PDS_VERSION_ID = PDS3 LABEL_REVISION_NOTE = "Michael Bird & Lyle Huber, June 2002; RDR, 2006-04-01" RECORD_TYPE = STREAM OBJECT = INSTRUMENT INSTRUMENT_HOST_ID = "HP" INSTRUMENT_ID = "DWE" OBJECT = INSTRUMENT_INFORMATION INSTRUMENT_NAME = "DOPPLER WIND EXPERIMENT" INSTRUMENT_TYPE = "RADIO SCIENCE" INSTRUMENT_DESC = " Instrument Overview =================== See [BIRDETAL1997]. The Doppler Wind Experiment (DWE) is a high-precision tracking investigation to determine wind direction and magnitudes in Titan's atmosphere. The prime science objective, a height profile of the wind velocity, will be derived from the Doppler shift of the Probe Relay Link (PRL) signal from the Huygens Probe to the Cassini Orbiter. ------Update 06/04/01, RDR--------- Due to a flaw in the sequence to be executed by Cassini during the Huygens mission, the Doppler shift could not be measured aboard Cassini. It was, however, also measured on Earth [BIRDETAL2005]. The prime science objective was achieved using those data. -------end of Update------------- After correcting for all known Doppler shifts due to orbit and propagation effects, the wind-induced motion of the Probe will be determined to an accuracy of better than 1 m/s, from parachute deployment at an altitude of ~ 160 km down to surface impact. In addition to the measurements of drift motions due to winds, DWE is capable of achieving two secondary scientific objectives: (a) measurement of Doppler fluctuations to determine the level and spectral index of turbulence and possible wave activity in Titan's atmosphere (b) measurement of Doppler and signal level modulation to monitor Probe descent dynamics, including its rotation rate and phase, parachute swing and post-impact status. By achieving the first of these objectives, DWE will contribute, along with Probe accelerometry and radio occultation measurements, to the assessment of atmospheric turbulence associated with vertical wave propagation, the buoyantly-driven surface layer or possible methane moist convection. The second objective represents a potential DWE 'service' to other Probe instruments to assist in the interpretation of their data. The known Probe-Orbiter geometry enables a determination of directions on Titan from the phase of the Doppler/amplitude modulation. The Probe's passively-controlled spin rate will also be determined to a high degree of accuracy from the same data. These precise measurements of velocity will be integrated to reconstruct the descent flight path, thereby providing the most accurate determination of the Probe's impact coordinates. The most severe constraints on the accuracy of the DWE wind measurement are trajectory errors and instability of the Probe oscillator used to generate the PRL signal. Present assessments of these limitations indicate that a zonal wind height profile u(z) can be recovered with a mean error delta u of less than +-1 m/s. This can be achieved only with a sufficiently stable PRL signal over the duration of the descent (delta f <~ 0.4 Hz at S-band) in order to exclude contamination of the measurement by oscillator drift. The transmitter's frequency stability is guaranteed by using an ultrastable oscillator (USO) to generate the PRL carrier signal. In addition to this transmitter USO (TUSO), it is necessary to incorporate an additional unit in the receiver (RUSO) of the Probe Support Avionics (PSA) on the Cassini Orbiter, where the frequency measurement is recorded. Scientific Objectives ===================== Expectations based on Voyager results ------------------------------------- Almost all presently available information about Titan's atmospheric dynamics and meteorology derives from Voyager 1's reconnaissance, with essential contributions from the infrared interferometry spectrometer (IRIS) and radio occultation (RSS) investigations [HANELETAL1981; TYLERETAL1981]. Titan's vertical temperature-pressure structure, retrieved from Voyager RSS data [LINDALETAL1983], is represented by the profile shown in Fig. 1. The occultation data extend all the way to the surface, where the temperature is approximately 97K (with a systematic uncertainty of about +-7K). Methane clouds may form above a few kilometres, depending on the relative humidity near the surface [TOONETAL1988]. The decrease of temperature with altitude below 40 km implies the absorption of sunlight at the surface and a weak tropospheric greenhouse. The radio occultation measurements indicate a nearly adiabatic lapse rate below 3-4 km, consistent with the inferred turbulent surface layer, but a statically stable profile above this height. [HINSON&TYLER1983] have found evidence for vertically propagating gravity waves at altitudes of 25-90 km, based on their analysis of radio occultation scintillations. The indirect inference of 100 m/s zonal cyclostrophic winds on Titan is based primarily upon the latitudinal contrast of temperature at infrared sounding levels observed by Voyager's IRIS [FLASARETAL1981; FLASAR&CONRATH1990]. As interpreted by the thermal wind equation under the assumption of hydrostatic gradient-balanced flow, the vertical variation of the Coriolis plus centripetal acceleration of the zonal velocity u is related to the latitudinal temperature gradient dT/dlambda by d/dz^ [u^2 tan lambda + 2 u Omega a sin lambda] = -R d/dlambda [T / mu] (1) where z^ = ln(P0/P) is the vertical log-pressure coordinate with P0 the surface pressure, lambda the latitude, a the planetary radius, Omega the planetary rotation velocity, mu the mean molecular weight (mass per mole) of the atmospheric gas, and R the gas constant. With the further plausible assumption of relatively weak winds near the surface [ALLISON1992], the thermal wind equation may be solved for the zonal velocity at all levels for which there are vertically continuous observations of the horizontal temperature gradient. The second term in the thermal wind equation (1) can be neglected, except very close to the surface, because of the slow Titan angular rotation. The resulting circulation is in cyclostrophic balance with Omega a / u << 1. However, the direction of the zonal wind (prograde or retrograde) cannot be uniquely determined from (1) under this condition. The vertical integration of the thermal wind equation for the practical inference of zonal motions requires the independent specification of the velocity at one or more levels where the thermal gradient can be accurately estimated. For sufficiently strong surface drag, the zonal velocity at the bottom of the atmosphere may be assumed to be nearly zero. On this basis, [FLASARETAL1981] estimated a zonal wind approaching 100 m/s in the upper stratosphere at a latitude of lambda = 45 deg. Although Titan's thermal spectrum is partly convolved with variable aerosol opacity, the 1304 cm^-1 region, which comes mainly from the 0.5 mbar pressure level (~ 230 km), is presumed to be relatively free of these effects. Measured brightness variations across the disc in this channel could thus be interpreted as a real difference in kinetic temperature between low and high latitudes of roughly 16K [FLASARETAL1981]. More recently, [FLASAR&CONRATH1990] retrieved stratospheric temperatures from a combination of the measured radiances at 1304 cm^-1 and the P and Q branches of methane at 1260-1292 cm^-1, essentially confirming the results of the earlier brightness temperature analysis. The data also indicated a hemispheric asymmetry in Titan's latitudinal temperature distribution at the time of the Voyager encounters. [COUSTENIS1991] has performed an independent retrieval of the latitudinal temperature structure as inferred from the 1304 cm^-1 channel, with similar results. IRIS measurements for a second thermal sounding channel at 200 cm^-1, corresponding to emission in the vicinity of 100 mbar (~40 km), suggest that latitudinal thermal contrasts at this level may be less than about 1K. Measurements for a third thermal channel at 530 cm^-1, corresponding to emission near Titan's surface, were interpreted by [FLASARETAL1981] as a latitudinal temperature difference of about 2K between the equator and 60 deg latitude. This interpretation has since been qualified, however, by the study of [TOONETAL1988], who pointed out that stratospheric aerosols may also contribute significantly to the brightness temperature at this wavenumber. The apparent latitudinal contrast may thus be best interpreted as a crude upper limit to actual variations in the kinetic temperature. A more comprehensive review of the evidence supporting our contemporary model of Titan's zonal winds has been compiled by [FLASARETAL1997]. As with Venus, the inferred cyclostrophic flow regime on Titan is not yet understood, representing a fundamental unresolved problem in the theory of atmospheric dynamics. The magnitudes of the meridional and vertical winds are also quite speculative. Assuming the poleward advection of heat is balanced by radiative cooling, [FLASARETAL1981] estimated a mean meridional motion of v<=0.04 cm/s in the lower troposphere. Invoking continuity, the associated mean vertical motion may be estimated as omega ~ vH/a <= 10^-3 cm/s, where H ~ 20 km is the pressure scale height. It is possible that much stronger vertical motions exist in locally turbulent regions of the atmosphere, such as a methane thunderhead, but probably only over a relatively small fraction of the total area. The associated depth of the Ekman planetary boundary layer (PBL), where the surface winds are rotated and strongly sheared with altitude until they match the thermal wind aloft, may be estimated as D_E ~ 0.7 km [ALLISON1992]. Because the PBL mediates the transfer of angular momentum from the surface to the atmosphere, it is an important target for observational characterisation. With strong surface drag and a global meridional thermal contrast no larger than assumed by [FLASARETAL1981], the vertically integrated thermal wind equation implies a geostrophic regime (Omega a/u >> 1) extending up to about 5 km altitude. In view of the various gaps and uncertainties in the thermal structure data, we cannot claim to know very much about the vertical and horizontal structure of Titan's zonal wind. Assuming that at least the 1304 cm^-1 measurements from Voyager's IRIS have been more or less correctly interpreted, however, it is difficult to escape the conclusion that the thermal wind balance will involve zonal motions of the order of (R DeltaT/mu)^0.5 ~ 70 m/s at stratospheric levels. The direct observational confirmation of the inferred zonal cyclostrophic motion of Titan's atmosphere by DWE will represent a fundamental contribution to the study of planetary meteorology. If its conjectured analogy to Venus is established, it would imply the probable ubiquity of atmospheric superrotation as a robust feature of slowly rotating, differentially heated planets. Atmospheric model simulations ----------------------------- Only recently has a general circulation model (GCM), originally developed for the simulation of the terrestrial weather and climate, begun to be adapted successfully to the study of atmospheric superrotation on Titan. [DELGENIOETAL1993] have shown that the addition of an optically thick, statically stable upper cloud layer to a terrestrial GCM, also modified with a 16-day planetary rotation period, results in the generation of an equilibrated zonal-mean flow of several tens of metres per second. Their diagnosis of the responsible flux transports indicate that this regime is supported by the horizontal mixing of quasi-barotropic eddies. Similar results have also been obtained with a Titan GCM under development by [HOURDINETAL1995]. Despite their preliminary state of development, these GCM experiments lend some confidence to the indirect inference of superrotational winds from the observations. [ALLISONETAL1994] have argued that the GCM experiments as well as the limited planetary observations are suggestive of the dynamical maintenance of these circulations by efficient 'potential vorticity' mixing. In the implied 'ZPV' (zero potential vorticity) limit for stable zonal flow, the latitudinal wind profile will be bounded by a maximum envelope of the form U_max = (U_e + Omega a) (cos lambda)^[(2/R_i)-1] - Omega a cos lambda (2) where R_i is the local Richardson number (the squared ratio of the Brunt-Vaisala frequency to the vertical wind shear), and U_e is the zonal velocity at the equator. Except where R_i is less than 2, this envelope implies an increase in the maximum possible velocity with latitude, and can therefore be expected to approximate the actual winds only between the equator and the latitudes of the 'jet' maxima. The Titan zonal wind profile presented by [HUBBARDETAL1993] is consistent with this prescribed envelope within 60 deg latitude, assuming a large value for the Richardson number, as appropriate for the statically stable stratosphere. The combination of the vertical wind shear measured by the Doppler Wind Experiment and the static stability inferred from the Huygens Atmospheric Structure Instrument (HASI) will provide an in situ determination for the vertical profile of R_i, which could then be interpreted with the ZPV constraint for comparison with independent remote sensing observations from the Cassini Orbiter. Recent and future ground-based observations ------------------------------------------- Some independent evidence for winds on Titan, based on ground-based stellar occultation measurements, has been presented by [HUBBARDETAL1993]. Their analysis of the indicated latitudinal deformation of isopycnic surfaces in the stratosphere implies a zonal wind at the 0.25 mbar level varying from some 80 m/s near the equator to >170 m/s at 60 deg latitude. Another promising new method for determining stratospheric wind motions is the IR heterodyne observation of Titan's ethane emission at 841 cm^-1, which originates primarily from the 1 mbar level. Initial measurements of the differential Doppler shift between east and west limb spectra indicate that the winds are prograde and have speeds of the order of 80 m/s. Fifty near-IR Titan images were obtained with the Hubble Space Telescope's refurbished planetary camera (WF/PC2) during 4-18 October 1994 [SMITHETAL1994; SMITHETAL1996]. A variety of continuum and methane filters were employed with the goal of obtaining cloud-tracked drift speeds associated with the differential motion of longitudinal structure of the type previously observed at 970 nm by [SMITHETAL1992]. A global surface map compiled from 14 processed images covering nearly the full Titan rotation period reveals a large bright feature in the leading hemisphere centred just south of the equator. Although some structure in the unaveraged images is suggestive of cloud features, attempts to derive cloud-tracked vectors have thus far proved unsuccessful. Further observations, obtained during Saturn's opposition of September/October 1995, are being evaluated [CALDWELLETAL1996]. The DWE will complement remote sensing observations of temperatures and winds from the Cassini Orbiter. It will provide ground-truth corroboration of the thermal wind retrievals from Composite Infrared Spectrometer (CIRS) measurements, also providing a check on their assumed aerosol and cloud opacities. The high vertical resolution of the DWE retrieval of vertical shear near impact will provide an important characterisation of the surface boundary layer unobtainable by thermal sounding. If the probe encounters sufficiently vigorous turbulence or vertical wave propagation, variations of the Doppler signal will provide information on the associated eddy momentum mixing and/or planetary waves. DWE Concept and Mission Planning ================================ Titan targeting --------------- The geometrical configuration and sequence of events during descent are of vital importance to DWE's execution. The present mission baseline [LEBRETON&MATSON1997] plans Probe release on the Orbiter's first inbound pass after Saturn Orbit Insertion (SOI). The nominal date for Probe descent is 27 November 2004, some 150 days after SOI and >7 years after launch. Titan, which orbits Saturn at a constant distance of 20.4 Saturn radii, is very close to Saturn's noon meridian on this date. Probe separation from the Orbiter occurs ~ 22 days before Titan encounter, at which point the Probe is targeted for entry into Titan's atmosphere. The final Orbiter Deflection Manoeuvre is performed 2 days after Probe separation, targeting the Orbiter so that it will fly very nearly over the Probe but at a safe 1500 km above Titan's surface. The Orbiter's closest approach is delayed to about 4 h after the Probe's atmospheric entry. This Orbiter Delay Time (ODT) is short enough to provide adequate margin of the PRL at the beginning of descent (maximum range) and yet long enough to avoid Orbiter High Gain Antenna (HGA) pointing problems toward the end of Huygens' mission (minimum range). The Probe's target is characterised by parameters in Titan's B-plane, defined by the asymptotic Probe approach velocity. Aim points in the B-plane are defined by the magnitude of the impact parameter vector B and its associated azimuthal angle with respect to the T-axis. Equivalently, the atmospheric entry angle gamma at a given altitude could be used instead of the impact parameter B. This alternative was adopted for Huygens mission planning purposes. It is planned to target the Probe at an entry angle gamma = 64 deg, and B-plane azimuth angle theta = -60 deg. The Probe delivery accuracy (3 sigma) in the B-plane is given by the ellipse encircling the tip of the target vector (theta = -60 deg, gamma = 64 deg). The semi-major and semi-minor axes of the ellipse are 452 km in the horizontal (longitudinal) direction and 59 km in the vertical (latitudinal), respectively. The targeted value of gamma, defined as the angle between the nadir direction and the Probe's velocity vector at a reference height of 1270 km, was selected to ensure a safe atmospheric entry and to guarantee successful Probe radio communications via the PRL. The Probe's targeted latitude is ~18 deg N. ------Update 06/04/01, RDR--------- After the discovery of a flaw in the radio system, it became necessary to redesign the targeting strategy. See [LEBRETONETAL2005] for more information. -------end of Update-------------- Owing to the extreme elongation of the error ellipse in the horizontal direction, dispersion in the entry angle is small for values of theta ~ -90 deg, but quite large for theta ~ 0 deg. In order to minimise entry angle dispersion, Huygens' B-plane target azimuth was selected as theta = -80 deg in the early stages of the mission planning process. Considerable trade-off analyses, however, were conducted to address scientific preferences for the impact latitude, solar zenith angle (SZA) and, specifically for DWE, the angle between the east-west direction on Titan and the line-of-sight direction from Probe to Orbiter. This angle, referred to as the Doppler Wind Component (DWC) angle, was very unfavourable to DWE for theta = -80 deg. The cosine of the angle DWC, which is the 'zonal wind projection' (ZWP) onto the PRL ray path, regulates the magnitude of the Doppler shift from zonal winds. A study of the DWE wind recovery algorithm under various Probe/Orbiter geometries showed that a very good representation of the input wind was derived for ZWP > 0.5. This is Region (a). Less precise, but still satisfactory, recoveries could be obtained in Region (b), where 0.3 < ZWP < 0.5. The discrepancy between the input wind and the recovered profile begins to increase dramatically, however, when ZWP < 0.3. The Region (c), where the zonal wind recovery error became unacceptably large, was thus declared a 'zone of avoidance' by DWE. The original Huygens target with gamma = 64 deg, theta = -80 deg was, in fact, located in Region (c). The DWE request to move the B-plane azimuth angle to theta = -60 deg was granted after carefully reassessing the consequences for the overall mission performance. Only a very small portion of the 3 sigma targeting ellipse is now located in Region (c). The calculations of ZWP were performed for the '100% nominal' input wind model (prograde, linearly increasing from zero at the surface to 100 m/s at 200 km altitude; [FLASARETAL1981]). Different contours are obtained for other wind models. A prograde zonal wind significantly shifts the touchdown site to the east of the Probe's atmospheric injection point, thereby improving the ZWP. The situation is the reverse for retrograde winds, which tend to increase the recovery errors. For all reasonable cases tested, however, the recovery algorithm never becomes indeterminate (i.e. when ZWP --> 0). Titan atmospheric descent: PRL Doppler effects ---------------------------------------------- As the Probe enters Titan's atmosphere, it is subjected to a severe deceleration at ~ 250 km altitude that could be as high as 16.1 g. A parachute is deployed near Mach 1.5, marking the beginning of the descent phase (time = t0). After the Probe slows to subsonic velocity, the heat shield is jettisoned (at t0+30 s) and the PRL is established for transmitting data to the Orbiter (no later than t0+150 s). The large initial parachute is released at t0 + 15 min and replaced by a smaller drogue in order to decrease the descent duration. Probe motion during descent --------------------------- The Probe's motion in Titan's atmosphere is determined by the action of two forces: gravity and wind drag. The acceleration due to wind drag a_d is taken from [FLURY1986] to be: a_d = - (rho(z) / 2 C_B) V_PW V_PW^ (3) where rho(z) is the atmospheric density, V_PW = V_P - V_W is the relative velocity of the Probe with respect to the atmospheric wind, and C_B is the ballistic coefficient (units: kg/m^2) defined by: C_B = m / (C_D A) (4) with m = probe mass, C_D = drag coefficient, and A = effective probe area to the flow. The wind velocities in the (x, y, z) directions of a Cartesian coordinate system on Titan's surface are given by: u = zonal wind (x-axis: positive towards east) v = meridional wind (y-axis: positive towards north) w = vertical wind (z-axis: positive upwards) The Probe's equation of motion during the Titan descent phase can thus be written: xdotdot = - (rho / 2 C_B) V_PW (xdot-u) (5) ydotdot = - (rho / 2 C_B) V_PW (ydot-v) (6) zdotdot = - (rho / 2 C_B) V_PW (zdot-w) - g(z) (7) where g(z) is Titan's gravitational acceleration, and V_PW = [(xdot-u)^2 + (ydot-v)^2 + (zdot-w)^2]^0.5 (8) The meridional and vertical winds, v and w, are assumed to be considerably weaker than the zonal wind u. In this case, the latitude of the Probe (y component) remains approximately constant. Knowing the Probe's velocity on Titan, it is not difficult to determine the Doppler shift projected on to the PRL ray path back to the Orbiter: Delta f = - (f / c) Delta V (9) where Delta V = (V_P - V_O) 'dot' DELTA (10) with V_P = Probe velocity wrt Titan centre V_O = Orbiter velocity wrt Titan centre DELTA = unit vector pointing from Orbiter to Probe The vector DELTA defines the line-of-sight of the PRL. Delta V is negative during descent, so that the received frequency is increased (blue shifted). The PRL Doppler shift (9) is ~37.6 kHz for the nominal Orbiter starting approach velocity of -5.53 km/s. The velocity of the Probe projected on to the line-of-sight can be written: DELTA 'dot' V_P = V_1 + V_2 + V_3 + V_4 (11) where V_1 = xdot sin (alpha) cos (beta) (12) V_2 = Omega (a + z) cos (lambda) sin (alpha) cos (beta) (13) V_3 = ydot sin (alpha) sin (beta) (14) V_4 = zdot cos (alpha) (15) Furthermore, we define the projection of the Orbiter velocity on to the line-of-sight by: DELTA 'dot' V_O = V_5 (16) The angles alpha and beta define the direction from the Probe to the Orbiter in a local Cartesian coordinate system on the Probe oriented along a natural Titan coordinate grid (x-axis positive towards east; y-axis positive towards north; z-axis positive upwards). The angle alpha, basically the zenith angle of the Orbiter as seen from the Probe [OTT1991], is also sometimes designated as the 'probe aspect angle' (PAA). The azimuthal angle beta of the Probe-to-Orbiter line-of-sight is labelled the 'line-of-sight azimuth' (LOSA) by some authors [ATKINSON1989; POLLACKETAL1992]. The term V_1 in (12), where the mean value of xdot ~ u, the drift velocity due to zonal winds, is the quantity to be determined by DWE. This velocity is co-aligned with the contribution V_2 in (13) due to Titan rotation. It is assumed for simulation purposes that the rotation is synchronous with Titan's orbital period (Omega a ~ 11.7 m/s). As noted above, the relative importance of V_1, and thus the quality of the reconstructed wind profile, is strongly dependent on the values of alpha and beta. The term V_3 in (14), arising from meridional drift, should be small because ydot is not expected to be important. The term V_4 in (15) contains the vertical descent velocity zdot. This quantity can be obtained to a high degree of accuracy either from the range rates deduced from the Probe's proximity sensor data, or using measurements of temperature T and pressure P from HASI. In the latter case, based on the assumption of hydrostatic equilibrium and ideal gas behaviour, the descent velocity is determined from: zdot = - (H / P) dP/dt (17) where variations in H = RT/mu g, the atmospheric scale height, are assumed to be negligible over the altitude range of Huygens' descent. The final error associated with the determination of the reconstructed Probe descent velocity is estimated to be of the order of 1%. Knowing zdot(t) from the proximity sensor or from HASI measurements, it is possible to reconstruct xdot(t) ~ u(t) from the PRL's Doppler shift (9). Using in situ measurements of the density rho(z), we can then extract the exact height profile u(z) from the motion equations (5) and (7). Comparative wind measurements are expected near the surface from two independent sources on board the Probe: - proximity sensor measurements using pendulum swing motion, or - inference of horizontal motion from the Descent Imager/Spectral Radiometer (DISR). This would provide verification of the DWE height profile for the last few data points before touchdown, when horizontal winds are likely to be weak. If the Probe survives the surface impact and continues to transmit, further DWE measurements might provide a frequency reference for zero wind. The post-impact data are not necessarily a reliable calibration, however, because the Probe oscillator may execute shock-induced frequency shifting at impact. In addition to the derivation of a zonal wind height profile from the large-scale drift, DWE may be able to provide valuable science and navigation services from the small-scale variations in frequency as a by-product of the analysis. Two such possibilities are: 1. knowledge of the zonal wind velocity enables an extremely accurate reconstruction of the flight path during descent, and therefore the best possible determination of the impact coordinates on Titan 2. rotational motion of the Probe during descent will produce a small modulation in the PRL frequency and signal level that will be exploited to derive the Probe's spin rate and spin phase, thereby effectively serving as a compass in the Titan landscape. The amplitude of the Doppler modulation from the rotation is proportional to the spin rate. The ultimate accuracy of the Doppler measurement, limited by the clock digitisation of the PSA receivers, is 60 mHz. This accuracy translates to an estimate of the spin rate error (1 sigma) of delta omega = +-0.27/(omega tau)^0.5 rpm, where omega is the spin rate in rpm and tau is the integration time in minutes. For a representative value of tau = 1 min, this yields a relative error of 27 % just before impact, when the Probe spin rate is expected to be ~1 rpm. The relative error is much smaller at higher altitudes where the spin rate is greater (e.g. ~0.9% for omega = 10 rpm, tau = 1 min). An improvement in this estimate from the Doppler data can be achieved by including the signal level data. As the radiation pattern of the Probe Transmitter Antenna (PTA) is asymmetric, the Probe's rotation phase can be determined from the phase of the received signal level modulation. The amplitude of the PTA's cyclic variation will be ~ +- 2 dB during each rotation. ------Update 06/04/01, RDR--------- The time resolution of the Doppler data obtained on Earth is not sufficient to derive a spin rate. Instead, signal level measurements obtained from channel B on Cassini were used. -------end of Update-------------- Vertical resolution: Probe response time ---------------------------------------- It is important to characterise the wind shear and turbulence in Titan's atmosphere with DWE measurements of the smaller-scale Probe motion due to wind gusts [STROBEL&SICARDY1997]. To a good approximation [ATKINSON1989], the Probe response to a horizontal wind gust of magnitude u0, applied at time t = 0, is given by: xdot(t) = u0 [1 - exp(-g t / V_T)] (18) where V_T is the terminal velocity determined from the force balance (7) between gravity and the drag force (for zdotdot = 0) given by V_T ~ [(2mg) / (C_D A rho)]^0.5 (19) Referring to (18), the Probe evidently adjusts itself to the wind on time scales of the order of: tau = V_T / g (20) The relaxation time constant tau at some level in the atmosphere is equal to the ratio of the Probe terminal velocity V_T to the acceleration of gravity g at that level. Both the descent velocity (19) and the Probe response time (20) vary with atmospheric density as rho^-0.5. If the Probe suddenly encounters a wind gust, the Probe horizontal velocity will adjust itself to a factor (1 - e^-1), or about 63% of the gust velocity, in a time equal to the time constant (20). In this time, the Probe has descended through a distance l_min = V_T tau = V_T^2 / g (21) Wind shears contained within a spatial layer smaller than l_min will have essentially no effect on the Probe. The spatial scale l_min thus represents the minimum (vertical) size of atmospheric structure detectable by monitoring the PRL Doppler profile. Invoking (19) and (20), it is seen that l_min is inversely proportional to the atmospheric density and independent of g: l_min = 2 C_B / rho (22) Using the Titan atmospheric pressures and densities from [LELLOUCHETAL1989], the Probe descent velocity V_T, response time tau and minimum scale sizes l_min are tabulated in Table 1, based on the nominal descent profile of duration 135 min. Under the initially large parachute (C_B ~ 8 kg/m^2) in the upper atmosphere (above ~ 115 km), the Probe motion will reflect atmospheric structure with kilometre size scales. Near the surface, where C_B > 50 kg/m^2 and V_T is much smaller, the minimum detectable structure size is of the order of 20 m. Table 1. Probe descent velocity, response time and vertical resolution. ----------------------------------------------------------------------------- time altitude P rho C_B V_T tau l_min (min) (km) (mbar) (kg/m^3) (kg/m^2) (m/s) (s) (m) ----------------------------------------------------------------------------- 8 130 4 0.009 8 47.0 38.6 1815 40 50 77 0.363 52 19.3 14.8 286 82 20 487 2.19 53 8.0 6.0 48 105 10 859 3.58 53 6.3 4.7 30 120 5 1120 4.43 53 5.7 4.2 24 135 0 1440 5.38 56 5.3 3.9 21 ----------------------------------------------------------------------------- Doppler wind recovery algorithm ------------------------------- A robust zonal wind recovery algorithm for the Cassini-Huygens scenario has been developed by [ATKINSONETAL1990]. Much of the formalism has been carried over from the experience gained from a very similar DWE investigation on Galileo [POLLACKETAL1992]. ------Update 06/04/01, RDR--------- Another alternative algorithm has been developed by [DUTTAROY2002]. -------end of Update-------------- An essential prerequisite to application of the algorithm ------Update 06/04/01, RDR--------- developed by [ATKINSONETAL1990] -------end of Update-------------- is an accurate reconstruction of the Probe-Orbiter relative geometry. It is initially assumed that the Probe's position is not affected by the integrated effect of the winds. Once a preliminary wind profile is calculated, the Probe descent position can be updated to reflect the integrated effect of the winds on the Probe descent longitude, and the wind profile is recalculated with the new time-varying Probe longitudes. An important assumption is that the zonal winds are dominant, with the possible exception of the last few kilometres above the surface. ------Update 06/04/01, RDR--------- The algorithm developed by [DUTTAROY2002], on the other hand, determines the zonal velocity and the corresponding longitude drift in one step. -------end of Update-------------- The zonal wind profile derived from the DWE measurements on Galileo is a relative (wind shear), rather than absolute, profile. This is because of the rather large uncertainty in the actually transmitted frequency from the quartz USO on the Galileo Probe. The 'constant' of the integration, i.e. a value of the wind u_N at a specific time t_N, must be determined by independent means. This problem does not exist with the Huygens DWE because the absolute fractional frequency uncertainties inherent to the TUSO and RUSO are of the order delta f/f ~ 2 x 10^-10. Under these circumstances, the accuracy with which the zonal winds can be recovered is determined by the imperfect knowledge of the Probe/Orbiter trajectory. Of lesser importance are second-order Taylor, Doppler and Special Relativistic terms that are usually dropped in order to keep the recovery problem linear, and environmental effects due to S-band signal propagation through a refracting, attenuating atmosphere (e.g. [BIRD1997]). Trajectory and oscillator drift errors introduce a small time-varying component into the Probe-Orbiter relative velocity that cannot be distinguished from atmospheric winds. A detailed treatment of the effects of these various errors for the recovery of the zonal wind height profile may be found in [ATKINSONETAL1990] ------Update 06/04/01, RDR--------- and [DUTTAROY2002]. -------end of Update-------------- In order to better understand the limitations of the recovery algorithm, we numerically simulated the wind recoveries for a variety of different trajectories, wind environments and errors. Here we summarise the results from several of those simulations and compare the recovered wind profile to the ideal case, where no frequency errors, trajectory uncertainties or anomalous wind regions exist. In all cases the input wind profile is a 'Flasar-type' model, a linearly increasing zonal wind with height, the direction and magnitude of which are variable parameters of the simulation. The largest error sources are those due to Probe entry longitude (0.7 deg), Probe entry latitude (0.1 deg), Probe descent velocity (0.005 zdot) and the maximum possible USO frequency drift (0.4 Hz in 2.5 h). The errors decrease during the Probe descent - in these simulations constrained to be zero at Titan's surface. The touchdown longitude phi_td is located about 6 deg (260 km) east of the entry longitude for the case with the nominal prograde wind profile. The Probe's mean speed in its easterly drift is ~ 100 km/h (28 m/s). Since the drift speed can be higher than the vertical descent velocity, the flight path to the surface can become rather flat. As expected from (7), the vertical descent time is only weakly dependent on the zonal wind u. The quality of the wind recovery is summarised in Table 2, which shows (a) the mean error in the determination of the zonal wind over the height interval 0-100 km (delta u in m/s), and (b) the error in the determination of the impact longitude (delta phi_td in degrees). Table 2. Mean zonal wind and impact longitude errors. ------------------------------------------------------------------------------ Input wind model delta u (m/s) delta phi_td (deg) ------------------------------------------------------------------------------ 1 x prograde 1.14 0.48 2 x prograde 2.34 0.22 1 x retrograde -2.57 1.23 2 x retrograde -6.42 1.96 ------------------------------------------------------------------------------ Additional simulations verified that the wind recovery algorithm is not affected by an input wind profile with a region of high shear, i.e. the wind recovery is not limited to smooth and slowly varying horizontal winds. As noted in the previous section, however, the detectability of fine wind structure is limited by the Probe response time (20). End-to-end measurement concept ------------------------------ DWE is the only Huygens investigation with 'science hardware' on both the Probe and within the Probe Support Avionics (PSA) of the Probe Support Equipment (PSE) on the Orbiter. The DWE-TUSO on the Huygens Probe is the primary signal generator used to drive the PRL of Transmitter A (Tx A). In case of failure, the driver signal can be switched to an internal temperature-controlled quartz oscillator (TCXO) with a frequency stability approximately 1000 times worse than the TUSO. The choice of oscillators for Tx A will be made before Probe-Orbiter separation, based on the performance of the TUSO and TCXO during the regular cruise phase checkouts. The 10 MHz output of the TUSO is upconverted to the PRL frequency of 2040 MHz, at which point it is transmitted via one of two routes to PSA Receiver A (Rx A) on the Orbiter. During cruise checkouts the signal is attenuated and sent via the radio frequency built-in-test equipment (RF-BITE) across the Umbilical Separation Mechanism (USM). During the Titan descent the signal is amplified for free-space RF transmission via the PTA to the Cassini HGA. The centre frequency of PSA Rx A is tuned to the nominal Tx A output frequency at 2040 MHz in checkout mode and is shifted by +38.5 kHz for descent mode. All timing and signal generator requirements for Rx A are controlled by the DWE-RUSO, which is virtually identical to the TUSO on the Probe. Similar to the Probe side, switching to a back-up TCXO is possible in case the RUSO fails. After downconversion, the PSA receiver phase locks onto the PRL signal, the loop control being governed by a numerically-controlled oscillator (NCO). The digitised values of the changes in NCO frequency required to maintain phase lock, an 'NCO control word' of 20 bits, is written to the PSE housekeeping (HK) data at 8 samples/s. The PRL signal level is monitored at the same sample rate by recording the PSA receiver automatic gain control (AGC; length 16 bits). The frequency resolution due to the data digitisation is governed by the internal PSA receiver clock and is given as ~ 60 MHz, which is very close to the least significant bit written into the NCO control word (48 MHz). The range of Doppler shifts accomodated by the 20-bit NCO control word is +-25 kHz (value corrected 06/04/01, RDR). The NCO control word and AGC data are routed to the solid state recorders on the NASA side of the Orbiter for later playback to Earth. The 'redundant' data from PSA receiver B (Rx B) are generated and received using standard TCXOs and thus not expected to yield information about the Probe's drift motions. Nevertheless, these data will also be recorded for the AGC information, thereby providing a control value for comparison with the data from Rx A. The only DWE data transmitted with the PRL telemetry are three temperature measurements (8 bits each) and one bi-level lock status bit, which are recorded every 16 s as part of the Probe housekeeping (HK) telemetry blocks. The associated DWE 'bit rate' of these HK data is thus only ~ 1.6 bit/s. Analogous HK-data are recorded at 25 bit/s on the PSE side from the RUSO. The quantities comprising DWE science data, the NCO control word for the frequency and the AGC for the signal level of the PRL are recorded at 288 bit/s. Assuming these data are recorded for the longest possible Huygens mission of 3 h (including 30 min on the surface), the total amount of cumulated DWE data for one chain will be ~425 kbyte. The Tx/Rx chain B, to be analysed primarily for its AGC information (no RUSO data), will generate another ~391 kbyte. Noting that these data will be stored redundantly for later playback to Earth, we arrive at a grand total for the DWE data volume of 1.632 Mbyte. The TUSO will be powered up before Titan atmospheric entry at the time t0 - 17 min 46 s, in order that it be given sufficient warm-up time to achieve the required frequency stability. The RUSO, which will be turned on at t0 - 30 min, is less critical because of its comparatively safe location onboard the Orbiter. The DWE-USO frequency stability will be monitored during the regularly scheduled Probe cruise checkouts en route to Saturn. The last such rehearsal will occur during the initial Saturn orbit, about 10 days before Probe separation. DWE Instrumentation: Ultrastable Oscillator (USO) ================================================= Transmitter and receiver USO programmes --------------------------------------- Two ultrastable oscillators, a TUSO for the transmitter on Huygens and a RUSO for the receiver on Cassini, have been constructed as identical units within one and the same USO programme in order to minimise costs. The contractor for this work is Daimler-Benz Aerospace (DASA), Satellite Systems Division, in Ottobrunn, Germany. The DASA design concept is based on a space-qualified rubidium oscillator Physics Package supplied by Ball Efratom Elektronik GmbH. The TUSO/RUSO combination represents the first use of rubidium oscillators on a deep space planetary mission. A total of six USO models have been built. The single Structural, Thermal and Pyrotechnic Model (STPM) was delivered for system testing on schedule in April 1994. The two Electrical Models (EM), one TUSO and one RUSO, were delivered in November 1994 to the System AIV (Assembly, Integration and Verification) Program at DASA. Three interchangable units of flight standard (FM) have been fabricated: one TUSO FM, one RUSO FM and one Qualification Flight Spare (QFS, a refurbished unit used for qualification testing at unit level). USO mechanical/electrical characteristics ----------------------------------------- The DWE USOs are designed to withstand the Cassini/Huygens launch and cruise phase (10 years in the event of a 1999 backup launch), as well as the Huygens atmospheric entry and descent on Titan. The TUSO on Huygens is exposed to higher mechanical loads than the RUSO on Cassini. The most critical factor for the TUSO, a major driver in the selection of an ultrastable oscillator based on rubidium technology, is the peak deceleration of up to 16.1 g during the Huygens entry phase. It could not be guaranteed that the required frequency stability of delta f0/f0 < 2 x 10^-10 (f0 = nominal output frequency) could be met after the Probe entry into Titan's atmosphere with a state-of-the-art quartz oscillator. The high mechanical load during entry might cause a deformation of the internal quartz fastening system in combination with an unpredictable frequency offset and an unknown frequency relaxation time. A similar problem with continuously varying mechanical stresses on the quartz box was foreseen in the subsequent pressure variation from 0 bar to 1.5 bar during descent. These adverse effects can be averted with a rubidium oscillator, for which the frequency source is also a quartz, because the nominal output frequency is locked to the very stable frequency of the rubidium ground-state hyperfine transition. The basic principle of a rubidium oscillator is to employ the two ground-state hyperfine levels A and B and a much higher optical level C of rubidium atoms to produce an error signal for the control circuit of a voltage-controlled quartz oscillator VCXO. IR light from a rubidium lamp is filtered and passes through the heated rubidium resonance cell with a frequency nu_CA = nu_C - nu_A, exciting transitions to state C of the rubidium gas. Atoms in level C drop back after a very short time either to state A or, with less probability, to state B. As atoms in state A are continuously re-excited to C, the population of level B steadily increases ('optical pumping'). When state A is depopulated, a maximum photocurrent is produced in the photocell as the light is no longer attenuated by excitation processes from level A to C. The HF-signal of the synthesiser, which is upconverted from the quartz output signal, also irradiates the rubidium resonance cell. The synthesiser is calibrated to generate the exact resonance frequency nu_BA = nu_B - nu_A ~ 6.835 GHz if the quartz has its nominal output frequency. At this frequency, atoms in level B de-excite to state A, thereby inducing a minimum photocurrent, because the rubidium vapour is no longer transparent to the frequency nu_CA. Deviations from the minimum current condition are produced by deviations in the nominal VCXO output frequency. The current dip, however, is very small (0.1% of the total photocurrent) and not suitable for a DC detection. To circumvent this shortcoming, the synthesiser frequency is modulated at an audio frequency nu_m = 127 Hz. The similarly modulated photocurrent can now be AC-detected by synchronous current demodulation. There is a positive correlation between the sign of the current error signal (positive or negative) and the offset from the nominal quartz output frequency. This error signal is used to lock the nominal quartz signal to the rubidium resonance frequency by varying the quartz control voltage. The USO consists of the Physics Package (rubidium resonance cell and lamp) and several printed circuit boards, which contain discrete electronic elements as well as integrated circuits. The USO electronics and the Physics Package are integrated into an aluminium box, which is constructed as a Faraday cage to avoid electromagnetic contamination of the USO environment. The USO box is attached to the experiment platform via four mounting studs, thereby providing a thin air buffer for better insulation from the mounting platform. This reduces the USO conductive heat loss, saving power that would otherwise be needed for heating. The box surface is plated with nickel and coated with Chemglaze Z 306. The total mass of the USO is ~1.9 kg. The electronics and Physics Package is surrounded by mu-metal shielding to minimise the effects of changes in the external magnetic fields that range from >10^4 nT on Earth to essentially zero at Titan. Variations in the ambient magnetic field induce a change in the Rb hyperfine resonance frequency and thus a frequency shift in the USO output signal. Radiation-sensitive USO components such as transistors and analogue ICs are radiation-hardened up to 10 krad. While the maximum radiation dose for the Probe TUSO is ~ 5 krad, the RUSO on the Orbiter will be exposed to about 18 krad in the event of a 1999 launch. In order to provide additional radiation protection for the RUSO, the critical components of both USO units are shielded with tantalum caps (thickness ~1 mm). Radiation shielding contributes about 10% to the total USO mass budget. The USO external supply voltage of 28 V (TUSO) or 30 V (RUSO), is transformed down to 5 V and 17 V by the DC/DC converter. The converter control signal is provided by the synthesiser, synchronised to the quartz output signal. The VCXO quartz, located on the Oscillator Board, provides the 10 MHz output through a buffer amplifier. The same signal is upconverted by the synthesiser to the rubidium resonance frequency. The photocurrent of the Physics Package is routed to the Servo Board, which generates the error signal for the voltage control of the VCXO. The heater control of the rubidium lamp within the Physics Package is located on a separate Lamp Board. Three analogue sensors monitor the temperatures of the rubidium lamp, rubidium cell and VCXO, respectively. A bi-level lock indicator flips from '0' to '1' whenever the quartz output signal is in lock with the rubidium resonance frequency. The temperatures and lock indicator signal are part of the Huygens HK data. Table 3 summarises the USO mechanical and electrical characteristics. Table 3. USO mechanical and electrical characteristics. ----------------------------------------------------------------------------- Mass 1.90 +- 0.15 kg Dimensions (length x width x hieght) 180 x 149 x 118 mm Warm-up Power P_wu = 18.2 W (<= 30 min) Steady State Power at baseplate temp. T P_ss = 10.375 - 0.0625 T (W) Output Frequency 10 MHz +- 0.1 Hz Output Signal Level 0 +- dBm into 50 Ohms Warm-up Time to delta f0/f0= 2x10^-10 <= 30 min Supply Voltage (TUSO) 28 (+0.35,-0.63) V Supply Voltage (RUSO) 30 (+1.5,-3.38) V ----------------------------------------------------------------------------- USO frequency characteristics ----------------------------- The USO long-term frequency stability requirement of delta f0/f0 <= 2x10^-10 is critical for a successful Doppler Wind Experiment. By 'long-term', it is meant that the uncertainty in the frequency must be constrained within these limits for the maximum expected duration of the atmospheric descent (2.5 h), i.e. the total frequency shift of the PRL signal is <0.4 Hz over the entire Huygens mission. Assuming a typical DWC angle of 65 deg, an unwanted frequency shift of this magnitude would be indistinguishable from a zonal wind with velocity of ~0.15 m/s. This USO frequency stability is thus mandatory for achieving the primary DWE goal, a determination of Titan's zonal wind height profile with an accuracy well below the +- 1 m/s level. This must be maintained throughout the entire Huygens mission, in spite of many rather severe changes in the environmental conditions (temperature T, pressure P, acceleration A and magnetic field B). Generally, this requirement can be expressed by: |delta f0/f0| = 1/f0 [(delta f0/delta T Delta T)^2 + (delta f0/delta P Delta P)^2 + (delta f0/delta A Delta A)^2 + (delta f0/delta B Delta B)^2 + (delta f0/delta X Delta X)^2]^0.5 <= 2x10^-10 (23) where 'X' collectively labels other factors, e.g. the variation of the USO supply voltage. Typical values for the expected fractional change in the USO output frequency are listed in Table 4. Although not as critical as the long-term frequency stability, the short-term frequency stability is another criterion for the frequency quality of the USO. The short-term frequency stability is characterised by the mean fractional frequency deviation sigma = Delta f0 / f0 (24) This quantity, sometimes called the 'Allan deviation', depends on the integration time tau (see also equation 26). The specified USO short-term frequency stabilities are shown in Table 5. These values are sufficient to detect PRL frequency modulations owing to atmospheric turbulence as well as the Probe's pendulum and rotational motion. Table 6 lists the specifications on the phase noise of the output signal spectrum for various displacements from the centre peak. Table 4. Fractional change of USO output frequency. ----------------------------------------------------------------------------- Environmental fractional frequency worst case: Huygens descent factor change delta f0/f0 TUSO RUSO ----------------------------------------------------------------------------- Temperature (per deg C) 3 x 10^-12 Delta T ~ 15 deg C ~ 5 deg C Pressure P (per bar) 1 x 10^-12 Delta P ~ 1.6 bar ~ 0 bar Acceleration A (per g) 4 x 10^-12 Delta A ~ +- 1.0 g ~ +- 0.1 g Magnetic field B (per G) 2 x 10^-12 Delta B ~ 1 mG ~ 0.1 mG ----------------------------------------------------------------------------- Table 5. USO short-term frequency stability. ----------------------------------------------------------------------------- Integration time tau (s) Delta f0/f0 ----------------------------------------------------------------------------- 0.1 6 x 10^-11 1 1 x 10^-11 10 5 x 10^-12 100 1 x 10^-12 ----------------------------------------------------------------------------- Table 6. USO phase noise. ----------------------------------------------------------------------------- Displacement (Hz) Phase noise (dBc) ----------------------------------------------------------------------------- 1 -90 10 -120 100 -140 1000 -150 ----------------------------------------------------------------------------- USO test programme ------------------ At this writing, only results from the EM test programme are available. A USO test assembly, developed at the University of Bochum, is used at the unit level to determine USO frequency characteristics. Of particular interest is the frequency stability of the USO on all time scales. It is also necessary to characterise the repeatability of the long-term drift of the USO in order to correct for this effect under the conditions of the Huygens descent. The frequency measurement procedure is based on a comparison of the test frequency f (DWE-USO) with a reference signal f0 (rubidium reference oscillator), the frequency stability of which is better than or equal to that of the test signal. Using a synthesiser, the reference frequency is offset from the nominal frequency of the test object f0 by a known amount f_offset. The frequency difference of the two signals is generated with the help of a mixer and then measured by a frequency counter. Following a fully automated procedure, both the long-term and short-term stabilities of the test object are determined by recording the frequency difference for an appropriately long interval (nominal test run: 3 h). The reference and offset frequencies, as well as the integration and sample times can be adjusted as necessary with respect to their default values for each individual run. The number of single measurements is defined by the ratio of the adjusted total measurement time to the mean duration of the single measurement cycle. A single measurement cycle consists of the adjusted fixed integration time and the counter register read-out time. The read-out phase requires about 20 ms. The mean duration of one measurement cycle is thus about 1.02 s for an integration time of tau = 1 s, and 10.02 s for tau = 10 s, respectively. The absolute frequency f(t) and the Allan variance sigma^2 are then derived as follows: f(t) = Delta f + f0 - f_offset (25) sigma^2(tau) = (1/N) SUM[k=1..N] { (ybar_k+1 - ybar_k)^2 / 2} (26) where ybar_k = Delta f_k / f0, Delta f_k = m_k/tau, and m_k is the kth cycle count over the integration time tau. Running measurements of sigma, defined by (26) are made for integration times tau = 1 s & 10 s. The mean values of the Allan deviation in this case were < sigma(1 s) > = 1.8 x 10^-11 and < sigma(10 s) > = 5.5 x 10^-12, respectively. After reaching its asymptotic value, the output frequency of both models stays within a band delta f/f < +- 2x10^-10. The model without magnetic compensation was found to require too much time to warm up and also exhibited undesireable variations in asymptotic output frequency with temperature. Significant improvement was achieved by implementation of a magnetic control loop, which exploits the sensitivity of the rubidium hyperfine transition to the strength of the magnetic field in the rubidium cell. The strength of the magnetic field applied to compensate for the thermal drift is determined via a feedback loop using the thermal sensor in the crystal oscillator. Preliminary results from recent thermal vacuum tests (FM Test Program) over an ambient temperature range from -30 deg C to +60 deg C indicate that: (a) the USO warm-up time is always <30 min, and (b) the variation in mean asymptotic frequency is <2x10^-9. " END_OBJECT = INSTRUMENT_INFORMATION OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "BIRDETAL1997" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "ALLISON1992" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "ALLISONETAL1994" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "ATKINSON1989" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "ATKINSONETAL1990" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "CALDWELLETAL1996" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "COUSTENIS1990" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "DELGENIOETAL1993" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "FLASARETAL1997" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "FLASARETAL1990" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "FLASARETAL1981" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "FLURY1986" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "HANELETAL1981" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "HINSON&TYLER1983" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "HOURDINETAL1995" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "HUBBARDETAL1993" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "LEBRETON&MATSON1997" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "LELLOUCHETAL1989" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "LINDALETAL1983" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "OTT1991" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "POLLACKETAL1992" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "SMITHETAL1992" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "SMITHETAL1996" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "SMITHETAL1994" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "STROBEL&SICARDY1997" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "TOONETAL1988" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "TYLERETAL1981" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "BIRDETAL2005" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "LEBRETONETAL2005" END_OBJECT = INSTRUMENT_REFERENCE_INFO OBJECT = INSTRUMENT_REFERENCE_INFO REFERENCE_KEY_ID = "DUTTAROY2002" END_OBJECT = INSTRUMENT_REFERENCE_INFO END_OBJECT = INSTRUMENT END