LCROSS Plume visibility analysis and imaging exposure recommendations
2 views
Skip to first unread message
cano...@yahoo.com
Sep 15, 2009, 4:30:28 AM9/15/09
to LCROSS_Observation
The following is an amateur note. Comments and criticisms on the same
are welcomed.
At url -
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/20090915LCROSSvisibilityanalysis_kaf.txt
http://tinyurl.com/mqylqf
- I have posted another in a series of ongoing discussions between
myself and Jim Mosher regarding the expected visibilty of the impact
plume. My response to Jim is overlength and goes into some technical
matters that may not be of interest to general group readers.
Therefore, I have posted by link to a separate text file.
My amateur analysis of the plume ends with my conclusions regarding
whether the plume can be observed visually and imaged. The extended
discussion also concludes with a description of what the plume will
look like for amateurs using scopes of 5 inches or larger and with
recommended strategies for imaging the plume. The support and
reasoning for my conclusions are contained in the linked extended
discussion file.
Those conclusions and recommendations may be of interest to general
readers and I replicate that "conclusions" portion of the extended
response below.
Clear skies - Kurt
-----------------------------
- Conclusions
On Sept. 11, Rick Baldridge noted that:
"NASA will provide professional and amateur observing groups more
detail regarding plume size and visibility in the coming weeks. The
plume will not extend above the lunar limb, and will not be situated
against a dark background such as a shadowed region between craters.
However, that does not mean the plume will not be visible. Video and
photographic observations must now focus on bringing out the
brightening caused by the eject plume in front of a lit lunar
surface."
September 11, 2009 LCROSS Science Team Announced Target Crater
Posted by rickbaldridge on September 11th, 2009.
NASA LCROSS Citizen Science Page Blog
http://apps.nasa.gov/lcross/
Implications from the foregoing discussion [in the linked extended
discussion file] for imaging the LCROSS impact are as follows.
If the ejecta plume follows its predicted apparent brightness _and_ if
the mpsas of the shadowed portion of crater Cabeus A1 is at least one
magnitude lower than 4 mpsas, the impact can be easily imaged and
observed using 5 inches of aperture or more. Earth based amateur
observers will see essentially a surface contrast effect. As the
brightest part of the 10km plume expands within and above the bowl of
17km Cabeus A1 and reaches a brightness of about 4.0 mpsas, the plume
will obscure the shadowed portion of Cabeus A1 from Earth observers.
The plane of the rim and/or just above the rim will take on a surface
brightness equal the surface brightness of the surrounding surface
terrain. That surface typically will have an mpsas between 4.0 and
6.0.
Normally, Caebus A1 has a bowl shaped appearance caused by crater
shadowing when viewed from Earth visually at 300x and/or when imaged,
as illustrated in this high resolution amateur image taken in
September by Stefan Lammel:
http://tinyurl.com/qjex6e
(I visually observed the same view a few hours after Stefan took his
image using a Meade ETX 125 with a TMB 4mm planetary eyepiece at
300x. The level of detail that I was to observe visually easily
exceed what Stefan was able to capture photographically. Stefan when
posting his image accompanied the comment that the Moon was at a low
altitude and that it was not his best work.)
During the first 30 secs, the brightest part of the ejecta plume will
create a contrast effect as the surface brightness of the brightest
part of the ejecta plume removes the contrast between the crater
shadow and surrounding terrain. For visual observers, the crater may
see a blurry mesa effect. Caebus A1 will look like a terresterial mesa
or will have a slight "raised muffin" appearance. Where the brightest
part of the ejecta cloud crosses the sunlit portion of the lunar
surface, it will not be visible due to lack of contrast between the
cloud and the lunar surface.
For low resolution imagers like myself, the crater will simply
"disappear" due to this contrast effect beginning near the 30 second
mark for a duration of about 30 seconds and then mysteriously reappear
on AVI frames. See my 9-10-2009 image for an example of low-
resolution image:
http://tinyurl.com/qw32fy
Although no images have been gathered that match the libration and
illuminated fraction of the impact, the reduction in libration in
latitude to -6.0 to -3.0 between now and the impact on October 9 will
only increase this contrast effect. The Cabeus A1 crater shadow will
be relatively thinner (in arscsecs) on the day of impact as compared
to that shown in Stefan's image.
Imaging of the LCROSS impact will be a fairly straight forward process
for amateurs. High focal length imaging is preferred in order to
minimize the percent of the sunlit lunar disk captured in a frame.
High focal lengths dictate that large pixel DSLR cameras and CCD
cameras are disfavored relative to small pixel sized fast moderate and
high-end lunar imaging cameras. See Sinnott's Effective Focal Length
to Pixel Size nomogram, url -
http://media.skyandtelescope.com/images/Linked.gif
- and the more detailed discussion in another message in the LCROSS
Observation newsgroup -
Post by K. Fisher 9-8-2009 LCROSS Observation Group
Efls for imaging the LCROSS impact
http://groups.google.com/group/lcross_observation/msg/764ceeede969207a
http://tinyurl.com/owkdf7
Pre-impact image calibration is an easy three-step process. The goal
of this process is to set the pixel value of the brightest edge of the
rim of Cabeus A1 to 75% of your camera's well capacity in ADUs. This
should assure that the full range of pixel values that can be captured
on a line profile across the major axis of crater Cabeus A1 are
recorded on images stored to your disk.
First, focus your imaging gear on the target crater without concern
for the exposure setting.
Second, slew to 2.6 stellar magnitude theta Auriga. On the morning of
the impact, the Moon will be between the horn stars of Taurus and just
next to 1.7 mag bet Taurus (Alnath). 2.6 magnitude theta Auriga is
one of the figure stars of Auriga and is about 11 degrees away.
Nearby alternative stars for exposure calibration around 2.5 stellar
magnitudes include: zeta Per 2.8mags B0.5V, delta Orion 2.2 mags
O9.5II, gamma Gem 1.9 mags AOIV, beta Auriga 1.9 mags A2IV.
Take some test images of theta Tau and adjust your exposure setting so
theta Tau's brightness peaks at 50% of your well capacity.
Note that the preview histogram in some image capture software _does
not accurately_ represent what is stored in captured images on a
disk. Open the test images stored on the disk and run a profile
measurement or histogram on your image of theta Aur using your image
processing software so you are sure your exposure setting captures the
right amount of well ADUs.
Keep this exposure setting and slew back onto the impact target Cabeus
A1.
Take some test frames and look at some of the raw images on your
disk. Use the profile measuring line tool (e.g. one is available in
AIP4WIN) and take a profile of the pixels that cross the major axis of
Cabeus A1.
Now adjust your exposure setting so that high pixel value of Cabeus A1
rim is at 75% of your well capacity. The minimum pixel value shown
for the Cabeus A1 crater line profile should also be within the range
of a histogram made of the entire test image.
You can slightly adjust back from this exposure setting so the Moon is
not overexposed _on images stored on your disk._ Again, _do not
trust the preview image and histogram_ in your image capture
software.
Post-image processing will favor software packages that offer region
masking like Photoshop. This way individual regions of the bright
lunar surface can be supressed in brightness, but pixels that
encompass the area within the Caebus A1 crater rim can be selectively
stacked and gamma stretched.
The impact will provide imagers interested in hobby science with an
opportunity to study plume kinematics using photometry measurements
from their images. The LCROSS ejecta plume will rise 5 kilometers
(5000 meters) to its maximum brightness in about 30 seconds. The
vertical plume speed is estimated at an average of 167 meters per
second (5000/30). For the first 2 kilometers, the ejecta cloud will be
masked from Earth view by Cabeus A1's crater rim. This trip above the
crater rim will occur between about 18 seconds after impact through
impact + 30 seconds. (30 seconds * 3000 meters / 5000 meters). For
the final three kilometers, the plume will be sunlight and the total
light from the cloud may have a changing photometric signature related
to its vertical travel that will be recorded by amateur video
imagers.
Such recordings might be examined to extract a plot of the total
brightness of the Earth visible ejecta curtain against time. The
process for making such a recording and reducing it is generally
described as follows. An LPI camera of video that records AVI files
including both an audio track and a video track will be needed.
For the time signal audio track, the video can be time stamped using a
digital metronome as the timed audio source. An inexpensive $30 Ibanez
model emits a good sharp tone at a maximum of 180 beats per minute
with a different second identifier signal and is available at many
local music and guitar stores. url: http://www.ibanez.com/electronics/product.aspx?m=MU40
. Alternatively, imagers can use a more expensive video time stamping
rig favored by lunar occulation observers - the KIWI OSDI video-GPS
timestamper. url: http://www.pfdsystems.com/ . The NIST WWW shortwave
time broadcast is another option for a timing audio signal, but since
there is no need to coordinate observations between observers and
clear reception of the NIST shortwave signal is usually problematic,
the Ibanez metronome may be the better inexpensive no-hassle option.
Post imaging, the AVI file is reviewed using movie making software.
The individual frames that contain identifable time beats and good
images of Caebus A1 are separated. Then each time-stamp identified
image is reviewed in image processing software. Most image processing
software (such as AIP4WIN) contain a region measuring tool. These
tools count the average value of pixels in an identified circular or
rectangular area.
Use the region measuring tool on each time-stamped image and surround
all of Cabeus A1. It will be important to use the same relative pixel
coordinates from the center of Cabeus A1 in each individual frame.
Note the time stamp and the average pixel value for the uniform
measuring region on each frame. Then plot the pixel values against
time.
Finally, compare your plot against the predicted increases in plume
brightness that presumably will be provided by the LCROSS Team.
In conclusion, the LCROSS impact can be easily imaged, assuming it
reaches the brightness of 4.0 mpsas stated in LCROSS pre-impact
modeling. The view may not be as dramatic as one might imagine, but
it appears certainly worth trying for.
The success of amateur imaging is dependent on the LCROSS Team
gathering and publishing for amateur use, the apparent brightness of
the surface area around Caebus A1 and the shadowed portion of Cabeus
A1 in both mpsas and V stellar magnitudes. If the shadowed floor of
Cabeus A1 is brighter than 4.0 mpsas, the impact cannot be observed or
imaged by amateurs.
By the NASA LCROSS Team calling for amateur imaging and by inducing,
through press releases stating that the impact is observable, the
public's attendence at private star parties, the LCROSS Team has
undertaken the business ethical obligation to gather and publish such
photometry data prior to September 27 and before Oct. 9. This ethical
obligation is also incured by their dual role as scientists and public
governmental employees.
September 27 represents the last date in which the south lunar pole
will be at 71% illuminated fraction and on which amateurs can make
useful "dry runs" of their imaging gear.
Advanced amateurs with photometric gear may wish to gather and share
their hobbyist studies of the apparent brightness of the shadows of
small craters on the opposite east side of the southern polar Moon on
September 27.
Again, this is an amateur note. Comments and criticisms on the same
are welcomed.
Clear Skies - Kurt
are welcomed.
At url -
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/20090915LCROSSvisibilityanalysis_kaf.txt
http://tinyurl.com/mqylqf
- I have posted another in a series of ongoing discussions between
myself and Jim Mosher regarding the expected visibilty of the impact
plume. My response to Jim is overlength and goes into some technical
matters that may not be of interest to general group readers.
Therefore, I have posted by link to a separate text file.
My amateur analysis of the plume ends with my conclusions regarding
whether the plume can be observed visually and imaged. The extended
discussion also concludes with a description of what the plume will
look like for amateurs using scopes of 5 inches or larger and with
recommended strategies for imaging the plume. The support and
reasoning for my conclusions are contained in the linked extended
discussion file.
Those conclusions and recommendations may be of interest to general
readers and I replicate that "conclusions" portion of the extended
response below.
Clear skies - Kurt
-----------------------------
- Conclusions
On Sept. 11, Rick Baldridge noted that:
"NASA will provide professional and amateur observing groups more
detail regarding plume size and visibility in the coming weeks. The
plume will not extend above the lunar limb, and will not be situated
against a dark background such as a shadowed region between craters.
However, that does not mean the plume will not be visible. Video and
photographic observations must now focus on bringing out the
brightening caused by the eject plume in front of a lit lunar
surface."
September 11, 2009 LCROSS Science Team Announced Target Crater
Posted by rickbaldridge on September 11th, 2009.
NASA LCROSS Citizen Science Page Blog
http://apps.nasa.gov/lcross/
Implications from the foregoing discussion [in the linked extended
discussion file] for imaging the LCROSS impact are as follows.
If the ejecta plume follows its predicted apparent brightness _and_ if
the mpsas of the shadowed portion of crater Cabeus A1 is at least one
magnitude lower than 4 mpsas, the impact can be easily imaged and
observed using 5 inches of aperture or more. Earth based amateur
observers will see essentially a surface contrast effect. As the
brightest part of the 10km plume expands within and above the bowl of
17km Cabeus A1 and reaches a brightness of about 4.0 mpsas, the plume
will obscure the shadowed portion of Cabeus A1 from Earth observers.
The plane of the rim and/or just above the rim will take on a surface
brightness equal the surface brightness of the surrounding surface
terrain. That surface typically will have an mpsas between 4.0 and
6.0.
Normally, Caebus A1 has a bowl shaped appearance caused by crater
shadowing when viewed from Earth visually at 300x and/or when imaged,
as illustrated in this high resolution amateur image taken in
September by Stefan Lammel:
http://tinyurl.com/qjex6e
(I visually observed the same view a few hours after Stefan took his
image using a Meade ETX 125 with a TMB 4mm planetary eyepiece at
300x. The level of detail that I was to observe visually easily
exceed what Stefan was able to capture photographically. Stefan when
posting his image accompanied the comment that the Moon was at a low
altitude and that it was not his best work.)
During the first 30 secs, the brightest part of the ejecta plume will
create a contrast effect as the surface brightness of the brightest
part of the ejecta plume removes the contrast between the crater
shadow and surrounding terrain. For visual observers, the crater may
see a blurry mesa effect. Caebus A1 will look like a terresterial mesa
or will have a slight "raised muffin" appearance. Where the brightest
part of the ejecta cloud crosses the sunlit portion of the lunar
surface, it will not be visible due to lack of contrast between the
cloud and the lunar surface.
For low resolution imagers like myself, the crater will simply
"disappear" due to this contrast effect beginning near the 30 second
mark for a duration of about 30 seconds and then mysteriously reappear
on AVI frames. See my 9-10-2009 image for an example of low-
resolution image:
http://tinyurl.com/qw32fy
Although no images have been gathered that match the libration and
illuminated fraction of the impact, the reduction in libration in
latitude to -6.0 to -3.0 between now and the impact on October 9 will
only increase this contrast effect. The Cabeus A1 crater shadow will
be relatively thinner (in arscsecs) on the day of impact as compared
to that shown in Stefan's image.
Imaging of the LCROSS impact will be a fairly straight forward process
for amateurs. High focal length imaging is preferred in order to
minimize the percent of the sunlit lunar disk captured in a frame.
High focal lengths dictate that large pixel DSLR cameras and CCD
cameras are disfavored relative to small pixel sized fast moderate and
high-end lunar imaging cameras. See Sinnott's Effective Focal Length
to Pixel Size nomogram, url -
http://media.skyandtelescope.com/images/Linked.gif
- and the more detailed discussion in another message in the LCROSS
Observation newsgroup -
Post by K. Fisher 9-8-2009 LCROSS Observation Group
Efls for imaging the LCROSS impact
http://groups.google.com/group/lcross_observation/msg/764ceeede969207a
http://tinyurl.com/owkdf7
Pre-impact image calibration is an easy three-step process. The goal
of this process is to set the pixel value of the brightest edge of the
rim of Cabeus A1 to 75% of your camera's well capacity in ADUs. This
should assure that the full range of pixel values that can be captured
on a line profile across the major axis of crater Cabeus A1 are
recorded on images stored to your disk.
First, focus your imaging gear on the target crater without concern
for the exposure setting.
Second, slew to 2.6 stellar magnitude theta Auriga. On the morning of
the impact, the Moon will be between the horn stars of Taurus and just
next to 1.7 mag bet Taurus (Alnath). 2.6 magnitude theta Auriga is
one of the figure stars of Auriga and is about 11 degrees away.
Nearby alternative stars for exposure calibration around 2.5 stellar
magnitudes include: zeta Per 2.8mags B0.5V, delta Orion 2.2 mags
O9.5II, gamma Gem 1.9 mags AOIV, beta Auriga 1.9 mags A2IV.
Take some test images of theta Tau and adjust your exposure setting so
theta Tau's brightness peaks at 50% of your well capacity.
Note that the preview histogram in some image capture software _does
not accurately_ represent what is stored in captured images on a
disk. Open the test images stored on the disk and run a profile
measurement or histogram on your image of theta Aur using your image
processing software so you are sure your exposure setting captures the
right amount of well ADUs.
Keep this exposure setting and slew back onto the impact target Cabeus
A1.
Take some test frames and look at some of the raw images on your
disk. Use the profile measuring line tool (e.g. one is available in
AIP4WIN) and take a profile of the pixels that cross the major axis of
Cabeus A1.
Now adjust your exposure setting so that high pixel value of Cabeus A1
rim is at 75% of your well capacity. The minimum pixel value shown
for the Cabeus A1 crater line profile should also be within the range
of a histogram made of the entire test image.
You can slightly adjust back from this exposure setting so the Moon is
not overexposed _on images stored on your disk._ Again, _do not
trust the preview image and histogram_ in your image capture
software.
Post-image processing will favor software packages that offer region
masking like Photoshop. This way individual regions of the bright
lunar surface can be supressed in brightness, but pixels that
encompass the area within the Caebus A1 crater rim can be selectively
stacked and gamma stretched.
The impact will provide imagers interested in hobby science with an
opportunity to study plume kinematics using photometry measurements
from their images. The LCROSS ejecta plume will rise 5 kilometers
(5000 meters) to its maximum brightness in about 30 seconds. The
vertical plume speed is estimated at an average of 167 meters per
second (5000/30). For the first 2 kilometers, the ejecta cloud will be
masked from Earth view by Cabeus A1's crater rim. This trip above the
crater rim will occur between about 18 seconds after impact through
impact + 30 seconds. (30 seconds * 3000 meters / 5000 meters). For
the final three kilometers, the plume will be sunlight and the total
light from the cloud may have a changing photometric signature related
to its vertical travel that will be recorded by amateur video
imagers.
Such recordings might be examined to extract a plot of the total
brightness of the Earth visible ejecta curtain against time. The
process for making such a recording and reducing it is generally
described as follows. An LPI camera of video that records AVI files
including both an audio track and a video track will be needed.
For the time signal audio track, the video can be time stamped using a
digital metronome as the timed audio source. An inexpensive $30 Ibanez
model emits a good sharp tone at a maximum of 180 beats per minute
with a different second identifier signal and is available at many
local music and guitar stores. url: http://www.ibanez.com/electronics/product.aspx?m=MU40
. Alternatively, imagers can use a more expensive video time stamping
rig favored by lunar occulation observers - the KIWI OSDI video-GPS
timestamper. url: http://www.pfdsystems.com/ . The NIST WWW shortwave
time broadcast is another option for a timing audio signal, but since
there is no need to coordinate observations between observers and
clear reception of the NIST shortwave signal is usually problematic,
the Ibanez metronome may be the better inexpensive no-hassle option.
Post imaging, the AVI file is reviewed using movie making software.
The individual frames that contain identifable time beats and good
images of Caebus A1 are separated. Then each time-stamp identified
image is reviewed in image processing software. Most image processing
software (such as AIP4WIN) contain a region measuring tool. These
tools count the average value of pixels in an identified circular or
rectangular area.
Use the region measuring tool on each time-stamped image and surround
all of Cabeus A1. It will be important to use the same relative pixel
coordinates from the center of Cabeus A1 in each individual frame.
Note the time stamp and the average pixel value for the uniform
measuring region on each frame. Then plot the pixel values against
time.
Finally, compare your plot against the predicted increases in plume
brightness that presumably will be provided by the LCROSS Team.
In conclusion, the LCROSS impact can be easily imaged, assuming it
reaches the brightness of 4.0 mpsas stated in LCROSS pre-impact
modeling. The view may not be as dramatic as one might imagine, but
it appears certainly worth trying for.
The success of amateur imaging is dependent on the LCROSS Team
gathering and publishing for amateur use, the apparent brightness of
the surface area around Caebus A1 and the shadowed portion of Cabeus
A1 in both mpsas and V stellar magnitudes. If the shadowed floor of
Cabeus A1 is brighter than 4.0 mpsas, the impact cannot be observed or
imaged by amateurs.
By the NASA LCROSS Team calling for amateur imaging and by inducing,
through press releases stating that the impact is observable, the
public's attendence at private star parties, the LCROSS Team has
undertaken the business ethical obligation to gather and publish such
photometry data prior to September 27 and before Oct. 9. This ethical
obligation is also incured by their dual role as scientists and public
governmental employees.
September 27 represents the last date in which the south lunar pole
will be at 71% illuminated fraction and on which amateurs can make
useful "dry runs" of their imaging gear.
Advanced amateurs with photometric gear may wish to gather and share
their hobbyist studies of the apparent brightness of the shadows of
small craters on the opposite east side of the southern polar Moon on
September 27.
Again, this is an amateur note. Comments and criticisms on the same
are welcomed.
Clear Skies - Kurt
cano...@yahoo.com
Sep 15, 2009, 5:22:53 PM9/15/09
to LCROSS_Observation
In follow-up to my post yesterday, I have a correction and a few after
thoughts:
I have been reminded that the Kwiki video time stamp vendor has ceased
operations. http://www.pfdsystems.com/sundown.htm . There is a
comparable replacement product, but it is expensive. http://www.horita.com/gpsvideo.htm
Pooling resources in group observation is preferred. This
recommendation relates to leverage the unique ability to the known
time of the impact. NASA TV will be broadcasting images from the
shepherding satellite and is offered as a channel on some commercial
satellite signal vendors. NASA TV can be accessed via a portable TV
dish setup. This will allow observers to time when start that video
imaging capture and minimize the amount of wasted disk space. This
compares to SMART-1 and Deep Impact, where many minutes of video
capture needed to be done, because the moment of impact could not be
determined precisely. This consideration weighs in favor of group
observing and pooling resources to support a common dish setup.
Another timing option seen on Deep Impact is to have one local club
member monitor NASA TV on a cable channel. Just before the impact,
cells phones can be used to alert remote site observers to start their
imagers.
Because of the recommended high magnification, box mounted DOB
observers may wish to consider buddying up with a tracking mount scope
owner.
Smaller 5 inch Alt Az mount owners may wish to consider switching to
equatorial mount mode. With my Meade ETX 125, even with tracking, alt
az tracking at very high magnification was okay, but bothersome.
Switching the Meade ETX over to equatorial polar alignment mode,
allowed for much smoothier and reliable tracking. With this improved
tracking, observing at 300x through a 4mm TMB planetary eyepiece was
much more relaxed and enjoyable.
Clear Skies - Kurt
thoughts:
I have been reminded that the Kwiki video time stamp vendor has ceased
operations. http://www.pfdsystems.com/sundown.htm . There is a
comparable replacement product, but it is expensive. http://www.horita.com/gpsvideo.htm
Pooling resources in group observation is preferred. This
recommendation relates to leverage the unique ability to the known
time of the impact. NASA TV will be broadcasting images from the
shepherding satellite and is offered as a channel on some commercial
satellite signal vendors. NASA TV can be accessed via a portable TV
dish setup. This will allow observers to time when start that video
imaging capture and minimize the amount of wasted disk space. This
compares to SMART-1 and Deep Impact, where many minutes of video
capture needed to be done, because the moment of impact could not be
determined precisely. This consideration weighs in favor of group
observing and pooling resources to support a common dish setup.
Another timing option seen on Deep Impact is to have one local club
member monitor NASA TV on a cable channel. Just before the impact,
cells phones can be used to alert remote site observers to start their
imagers.
Because of the recommended high magnification, box mounted DOB
observers may wish to consider buddying up with a tracking mount scope
owner.
Smaller 5 inch Alt Az mount owners may wish to consider switching to
equatorial mount mode. With my Meade ETX 125, even with tracking, alt
az tracking at very high magnification was okay, but bothersome.
Switching the Meade ETX over to equatorial polar alignment mode,
allowed for much smoothier and reliable tracking. With this improved
tracking, observing at 300x through a 4mm TMB planetary eyepiece was
much more relaxed and enjoyable.
Clear Skies - Kurt
Jim Mosher
Sep 17, 2009, 1:39:42 PM9/17/09
to LCROSS_Observation
Kurt,
Many thanks for your detailed critique of what you say is my flawed
understanding of the terms and concepts necessary to understand plume
brightness.
Although it cannot fully compensate for the confusion I may have
created, it is gratifying to see, as best I understand your note, that
despite my "misinterpretation of controlling photometry equations" my
conclusion that Dr. Wooden's red curve:
http://01227941410742638900-a-g.googlegroups.com/web/ejecta_flux_predict_09sep02.pdf?gda=rXQxpFEAAACwPjh7SItssvxptLN2TxYMy0vBI807etRS9lz8kBQ5Jcf6BnDYuLVmMAUWOblHGowMBYBYm1nev0Mv5LzGbk8iUwk_6Qi3BU8HCN0q6OYwM5VxXgp_nHWJXhfr7YhqVgA
indicates a surface brightness of 4.0 mpsas (in the "V" band) at
points in the impact cloud where the observed density of "a=35 micron"
grains is 1x10^7 particles/m^2 agrees with your understanding of that
curve.
I also agree with you that a 1 square arc-sec piece of an extended
source with a uniform surface brightness of 5 magnitudes per square
arc-sec has stellar magnitude of 5. 5 stellar magnitudes would also
be the magnitude of a 1 square arc-min piece of an extended source
with a uniform surface brightness of 5 magnitudes per square arc-min;
or of a 1 steradian piece of a source with a uniform brightness of 5
magnitudes per steradian, etc. For this reason, I was trying to say
that a *formula* that successfully relates irradiance (total "flux")
to magnitudes will also relate radiance ("flux" per unit solid angle)
to magnitudes per unit solid angle, provided the same "per ..." unit
is used on both sides of the equation. I regret that my awkward way
of expressing that idea may have led to the impression that I believed
stellar magnitudes and mpsas were identical. On the contrary,
although my understanding of photometric concepts is weak, I believe
they are fundamentally different (but related) concepts, much like the
luminosity, or "absolute magnitude", of a star (an intrinsic property
unaffected by viewing distance) is a different (but related) concept
from its "apparent magnitude" outside the Earth's atmosphere (affected
by viewing distance), and that "apparent magnitude" is a different
(but related) concept from detectable photons/sec after passage
through the atmosphere and a light-collecting system.
I also believe that focusing so much attention on the "per ..." solid
angle unit used for expressing surface brightness obscures the fact
that surface brightness is a *differential* quantity: the unit used is
quite arbitrary and the number quoted, whatever its unit of solid
angle, is guaranteed to apply only in an infinitesimally small range
of angle around a particular viewing direction. The use of "per
square arc-sec" in expressing a surface brightness of "5 mpsas" is
such an arbitrary choice. "5 mpsas" could equally well be stated (if
I have done the math right) as "-3.9 magnitudes per square arc-min" or
"-12.8 magnitudes per square degree" or "-21.6 magnitudes per
steradian". In no case does the choice of unit imply that the stated
brightness exists over the full unit of solid angle chosen any more
than saying that a car is moving at "5 miles per hour" means it is
moving at exactly that speed for exactly that time. The car may be
moving at 5 mph for only a fraction of second, or for an arbitrarily
long time. Likewise, a surface brightness of 5 mpsas can exist over a
range of angle much less than 1 square arc-sec or over one much
greater than that. The choice of square arc-seconds is admittedly
convenient and somewhat intuitive since it is in the same ballpark as
the resolution limit of most telescopes, but there is nothing
fundamental about it, and "one square arc-second" is not meant to
represent some kind of standardized star size or any other such thing.
For those less familiar with these concepts, radiance, or surface
brightness (mpsas), tends to be useful for discussing extended objects
(ones we see spread out as two dimensional arrays with the possibility
of different brightnesses at different positions), while irradiance
(apparent stellar magnitudes) tends to be useful for discussing light
sources regarded as unresolved points to which a single number is
assigned. In astronomy, the borderline between the applicability of
these two concepts depends very much on the instrument being used for
the observation: planets that are extended objects in a telescope will
appear as point sources to the naked eye; what appears as an extended
object in a large telescope may be unresolved in a small one; and in
any telescope, what appears as an extended object in good seeing may
be unresolved in bad seeing. In general optical systems tend to
preserve radiances, but to enhance point sources by increasing the
collected energy in proportion to the area of the entrance aperture,
making statements about the relative signal levels expected from point
and extended objects very dependent on the assumed system diameter.
Both can be diminished if there are losses or if the exit pupil of the
system incompletely fills the entrance aperture of the detector (as
can happen in visual observing with excessive power).
I have refrained from commenting on your many recent and very lengthy
postings about glare, sky brightness and plume detectability because,
aside from many math and numeric errors, I find these concepts mixed
and confused in ways that I (at least) have been unable to follow.
--
In the current posting I believe you have somehow added up the mpsas
over what you believe is the expected area of the LCROSS plume and
come to the conclusion that the total light added to the lunar
landscape will be equal to that from an apparent magnitude 2.5 star.
You are, therefore, as I understand it, recommending that amateurs
photograph such a star and compare it (in some way not entirely
defined) to the total added brightness observed during the impact.
I would caution, as you do in the opening slides of your accompanying
presentation:
http://groups.google.com/group/lcross_observation/msg/9611758825327ea0
that there is great uncertainty in all such predictions. Dr. Wooden,
in her characteristically cryptic posting:
http://groups.google.com/group/lcross_observation/msg/96ef20c316a23d27
mentions that her curve is based on grain densities from a Goldstein
(2008) reference which you were kind enough to track down; repeated in
truncated form in the more accessible Summy reference:
http://groups.google.com/group/lcross_observation/msg/586e54f5b274a3e4
I am admittedly a poor reader, but I find no claims in these articles
that they show grain densities or plume sizes expected with "95%
confidence".
Not only is the model referred to by Dr. Wooden not specific to Cabeus
A1, but the authors admit they have little idea of what the actual
dust densities will be. They say they are displaying one particular
set of predictions (for an impact at a site in 2-km deep Shoemaker
"just beyond the limb") based on the seemingly arbitrary assumption
that 10^6 kg of material will be lofted in the form of "70 micron"
particles, of which 19,000 kg "rises high enough to be exposed to
sunlight." Earlier in the article they mention that the details of
the impact and what it will produce are a "difficult problem", and
rather than trying to estimate the size and volume of ejecta
themselves, they rely on the word of impact expert Don Korycansky:
http://es.ucsc.edu/personnel/Korycansky/
who, based on unspecified assumptions, privately guesstimates the
LCROSS impact might eject "O(10^6 kg)", and the lunar soil studies of
Kring, who finds typical particle sizes of 70 microns.
The notation "O(10^6 kg)" means "on the order of", which to most
scientists means it would not be surprising to find the actual answer
to be a factor of 10 higher or lower: that is 10^7 kg or 10^5 kg.
However intricately detailed Goldstein's model may be (and from the
description it sounds incredibly intricate), the predicted grain
densities would presumably have to scale in proportion to the ejected
mass, making them uncertain by at least as much as the uncertainty in
that mass. I would assume there are vast additional uncertainties in
the fraction of ejecta that will rise to any particular elevation,
which might in turn be dependent on the somewhat arbitrary assumption
about grain size.
Even assuming one knew with precision what the observed grain density
will be (which it certainly sounds like we do not), I have, for the
past many months, been trying to make the point that there may be a
systematic error in the conversion of grain densities to predictions
of reflected light: the brightnesses quoted by the LCROSS team seem
much higher than one would naively expect for the stated number and
size of reflecting spheres: on the simple, but perhaps overly
simplistic assumption that the grains act like mini-Moons (with
exactly the same reflectance as the sphere of the Moon as a whole) I
would, for example, expect something like 8.3 mpsas for the 1x10^7
particles/m^2 to which Dr. Wooden assigns a surface brightness
equivalent to 4.0 mpsas -- a factor of 50 discrepancy in brightness.
In addition, as Dr. Wooden appears to be pointing out, for a given
mass of suspended material, the reflectance depends strongly on the
assumed particle size, and presumably other properties of those
particles, none of which are well known. The predicted reflectance
will increase or decrease in roughly inverse proportion to the assumed
size.
For all these reasons, I find it hard to understand how one can say
with any certainty that the total intensity of light collected from
the LCROSS impact plume at its point of maximum visibility will equal
that collected from a magnitude 2.5 star, and recommend a detailed
observing strategy based on that proposition.
But again, I may have missed the place where the LCROSS scientists say
a specific size and brightness (or larger) plume will be achieved with
95% confidence.
--
Regarding the surface brightness of the shadowed area in Cabeus A1,
which you have asked repeatedly about, I have, of course, not studied
this subject in as great a depth as you, nor do I wish to pose as an
expert on anything, but my understanding is that there would be three
main components that would need to be added in physical brightness
units (W m^-2 sr^-1, or W m^2 arc-sec^-2 if you prefer) before
converting to mpsas. Listed in order of increasing importance, they
are:
1. Earthshine : a uniform, nearly flat Full Moon like surface
brightness added to all features. The strength of this is a well-
documented function of the Earth's phase as seen from the Moon, but
would have to corrected for the albedo of Cabeus A1's floor as seen at
Full Moon.
2. Lunar surface reflections : this comes from any sunlit features
visible from the crater floor, such as the sunlit inner south wall.
The intrinsic radiance of these features is slightly smaller than that
of the Earth, but the irradiance they contribute on the crater floor
can exceed that from Earthshine if the solid angle they present is
larger, as it is likely to be. The strength of this contribution can
be estimated; however, although it may well exceed the Earthshine it
is likely, in turn to be dominated by the following sources of surface
brightness.
3. Glare : this itself comes from three main sources –
a. Light pollution : the glow the sky would have in the absence of
the Moon. This is moderately well documented, but depends strongly on
the locality and state of the sky at the moment of impact.
b. Atmospheric scattering : the aureole seen around bright features
(such as the Moon as whole, and inidvidual parts of it) as a result of
molecular and particulate scattering in the Earth's atmosphere. In
the most perfect of skies the veiling glare over the Moon will be
several millionths the surface brightness of the Moon. In average
skies it will be much higher.
c. Instrumental scattering : an additional aureole produced by
scattering and reflections within the observer's telescope and
detector. I would guess the veiling glare from instrumental scatter
can easily be a few tenths of a percent of the surface brightness of
the Moon.
I believe all of these, with the possible exception of lunar surface
reflections, are dealt with in Dr. Schaefer's extended article on the
visibility of lunar occultations:
http://adsabs.harvard.edu/abs/1992Icar..100...60S
Despite his claims that his model has been thoroughly tested and is
nearly perfect, I would not at all vouch for the accuracy or
applicability of his results, particularly as it seems to be geared
towards visual observations on perfect nights at exceptional sites.
For most amateurs I would suspect that instrumental glare will be by
far the dominant source of light in the shadowed bowl of Cabeus A1 on
impact night, followed by scattering from a less-than-perfect sky.
Since instrumental glare is highly dependent on the individual
instrument it seems impossible to make any sweeping statement about
it.
--
Regarding the brightness needed for the plume to be visible against
the shadowed bowl, again I think no hard-and-fast answer is possible.
The plume will add to the surface brightness, just like the six other
sources listed above. Whether such an addition can be detected, or
not, depends on the instrumental noise.
--
Regarding the visibility of the upper parts of the plume against the
sunlit south wall and terrain beyond, it is not entirely obvious to me
if it will appear bright or dark. As mentioned long ago, the closest
analogy I can think of is a dust-devil in an arid land (or on Mars).
Although I have not checked how the particle densities compare to
those expected in the LCROSS impact, the suspension of particles will
both remove and add light to that which we would see in its absence.
Those parts that add more light than they remove will lower the
surface brightness (and appear darker than pre-impact), while those
that add more light than they remove will appear bright. Again, I am
no expert on this, but the addition of light comes from the reflection
of the sunlight striking the particles. If they act like diffuse
reflectors ("Lambertian spheres") they could be thought of as mini-
Moons, lit at the same phase angle, but with somewhat higher
reflectance. The removal of light comes from interposing the
particles in front the sunlit lunar landscape. They will both block
and bend that light. A rather strange, but well documented, physical
principle states that the total amount of light blocked and bent by a
sphere is twice what one would expect from its geometric cross-section
(equal amounts being blocked and bent). However, of the half of the
light that is bent, most of it is bent at very small angles, so in
many practical situations (probably including the present one) the
observer misses only the light physically blocked by the particle.
My guess would be that in the present case the ejected particles will
have higher albedo than the space-weathered background against which
they are seen, and hence will add more reflected sunlight than they
block. I would, therefore, expect the plume to look bright, but like
trying to guess whether a dust devil will appear light or dark
relative to some unknown background, that conclusion, like all others
of mine, is uncertain.
Assuming the plume brightness inferred from presentations like Dr.
Wooden's can simply be added to the pre-impact bright background is
unlikely to give a correct result.
--
I hope the above comments can be regarded as an attempt at making a
constructive contribution, but I cannot guarantee the accuracy of any
of it. To the extent these conclusions are based on my frequently
incorrect understanding of fundamental physical principles, and
untempered by the antidote of practical experience, I apologize in
advance for the confusion my remarks will undoubtedly create.
-- Jim
P.S.: I'm not sure I understand the advice to commence observations
only when the observability of a sunlit plume is apparent on NASA TV.
Particularly in view of the possibility that the unexpected may
happen, I would think most amateurs would be inclined to commence
observations at the stated instant of impact, if not slightly before,
which should require nothing more than an ordinary watch (or computer
clock) synchronized to WWV or an internet time signal.
I know very little about broadcast technology, but it would seem the
LCROSS impact images appearing on NASA TV must have to be late by at
least 1 second due to the light travel time from the Moon, and the
actual delay may be greater. I know nothing at all about how the NASA
network functions, but I do know that the all-news (CBS) radio station
in Los Angeles broadcasts a "the time is" tone at the start of each
hour which is heard over the air something like 9-12 seconds *after*
the actual start of the hour. Hopefully NASA TV does not introduce
such additional delays, but if they do, it might be a very unreliable
source of timing information.
On Sep 15, 1:30 am, "canopu...@yahoo.com" <canopu...@yahoo.com> wrote:
> The following is an amateur note. Comments and criticisms on the same
> are welcomed.
>
> At url -
>
> http://members.csolutions.net/fisherka/astronote/observed/LCROSS/2009...
> NASA LCROSS Citizen Science Page Bloghttp://apps.nasa.gov/lcross/
> Efls for imaging the LCROSS impacthttp://groups.google.com/group/lcross_observation/msg/764ceeede969207ahttp://tinyurl.com/owkdf7
> timestamper. url:http://www.pfdsystems.com/. The NIST WWW shortwave
>
> read more »
Many thanks for your detailed critique of what you say is my flawed
understanding of the terms and concepts necessary to understand plume
brightness.
Although it cannot fully compensate for the confusion I may have
created, it is gratifying to see, as best I understand your note, that
despite my "misinterpretation of controlling photometry equations" my
conclusion that Dr. Wooden's red curve:
http://01227941410742638900-a-g.googlegroups.com/web/ejecta_flux_predict_09sep02.pdf?gda=rXQxpFEAAACwPjh7SItssvxptLN2TxYMy0vBI807etRS9lz8kBQ5Jcf6BnDYuLVmMAUWOblHGowMBYBYm1nev0Mv5LzGbk8iUwk_6Qi3BU8HCN0q6OYwM5VxXgp_nHWJXhfr7YhqVgA
indicates a surface brightness of 4.0 mpsas (in the "V" band) at
points in the impact cloud where the observed density of "a=35 micron"
grains is 1x10^7 particles/m^2 agrees with your understanding of that
curve.
I also agree with you that a 1 square arc-sec piece of an extended
source with a uniform surface brightness of 5 magnitudes per square
arc-sec has stellar magnitude of 5. 5 stellar magnitudes would also
be the magnitude of a 1 square arc-min piece of an extended source
with a uniform surface brightness of 5 magnitudes per square arc-min;
or of a 1 steradian piece of a source with a uniform brightness of 5
magnitudes per steradian, etc. For this reason, I was trying to say
that a *formula* that successfully relates irradiance (total "flux")
to magnitudes will also relate radiance ("flux" per unit solid angle)
to magnitudes per unit solid angle, provided the same "per ..." unit
is used on both sides of the equation. I regret that my awkward way
of expressing that idea may have led to the impression that I believed
stellar magnitudes and mpsas were identical. On the contrary,
although my understanding of photometric concepts is weak, I believe
they are fundamentally different (but related) concepts, much like the
luminosity, or "absolute magnitude", of a star (an intrinsic property
unaffected by viewing distance) is a different (but related) concept
from its "apparent magnitude" outside the Earth's atmosphere (affected
by viewing distance), and that "apparent magnitude" is a different
(but related) concept from detectable photons/sec after passage
through the atmosphere and a light-collecting system.
I also believe that focusing so much attention on the "per ..." solid
angle unit used for expressing surface brightness obscures the fact
that surface brightness is a *differential* quantity: the unit used is
quite arbitrary and the number quoted, whatever its unit of solid
angle, is guaranteed to apply only in an infinitesimally small range
of angle around a particular viewing direction. The use of "per
square arc-sec" in expressing a surface brightness of "5 mpsas" is
such an arbitrary choice. "5 mpsas" could equally well be stated (if
I have done the math right) as "-3.9 magnitudes per square arc-min" or
"-12.8 magnitudes per square degree" or "-21.6 magnitudes per
steradian". In no case does the choice of unit imply that the stated
brightness exists over the full unit of solid angle chosen any more
than saying that a car is moving at "5 miles per hour" means it is
moving at exactly that speed for exactly that time. The car may be
moving at 5 mph for only a fraction of second, or for an arbitrarily
long time. Likewise, a surface brightness of 5 mpsas can exist over a
range of angle much less than 1 square arc-sec or over one much
greater than that. The choice of square arc-seconds is admittedly
convenient and somewhat intuitive since it is in the same ballpark as
the resolution limit of most telescopes, but there is nothing
fundamental about it, and "one square arc-second" is not meant to
represent some kind of standardized star size or any other such thing.
For those less familiar with these concepts, radiance, or surface
brightness (mpsas), tends to be useful for discussing extended objects
(ones we see spread out as two dimensional arrays with the possibility
of different brightnesses at different positions), while irradiance
(apparent stellar magnitudes) tends to be useful for discussing light
sources regarded as unresolved points to which a single number is
assigned. In astronomy, the borderline between the applicability of
these two concepts depends very much on the instrument being used for
the observation: planets that are extended objects in a telescope will
appear as point sources to the naked eye; what appears as an extended
object in a large telescope may be unresolved in a small one; and in
any telescope, what appears as an extended object in good seeing may
be unresolved in bad seeing. In general optical systems tend to
preserve radiances, but to enhance point sources by increasing the
collected energy in proportion to the area of the entrance aperture,
making statements about the relative signal levels expected from point
and extended objects very dependent on the assumed system diameter.
Both can be diminished if there are losses or if the exit pupil of the
system incompletely fills the entrance aperture of the detector (as
can happen in visual observing with excessive power).
I have refrained from commenting on your many recent and very lengthy
postings about glare, sky brightness and plume detectability because,
aside from many math and numeric errors, I find these concepts mixed
and confused in ways that I (at least) have been unable to follow.
--
In the current posting I believe you have somehow added up the mpsas
over what you believe is the expected area of the LCROSS plume and
come to the conclusion that the total light added to the lunar
landscape will be equal to that from an apparent magnitude 2.5 star.
You are, therefore, as I understand it, recommending that amateurs
photograph such a star and compare it (in some way not entirely
defined) to the total added brightness observed during the impact.
I would caution, as you do in the opening slides of your accompanying
presentation:
http://groups.google.com/group/lcross_observation/msg/9611758825327ea0
that there is great uncertainty in all such predictions. Dr. Wooden,
in her characteristically cryptic posting:
http://groups.google.com/group/lcross_observation/msg/96ef20c316a23d27
mentions that her curve is based on grain densities from a Goldstein
(2008) reference which you were kind enough to track down; repeated in
truncated form in the more accessible Summy reference:
http://groups.google.com/group/lcross_observation/msg/586e54f5b274a3e4
I am admittedly a poor reader, but I find no claims in these articles
that they show grain densities or plume sizes expected with "95%
confidence".
Not only is the model referred to by Dr. Wooden not specific to Cabeus
A1, but the authors admit they have little idea of what the actual
dust densities will be. They say they are displaying one particular
set of predictions (for an impact at a site in 2-km deep Shoemaker
"just beyond the limb") based on the seemingly arbitrary assumption
that 10^6 kg of material will be lofted in the form of "70 micron"
particles, of which 19,000 kg "rises high enough to be exposed to
sunlight." Earlier in the article they mention that the details of
the impact and what it will produce are a "difficult problem", and
rather than trying to estimate the size and volume of ejecta
themselves, they rely on the word of impact expert Don Korycansky:
http://es.ucsc.edu/personnel/Korycansky/
who, based on unspecified assumptions, privately guesstimates the
LCROSS impact might eject "O(10^6 kg)", and the lunar soil studies of
Kring, who finds typical particle sizes of 70 microns.
The notation "O(10^6 kg)" means "on the order of", which to most
scientists means it would not be surprising to find the actual answer
to be a factor of 10 higher or lower: that is 10^7 kg or 10^5 kg.
However intricately detailed Goldstein's model may be (and from the
description it sounds incredibly intricate), the predicted grain
densities would presumably have to scale in proportion to the ejected
mass, making them uncertain by at least as much as the uncertainty in
that mass. I would assume there are vast additional uncertainties in
the fraction of ejecta that will rise to any particular elevation,
which might in turn be dependent on the somewhat arbitrary assumption
about grain size.
Even assuming one knew with precision what the observed grain density
will be (which it certainly sounds like we do not), I have, for the
past many months, been trying to make the point that there may be a
systematic error in the conversion of grain densities to predictions
of reflected light: the brightnesses quoted by the LCROSS team seem
much higher than one would naively expect for the stated number and
size of reflecting spheres: on the simple, but perhaps overly
simplistic assumption that the grains act like mini-Moons (with
exactly the same reflectance as the sphere of the Moon as a whole) I
would, for example, expect something like 8.3 mpsas for the 1x10^7
particles/m^2 to which Dr. Wooden assigns a surface brightness
equivalent to 4.0 mpsas -- a factor of 50 discrepancy in brightness.
In addition, as Dr. Wooden appears to be pointing out, for a given
mass of suspended material, the reflectance depends strongly on the
assumed particle size, and presumably other properties of those
particles, none of which are well known. The predicted reflectance
will increase or decrease in roughly inverse proportion to the assumed
size.
For all these reasons, I find it hard to understand how one can say
with any certainty that the total intensity of light collected from
the LCROSS impact plume at its point of maximum visibility will equal
that collected from a magnitude 2.5 star, and recommend a detailed
observing strategy based on that proposition.
But again, I may have missed the place where the LCROSS scientists say
a specific size and brightness (or larger) plume will be achieved with
95% confidence.
--
Regarding the surface brightness of the shadowed area in Cabeus A1,
which you have asked repeatedly about, I have, of course, not studied
this subject in as great a depth as you, nor do I wish to pose as an
expert on anything, but my understanding is that there would be three
main components that would need to be added in physical brightness
units (W m^-2 sr^-1, or W m^2 arc-sec^-2 if you prefer) before
converting to mpsas. Listed in order of increasing importance, they
are:
1. Earthshine : a uniform, nearly flat Full Moon like surface
brightness added to all features. The strength of this is a well-
documented function of the Earth's phase as seen from the Moon, but
would have to corrected for the albedo of Cabeus A1's floor as seen at
Full Moon.
2. Lunar surface reflections : this comes from any sunlit features
visible from the crater floor, such as the sunlit inner south wall.
The intrinsic radiance of these features is slightly smaller than that
of the Earth, but the irradiance they contribute on the crater floor
can exceed that from Earthshine if the solid angle they present is
larger, as it is likely to be. The strength of this contribution can
be estimated; however, although it may well exceed the Earthshine it
is likely, in turn to be dominated by the following sources of surface
brightness.
3. Glare : this itself comes from three main sources –
a. Light pollution : the glow the sky would have in the absence of
the Moon. This is moderately well documented, but depends strongly on
the locality and state of the sky at the moment of impact.
b. Atmospheric scattering : the aureole seen around bright features
(such as the Moon as whole, and inidvidual parts of it) as a result of
molecular and particulate scattering in the Earth's atmosphere. In
the most perfect of skies the veiling glare over the Moon will be
several millionths the surface brightness of the Moon. In average
skies it will be much higher.
c. Instrumental scattering : an additional aureole produced by
scattering and reflections within the observer's telescope and
detector. I would guess the veiling glare from instrumental scatter
can easily be a few tenths of a percent of the surface brightness of
the Moon.
I believe all of these, with the possible exception of lunar surface
reflections, are dealt with in Dr. Schaefer's extended article on the
visibility of lunar occultations:
http://adsabs.harvard.edu/abs/1992Icar..100...60S
Despite his claims that his model has been thoroughly tested and is
nearly perfect, I would not at all vouch for the accuracy or
applicability of his results, particularly as it seems to be geared
towards visual observations on perfect nights at exceptional sites.
For most amateurs I would suspect that instrumental glare will be by
far the dominant source of light in the shadowed bowl of Cabeus A1 on
impact night, followed by scattering from a less-than-perfect sky.
Since instrumental glare is highly dependent on the individual
instrument it seems impossible to make any sweeping statement about
it.
--
Regarding the brightness needed for the plume to be visible against
the shadowed bowl, again I think no hard-and-fast answer is possible.
The plume will add to the surface brightness, just like the six other
sources listed above. Whether such an addition can be detected, or
not, depends on the instrumental noise.
--
Regarding the visibility of the upper parts of the plume against the
sunlit south wall and terrain beyond, it is not entirely obvious to me
if it will appear bright or dark. As mentioned long ago, the closest
analogy I can think of is a dust-devil in an arid land (or on Mars).
Although I have not checked how the particle densities compare to
those expected in the LCROSS impact, the suspension of particles will
both remove and add light to that which we would see in its absence.
Those parts that add more light than they remove will lower the
surface brightness (and appear darker than pre-impact), while those
that add more light than they remove will appear bright. Again, I am
no expert on this, but the addition of light comes from the reflection
of the sunlight striking the particles. If they act like diffuse
reflectors ("Lambertian spheres") they could be thought of as mini-
Moons, lit at the same phase angle, but with somewhat higher
reflectance. The removal of light comes from interposing the
particles in front the sunlit lunar landscape. They will both block
and bend that light. A rather strange, but well documented, physical
principle states that the total amount of light blocked and bent by a
sphere is twice what one would expect from its geometric cross-section
(equal amounts being blocked and bent). However, of the half of the
light that is bent, most of it is bent at very small angles, so in
many practical situations (probably including the present one) the
observer misses only the light physically blocked by the particle.
My guess would be that in the present case the ejected particles will
have higher albedo than the space-weathered background against which
they are seen, and hence will add more reflected sunlight than they
block. I would, therefore, expect the plume to look bright, but like
trying to guess whether a dust devil will appear light or dark
relative to some unknown background, that conclusion, like all others
of mine, is uncertain.
Assuming the plume brightness inferred from presentations like Dr.
Wooden's can simply be added to the pre-impact bright background is
unlikely to give a correct result.
--
I hope the above comments can be regarded as an attempt at making a
constructive contribution, but I cannot guarantee the accuracy of any
of it. To the extent these conclusions are based on my frequently
incorrect understanding of fundamental physical principles, and
untempered by the antidote of practical experience, I apologize in
advance for the confusion my remarks will undoubtedly create.
-- Jim
P.S.: I'm not sure I understand the advice to commence observations
only when the observability of a sunlit plume is apparent on NASA TV.
Particularly in view of the possibility that the unexpected may
happen, I would think most amateurs would be inclined to commence
observations at the stated instant of impact, if not slightly before,
which should require nothing more than an ordinary watch (or computer
clock) synchronized to WWV or an internet time signal.
I know very little about broadcast technology, but it would seem the
LCROSS impact images appearing on NASA TV must have to be late by at
least 1 second due to the light travel time from the Moon, and the
actual delay may be greater. I know nothing at all about how the NASA
network functions, but I do know that the all-news (CBS) radio station
in Los Angeles broadcasts a "the time is" tone at the start of each
hour which is heard over the air something like 9-12 seconds *after*
the actual start of the hour. Hopefully NASA TV does not introduce
such additional delays, but if they do, it might be a very unreliable
source of timing information.
On Sep 15, 1:30 am, "canopu...@yahoo.com" <canopu...@yahoo.com> wrote:
> The following is an amateur note. Comments and criticisms on the same
> are welcomed.
>
> At url -
>
> time broadcast is another option for a timing audio signal, but since
> there is no need to coordinate ...
>
> read more »
cano...@yahoo.com
Sep 17, 2009, 10:58:32 PM9/17/09
to LCROSS_Observation
Jim,
> I also agree with you that a 1 square arc-sec piece of an
> extended source with a uniform surface brightness of 5
> magnitudes per square arc-sec has stellar magnitude of 5.
I did not say that, I said the opposite of that. I said that you
fundamentally misinterpert the physics of the situation by equating
stellar magnitudes to mpsas based on a misapplication of the concepts
of irradiance and radiance.
Thanks for your comments and observations. At this point, we will
have to agree to disagree, since we both feel the other fundamentally
misunderstands the physics of the situation.
I recommend that you ask yourself why the photograph I took of the
lunar surface came out properly exposed when calibrated to a steller
magnitude 2.8 star. This is a physical test and illustrates that your
notion that a 5.0 stellar magnitude is equivalent to 5.0 mspas is
simply wrong. Had I exposed calibrated to a 5 V mag star, the image
would have been horribly underexposed.
Your notion that mpsas = V mag is inconsistent with physical
observation that any amateur can verify with their own LPI or CCD
cameras and a few minutes of imaging.
> Assuming the plume brightness inferred from presentations like Dr.
> Wooden's can simply be added to the pre-impact bright background is
> unlikely to give a correct result.
I did not state that anywhere. My discussion went to whether there
would be a sufficient contrast between the shadowed portion of the
crater and the foreground plume. Nowhere did I state or imply that
plume brightness would be summed to the mpsas of the shadowed crater.
In this context, plume brightness inherently means brighter and
references the concept of "contrast index" - a term well-known and
understood by amateur astrophotography imagers.
> Not only is the model referred to by Dr. Wooden not specific to Cabeus
> A1, but the authors admit they have little idea of what the actual
> dust densities will be.
Clarifying the degree of uncertainty involved was the motivation for
my additional question to Dr. Wooden about the team's "estimates for
the probabilty of the parameters."
As a consumer of expert level information issued by the NASA LCROSS
Team, I feel like I should be able to rely on their warrant that the
plume will be visible in a 10 inch telescope. We have 50+ people
showing up at our club on Oct. 9 and 2 TV stations. The equivalent
social economic value for that activity is not much - maybe $2,000,
but it is not zero. Most amateur astronomer club members happily
accept the notion of experiemental uncertainty - assuming the degree
of the risk is laid out in advance - and assume the risk there will be
nothing to see.
As an amateur, all I can do is accept the LCROSS Team expert opinion
regarding a 4.0-6.0 peak mpsas plume and plan to image based on that
recommendation - after reasonably exploring the basis for their
opinion on my own. I feel that, as an amateur, I have done my "due
diligence" on the background matter that underlies the LCROSS Team
claims. The remaining area of uncertainty is that no one has really
done any site specific photometry defining mpsas values for the
surface around Cabeus A1 or the dark shadowed portion of the crater at
71% illumination. If they are brighter than 4.0 mpsas, there will be
an insufficient _contrast index_ between the background dark shadowed
portion of the crater and the plume. Hence, nothing to see.
My conclusion is that there is an underlying level of experimental
uncertainty to the brightness and size of the ejecta plume, but that
the level of experimental uncertainty is ordinary and not in the
category of "very unlikely" to be visible. I would say among the
amateurs that I talked with, they most often assign LCROSS impact
plume visibility to the categories "extremely unlikely" or "ain't
happening."
At this point, I feel reasonably justified in hauling my butt out of
bed at 3:00am, dragging 100lbs of telescope gear over 30 miles of
highway to try to take a picture of the plume and suggesting to my
fellow local club members that they do the same.
Clear Skies and Best wishes. I have enjoyed your company and our
dialogue over the last months.
- Kurt
> I also agree with you that a 1 square arc-sec piece of an
> extended source with a uniform surface brightness of 5
> magnitudes per square arc-sec has stellar magnitude of 5.
fundamentally misinterpert the physics of the situation by equating
stellar magnitudes to mpsas based on a misapplication of the concepts
of irradiance and radiance.
Thanks for your comments and observations. At this point, we will
have to agree to disagree, since we both feel the other fundamentally
misunderstands the physics of the situation.
I recommend that you ask yourself why the photograph I took of the
lunar surface came out properly exposed when calibrated to a steller
magnitude 2.8 star. This is a physical test and illustrates that your
notion that a 5.0 stellar magnitude is equivalent to 5.0 mspas is
simply wrong. Had I exposed calibrated to a 5 V mag star, the image
would have been horribly underexposed.
Your notion that mpsas = V mag is inconsistent with physical
observation that any amateur can verify with their own LPI or CCD
cameras and a few minutes of imaging.
> Assuming the plume brightness inferred from presentations like Dr.
> Wooden's can simply be added to the pre-impact bright background is
> unlikely to give a correct result.
would be a sufficient contrast between the shadowed portion of the
crater and the foreground plume. Nowhere did I state or imply that
plume brightness would be summed to the mpsas of the shadowed crater.
In this context, plume brightness inherently means brighter and
references the concept of "contrast index" - a term well-known and
understood by amateur astrophotography imagers.
> Not only is the model referred to by Dr. Wooden not specific to Cabeus
> A1, but the authors admit they have little idea of what the actual
> dust densities will be.
my additional question to Dr. Wooden about the team's "estimates for
the probabilty of the parameters."
As a consumer of expert level information issued by the NASA LCROSS
Team, I feel like I should be able to rely on their warrant that the
plume will be visible in a 10 inch telescope. We have 50+ people
showing up at our club on Oct. 9 and 2 TV stations. The equivalent
social economic value for that activity is not much - maybe $2,000,
but it is not zero. Most amateur astronomer club members happily
accept the notion of experiemental uncertainty - assuming the degree
of the risk is laid out in advance - and assume the risk there will be
nothing to see.
As an amateur, all I can do is accept the LCROSS Team expert opinion
regarding a 4.0-6.0 peak mpsas plume and plan to image based on that
recommendation - after reasonably exploring the basis for their
opinion on my own. I feel that, as an amateur, I have done my "due
diligence" on the background matter that underlies the LCROSS Team
claims. The remaining area of uncertainty is that no one has really
done any site specific photometry defining mpsas values for the
surface around Cabeus A1 or the dark shadowed portion of the crater at
71% illumination. If they are brighter than 4.0 mpsas, there will be
an insufficient _contrast index_ between the background dark shadowed
portion of the crater and the plume. Hence, nothing to see.
My conclusion is that there is an underlying level of experimental
uncertainty to the brightness and size of the ejecta plume, but that
the level of experimental uncertainty is ordinary and not in the
category of "very unlikely" to be visible. I would say among the
amateurs that I talked with, they most often assign LCROSS impact
plume visibility to the categories "extremely unlikely" or "ain't
happening."
At this point, I feel reasonably justified in hauling my butt out of
bed at 3:00am, dragging 100lbs of telescope gear over 30 miles of
highway to try to take a picture of the plume and suggesting to my
fellow local club members that they do the same.
Clear Skies and Best wishes. I have enjoyed your company and our
dialogue over the last months.
- Kurt
Jim Mosher
Sep 19, 2009, 4:52:39 PM9/19/09
to LCROSS_Observation
Kurt,
Thank you, again, for your patience with my misunderstanding of your
postings regarding photometry. This is probably a problem unique to
me, for I see the resulting imaging recommendations have been widely
and warmly embraced on at least some of the other forums where you
have posted them:
http://www.cloudynights.com/ubbthreads/showthreaded.php/Number/3339253/
I hope, too, that my remarks about possible inconsistencies in the
information offered by the LCROSS science team about the size and
brightness of the ejecta plume, and ways that information has been
interpreted, will not discourage anyone from looking. That has never
been their intention, and I, personally, have no idea if the LCROSS
impact event will be easy or difficult to observe from Earth.
And finally I hope that an extended discussion of the following
question will not distract too much from efforts to prepare for
observing and imaging the event.
> I recommend that you ask yourself why the photograph
> I took of the lunar surface came out properly exposed
> when calibrated to a steller magnitude 2.8 star. ...
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/2009_8_14_0246UT_KafTestPanel.jpg
mentioned in:
http://groups.google.com/group/lcross_observation/msg/d9b0b36ce9f2c59e
showing what looks (on my CRT monitor) like a rather dim image of the
Moon next to a what looks like a rather overexposed image of "2.87v
Alcyone".
I am not disputing that you were happy with the result you achieved,
but I think the advice to adjust exposures for the Moon (an extended
object) based on the appearance of star images (point sources) lacks
the generality needed to make it helpful to others.
If I am not mistaken, Tom Bash was trying to point out this
fundamental problem in response to a copy of one of your postings
submitted to another forum:
http://tech.groups.yahoo.com/group/lunar-observing/message/25356
As Tom says, the relative intensities observed for stars (point
sources) and planets (extended sources) differ with aperture (and
seeing). What gives a good result with one aperture and seeing
condition may not give a good result with a different apertures and/or
seeing.
This is not to say star images are not helpful for calibration; they
just don't seem to be so for setting exposures unless one has reason
to think the impact plume will have a star-like appearance with known
total magnitude.
--
The concepts and math related to the imaging of extended versus point
sources by a simple telescope/detector combination are quite
straightforward (see below), and were partially presented by Clif
Ashcraft in some of the earliest threads on this form:
http://groups.google.com/group/lcross_observation/msg/fc2a2eb642d740f6
They are much simpler than the ideas needed to grasp the limits of
visual observing, which are the subject of Dr. Schaefer's many
articles.
It would be easier to comment on your specific statement about the
appearance of a "magnitude 2.8 star" relative to the 71% illuminated
Moon if I was sure I understood what you mean by "calibrate to".
In an early version of your slide presentation and imaging protocol:
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/20090916LCROSSImpactUpdate.pdf
mentioned in:
http://groups.google.com/group/lcross_observation/msg/9611758825327ea0
I had the impression that you were recommending that amateurs set
their exposures based on the bright rim of Cabeus A1, but supplement
this with "calibration images" of stars of various magnitudes,
possibly at the same exposure, or possibly with the exposure time
modified to avoid saturation or blackness. I assumed, apparently
incorrectly, that your intention was to add up the total detector
counts resulting from a star of known magnitude, apparently on the
assumption that the light output from these stars is better known than
that from the Sun or Moon.
A more recent version (9-18-2007rev, p. 46) seems to recommend setting
the camera's gain and exposure time so that the peak detector count
(at the center pixel of the star image?) comes out at some comfortable
level, then imaging the Moon with same setting, as an alternative
procedure.
In the present posting you seem also to be saying that when the center
pixel of the image of a magnitude 2.8 (or 2.87?) star is in a
comfortable range, the sunlit parts of the Moon will be in a similarly
comfortable range. This interpretation seems to be supported by page
48 in your imaging protocol, where you appear to be assigning a pixel
value to each star that possibly indicates the count in the central
pixel of the image (although it could be the sum of counts over all
the pixels of the stellar image?). It is contradicted by your present
statement that if you "calibrated to a 5 V mag star" (as opposed to a
magnitude 2.8 one) the lunar image would be *underexposed*: I would
have thought a fainter test star would dictate higher gain and
exposure time settings, leading to an *overexposed* lunar image.
Tom Bash, too, seems to have been confused as whether you are talking
about peak ("ADU") or total ("stellar flux") detector counts
http://tech.groups.yahoo.com/group/lunar-observing/message/25387
So I am not at all sure I understand what "calibrating to" a star
means, but I will proceed as best I can. The following commentary
consists of my thoughts on this subject, and may contain errors. My
reading is not as wide as it should be, but I assume similar thoughts
can be found expressed with greater clarity and accuracy in standard
texts on astronomical imaging.
--
To the best of my knowledge, the only concepts needed to understand
the relative counts per pixel expected in the image plane of a
telescope observing point and extended sources are:
1. In any given spectral band, a point source bathes the front of the
telescope in certain number of Watts m^-2, loosely indicated by its
"stellar magnitude".
2. The total energy collected is that flux per unit area times the
clear area of the entrance aperture.
3. After correction for losses in transmission and reflection, that
total energy is distributed over an image area that in a ideal system
is the diffraction pattern of the entrance aperture, but in real life
is likely to be fuzzed out by the imperfect focusing of the optics,
and, usually most importantly, by seeing. This is in turn converted
to "counts" according to the area and quantum efficiency of the
detector element at the wavelength of interest.
4. For an extended source in the same spectral band, each unit solid
angle of the source bathes the front of the telescope in certain
number of Watts m^-2, a quantity that can be indicated by a number of
magnitudes per unit solid angle.
5. The total energy collected *per unit angle* of source is this flux
per unit area and per unit solid area times the clear area of the
entrance aperture.
6. This total energy collected by the entrance aperture, after
correction for the same losses, is distributed over an area in the
image plane equal to the unit of solid angle operating over the
effective focal length of the system. The portion of this flux falling
on a single detector element is converted to counts in the same way as
for a point source.
Any desired relationship regarding how the relative appearance of the
two kinds of sources will vary with aperture and focal length can be
derived from these simple and intuitive principles, although trying to
translate the answers to the sign-reversed logarithmic system of
stellar magnitudes can introduce additional levels of confusion (I,
for example, am continually tempted to sum magnitudes, or to multiply
mpsas by areas in square-arc seconds to obtain a total magnitude, but
these are not valid operations).
Looking at the six intuitive principles list above one can readily see
that the intensity per pixel in the image of a point source is going
to be affected by variables of resolution and seeing that do not
affect an extended source. For this reason, point sources cannot be
used as reliable guides to the exposure of extended objects unless one
employs some complicated scheme involving summing the total counts
received from the point source and correcting that to the expected
area of the extended source.
In poor seeing, the count at the center pixel of a stellar image is
entirely at the mercy of the current smearing. In good seeing, larger
apertures and more optically perfect telescopes stuff the captured
light into a small angular size than do smaller and less perfect ones
(the angular size of the diffraction disk *decreases* with increasing
aperture).
--
For the sake of discussion, I will take your statement to mean that
with your observing set-up you found the central pixel of the image of
a magnitude 2.8 star to exhibit the same count as an average over the
sunlit portion of the Moon, which (again purely for the sake of
discussion) we can assign a surface brightness of 5 mpsas. Under what
circumstances would this be expected?
Again, not to be overly mathematical, but in the absence of any other
information one might reasonably guess that the light from the star
would be spread over a vaguely bell-shaped or "Gaussian"
distribution. Mathematically, this can be written:
I(r) = Io * exp(-(r/a)^2)
where Io is the count (per pixel) at the center of the image, I(r) is
the count (per pixel) at distance r from the center, "a" is a scaling
factor, and "exp()" means raising "e = 2.718..." to the power
indicated in parenthesis (labeled "e^x" on many scientific
calculators). The intensity at r = a is Io/e = 0.368*Io, and the
full width at half maximum (FWHM) is 1.665*a . A table of integrals
indicates that the total counts in such a pattern is the same as if
the central intensity (Io) existed uniformly over a disk of area
Pi*a^2 (where Pi = 3.142...).
Without trying to be intentionally obscure, if I quite arbitrarily
assume a = Sqrt(1/Pi) arc-sec = 0.564 arc-sec (or 0.939 arc-sec FWHM)
the central intensity of the magnitude 2.8 Gaussian stellar disk would
be expected to be the equivalent of spreading the total counts
uniformly over a disk with an area of exactly 1 square arc-sec, which
is to say 2.8 mpsas. You are telling me (as I interpret your
statement) that the actual observed intensity at the center of the
stellar disk is dimmer than this: in fact, the same as the intensity
observed for a 5 mpsas extended source. Since by definition 5
astronomical magnitudes is a factor of 100 in brightness, the
difference between 2.9 mpsas and 5 mpsas is a factor of 100^((5-2.8)/
5) = 7.59x, meaning the actual stellar disk must have been 7.59x
larger in area than I have arbitrarily assumed. Therefore, based on
your observation, I conclude the actual stellar disk, if Gaussian, had
a = 0.564*Sqrt(7.59) = 1.55 arc-sec, or FWHM = 2.59 arc-sec. This
FWHM seems roughly consistent with the image of Alcyone that you
present, although the pixel scale is probably non-linear so I can't be
sure where the 50% intensity point is, or how the count pattern would
look if averaged over azimuth.
The large observed stellar disk size (assuming I am interpreting your
statement correctly) could be the result of seeing and/or focus
errors. Potentially, an unobstructed telescope with an entrance
aperture of 125 mm diameter and operating at a wavelength of 0.5
micron, can focus 84% of the light it collects from a point source
into a pseudo-Gaussian Airy central disk with a FWHM of about 0.85 arc-
sec (diameter of first dark ring = 1.01 arc-sec). By this very rough
reasoning, by concentrating the same collected light into this smaller
space, the present telescope (if that were its description), under
perfect conditions, could produce a star image with a count at the
central pixel 0.84*(2.59/0.85)^2 = 7.8x higher than in the present
image. Had the camera gain been reduced to produce a comfortable level
at the center of the star image under those ideal conditions, the
lunar image (whose smeared out brightness is not affected by these
considerations of resolution) would presumably have been underexposed
by that factor. Conversely, if the seeing had deteriorated to where
stellar images had an observed FWHM of 5.2 arc-sec (twice the 2.6 arc-
sec), the central count would decrease in inverse proportion to the
increase in area (that is, by a factor of 4): the gain would have to
increased to recover the intensity in the stellar image, giving an
overexposed Moon.
Again, I may still not be correctly understanding what you are trying
to say, but the recommendation to set exposures based on the peak
count in the image of a magnitude 2.8 star seems equally problematic
if applied to telescopes of different apertures. It *would*, however,
work if all observers were experiencing uniformly bad seeing
conditions. If the seeing disk could be relied upon to always give
stellar disks of 2.6 arc-sec FWHM in telescopes of all sizes, then the
stars could be thought of as extended sources of that size, and such
star images could be used for setting lunar exposures just as setting
the exposure to capture Mars at a comfortable level will also capture
a 71% illuminated Moon at a comfortable level.
Your recommendation to use stars to set lunar exposures might also
work if you are talking about the total count from the star in the
image plane, although this would require fore-knowledge of the lunar
brightness and some rather elaborate considerations about image scale,
and I have trouble seeing the advantage of doing this over simply
setting the gain to give a nice lunar image. As an example, on page
47 of your imaging protocol you mention an image scale of 0.32 arc-sec/
pixel. In that particular case a single pixel would subtend an area
of 0.32^2 = 0.10 square arc-sec. If we hypothesize the Moon we want
to expose has a surface brightness of 5 mpsas, since a 0.10 square arc-
sec area will capture 1/10th the amount of light captured by a 1
square arc-sec area, it will see the equivalent of the light from a 5
- 2.5*log(0.1) = magnitude 7.5 star (if I have done the math right).
In principle, then, one could set the exposure so that the total
counts from (all the pixels recording) a magnitude 7.5 star equaled
the count desired for 1 pixel in the lunar image. This would, of
course, require dark level subtraction and a detector with a linear
response of counts to light level in W m^-2.
Your recommendation to set exposures based on stars rather than the
lunar disk would also seem appropriate if there is reason to believe
the most interesting part of the LCROSS impact will be seen as an
unresolved star-like point of known total magnitude overlying the
shadowed interior of the target crater. If for example, it is
confidently known that such a dot will have a total visual magnitude
of 2.5, then setting the exposure on the basis of an actual visual
magnitude 2.5 star (as seen under the current seeing conditions) would
be extremely sensible.
--
For the few who have read this far, let me again applaud your interest
and enthusiasm in attempting to observe this event, both personally
and as an organizer of others; and your hard work on your slide
presentation:
http://groups.google.com/group/lcross_observation/msg/9611758825327ea0
I do think the bulk of that presentation, despite its protestations of
uncertainty at the start (and now on page 31), invites amateurs to
prepare for a single easily observable impact scenario which may or
may not be the correct one.
I would think that at least some people may wish to prepare for the
possibility of a very difficult to observe event. Such folks would
have the advantage of being happily surprised if it turned out easy;
as opposed to being disappointed by setting up for what they supposed
would be a very easily seen event, only to discover additional
measures were necessary to observe it. As previously mentioned in a
different thread, for the former group I would think that useful
experiments between now and impact night might include testing on
terrestrial targets (such as lights at night) the effectiveness of
possible measures to reduce glare in their instruments, and
experimenting on the Moon to find the conditions and powers that most
extend their ability to see, for example, the first and last
appearance of sunlit peaks and ridges out of shadows in small craters
-- something Chris Kitting has shown to be a non-trivial exercise, and
an area in which much improvement would seem possible. Perfection of
techniques for detecting small changes in the edges of shadows, and
rejecting distortions due to seeing (if that is possible), may also be
helpful, for recent descriptions suggest the LCROSS impact may (to
minimize the vertical distance to sunlight) be directed to the outer
edge of the shadow in the target crater, and hence may be seen mostly
as a pair of brief indentations in the edge of that shadow (from the
dense sides of the lampshade ejecta curtain moving laterally out), and
a faint wave moving across the shadow (from the higher and thinner
central part of the curtain expanding towards the shadow-casting
rim).
-- Jim
Thank you, again, for your patience with my misunderstanding of your
postings regarding photometry. This is probably a problem unique to
me, for I see the resulting imaging recommendations have been widely
and warmly embraced on at least some of the other forums where you
have posted them:
http://www.cloudynights.com/ubbthreads/showthreaded.php/Number/3339253/
I hope, too, that my remarks about possible inconsistencies in the
information offered by the LCROSS science team about the size and
brightness of the ejecta plume, and ways that information has been
interpreted, will not discourage anyone from looking. That has never
been their intention, and I, personally, have no idea if the LCROSS
impact event will be easy or difficult to observe from Earth.
And finally I hope that an extended discussion of the following
question will not distract too much from efforts to prepare for
observing and imaging the event.
> I recommend that you ask yourself why the photograph
> I took of the lunar surface came out properly exposed
> Had I exposed calibrated to a 5 V mag star, the image
> would have been horribly underexposed.
I am guessing this is in reference to the "test panel":
> would have been horribly underexposed.
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/2009_8_14_0246UT_KafTestPanel.jpg
mentioned in:
http://groups.google.com/group/lcross_observation/msg/d9b0b36ce9f2c59e
showing what looks (on my CRT monitor) like a rather dim image of the
Moon next to a what looks like a rather overexposed image of "2.87v
Alcyone".
I am not disputing that you were happy with the result you achieved,
but I think the advice to adjust exposures for the Moon (an extended
object) based on the appearance of star images (point sources) lacks
the generality needed to make it helpful to others.
If I am not mistaken, Tom Bash was trying to point out this
fundamental problem in response to a copy of one of your postings
submitted to another forum:
http://tech.groups.yahoo.com/group/lunar-observing/message/25356
As Tom says, the relative intensities observed for stars (point
sources) and planets (extended sources) differ with aperture (and
seeing). What gives a good result with one aperture and seeing
condition may not give a good result with a different apertures and/or
seeing.
This is not to say star images are not helpful for calibration; they
just don't seem to be so for setting exposures unless one has reason
to think the impact plume will have a star-like appearance with known
total magnitude.
--
The concepts and math related to the imaging of extended versus point
sources by a simple telescope/detector combination are quite
straightforward (see below), and were partially presented by Clif
Ashcraft in some of the earliest threads on this form:
http://groups.google.com/group/lcross_observation/msg/fc2a2eb642d740f6
They are much simpler than the ideas needed to grasp the limits of
visual observing, which are the subject of Dr. Schaefer's many
articles.
It would be easier to comment on your specific statement about the
appearance of a "magnitude 2.8 star" relative to the 71% illuminated
Moon if I was sure I understood what you mean by "calibrate to".
In an early version of your slide presentation and imaging protocol:
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/20090916LCROSSImpactUpdate.pdf
mentioned in:
http://groups.google.com/group/lcross_observation/msg/9611758825327ea0
I had the impression that you were recommending that amateurs set
their exposures based on the bright rim of Cabeus A1, but supplement
this with "calibration images" of stars of various magnitudes,
possibly at the same exposure, or possibly with the exposure time
modified to avoid saturation or blackness. I assumed, apparently
incorrectly, that your intention was to add up the total detector
counts resulting from a star of known magnitude, apparently on the
assumption that the light output from these stars is better known than
that from the Sun or Moon.
A more recent version (9-18-2007rev, p. 46) seems to recommend setting
the camera's gain and exposure time so that the peak detector count
(at the center pixel of the star image?) comes out at some comfortable
level, then imaging the Moon with same setting, as an alternative
procedure.
In the present posting you seem also to be saying that when the center
pixel of the image of a magnitude 2.8 (or 2.87?) star is in a
comfortable range, the sunlit parts of the Moon will be in a similarly
comfortable range. This interpretation seems to be supported by page
48 in your imaging protocol, where you appear to be assigning a pixel
value to each star that possibly indicates the count in the central
pixel of the image (although it could be the sum of counts over all
the pixels of the stellar image?). It is contradicted by your present
statement that if you "calibrated to a 5 V mag star" (as opposed to a
magnitude 2.8 one) the lunar image would be *underexposed*: I would
have thought a fainter test star would dictate higher gain and
exposure time settings, leading to an *overexposed* lunar image.
Tom Bash, too, seems to have been confused as whether you are talking
about peak ("ADU") or total ("stellar flux") detector counts
http://tech.groups.yahoo.com/group/lunar-observing/message/25387
So I am not at all sure I understand what "calibrating to" a star
means, but I will proceed as best I can. The following commentary
consists of my thoughts on this subject, and may contain errors. My
reading is not as wide as it should be, but I assume similar thoughts
can be found expressed with greater clarity and accuracy in standard
texts on astronomical imaging.
--
To the best of my knowledge, the only concepts needed to understand
the relative counts per pixel expected in the image plane of a
telescope observing point and extended sources are:
1. In any given spectral band, a point source bathes the front of the
telescope in certain number of Watts m^-2, loosely indicated by its
"stellar magnitude".
2. The total energy collected is that flux per unit area times the
clear area of the entrance aperture.
3. After correction for losses in transmission and reflection, that
total energy is distributed over an image area that in a ideal system
is the diffraction pattern of the entrance aperture, but in real life
is likely to be fuzzed out by the imperfect focusing of the optics,
and, usually most importantly, by seeing. This is in turn converted
to "counts" according to the area and quantum efficiency of the
detector element at the wavelength of interest.
4. For an extended source in the same spectral band, each unit solid
angle of the source bathes the front of the telescope in certain
number of Watts m^-2, a quantity that can be indicated by a number of
magnitudes per unit solid angle.
5. The total energy collected *per unit angle* of source is this flux
per unit area and per unit solid area times the clear area of the
entrance aperture.
6. This total energy collected by the entrance aperture, after
correction for the same losses, is distributed over an area in the
image plane equal to the unit of solid angle operating over the
effective focal length of the system. The portion of this flux falling
on a single detector element is converted to counts in the same way as
for a point source.
Any desired relationship regarding how the relative appearance of the
two kinds of sources will vary with aperture and focal length can be
derived from these simple and intuitive principles, although trying to
translate the answers to the sign-reversed logarithmic system of
stellar magnitudes can introduce additional levels of confusion (I,
for example, am continually tempted to sum magnitudes, or to multiply
mpsas by areas in square-arc seconds to obtain a total magnitude, but
these are not valid operations).
Looking at the six intuitive principles list above one can readily see
that the intensity per pixel in the image of a point source is going
to be affected by variables of resolution and seeing that do not
affect an extended source. For this reason, point sources cannot be
used as reliable guides to the exposure of extended objects unless one
employs some complicated scheme involving summing the total counts
received from the point source and correcting that to the expected
area of the extended source.
In poor seeing, the count at the center pixel of a stellar image is
entirely at the mercy of the current smearing. In good seeing, larger
apertures and more optically perfect telescopes stuff the captured
light into a small angular size than do smaller and less perfect ones
(the angular size of the diffraction disk *decreases* with increasing
aperture).
--
For the sake of discussion, I will take your statement to mean that
with your observing set-up you found the central pixel of the image of
a magnitude 2.8 star to exhibit the same count as an average over the
sunlit portion of the Moon, which (again purely for the sake of
discussion) we can assign a surface brightness of 5 mpsas. Under what
circumstances would this be expected?
Again, not to be overly mathematical, but in the absence of any other
information one might reasonably guess that the light from the star
would be spread over a vaguely bell-shaped or "Gaussian"
distribution. Mathematically, this can be written:
I(r) = Io * exp(-(r/a)^2)
where Io is the count (per pixel) at the center of the image, I(r) is
the count (per pixel) at distance r from the center, "a" is a scaling
factor, and "exp()" means raising "e = 2.718..." to the power
indicated in parenthesis (labeled "e^x" on many scientific
calculators). The intensity at r = a is Io/e = 0.368*Io, and the
full width at half maximum (FWHM) is 1.665*a . A table of integrals
indicates that the total counts in such a pattern is the same as if
the central intensity (Io) existed uniformly over a disk of area
Pi*a^2 (where Pi = 3.142...).
Without trying to be intentionally obscure, if I quite arbitrarily
assume a = Sqrt(1/Pi) arc-sec = 0.564 arc-sec (or 0.939 arc-sec FWHM)
the central intensity of the magnitude 2.8 Gaussian stellar disk would
be expected to be the equivalent of spreading the total counts
uniformly over a disk with an area of exactly 1 square arc-sec, which
is to say 2.8 mpsas. You are telling me (as I interpret your
statement) that the actual observed intensity at the center of the
stellar disk is dimmer than this: in fact, the same as the intensity
observed for a 5 mpsas extended source. Since by definition 5
astronomical magnitudes is a factor of 100 in brightness, the
difference between 2.9 mpsas and 5 mpsas is a factor of 100^((5-2.8)/
5) = 7.59x, meaning the actual stellar disk must have been 7.59x
larger in area than I have arbitrarily assumed. Therefore, based on
your observation, I conclude the actual stellar disk, if Gaussian, had
a = 0.564*Sqrt(7.59) = 1.55 arc-sec, or FWHM = 2.59 arc-sec. This
FWHM seems roughly consistent with the image of Alcyone that you
present, although the pixel scale is probably non-linear so I can't be
sure where the 50% intensity point is, or how the count pattern would
look if averaged over azimuth.
The large observed stellar disk size (assuming I am interpreting your
statement correctly) could be the result of seeing and/or focus
errors. Potentially, an unobstructed telescope with an entrance
aperture of 125 mm diameter and operating at a wavelength of 0.5
micron, can focus 84% of the light it collects from a point source
into a pseudo-Gaussian Airy central disk with a FWHM of about 0.85 arc-
sec (diameter of first dark ring = 1.01 arc-sec). By this very rough
reasoning, by concentrating the same collected light into this smaller
space, the present telescope (if that were its description), under
perfect conditions, could produce a star image with a count at the
central pixel 0.84*(2.59/0.85)^2 = 7.8x higher than in the present
image. Had the camera gain been reduced to produce a comfortable level
at the center of the star image under those ideal conditions, the
lunar image (whose smeared out brightness is not affected by these
considerations of resolution) would presumably have been underexposed
by that factor. Conversely, if the seeing had deteriorated to where
stellar images had an observed FWHM of 5.2 arc-sec (twice the 2.6 arc-
sec), the central count would decrease in inverse proportion to the
increase in area (that is, by a factor of 4): the gain would have to
increased to recover the intensity in the stellar image, giving an
overexposed Moon.
Again, I may still not be correctly understanding what you are trying
to say, but the recommendation to set exposures based on the peak
count in the image of a magnitude 2.8 star seems equally problematic
if applied to telescopes of different apertures. It *would*, however,
work if all observers were experiencing uniformly bad seeing
conditions. If the seeing disk could be relied upon to always give
stellar disks of 2.6 arc-sec FWHM in telescopes of all sizes, then the
stars could be thought of as extended sources of that size, and such
star images could be used for setting lunar exposures just as setting
the exposure to capture Mars at a comfortable level will also capture
a 71% illuminated Moon at a comfortable level.
Your recommendation to use stars to set lunar exposures might also
work if you are talking about the total count from the star in the
image plane, although this would require fore-knowledge of the lunar
brightness and some rather elaborate considerations about image scale,
and I have trouble seeing the advantage of doing this over simply
setting the gain to give a nice lunar image. As an example, on page
47 of your imaging protocol you mention an image scale of 0.32 arc-sec/
pixel. In that particular case a single pixel would subtend an area
of 0.32^2 = 0.10 square arc-sec. If we hypothesize the Moon we want
to expose has a surface brightness of 5 mpsas, since a 0.10 square arc-
sec area will capture 1/10th the amount of light captured by a 1
square arc-sec area, it will see the equivalent of the light from a 5
- 2.5*log(0.1) = magnitude 7.5 star (if I have done the math right).
In principle, then, one could set the exposure so that the total
counts from (all the pixels recording) a magnitude 7.5 star equaled
the count desired for 1 pixel in the lunar image. This would, of
course, require dark level subtraction and a detector with a linear
response of counts to light level in W m^-2.
Your recommendation to set exposures based on stars rather than the
lunar disk would also seem appropriate if there is reason to believe
the most interesting part of the LCROSS impact will be seen as an
unresolved star-like point of known total magnitude overlying the
shadowed interior of the target crater. If for example, it is
confidently known that such a dot will have a total visual magnitude
of 2.5, then setting the exposure on the basis of an actual visual
magnitude 2.5 star (as seen under the current seeing conditions) would
be extremely sensible.
--
For the few who have read this far, let me again applaud your interest
and enthusiasm in attempting to observe this event, both personally
and as an organizer of others; and your hard work on your slide
presentation:
http://groups.google.com/group/lcross_observation/msg/9611758825327ea0
I do think the bulk of that presentation, despite its protestations of
uncertainty at the start (and now on page 31), invites amateurs to
prepare for a single easily observable impact scenario which may or
may not be the correct one.
I would think that at least some people may wish to prepare for the
possibility of a very difficult to observe event. Such folks would
have the advantage of being happily surprised if it turned out easy;
as opposed to being disappointed by setting up for what they supposed
would be a very easily seen event, only to discover additional
measures were necessary to observe it. As previously mentioned in a
different thread, for the former group I would think that useful
experiments between now and impact night might include testing on
terrestrial targets (such as lights at night) the effectiveness of
possible measures to reduce glare in their instruments, and
experimenting on the Moon to find the conditions and powers that most
extend their ability to see, for example, the first and last
appearance of sunlit peaks and ridges out of shadows in small craters
-- something Chris Kitting has shown to be a non-trivial exercise, and
an area in which much improvement would seem possible. Perfection of
techniques for detecting small changes in the edges of shadows, and
rejecting distortions due to seeing (if that is possible), may also be
helpful, for recent descriptions suggest the LCROSS impact may (to
minimize the vertical distance to sunlight) be directed to the outer
edge of the shadow in the target crater, and hence may be seen mostly
as a pair of brief indentations in the edge of that shadow (from the
dense sides of the lampshade ejecta curtain moving laterally out), and
a faint wave moving across the shadow (from the higher and thinner
central part of the curtain expanding towards the shadow-casting
rim).
-- Jim
Reply all
Reply to author
Forward
0 new messages