radar anomalies; that is, they have enhanced eclipse temperatures and are strong radar scatterers. With increasing crater age, the associated thermal and radar responses become progressively less noticeable until they assume values for the average lunar surface. The last type of anomaly to disappear is radar enhancement at longer wavelengths. A few craters, however, have infrared and radar behaviors not predicted by the aging model. One previously unknown feature - a field strewn with centimeter-sized rock fragments - has been identified by this technique of comparing maps at the infrared, radar, and visual wavelengths.     

Subsurface radar sounding of the martian polar cap: radiative transfer approach

The problem of subsurface radar sounding of the martian polar caps [Ilyushin, 2004. Martian northern polar cap: layering and possible implications for radar sounding. Planet. Space Sci. 52, 1195–1207] is considered from the point of view of incoherent radiative transfer theory. Since it has been previously shown that the radar signal field within the polar cap has diffuse structure, there is a need for a statistical approach to the problem. Radiative transfer theory, which is now well developed, seems to be the most appropriate formalism for this approach.Several physical models of polar caps have been formulated. The asymptotic solutions for all proposed models are derived here. In the present paper only the case of orbital ground penetrating radar is considered, because it is of great interest in relationship to currently developed radar experiments. In principle, the approach is believed to be applicable to a wide class of short pulse and compressed chirp radar experiments, including both orbital and landed instruments and media more complicated than a simple plane parallel geometry. This work, however, is postponed to future papers.Techniques for retrieval of physical properties of polar caps from the radar measurements are proposed. From the observational data, the macroscopic parameters of the medium appearing in radiative transfer theory, i.e. the single scattering albedo and volume extinction coefficient can be estimated. These estimates put certain constraints on the physical parameters of the medium model introduced in the paper. With some additional information, known a priori or from other observations, these estimates can be used to retrieve physically meaningful information, for example, the average content of impurities in the ice.     

Remote sensing of lunar pyroclastic mantling deposits

Mantling deposits on the Moon are considered to be pyroclastic units emplaced on the lunar surface as a result of explosive fire fountaining. These pyroclastic units are characterized as having low albedos, having smooth fine-textured surfaces, and consisting in part of homogeneous, Febearing volcanic glass and partially crystallized spheres. Mantling units exhibit low returns on depolarized 3.8-cm radar maps, indicating an absence of surface scatterers in the 1- to 50-cm-size range. A number of reflectance spectra from several regional pyroclastic deposits are presented for the first time; these data support a previous interpretation that mantling units have a unique spectral signature which is indicative of the presence of a significant Fe-bearing volcanic glass component. The Rima Bode region is discussed as an example of an area in which several types of remote sensing data (including 3.8-cm radar, spectral reflectance, and multispectral vidicon data) were used to reconstruct the geologic events surrounding the emplacement of a regional pyroclastic mantling deposit. The recognition of numerous varieties of volcanic glass samples, especially relatively high-albedo (e.g., green, yellow) glasses, suggests the existence of additional, unrecognized mantling deposits with albedos higher than those studied to date. On the basis of the remote sensing data summarized and presented, five new areas have been identified which may represent higher-albedo regional pyroclastic deposits.     

The effect of gravitational and pressure torques on Titan's length-of-day variations

Cassini radar observations show that Titan's spin is slightly faster than synchronous spin. Angular momentum exchange between Titan's surface and the atmosphere over seasonal time scales corresponding to Saturn's orbital period of 29.5 year is the most likely cause of the observed non-synchronous rotation. We study the effect of Saturn's gravitational torque and torques between internal layers on the length-of-day (LOD) variations driven by the atmosphere. Because static tides deform Titan into an ellipsoid with the long axis approximately in the direction to Saturn, non-zero gravitational and pressure torques exist that can change the rotation rate of Titan. For the torque calculation, we estimate the flattening of Titan and its interior layers under the assumption of hydrostatic equilibrium. The gravitational forcing by Saturn, due to misalignment of the long axis of Titan with the line joining the mass centers of Titan and Saturn, reduces the LOD variations with respect to those for a spherical Titan by an order of magnitude. Internal gravitational and pressure coupling between the ice shell and the interior beneath a putative ocean tends to reduce any differential rotation between shell and interior and reduces further the LOD variations by a few times. For the current estimate of the atmospheric torque, we obtain LOD variations of a hydrostatic Titan that are more than 100 times smaller than the observations indicate when Titan has no ocean as well as when a subsurface ocean exists. Moreover, Saturn's torque causes the rotation to be slower than synchronous in contrast to the Cassini observations. The calculated LOD variations could be increased if the atmospheric torque is larger than predicted and or if fast viscous relaxation of the ice shell could reduce the gravitational coupling, but it remains to be studied if a two order of magnitude increase is possible and if these effects can explain the phase difference of the predicted rotation variations. Alternatively, the large differences with the observations may suggest that non-hydrostatic effects in Titan are important. In particular, we show that the amplitude and phase of the calculated rotation variations are similar to the observed values if non-hydrostatic effects could strongly reduce the equatorial flattening of the ice shell above an internal ocean.