J. T. Schofield, J. R. Barnes, D. Crisp, R. M. Haberle, S. Larsen, J. A. Magalhães, J. R. Murphy, A. Seiff, G. Wilson

The Mars Pathfinder atmospheric structure investigation/meteorology (ASI/MET) experiment measured the vertical density, pressure, and temperature structure of the martian atmosphere from the surface to 160 km, and monitored surface meteorology and climate for 83 sols (1 sol = 1 martian day = 24.7 hours). The atmospheric structure and the weather record are similar to those observed by the Viking 1 lander (VL-1) at the same latitude, altitude, and season 21 years ago, but there are differences related to diurnal effects and the surface properties of the landing site. These include a cold nighttime upper atmosphere; atmospheric temperatures that are 10 to 12 degrees kelvin warmer near the surface; light slope-controlled winds; and dust devils, identified by their pressure, wind, and temperature signatures. The results are consistent with the warm, moderately dusty atmosphere seen by VL-1.

J. T. Schofield and D. Crisp, Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA.
J. R. Barnes, College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331, USA.
R. M. Haberle, Ames Research Center, Moffett Field, CA 94035-1000, USA.
S. Larsen, Wind Energy and Atmospheric Physics, Risoe National Laboratory, AMV-125, Post Office Box 49, Roskilde, Denmark DK-4000.
J. A. Magalhães, J. R. Murphy, A. Seiff, San Jose State University Foundation, San Jose, CA, and Ames Research Center, Moffett Field, CA 94035-1000, USA.
G. Wilson, Arizona State University, Tempe, AZ, and Ames Research Center, Moffett Field, CA 94035-1000, USA.

The ASI/MET experiment combined accelerometer and MET instruments (2). The accelerometer instrument contained science and engineering accelerometers that each monitored accelerations along three orthogonal axes. In each axis, the maximum sensitivity was 20 µm/s² [2 × 10^-6 Earth gravities (g)], and the accelerations expected during EDL were covered by commandable measurement ranges of 16mg, 800mg, and 40g full-scale. The MET instrument consisted of pressure, temperature, and wind sensors. Pressure was measured through a 1-m inlet tube that was exposed to the atmosphere during parachute descent as well as after landing (1, 2). The pressure measurements have a maximum sensitivity of 0.25 µbar, which is more than a factor of 100 better than that available to the Viking landers (5). All the MET temperature and wind sensors are mounted on a mast 1.1 m high, deployed at the end of a lander petal to isolate it from spacecraft thermal contamination (1, 2). Atmospheric temperature was measured by four thermocouples: one designed to measure temperature during parachute descent and three designed for surface boundary layer measurements 25, 50, and 100 cm above the base of the mast. All four thermocouples have time constants of 1 to 2 s and sensitivities of 0.01 K. Wind was measured by a six-segment hot-wire sensor at the top of the mast, 1.1 m above the mast base. The wires are heated by a current passed in series through all six segments, and the temperature differences between low and high current modes for each segment are used to determine wind speed and direction.

The accelerometer and MET instruments recorded data continuously throughout EDL until about 1 min after impact at about 03:00 local solar time (LST). Regular surface pressure, temperature, and wind measurements by the MET instrument began about 4 hours after impact at 07:00 LST on sol 1 (1 sol = 1 martian day = 24.7 hours), and the MET mast was deployed at 13:30 LST.

The science accelerometer detected the upper atmosphere 160 km above the landing site when the entry vehicle had a velocity of 7.4 km/s relative to the atmosphere and a flight path angle 14.8° below the local horizontal. 1.5 min later, the entry vehicle experienced a peak deceleration of 15.9g at an altitude of 33 km. After 3 min (9 km) the parachute deployed, and at 3.4 min (7.4 km) the heat shield separated from the lander, allowing the pressure sensor to begin unobstructed measurements of the atmosphere. The inflation of shock-absorbing airbags at 5.1 min (0.3 km) terminated the unobstructed pressure measurements, and descent rocket firing at 5.2 min (0.1 km) ended the direct measurement of aerodynamic decelerations. The first impact of the probe with the martian surface occurred 5.3 min after it entered the atmosphere. In the first minute after impact, the lander bounced 15 times and pressure sensor data indicated that it rolled 10 m vertically downhill. It came to rest about a minute later at a site 3389.7 km from the center of mass of Mars (6). Surface acceleration measurements of 3.716 m/s² agree with values of 3.717 m/s² calculated for the lander location and height (7), providing a verification of accelerometer gain calibration.

Because the engineering accelerometers were used to control parachute deployment and remained in their least sensitive 40g scale, atmospheric profiles were derived from science accelerometer data only, which were logged at 32 Hz throughout EDL. MET pressure and temperature data were collected at 2 Hz during the parachute descent and landing phases of EDL. However, spacecraft design constraints did not allow the descent temperature sensor to be exposed to the free flow around the lander, so that direct atmospheric temperature measurements were not possible (2, 8).

For the 30-day primary landed mission, 51 equally spaced MET measurement sessions were made each day to monitor the atmospheric diurnal cycle and synoptic (day-to-day) variations. These 3-min sessions sampled the atmosphere at 4-s intervals and were interspersed with 15-min and 1-hour sessions of 1-s sampling to monitor the surface boundary layer. The boundary layer sessions were repeated as frequently as allowed by data volume, in a pattern that sampled all local times of day as rapidly as possible. Finally, on sol 25, a 24-hour session was executed that sampled science data continuously at 4-s intervals for a complete daily cycle. The continuous MET observations desired during the primary mission were interrupted by spacecraft computer resets. One of these resets was associated with errors in MET software, and the measurements were limited to the nighttime from sols 12 to 17, until the errors were corrected.

Entry, descent, and landing. Atmospheric density, pressure, and temperature profiles were derived from ASI/MET deceleration measurements during Pathfinder EDL and compared with VL-1 results (Figs. 1 and 2). The mean solar forcing of the martian atmosphere at the time of the Pathfinder and VL-1 profiles was similar. Solar longitudes (L_s) of 98° (VL-1) and 142° (Pathfinder) correspond to a seasonal difference from mid- to late summer and to a smaller Mars-sun distance at the time of the Pathfinder profile. Solar activity, as measured by sunspot counts, was near-minimum for both profiles, the lower atmosphere had comparable amounts of dust (9), and the two landings were at similar latitudes and longitudes on Mars. However, the different time of day (03:00 LST for Pathfinder versus 16:15 LST for VL-1) and the 21 years separating the two entry profiles suggest that diurnal and secular effects could be important in understanding their differences.

For an entry vehicle, atmospheric density is directly related to aerodynamic deceleration, velocity relative to the atmosphere, and aerodynamic characteristics (2). Deceleration was measured directly; velocity and position can be reconstructed by integrating the equations of motion using the observed decelerations and an initial velocity and position; and Pathfinder entry vehicle aerodynamic characteristics are known from computational aerodynamic simulations and laboratory experiments.

Atmospheric densities measured by Pathfinder varied from ~5 × 10^-11 kg/m³ at the threshold of detection to 8 × 10³ kg/m³ immediately before parachute release at 9 km (Fig. 1). From 160 to 90 km, densities range from a factor of 5 lower than VL-1 values near and above 120 km to a factor of 2.5 lower near 90 km. The increase in density between 90 and 80 km, which corresponds to a deep temperature minimum (Fig. 2), raises the Pathfinder densities at lower altitudes to values slightly lower than VL-1 densities. The lower values of density, and therefore pressure, encountered by Pathfinder below 30 km are generally consistent with the lower overall mass and surface pressure of the martian atmosphere at the time of the Pathfinder landing (10). The decreased surface pressure results from the annual variation in atmospheric mass caused by condensation and sublimation from the polar caps (11).

The martian thermosphere, where temperature increases rapidly with altitude because of heating by solar extreme ultraviolet radiation, is evident above 125 km in the Pathfinder profile (Fig. 2). Although the uncertainties in the derived temperatures at these altitudes are large, it appears that Pathfinder temperatures are close to or slightly higher than those measured by VL-1.

From 65 to 125 km, observed temperatures were, on average, 20 K lower than those observed by VL-1 (Fig. 2). This contrast is responsible for the lower Pathfinder densities above 90 km. Because radiative time constants reach a minimum in this altitude range, a large response to diurnal forcing is a likely explanation for this difference (12). The temperature minimum of 92 K at 80 km is the lowest temperature ever measured in the martian atmosphere and may result from the superposition of waves, such as thermal tides, on the overall nighttime cooling. Tides propagate from the lower atmosphere with amplitudes that increase with altitude.

At 80 km, the Pathfinder temperature profile is lower than the CO₂ condensation temperature (Fig. 2). The lower temperature may result from supercooling or from inaccuracies in the vapor pressure curve for CO₂ (13), but it is possible that CO₂ could be condensing at these levels to form high-altitude clouds. Observations by the imager for the Mars Pathfinder (IMP) of sky brightening well before the expected start of dawn may indicate the presence of such clouds, although the substantial morning clouds identified in images of the eastern horizon before sunrise are almost certainly water clouds formed at much lower levels (9).

Below 60 km, temperatures measured by Pathfinder are higher than those measured by VL-1 down to 35 km and are similar or slightly lower at lower altitudes down to 16.5 km (Fig. 2). These differences are consistent with minor variability in dust content and are also within the amplitudes expected from vertically propagating atmospheric waves such as tides (10). In view of the similar solar forcing, it is not surprising that the Pathfinder and VL-1 profiles are comparable in the lower atmosphere. However, Earth-based, disk-averaged microwave observations over the past 10 to 15 years have been used to suggest that the martian lower atmosphere undergoes episodes of cooling, which are characterized by reduced solar-absorbing dust content and 20 K lower temperatures relative to the conditions observed during the Viking mission (4). The Pathfinder entry profile shows no evidence for substantial cooling of the lower atmosphere (Fig. 2), although microwave profiles obtained before and after landing suggested that a cool atmosphere would be found (14). Furthermore, IMP observations yield atmospheric dust opacities of 0.5 (9), which are comparable to those found by VL-1 at the same season. These opacities are consistent with our temperature profile observations. Finally, it is unlikely that Pathfinder entered a locally dusty and anomalously warm region, as dust opacities remained stable throughout the 30-day primary mission (9).

Below 16.5 km a strong thermal inversion is present in the Pathfinder data, where temperature decreases from 200 to 181 K at 10 km (Fig. 2). This inversion is at too high an altitude to be the strong thermal inversion in the lowest few kilometers of the nighttime diurnal thermal boundary layer on Mars predicted by radiative-convective models and one-dimensional dynamical boundary layer models (15, 16). At the base of the Pathfinder profile, temperature appears to increase with decreasing altitude again and can easily achieve the observed surface temperature, including predicted near-surface thermal inversions, without exceeding the adiabatic lapse rate.

The temperature minimum in the 10-km inversion is well below the condensation temperature of water vapor in the martian atmosphere, assuming that the 10 precipitable micrometers of water derived from IMP measurements are uniformly mixed (9). This inversion may mark the altitude of clouds seen in IMP images before sunrise (9) and near the low-latitude morning terminator in Hubble Space Telescope images (17), although the height of these clouds is not well known. After sunrise, the clouds burned off rapidly, and the inversion may have disappeared. What triggered the formation of a strong temperature inversion is not known, but possible mechanisms include the horizontal advection of cooler air at 10 km and vertically propagating finite-amplitude gravity waves excited by surface topography in the strongly stratified near-surface nighttime boundary layer (18). Once clouds formed, thermal emission from the cloud tops could have enhanced the intensity of the inversion at night.

Measurements made after landing. The pressure, temperature, and wind velocity data acquired by ASI/MET during the landed mission allow the variability of the martian atmosphere at the Ares Vallis landing site to be studied during the midsummer season on short, daily, synoptic (day-to-day), and seasonal time scales. This reveals not only the local properties of the atmosphere and its interaction with the surface but also more global information on atmospheric dust loading, circulation, and the seasonal CO₂ cycle. Comparisons with VL-1 data taken at the same time of year can be used to identify longer-term climate changes.

Pressure data. During sols 1 through 30, surface pressure at the landing site underwent substantial daily variations of 0.2 to 0.3 mbar, which were associated primarily with the large thermal tides in the thin martian atmosphere (11) (Fig. 3A). Daily pressure cycles were characterized by a strong semidiurnal oscillation, with two minima and two maxima per sol, together with diurnal and higher-order components, although there was considerable day-to-day variability (Fig. 3B). The presence of a large semidiurnal tidal oscillation is indicative of atmospheric dustiness over broad regions of Mars and over an altitude range of at least 10 to 20 km (19).

A long-term trend in daily mean pressure was also seen. A third-order polynomial fit to the data shows that mean pressure fell slowly at the beginning of the period and rose at the end, with a minimum just under 6.7 mbar near sol 20 (L_s ~ 153°) (Fig. 3A). This time corresponds to the annual deep minimum in the seasonal pressure cycle associated with CO₂ condensation and sublimation in the polar regions of Mars and was seen previously by the Viking landers (11).

The ASI/MET pressure sensor detected a variety of pressure variations on relatively short time scales. These ranged from seconds to hours and had magnitudes of 1 to 50 µbar. The shorter time-scale variations (<10 to 15 min) appear to be correlated with wind and temperature fluctuations and tend to be largest during late morning and early afternoon, when the boundary layer is most turbulent. The most dramatic pressure features were minima of 10 to 50 µbar, usually less than a minute in duration, associated with vortices (dust devils) passing over the lander. A particularly good example was seen during the continuous sampling of sol 25.

Temperature data. In common with the single-level observations made by VL-1 at the same season, the diurnal temperature variations at the three Pathfinder levels repeat from day to day with a high degree of consistency. The diurnal cycle was sampled particularly well on sol 25 (Fig. 4). For the top mast thermocouple, a typical maximum temperature was 263 K at 14:15 LST and a typical minimum was 197 K at 05:15 LST, shortly before sunrise. Because of the low density of the martian atmosphere, near-surface atmospheric temperatures are influenced by the surface temperature cycle, which is driven by solar heating during the day and infrared cooling at night. For this reason, diurnal temperature extremes at the bottom of the mast exceeded those at the top (Fig. 4). Temperatures at the top of the mast (1.1 m) were greater by 10 K during the day and 12 K at night than those seen at 1.6 m by VL-1, probably because of the lower albedo and greater thermal inertia of the surface at the Pathfinder landing site.

The diurnal variation of the vertical temperature gradient was also consistent from day to day. At sunrise, the atmosphere is typically stably stratified, and cool dense air lies near the surface (Fig. 5). As the surface warms, the air mass is heated from below. By about 06:30 LST, all three mast temperatures are about equal, indicating that the near-surface atmosphere is neutrally stable, because isothermal and adiabatic lapse rates are indistinguishable over a vertical range of 75 cm. By 07:30 LST, ground heating exceeds the ability of the atmosphere to transfer surface heating by conduction, the temperature gradient reverses, the atmosphere becomes unstable, and convection begins. Turbulent mixing carries heat from the surface and is revealed by large rapid temperature fluctuations of up to 15 to 20 K observed during the remainder of the morning and early afternoon. Later in the afternoon, the surface cools and turbulent mixing diminishes (Fig. 4). By 16:45 LST, the thermal profile is neutral and what surface winds are observed convect heat quickly. Shortly afterward, surface cooling causes the temperature gradient to invert, and the surface boundary layer becomes stably stratified for the duration of the night. Major nighttime temperature fluctuations are caused by downslope winds that disturb the surface boundary layer (Fig. 4).

Large near-surface temperature gradients of 10 to 15 K are probably a common feature of the martian daytime boundary layer. Because of low atmospheric densities, the convective heat flux is unable to cool the surface as efficiently as on Earth, where atmospheric fluxes typically remove 80 to 90% of the net surface radiative flux under convective conditions.

Wind data. The ASI/MET wind sensor measured wind speed and direction 1.1 m above the base of the mast. Accurate speed determination requires further calibration of the relation between wind speed, air temperature, and sensor hot wire overheat under Mars surface conditions and is not yet available. Here we discuss wind direction and approximate speed based on azimuthal variations of wire overheat, which are a function of both wind speed and direction.

For sols 1 to 30, wind direction generally rotated in a clockwise manner through all the points of the compass during the course of 1 sol (Fig. 6). This rotation was not uniform. Winds were consistently from the south in the late night and early morning hours and then rotated steadily through west, north, and east during the day. Over the 30-day period studied, nighttime wind direction was remarkably constant, but considerably more scatter was seen during the day (Fig. 6). The most anomalous wind variations were observed on sols 8 through 10, when northeasterly winds were absent from the daily cycle. A pronounced reduction in the daily variation of surface pressure was seen in the same time interval (Fig. 3).

The recurrent southerly wind from late evening through morning is consistent with a drainage flow down Ares Vallis, which rises to the south of the lander (1); and the northerly wind seen in the afternoon is indicative of flow up Ares Vallis. Although the clockwise rotation of the wind vector agrees with that expected from the westward-migrating classical diurnal thermal tide, the time-phasing of wind direction throughout the day does not. The classical tidal drive would generate a westerly maximum near 18:00 LST and a southerly maximum at about 12:00 LST. It therefore appears that local topography, or possibly nonclassical tides, are controlling the wind direction at the Pathfinder location during midsummer.

Winds measured at the VL-1 site at this same season 11 Mars years earlier (1976) were generally weak (<6 m/s), exhibited a time-averaged northwesterly direction, and were approximately upslope during the afternoon and downslope during the night and early morning. Increased wind speeds accompanied the increased pressure oscillations seen during the several sols surrounding L_s = 150 (20). If the winds are slope-driven at this season, the differences between VL-1 and Pathfinder winds are expected to reflect differences in the magnitude and direction of the slope at the two sites. Finally, preliminary estimates suggest that wind speeds were comparable with or lower than those measured by VL-1 at the same time of year. Speeds were generally less than 5 to 10 m/s, except during the passage of dust devils, and were often less than 1 m/s in the morning hours. This may be consistent with the lower slope at the Pathfinder site.

Synoptic and seasonal variations. Synoptic and seasonal variability in the martian atmosphere at the Ares Vallis landing site are seen most clearly in the pressure data (Fig. 3). In particular, the annual pressure cycle reached a minimum at about sol 20, corresponding to the greatest mass of the southern seasonal polar cap (11), although the precise timing of the pressure minimum can be up to 5 sols earlier and 1 sol later, depending on exactly how the pressure data are analyzed. The Pathfinder minimum appears to have occurred about 6 sols later than that seen during the first year at the VL-1 site (21), which seems to imply more or longer lasting CO₂ condensation on the south polar cap. Variations from year to year of at least several sols in the timing of the annual pressure minimum were observed by the two Viking landers (22).

Small synoptic variations in amplitude of less than 3 to 4 K in temperature and 20 to 30 µbar in pressure are present in the ASI/MET data, after diurnal fluctuations and longer term trends are removed. Variations of comparable magnitude were found in the VL-1 data for this season; the greatest variance was at relatively long (10- to 30-sol) periodicities (21, 23). Missing data and the limited temporal extent of the available Pathfinder observations make the estimation of periodicities difficult. On the basis of the VL-1 data and pre-landing general circulation model predictions for the Pathfinder site, it is expected that synoptic variations of much larger amplitude will begin as the transition to the fall and winter weather regime takes place (10, 21). There is evidence that this is beginning to happen at the end of the Pathfinder pressure record.

Thermal tides. For the first 30 days of the landed mission, the semidiurnal tide was the largest component of the daily variation in surface pressure (Fig. 7). Its amplitude is generally similar to that found by VL-1 and is consistent with a visible optical depth of global dust of order 0.5 (19). This value is comparable to that measured by the IMP (9). Given the long vertical wavelengths associated with the semidiurnal tide, its measured amplitudes further suggest that the dust was deeply distributed in the atmosphere and was not confined to a shallow layer near the surface. This inference is consistent with the dust scale height of about 13 km deduced from IMP measurements (9) and with our entry profiles, which show a relatively warm lower atmosphere below 60 km. However, as mentioned earlier, a deeply distributed global dust haze is not consistent with Earth-based microwave measurements for this same period (14).

The amplitude of the diurnal tide was smaller and more variable. On sols 5 and 9, normalized amplitudes were down to 0.002 (Fig. 7), which is more than a factor of 2 smaller than anything measured by Viking. The variation seen by Pathfinder is somewhat systematic in that diurnal amplitudes were low during the first 5 to 10 sols, gradually rose to values of around 0.010 by sol 20, then declined to rather low values again by sol 27. Phase also varied during this period, retarding during the first 15 sols then advancing during the last 15 sols. The amplitude of the diurnal tide is sensitive to the calibration of the pressure sensor, which is still preliminary. However, the time variation of amplitude and phase is relatively insensitive to calibration.

These variations in the diurnal tide suggest that it was modulated by interference effects between the sun-following westward tide and the topographically induced, resonantly enhanced, eastward traveling, diurnal kelvin mode (19, 20, 24). One possibility is that the interaction is between eastward and westward modes of comparable amplitude, where the eastward mode period is first slightly greater than and later slightly less than 1 sol. Alternatively, the eastward and westward modes could have similar periods, and amplitudes may vary with time. The modeling of this interaction may help to distinguish between these two possibilities.

Nighttime temperature inflections. ASI/MET temperature measurements show that, on average, atmospheric temperatures fall monotonically through the evening and night until sunrise. On most nights, however, the steady decline in temperature was interrupted by one or more inflections in the gradient, where the temperature fell less rapidly or even rose before resuming its downward trend. Striking features of this kind are seen in Fig. 4 at 18:15, 19:00, 20:40, and 03:05 LST and are particularly marked in top mast thermocouple temperatures.

Temperature inflections seen in the VL-1 temperature data set were attributed to radiative effects associated with the formation of ground fogs (25). At the Pathfinder landing site, temperatures are about 10 K warmer than those of VL-1 at night, and condensation is unlikely, given the 10 precipitable micrometers of water measured by the IMP (9). However, there was a positive correlation between increased wind speed and the observed thermocouple temperature inflections. Increased horizontal flow may have disturbed the strong temperature inversion that developed at night, bringing warmer air from aloft down to thermocouple heights through enhanced vertical mixing. After the nighttime warming episodes, temperatures fell rapidly in conjunction with substantial reductions in wind speed. The nighttime temperature inflections occurred in the presence of southerly winds and cannot be produced by thermal contamination from the body of the spacecraft, which is located to the northwest of the MET mast.

Dust devils. Short-term variations in measured surface pressure, wind velocity, and air temperature over periods of tens of seconds to minutes suggest that small-scale convective vortices passed through the Pathfinder lander site. During the passage of one of these features, pressure fell and recovered rapidly, accompanied by abrupt shifts in wind direction. These events, which we refer to as dust devils, are probably similar in character to features noted in the VL-1 and VL-2 meteorology data (26) and may have been seen by the Viking orbiter cameras (27). The narrow, well-defined pressure minimum is the most characteristic feature of these vortices, but it was not seen in the Viking lander pressure record because of poor resolution and coarse temporal sampling. There are currently no Pathfinder lander images to indicate whether the apparent vortices entrain dust, but the passage of a particularly large feature on sol 62 was correlated with a short-lived reduction of about 1.5% in the power generated by the Pathfinder solar panels.

A dust devil event at 14:15 LST on sol 25 has the characteristics of a clockwise-rotating vortex traveling toward the south-southeast (160°), with the vortex center passing west and south of the lander (Fig. 8). During this event, pressure fell 0.028 mbar in 24 s to a minimum, followed by a more rapid rise of 0.027 mbars in 8 s. The pressure minimum was accompanied by a local temperature maximum at all three thermocouple heights. These temperature maxima are themselves embedded in a pronounced temperature minimum, which extended 30 to 40 s before and after the pressure minimum. Wind direction rotated from northwesterly, which was the mean flow direction before the vortex arrived, through north to northeasterly in the 40 s preceding the pressure minimum, before shifting abruptly through northerly to westerly at and just after the pressure minimum. Wind direction slowly returned to north-northwesterly over the minutes following the pressure minimum. This pattern is consistent with a clockwise-rotating vortex embedded in an ambient flow that has a speed less than the vortex rotational speed (26).

Boundary layer results. The near-surface boundary layer of the martian atmosphere can be described by "universal" scaling laws involving dimensionless coefficients determined from measurements in Earth's atmosphere. The Viking and Pathfinder measurements provide us with an opportunity to test this universality by using measurements in another atmosphere. Although the martian atmosphere differs in many respects from our own, most of the universal relations are expected to be valid, based on the analysis of Viking data (28).

Relative to Viking, the Pathfinder measurements have the following advantages for boundary layer studies. Temperature measurements at several heights allow for a more direct determination of the vertical temperature gradient, and the successful MET sampling strategy has ensured that turbulence and profile data with high temporal resolution (1 Hz or better) are available over a wide variety of conditions, with good systematic coverage of diurnal and interdiurnal variability.

As an example of the analysis of turbulence characteristics from Pathfinder, we show an average temperature power spectrum obtained from the Fourier decomposition of several temperature time series (Fig. 9). Individual spectra are scaled by the appropriate temperature scale T_*, which, following the Monin-Obuchov formulation, is the most important scaling parameter for both vertical profiles and spectra (28). The T_* values scaling the spectra were derived from the measured temperature gradients, assuming close to neutral thermal stability due to the proximity of the ground (0.25 m), although T_* ranges from -5 K to +0.2 K. Although simplified, the scaling does a reasonable job in collapsing individual spectra and hence in supporting the applicability of the Monin-Obuchov scaling laws.

REFERENCES AND NOTES

1. M. P. Golombek, et al., Science 278, 1743 (1997) [Abstract/Free Full Text] .
2. A. Seiff, et al., J. Geophys. Res 102, 4045 (1997) [CrossRef].
3. A. Seiff and D. B. Kirk, ibid. 82, 4364 (1977).
4. R. T. Clancy, et al., Icarus 122, 36 (1996) [CrossRef] [ISI].
5. S. L. Hess, R. M. Henry, C. B. Leovy, J. A. Ryan, J. E. Tillman, J. Geophys. Res. 82, 4559 (1977) .
6. W. M. Folkner, C. F. Yoder, D. N. Yuan, E. M. Standish, R. A. Preston, Science 278, 1749 (1997) [Abstract/Free Full Text] .
7. W. M. Folkner, personal communication.
8. T. Rivell, et al., J. Spacecr. Rockets 34, 265 (1997) [ISI].
9. P. H. Smith, et al., Science 278, 1758 (1997) [Abstract/Free Full Text] .
10. R. M. Haberle, J. R. Barnes, J. R. Murphy, M. M. Joshi, J. Schaeffer, J. Geophys. Res. 102, 13301 (1997) [CrossRef].
11. R. W. Zurek et al., in Mars, H. H. Kieffer, B. M. Jakosky, C. W. Snyder, M. S. Matthews, Eds. (Univ. of Arizona Press, Tucson, 1992), pp. 835-933.
12. M. Lopez-Puertas and M. A. Lopez-Valverde, Icarus 114, 113 (1995) [CrossRef] [ISI].
13. P. B. James, H. H. Kieffer, D. A. Paige, in Mars, H. H. Kieffer, B. M. Jakosky, C. W. Snyder, M. S. Matthews, Eds. (Univ. of Arizona Press, Tucson, 1992), pp. 934-968.
14. R. T. Clancy, unpublished data.
15. P. J. Gierasch and R. M. Goody, Planet. Space Sci. 16, 615 (1968) [CrossRef].
16. R. M. Haberle, H. C. Houben, R. Hertenstein, T. Herdtle, J. Atmos. Sci. 50, 1544 (1993) [CrossRef].
17. M. J. Wolff, R. T. Clancy, P. B. James, S. W. Lee, J. F. Bell III, Bull. Am. Astron. Soc., in press.
18. J. A. Magalhães and R. E. Young, Icarus 113, 277 (1995) [CrossRef] [ISI].
19. R. W. Zurek and C. B. Leovy, Science 213, 437 (1981) [Abstract/Free Full Text] .
20. J. E. Tillman, J. Geophys. Res. 93, 9433 (1988) .
21. J. R. Murphy, C. B. Leovy, J. E. Tillman, ibid. 95, 14555 (1990).
22. J. E. Tillman, N. C. Johnson, P. Guttorp, D. B. Percival, ibid. 98, 10963 (1993).
23. R. D. Sharman and J. A. Ryan, J. Atmos. Sci. 37, 1994 (1980) [CrossRef].
24. R. J. Wilson and K. Hamilton, ibid. 53, 1290 (1996).
25. B. M. Jakosky, A. P. Zent, R. W. Zurek, Icarus, in press.
26. J. A. Ryan and R. D. Lucich, J. Geophys. Res. 88, 11005 (1983) .
27. P. Thomas and P. J. Gierasch, Science 230, 175 (1985) [Abstract/Free Full Text] .
28. J. E. Tillman, L. Landberg, S. E. Larsen, J. Atmos. Sci. 51, 1709 (1994) [CrossRef].
29. A. O. Nier and M. B. McElroy, J. Geophys. Res. 82, 4341 (1977) .
30. The ASI/MET science team would like to acknowledge the enthusiasm, dedication, and hard work of the engineering teams who built the accelerometer and MET instruments; the Pathfinder spacecraft engineering team who got them to Mars; and the Pathfinder mission operations team who made the instrument do what we wanted and got the data back to Earth. We also thank T. Clancy for supplying unpublished microwave profiles to the ASI/MET team within days of the original observations. The research described in this paper was carried out under contract with NASA.