On the evening of September 2nd 2006, I began setting up my 8" Meade LX-90 UHTC with a stock wedge and tripod in the driveway of my Santa Fe home. I was hurrying to get everything ready and aligned in time to try and record images of the SMART-1 impact area on the lunar surface. Impact was predicted to occur at approximately 11:42PM local time (Mountain Daylight Time) on September 3rd at 05:42AM Universal Time.

Conditions had been partly cloudy during the early evening and were improving as impact time neared. There was a slight breeze out of the west, but it was steady with no gusts. It was probably about 5-7 miles per hour. At my location, the Moon was to be just about 13 degrees above the horizon as impact time approached. It showed a waxing gibbous phase, with 73% of the surface illuminated while it steadily dropped toward the southwestern horizon in the constellation Sagittarius. A stream of patchy clouds billowed across the face of the Moon adding an element of suspense.

I brought out my laptop and Philips ToUcam Pro webcam fitted with an Infrared blocking filter to use for image capture. I mounted the webcam to a diagonal for prime focus imaging at the LX-90's native focal ratio of f/10. I wasn't that familiar with area of lunar surface near where impact was predicted to occur, but on my laptop display I could make out ray features near the southern end of the terminator related to the crater Tycho. I used those rays as guides to help me know where I needed to "step off" into the earthshine.

I used K3CCD Tools as my capture software. Expecting a dim event, I placed the webcam settings to a fairly high gain with a medium saturation. I used a slow shutter speed and set the recording to 5 frames per second.

I worked to frame the area of the lunar surface where I thought SMART-1 was going to make impact. On my live display, I noticed that the area of earthshine was very dark. Few surface features were visible. I didn't want the capture to be too noisy (it seemed like there was enough noise already) so I resisted further adjustments to the gain setting. As I used my Autostar hand paddle to make small movements towards final placement, I could see the outline of an ancient ridge near the bottom right corner of my field of view. I decided to move a little farther into the earthshine in an easterly direction. The ridge disappeared out of the frame.

I began recording AVI video frames a few minutes before impact time to ensure everything was working smoothly. My image scale is .58 arc seconds per pixel with a field of view 3.08 x 2.31 arc minutes. I had no problem with the first couple of attempts capturing AVI data, so I kept on recording. After a few captures had passed and while recording another file, I was startled by what seemed to be a faint streak near the bottom right side of my capture frame. It was almost like a small bug had flown into the glow of my laptop screen. I recorded 5 more captures with durations ranging from about 30-60 seconds to be sure I had covered the impact interval. Seeing nothing more than patches of cloud moving across the frame, I began to shutdown and pack the gear back into the house.

Once back inside, I began to examine the capture data. I was excited to see if I could find the file that had the 'streak' in it. I didn't see anything like that, but in the fourth capture file I did see a small blue dot appear for an instant near the bottom right corner of the recording. Enlarging the view from 320 x 240 to 640 x 480, I ran it again and I could see that there was a dim afterimage in the frame just following the blue dot frame.

I worked to convert the AVI file to BMP single frames for assembly into an animated GIF. I used AVITricks Classic to convert to BMP frames and Adobe ImageReady to assemble the animated GIF and add titles. The animated GIF is comprised of a 9 frame loop with a .5 second delay between each frame. A gradient from the terminator can be seen from right to left. Other than enlarging the frame and converting and compiling the data, no other processing or image enhancement was done.

After I completed the animated GIF. I posted it to my server and added a link for it to some online discussion groups. I was eager to see what others who attempted to observe or capture the event would report. Over the next few hours and days, my server began to get hits from all over the world. A number of people posted that they were not successful imaging or observing the impact. I started to work my data again to see what else was captured.

By browsing around the discussion groups, I learned that the Canada-France-Hawaii Telescope (CFHT) team had imaged the event. So, I visited their site to see what information had been posted. There I found a flash contour plot that I thought might be interesting to use for comparison with my flash capture. A contour plot is a way to show shape and values of brightness. Similar to how a topographic map shows the contour of a mountain with closer lines indicating steeper grade, the CFHT plot shows brighter parts of the flash as areas with lines closer together. The central part of the CFHT plot shows a loss of data due to saturation of their detector. In Adobe Photoshop CS, I worked to isolate the part of the frame where the blue dot appeared and enlarged it by 600x from the 640 x 480 size image. Then, I rotated it counter-clockwise to give an orientation that appeared to match what was posted on the CFHT site. The blue dot spans a number of pixels and shows evidence of shape and different levels of brightness.

 

I opened Patrick Chevalley's and Christian Legrand's Virtual Moon Atlas Pro (VMA), and started to work on comparing the position of my capture frame with the surface features known to be in the area of the impact site. Using VMA, I was able to approximately match the position of impact area with what I had framed before beginning my capture. I used a blue overlay to get a better sense of how the webcam detector was oriented. The surface features of the crater rim near Clausius 'H' and the ridge near Lee 'A' and Lee 'S' that I slewed out of the frame are identified by yellow target marks.

Going back to the CFHT site I looked at the 15 frame mosaic posted there to get a sense of how the shape of the faint glow in the afterimage frame compares with the dust cloud imaged by the 3.58 meter CFHT. The elongated 'C' shape looks pretty close to what is captured in my data.

With the kind assistance of Raffaello Lena of the Geological Lunar Research Group in Italy, a noise analysis of the flash frame and the after image frame was conducted. He took both frames and averaged the noise profile resulting in the image shown to the left. The flash frame is at center and shows a signature clearly above the background noise profile. To the right, is the faint after image frame which shows a lower amplitude signature, but is still above the background noise profile.

I have shared information and copies of my original capture data for further analysis and review with a number of parties around the world including Detlef Koschny of the European Space Agency and Christian Veillet of CFHT. I welcome additional questions and comments.

 ***

A note I received from Dr. Pierre Martin of CFHT

***

Links to cached pages showing discussion and news about the capture

Cloudy Nights Forum

Lunar Picture of the Day

Sky Tonight News

Sky & Telescope Magazine

I would like to acknowledge and say 'Thank You' to Meade Instruments Corporation and Astronomics for donating the 8" LX-90 UHTC telescope which was given away as door prize for the 2003 Texas Star Party. I ended-up being the lucky guy who took it home. That telescope is the same one I used for the capture.

Probe images at top of page © European Space Agency