HRDI Measures winds by determining the Doppler shiftof absorption and emission lines in the O2 Atmospheric band. Emission lines in the (0-0) A band are used to make measurements in the mesosphere and lower thermosphere (MLT), while absorption lines in the B(1-0) and gamma (2-0) are used in the stratosphere. The Atmospheric band of O2 is one that is particularly useful for probing the MLT region, since during the day, the entire atmosphere from ~50 to ~110 km emits this band very strongly. As shown in Fig. 1, several excitation sources result in a nearly constant brightness over the MLT region. See Bucholtz et al. (1986) for a further discussion of these mechanisms. At night all of these sources disappear except for the weak chemical source near 95 km. This is of sufficient brightness to allow wind measurements at this altitude. Since the vibrational band has its lower state in the ground state, the individual rotational lines can be significantly self-absorbed. For this reason altitude scans in the atmosphere require the use of different strength lines.

In addition, since it is desired to obtain a measurement of the rotational temperature as well as the wind, measurements at each altitude consist of at least 2 different strength rotational lines. These lines are alternated as the telescope scans up and down through the atmosphere. Near the top of the atmosphere lines that have large and medium strengths are used. Since the atmosphere is optically thin at high altitudes, this results in the largest signal. At a lower altitude (82.5 km) the pair switches to a medium and weak line. These are required to penetrate to the lowest altitudes, although there is some absorption in even the weakest line which requires compensation in the data analysis. Figure 2 shows the observed spectra from an up and down scan. In scan 1 a strong line is used from 110 to 120 km while the medium strength line is used from 65 to 110 km. In scan 2 the weak line is used from 50 to 82.5 km and the strong from 85 to 110 km. Note that above 110 km and below 65 km the lines are not paired. This means it is not possible to retrieve rotational temperatures outside this region, but it is possible to recover winds. The viewing described here corresponds to one look direction (forward or back). In order to form a vector wind it is necessary to view approximately the same volume in space from two nearly orthogonal look directions. This is accomplished by using the spacecraft motion to allow viewing from different locations along the orbit track. This occurs at intervals of approximately 9 minutes.

Figure 1. Excitation mechanisms for the O2 Atmospheric band. S-R=Schumann Runge. Only the chemical source operates at night.

Figure 2. Example of the spectra observed by HRDI during an up and down scan. Breaks in the data indicate a change in rotational lines.

The measured spectra are converted to a wind vector in a multitple step process. First, the centroid of spectrum is located on the detector. This is converted to a line of sight apparent wind by subtracting the location of the line in the absence of a Doppler shift, correcting for instrument thermal and long-term drift effects, and removing Earth rotation effects. This apparent line of sight wind is inverted using kernels similar to those shown in Fig. 3. The inversion is combined with a sequential estimator (Ortland et al., 1995) which filters out the high-frequency noise components and combines the two look directions into a vector wind.

One of the important issues concerning HRDI data is the size of the footprint the data represents. Along the line of sight, this is dictated by the observational geometry and the emission rate profile. Since the Atmospheric band is so bright over such a large altitude region, the effective length through the atmosphere is somewhat larger than it would be if the emission source were a narrow layer. An examination of these factors reveals that 50% of the total signal comes from about ±300 km of the tangent point at 50 km, decreasing to ±150 km at 110 km altitude, and 90% comes from ~750 km at 50 km and 400 km at 110 km altitude. The direction perpendicular to the line of sight is determined by the field of view and spacecraft motion. The horizontal field of view is 1.3 degrees, which corresponds to ~75 km. An individual measurement from any altitude employs an integration period of 1 to 3 seconds, during which time the spacecraft moves 7.5 to 22 km. This individual measurement is combined with the scan of which it is a part and the corresponding up or down scan. This entire up and down sequence takes approximately 1 minute, and the spacecraft moves about 450 km in this time. Data in the vertical are collected every 2.5 km. Table I summarizes the data sampling.

              Table I. Data grids

__________________________________________________________________
   Parameter                   Value
__________________________________________________________________
   Altitudes                  50-115 km (mesosphere)
                               10-40 km (stratosphere)
   Altitude resolution	         2.5 km
   Along the track resolution  ~2000 km
   Cross track resolution      ~2650 km (at equator)
   Footprint                    ~500 km x~500 km x 2.5 vertical
__________________________________________________________________

Coverage

The orbital inclination and altitude, plus the fact that HRDI can take measurements throughout most of the MLT region only when it is in sunlight, place restrictions on the latitude and local time coverage that is possible. In addition, the instrument observation time must be divided between the stratosphere and MLT regions. The limitations due to the spacecraft inclination are shown in Figs. 4 (a-f). These figures show the tangent point orbit track for both the warm (side of the spacecraft facing the sun) and cold (side of spacecraft away from the sun) sides. The terminator is shown as a solid line and the 80° solar zenith angle denoted as a dashed line. The plot is shown centered around noon universal time, but the x axis could readily be interpreted as local time, with 0° longitude equivalent to noon local time and 180° local midnight. The location of the sub-solar point is shown by the asterisk at about 0° longitude. HRDI changes from the daytime mode to the nighttime mode, and visa versa, when the solar zenith angle at the spacecraft reaches 90 degrees. The large symbols indicate measurements made by the daytime mode and the small symbols the nighttime. There is occasionally an overlap between the two, which means that one of the two views (forward or backward) would see a given location in the daytime mode, while the other view would be in the nighttime mode.

In order to form a meaningful vector wind, it is necessary to have both views in the same mode. This serves to slightly limit the latitude coverage. In Fig. 4(a) the situation is shown for 11 Dec. 1994. This is what is referred to as the "solstice condition." On this particular day, the warm side observation are always illuminated, albeit at times with a very large solar zenith angle, and the cold side is always dark. For daytime operations there is no sense in using the cold side, and HRDI operations in this time period are limited to the warm side. The latitude coverage during this time period is also limited, with vectors only possible from ~40°N to ~15°S. There are some latitudes that are double noded, that is they are viewed twice at different local times. Figure 4 (b) shows the situation approximately 2 weeks later. The orbit has shifted so the beta angle (the angle between the orbit plane and the sun) is much smaller. The warm side now views the atmosphere from ~72°S to ~30°N. The solar zenith angle will vary from nearly zero to 180°. The cold side permits a view a little more northward with views from 40°S to 50°N. The northward view is limited by the terminator, which extends fairly far south in this, the northern hemisphere winter. By 16 Jan [Fig. 4 (c)] the view is somewhat opposite to that on 11 Dec. Both sides of the spacecraft provide good views of the south, with the cold side providing measurements to ~72°S. Views to the north are now limited to ~15°. Note that as the orbit precesses, the cold side will have a good view of the high southern latitudes throughout this period. It turns out that extended views of high latitudes can only be achieved by viewing the cold side. By 20 Feb. [Fig. 4 (d)], the seasons are starting to change as northern hemisphere winter turns to spring. The viewing geometry is the same as it was for 11 Dec., except the cold side can be used to extend the coverage to higher northern latitudes. Figure 4(e) (10 March) shows the situation near both equinox and zero beta. The warm side views to the south (72°S to 40°N) and the cold side to the north (72°N to 40°S). Note the regions between 40°S and 40°N are well sampled by viewing on either side. The latitude range is limited by the orbital inclination and not by the terminator. Equinox periods are the best times for viewing the maximum latitude range. Finally, Fig. 4 (f) shows the viewing on 29 March. The cold side provides a good view of the south, although shortly the terminator line will progress north and bring the cold side into darkness. The warm side gives a view from ~40°S to just about the equator. These results are summarized in Fig. 5, which shows the latitude extent possible when viewing the warm and cold side. The following general conclusions can be reached:

Figure 4. Viewing Geometry. Warm side tangent points (squares), cold side tangent points (circles).

1. Extended views of high latitudes are possible only on the cold side and only when the hemisphere of interest is near summer solstice.
2. High beta angles, which occur about every 36 days, allow good views of one hemisphere, but not both. The same hemisphere can be viewed at approximately every other month. Multiple local time observations can sometimes be achieved under these conditions.
3. Low beta angles, which also occur about every 36 days, provide the maximum latitude coverage. The greatest extent, ±72° occurs near equinox.
4. Because both warm and cold sides give views between ±40° on most occasions, the largest concentration of data is between these latitudes. Figure 6 shows the local time covered during the month of January 1995. There is reasonable coverage in local time, particularly in the southern hemisphere.

Figure 7. Coverage maps for two consecutive days. This illustrates alternate days of stratosphere and MLT measurements.

The philosophy of how best to use the HRDI instrument has changed somewhat since the beginning of the mission. Initially, measurements on the spacecraft warm side were emphasized, as this maximized the number of photons that could be collected in a day. However, this limits the latitude coverage. Figure 7 shows the latitude coverage for two consecutive days. There are 15 orbits per day and the latitude coverage for each is nearly the same. The cold side was observed mainly near the times of small solar beta angles (angle between the orbit plane and the sun), when the illumination was nearly the same on both sides of the spacecraft. Modes were usually (but not exclusively) selected so that the MLT region was observed one day, followed by the stratosphere the next. Later, it was decided to extend the latitude coverage as much as possible which meant observing the cold side more often. This complicated the division of time between the stratosphere and MLT region somewhat. On most occasions, the instrument is operated so that on alternate orbits the stratosphere and MLT regions are viewed and on alternate days the warm or cold sides are observed This is illustrated in Fig. 8. Note here that for each region only half as many orbits per day are obtained as for the previous sequence, but over a two-day period exactly the same number of orbits are provided. The two-sided viewing effectively extends the latitude coverage, but there is a local time difference between the two. Except at high latitudes, the local time at any latitude is very nearly the same throughout a day. The precession rate of the orbit is such that it takes 36 days to sample a complete 24 hours of local time and hence the local time varies by 20 minutes a day.

Figure 8. Coverage map showing coverage for two consecutive days. This shows the effect of looking on alternate sides of the spacecraft to extend latitude coverage.

Last Update: 22-Apr-2003
Updated by: meyerkm@umich.edu