DATA_SET_DESCRIPTION |
Data Set Overview : The LRO/LCROSS spacecraft was launched on June 18, 2009, and crashed into the Cabeus crater near the Moon's south pole on October 9, 2009. The spent upper stage, called the Centaur, impacted the moon at 11:31:19.51 UTC at -84.68 deg latitude, -48.69 deg longitude, Mean Earth frame, and the LCROSS shepherding spacecraft impacted the surface at 11:35:34 UTC. This dataset consists of frame-transfer CCD observations of the LCROSS impact site as well as unresolved images of the primary impactor in the final hours before impact. The instrument is described in Young et al. (2010) [YOUNGETAL2010]. For this event two micro-Max cameras were deployed on the dichroic holder assembly so that two simultaneous imaging streams could be collected. All timing for the system is controlled by custom GPS-based units that provide timing accuracy good to ~3 milliseconds. Our observations were taken at the 2.4-m telescope on Magdalena Ridge. This facility is owned and operated by New Mexico Tech under the direction of Eileen Ryan. This is a new alt-az mount telescope system designed for fast and accurate tracking. It provides a Naysmith port with integral field de-rotator and filter wheel. The instrument was mounted on this port and rotated so that North was up in the raw imaging data. Telescope and weather status information were collected by separate programs and were interpolated to populate the final data headers for each image. The dichroic splits the range of a CCD detector into two roughly equal components. The blue light is reflected and the red light is transmitted. Internal reflections in the dichroic create low-level ghost images that are readily apparent on very bright objects. These reflections, while present in the primary data, do not affect the system sensitivity to the LCROSS impacts. During the pre-event imaging this was the only filtering element in the beam. The data from the Gjon camera was the blue channel and the data from the Doc camera was the red channel. The lunar observations were completely saturated with just the dichroic. Event night was the only night we were able to take data on the Moon without clouds and that was the night we discovered the saturation problem. The only tool available to cut the flux down was the standard filter wheel on the telescope. It was hoped that the Bessel U filter had sufficient red-leak to serve as a UV and Red attenuator. At the telescope we certainly did see the flux in both channels drop into a useful range. The flat fielding pattern also looked different giving some confidence to the interpretation that we were seeing different wavelengths. Further research into the characteristics of the dichroic and the filters revealed that the dichroic was not perfectly reflecting short-ward of 400 nm. As a result, the throughput for each channel was almost exactly the same with the same effective wavelength. Processing : The images have been processed to remove bias, dark-current, and pixel-to-pixel quantum efficiency variations. The native unit of the data is ADU or detected electrons/photons. The conversion to absolute flux is made by using calibrated images of the Moon from the ROLO project provided by T. C. Stone at USGS (ROLO data described in Kieffer and Stone, 2005) [KIEFFER&STONE2005]. Data : As soon as the impactor rose above the local horizon we began imaging observations for the purposes of extracting a light curve. These observations reveal a very strong light curve, in excess of 4-5 magnitudes, caused by the tumbling motion of the cylindrical booster stage. The object is known to be in non-principal axis rotation and the data are consistent with this in that details of the light curve features do not closely repeat on successive rotations. The images are flat-fielded as well as could be done but the later images suffer from severe scattered light as the projectile got ever closer to the Moon. Simple photometry from the images shows a strong light curve but the data were not calibrated due to other constraints on our time on event night. However, the light curve has such a large amplitude that calibration corrections are small. Absolute calibration is also not as important for an object of a known size and albedo (color). Data were taken on standard regions of the Moon as well as the impact site at a variety of airmasses through the night. All data were taken as high-speed frame-transfer image cubes. The read-out rate limit is determined by the speed of the A/D converter as well as the chip clocking speed. The rates in the data correspond to the fastest rate possible for the given sub-array size. As the night progressed we worked with ever smaller sub-arrays in order to get the fastest possible imaging sequence. The impact observations did not reveal any observable effect from the impact. Quick-and-dirty image differencing immediately after the event did not reveal any signature. Later, more detailed, efforts to subtract a template lunar image to find a trace of the impact flash was also unsuccessful. Given the brightness observed by the LCROSS spacecraft we would have been able to see the flash if we did not also have fully illuminated portions of the lunar surface in the same field of view. The 16-bit A/D converter on the CCD system simply has insufficient dynamic range to get the flash and the illuminated surface at the same time. The ratio between the two was so large that we were unable to work around a saturated lunar surface being in the field-of-view. Once the filter was introduced the signal was cut to a manageable range but the flash would have been much too faint to see. Media/Format : All data are stored in FITS format.
|