DATA_SET_DESCRIPTION |
Data Set Overview
=================
This data set contains two-dimensional infrared thermal maps of the
surface of comet 9P/Tempel 1. The maps were derived from three
spatially resolved scans of the nucleus acquired by the Deep Impact
High Resolution Infrared Spectrometer (HRII) about 19, 12, and 5
minutes before the impact on 4 July 2005. A high-resolution,
120-m/pixel, thermal composite map is also included. Surface
temperatures were derived from 1.0- to 4.0-micron data and ranged
from 272K to 336K +/- 7K. This data set also includes the incidence
and emission angle maps (mu0 and mu) associated with each thermal map
and a table of temperatures assigned to plates in the Tempel 1 shape
model that were illuminated and visible around the time of collecting
the thermal data. For a thorough discussion of the maps, see
Groussin, et al. 2007 [GROUSSINETAL2007].
Observations
------------
The characteristics of the three HRII scans acquired of the nucleus
and used to derive the three temperature maps were extracted from
Groussin, et al. 2007 [GROUSSINETAL2007] and provided below. The
given values are with respect to the mid-point of a scan. An impact
time of 05:44:34.265 UT 4 July 2005, at the flyby spacecraft, was
used to calculate the time before impact. During the scans,
9P/Tempel 1 had a heliocentric distance of 1.5 AU.
Exposure (scan) ID : 9000031 9000036 9000039
Image mode : 3 1 2
Number of frames in scan : 32 40 84
Percentage of visible nucleus covered : 90 100 50
Time at spacecraft (h:m:s UT) : 05:25:36 05:32:55 05:39:18
Time before impact (min) : 19.0 11.7 5.3
Scan duration (s) : 23 113 120
Slit length* (km) : 13 41 15
Minimum spectral resolution : 216 216 216
Target distance (km) : 20400 15900 12000
Spatial resolution (m/pixel) : 204 159 120
Phase angle (deg) : 62.88 62.89 62.91
* There is not correlation between the slit length and the
resolution. The slit length only refers to the number of
rows that are read out from the IR focal plane array (FPA);
the physical length of the slit never changes. Therefore,
as the spacecraft moved closer to the target, the spatial
resolution improved. The value for slit length changes
between the scans because the image mode is different
for each set of frames. The image mode specifies the
number of rows readout from the IR FPA in the along-slit
(spatial) dimension. For image mode 1 the slit width is
256 2x2-binned pixels, for mode 2 it is 128 2x2-binned
pixels, and for mode 3 it is 64 2x2 binned pixels.
Additionally, the physical width of the slit is fixed at
10 microradians and cannot be changed.
** Please note the map associated with exposure ID 9000039
is truncated because of impact activities by the spacecraft.
Related Data Sets
-----------------
The following PDS data sets are related to this one:
DIF-C-HRII-3-4-9P-ENCOUNTER-V2.0 : Calibrated HRII encounter
data, version 2.0
DIF-C-HRIV/ITS/MRI-5-TEMPEL1-SHAPE-V1.0 : Tempel 1 shape model
DI-C-HRII/HRIV/MRI/ITS-6-DOC-SET-V1.0 : Mission documentation
Please note that the temperature maps were derived from version 2.0
of the calibrated HRII data. Version 2.0 includes improved geometric
parameters, dark subtraction, and bad pixel identification.
Data
====
Individual Thermal Maps and Related Geometry Maps
-------------------------------------------------
Three individual thermal maps of the surface of Tempel 1 along
with the related incidence- and emission-angle maps were generated
from data for HRII scans 9000031, 9000036, and 9000039. Please note
the map for scan 9000039 is truncated because of impact activities
by the spacecraft. These two-dimensional maps are presented in the
FITS image format. The data values in the maps have units of kelvin.
The data files are listed here:
map_9000031.fit : Temperature maps derived from the
map_9000036.fit scans with exposure IDs of 9000031,
map_9000039.fit 9000036, and 9000039
map_9000031_mu0.fit : Registered maps of the values of mu0
map_9000036_mu0.fit (cosine of the incidence angle) for
map_9000039_mu0.fit each temperature map listed above
map_9000031_mu.fit : Registered maps of the values of mu
map_9000036_mu.fit (cosine of the emission angle) for
map_9000039_mu.fit each temperature map listed above
For context, a deconvolved image acquired by the HRI visible CCD
is included in the documentation for this data set.
Composite Thermal Map and Related Geometry Maps
-----------------------------------------------
The composite thermal map and its related geometry maps for mu0
(cosine of the incidence angle) and mu (cosine of the emission
angle) were generated from the highest-resolution thermal maps
derived from data for scans 9000036 and 9000039. The
two-dimensional maps are presented in the FITS image format.
The data values in the maps have units of kelvin.
thermal_composite.fit : Composite thermal map at 120 m/pixel
made from the highest-resolution
thermal maps (map_9000036.fit and
map_9000039.fit)
thermal_composite_mu0.fit : Registered map of values of mu0
(cosine of the incidence angle) for
the composite thermal map
thermal_composite_mu.fit : Registered map of values of mu
(cosine of the emission angle) for
the composite thermal map
Plate Temperature Table
-----------------------
The plate_temp.tab file is a fixed-width, ASCII table that provides
a temperature for each plate of the Tempel 1 shape mode that was
visible near impact. The first column in the table identifies the
plate index (0-based) within the archived shape model for Tempel 1.
The second column provides the corresponding temperature in units
of Kelvin.
The PDS data set, DIF-C-HRIV/ITS/MRI-5-TEMPEL1-SHAPE-V1.0, includes
the Tempel 1 shape model used for this analysis. It is available
online at the Small Bodies Node.
Map Orientation
---------------
When displaying the thermal or geometry maps, a true-sky view can
be achieved by displaying the image using the standard FITS
convention where the fastest-varying axis (samples) increase to
the right in the display window and the slowest-varying axis
(lines) increases to the top. This convention is also defined by
these keywords in the PDS data labels:
SAMPLE_DISPLAY_DIRECTION = RIGHT
LINE_DISPLAY_DIRECTION = UP
When using this convention to display the maps, the Sun is to the
right, ecliptic north is toward the right, celestial north is toward
the upper-right, and celestial east is toward the upper-left. The
projected positive spin axis of the nucleus points toward the top.
This is the orientation used by Groussin, et al. (2007)
[GROUSSINETAL2007].
Processing
==========
Thermal Maps
------------
Frames for each scan (9000031, 9000036, and 9000039) were
selected from version 2.0 of HRII reversibly calibrated data in
units of radiance (W/(m**2 micron steradian)). Because scattered
light is not negligible for the HRII spectrometer, the calibrated
data were then manually corrected (cleaned) for bad pixels and
scattered light for this analysis. The contribution from the coma
was negligible compared to that of the nucleus and was therefore
not subtracted. See Groussin, et al. (2007) [GROUSSINETAL2007]
for details about the scattered light model and the analysis of
the coma contribution. See the Deep Impact Instrument Calibration
document by Klaasen, et al. (2006) [KLAASENETAL2006] for a
discussion of the stray light for the HRII instrument and the
calibration pipeline.
To derive the temperature maps from the spectra of the nucleus, a
process that combined the scattered light from the Sun with the thermal
infrared emission of the nucleus was used to fit the data. The best
fit was achieved by using a least squares method in the thermal
infrared range of 3.0 to 4.0 microns. This method was applied to each
frame of one scan to derive a temperature map of the nucleus for that
scan. Details about the process and the analysis of the maps using
a thermal model are described by Groussin, et al. (2007)
[GROUSSINETAL2007]
Geometry Maps (mu0 and mu) for the Thermal Maps
-----------------------------------------------
The mu0 and mu maps were calculated with the shape model of Tempel
1 archived in PDS for the time of impact. The sub-spacecraft
longitude and latitude are 295 deg and -26 degrees, respectively.
The sub-solar longitude and latitude are 349 deg and 9.2 degrees,
respectively. The thermal data used to construct the thermal maps
were collected 19, 12, and 5 minutes before the impact, when the
change in the viewing geometry was less than 2 degrees. To
register the geometry maps with thermal maps as precisely as
possible, the geometry maps were first rendered at a pixel scale of
16m/pix, and the thermal maps were rescaled to this resolution for
the purpose of registration. The geometry maps and thermal maps
were then registered by shifting and slightly rotating according to
the limb/terminator and a few large features. Finally the
registered geometry maps were scaled down to the original pixel
scale of their corresponding thermal maps. If a pixel was in
shadow, its value was set to zero.
The uncertainty of these geometry maps was affected by the
uncertainty of the shape model and registration error. Close to
the limb and terminator, the uncertainty of angles can be large.
Thus pixels with value less then 0.17 (80 deg angle) should
probably be discarded for practical purposes.
Composite Thermal Map
---------------------
The composite thermal map was constructed from the two
highest-resolution thermal maps for scans 9000036 and 9000039.
First, the thermal map for 9000036 at 159 m/pixel was scaled to
120 m/pixel, the scale of the thermal map for exposure 9000039.
Then the rescaled map for 9000036 was registered against the
thermal map for 9000039 by shifting and rotating to align the limb
and terminator profile and some large surface thermal features.
Six lines (35-40) of thermal map 9000039 appeared to be bad during
a visual inspection and were therefore discarded. Finally the
upper half of the nucleus in thermal map 9000036 that was covered
by map 9000039 was replaced by lines from the latter and were
therefore not used. The geometry maps (mu0 and mu) for the
composite thermal map were calculated using the procedure as
described above for each individual thermal map.
Plate Temperature Table
-----------------------
As in the plate shape model, the surface of the nucleus of Tempel 1
was represented by a collection of triangle plates. From the
thermal maps, the temperatures of the illuminated and visible
plates at the time of impact could be calculated. To do this, the
composite thermal map with pixel scale 120 m/pixel was used. A map
of plate numbers was rendered using the geometry of impact, where
the value of a pixel represented the index of the plate that was
projected onto this pixel in the composite thermal map. The pixel
scale of the resulting plate map was 16m/pix, a scale that was
finer than the sizes of most plates (so that every pixel fell into
one plate). The composite thermal map was then rescaled to this
resolution, so that the plate map could be registered to it by
comparing the limb/terminator profile and a few large features.
Comparing the composite thermal map with the registered plate map,
the temperatures of illuminated and visible plates were extracted.
Due to the uncertainty of the plate shape model and the low signal
close to the limb and terminator of the nucleus in the thermal map,
the edge of the plate map and thermal map could not be aligned
everywhere. Therefore, all plates with their average values of mu0
or mu smaller than 0.17 (80 degrees) were discarded from this data
table. For the plates included, the uncertainties in the
temperature were dominated by the composite thermal map, which was
in turn affected by the registration error during composition, as
well as by the uncertainty of original thermal maps.
This data set overview was provided by O. Groussin and J.-Y. Li.
|