ASTEROID CCD PHOTOMETRY AND LIGHTCURVE ANALYSIS PERFORMED BY AMATEURS TO PROFESSIONAL STANDARDS [Figure 1] The title of this presentation is "Asteroid CCD Photometry and Lightcurve Analysis Performed by Amateurs to Professional Standards." A majority of all papers published in the Minor Planet Bulletin in recent years have been on the topic of asteroid lightcurves and rotation periods, and the majority of these authors have used the MPO Canopus program to analyze their photometric data. I consider it useful to show an example of the procedures used in MPO Canopus to construct asteroid lightcurves and measure their rotation periods and amplitudes. ALPO members outside the Minor Planets Section may gain much better understanding of what the Minor Planets Section is contributing. The observational procedures are simple. Turn on the CCD, take a set of darks, flats, and flatdarks, slew to target, take images of the same field all night while you sleep. In the morning merge these hundreds of target frames with the darks and flats, then measure the processed images photometrically. The image acquisition and dark and flat merging processes can be done with any of several software packages. This presentation will illustrate the photometric analysis of the processed images and construction of lightcurves with MPO Canopus software. Asteroid rotation periods, as well as constraints on shapes, are found by photometry, making frequent comparisons of the magnitude of the asteroid with that of nearby stars [#2]. The CCD and powerful software has made CCD photometry ten times easier and cheaper than it was in the days of the 1P21 photoelectric photometer. Amateurs are entering the field in increasing numbers. Asteroid lightcurve research requires a great deal of telescope time. Amateur astronomers with observatories dedicated to lightcurve research are publishing lightcurve papers just as good as those of the professionals of twenty years ago. They do not have to compete for telescope time and often have much more extensive data sets. I personally have about 250 nights per year suitable for astronomical observation, and use all of these nights for asteroid lightcurve data acquisition. The top professionals in the field are very supportive of this amateur research, as much as professional variable star researchers utilize AAVSO observations. Currently more than 90% of all asteroid lightcurves are obtained by by amateurs [#3]. The list of asteroids with reliably found (henceforth termed secure) periods is increasing rapidly. The preferred publication of lightcurve papers is the Minor Planet Bulletin of the Minor Planets Section of the Association of Lunar and Planetary observers. In the year 2009, 90% of all formally published asteroid lightcurve papers have been in the Minor Planet Bulletin [#4]. In most cases for an elongated asteroid, there are two maxima and minima per rotation, and irregularities in these bimodal lightcurves indicate shape irregularities. A large amplitude (magnitude difference between maximum and minimum light) indicates a highly elongated shape. A small amplitude might mean a nearly spherical shape. Or it might mean the pole of rotation is directed nearly to the Earth. Observations at a variety of locations distributed around the sky are needed to distinguish these possibilities. It may be interesting to add a personal touch, explain how I entered the field and how I select my targets, both of which are uniquely Frederick Pilcher. Following retirement from my career as a physics professor at Illinois College and given my lifelong enthusiasm for asteroid studies, it seemed proper for me to engage in asteroid research. Of the several lines of research in which I as an amateur astronomer could make a meaningful contribution, I preferred CCD photometry and lightcurve analysis. I had for thirty years been reading all the papers in the field which I could find, and knew all the opportunities and challenges. Furthermore with my interest in specific asteroids as individuals, I knew of many lacking any photometry, or secure rotation periods, or spin/shape models, for which I desired the answers. Frustrated over the years as others did not make and publish the relevant observations, I would seek them myself. Many of the specific asteroids for which I have obtained lightcurves, or plan to obtain in the months and years ahead, I had already selected years ago. Each spring Illinois College holds an employee recognition dinner for all employees, clerical and maintenance as well as faculty, with special recognition for those soon to retire. At my turn to recieve a retirement plaque and traditional rocking chair, I addressed all in attendance: "I don't consider this my retirement. I consider it my big mid-life career change. I am leaving physics teaching and entering astronomy research, in the specialty of my choice, of course," which I left unmentioned. I have always been proud of my amateur status in astronomy. [#5] Although I had to pay for my observatory and equipment myself, I am pleased that there are no grants, therefore no strings attached. I can choose my own projects and methods. With no pressure to publish or perish, I can wait until I am confident of my results before publishing. But I do have a reputation to preserve. That specialty was, of course, asteroid CCD photometry and lightcurve analysis. My move to the clear skies of southern New Mexico, purchase of a lot a few miles from town with darker, but not really dark, skies, and design of the observatory were all planned for that goal [#6]. Studies of unresolved sources are less severely affected by light pollution than those of extended faint sources, and the burden of living far from town with long drives and excessive gasoline usage to obtain supplies can be avoided. Because I would be tracking a single target all night, a sliding roof in which I did not worry about moving the slit in the dome seemed prudent. The same consideration recommends a fork mount rather than a German Equatorial to eliminate problems associated with the meridian flip and inversion of field subsequent to the flip. Good lightcurves of asteroids as faint as magnitude 15 obtained with 12 to 14 inch telescopes have been published. Hence a 14 inch telescope seemed adequate, and I obtained through the Astro Mart a good Meade 14 inch LX200 GPS Schmidt-Cassegrain. The CCD should have a large field of view for a wider selection of magnitude comparison stars and for greater tolerance of imprecise tracking during an all night unattended session, and also to eliminate the need for a focal reducer. From comparison of a large number of CCDs presented on the SBIG and APOGEE web sites I selected the SBIG STL-1001E CCD with 1024x1024 pixels of 24.7 micrometers each. At the Cassegrain focus of the 14 inch Meade telescope this has provided a field of about 25x25 arcminutes. Four custom filters were included for possible future color work. A clear filter is necessary for maximum possible light from very faint targets. Bessel equivalents of the Johnson B, V, and R filters, plus a clear filter, were also obtained. To date I have only used the R and clear filters. The R filter has been useful for brighter targets and has the additional advantage of reducing second order extinction when target asteroid and comparison stars have different effective colors and extinction is greater at lower altitudes for short wavelengths. The most familiar example of this is that the rising and setting Moon and Sun look red because of differential extinction. The clear filter is needed for faint objects for which I need as much light as possible. Finally an Optec Temperature-Controlled Focuser maintains a sharp focus by moving the focal plane outward, relative to the tube, as the tube contracts thermally with temperature drops during the night. The goal of sleeping through the night of unattended operation of the equipment has been largely achieved. [#7] Brian Warner is both an expert computer programmer and prolific observer of asteroid lightcurves. His MPO Connections program is ideally suited to asteroid data acquisition (telescope and CCD control) and his MPO Canopus program for photometric measurement of CCD images and the construction of lightcurves. This software package has written especially for asteroid observation. It has proved ideal and I need no other software. I now show several images of the observatory and equipment. The first picture [#8] shows the house and observatory from the outside, where a walk of 30 feet from the house brings one to the door of the observatory. It consists of two rooms, an office with a heater/air condiioner housing the control computer and a small library, in the same architectural style as the house. The telescope room with stone walls looks from the outside just like the stone walls which are in the yards of every residence in the development. The sliding roof with rails is scarcely noticeable. The next [#9] shows from the northwest the stone walls of the telescope enclosure, with the office section at right. The next two images [10-11] show the telescope, CCD, and Optec focuser, with the rails and sliding roof in the background. The final image [#12] is of the office with the control computer turned onto MPO Connections and the control for the Optec temperature-controlled focuser. Long before the observatory became operational I devised a standard procedure for lightcurve acquisition [#13]. My goal would be to observe each single target for as many hours in each night as possible and repeat for however many nights were required to obtain a secure, reliable period, and not give up for difficult cases. Observations should begin about a month before opposition and be continued well after opposition. The long interval decreases the +/- error in any derived period. If the target should have a period which is very long or commensurate with Earth's period, then not all of the lightcurve can be found at a single location. These missing segments can then be obtained by collaboration with other observers at widely different longitudes. My experience for four years has proved the validity of this approach. The Asteroid Lightcurve Data Base [#14] is a listing which includes the "best" available rotation period, amplitude or if observations have been obtained at more than one opposition, range of amplitudes found, and reliability of these results. It is downloadable from http://MinorPlanet.info and is updated about twice each year. Several of my own period determinations have been included. I now peruse this list for asteroids for which the period is not secure, or for which no previous lightcurve studies have been published. These become my targets. A small part of the summary page of the Asteroid Lightcurve Data Base is shown [#15]. On the left are number and name; toward the right are listed the rotation period in hours and reliability of this determination. Secure results are assigned 3; reliability 2 indicates the listed period is based on less than complete data and may be wrong. Most of these low numbered asteroids have secure period determinations, reliability 3, but we note for 65 Cybele the reliability is only 2, with most likely period 4.036 hours, amplitude 0.04 - 0.12 magnitudes depending on location in the sky. At this stage I place high priority on observing Cybele to obtain a secure period. For about half of all the asteroids with reliability 2 which I have targeted, my exhaustive investigation showed that the data base period was wrong, and I corrected it. All of my results published before presentation of the most recent version of the Lightcurve Data Base have been included therein. Next I examine a more detailed listing which includes all period determinations [#16]. For Cybele we find the first two studies showed a period of 6.07 hours and all others near 4.03 hours, none with reliability exceeding 2. References to the publications are also provided, and during my new investigation I visit the Branson Library at New Mexico State University to read all of these and make my own assessments of their published values. Starting years ago and requiring an enormous amount of number crunching I prepared a tabulation of maximum elongations of thousands of asteroids over many decades [#17]. Asteroids, like Mars, are best observed near opposition, or, equivalently, maximum elongation from the Sun. I did this to find times of opportunities for observation, brightness, and especially those maximum elongations at which the asteroid is much brighter than usual. It is commonly known that Mars had unusually favorable oppositions, brighter and closer to Earth than usual, in 1971, 1988, and 2003. An analogous situation holds for all asteroids except those with nearly circular orbits. I consulted my tabulation for minor planet 65 and note a maximum elongation for 2009 Sept. 7. Columns in this tabulation, from left to right, are interpreted as follows: asteroid number, date of maximum elongation in format yyyy/mm/dd, maximum elongation from the Sun in degrees, right ascension on date of maximum elongation, declination on date of maximum elongation, date of brightest magnitude in format yyyy/mm/dd, brightest V magnitude, date of minimum distance from earth in format yyyy/mm/dd, and minimum distance from Earth in astronomical units. From my practice of starting my series of observations well before maximum elongation, I made my first lightcurve 2009 July 31. Results from the first nights were inconclusive. I obtained another session on 2009 Aug. 28 UT. In the morning the images were dark and flat corrected, and I now describe the measurement and lightcurve construction procedures. I now set up a New session for the night's observations [#18]. This includes specifying the date, target, telescope, CCD, exposure time and temperature, and calculating the distance between Earth and target. It is standard in variable star work to correct all observed times caused by varying Earth distances to the Sun (heliocentric) and in asteroid work to the asteroid itself (light time). MPO Canopus contains a provision for either, but as I do only asteroids I use the light time correction between Earth and asteroid. When the Session is established I call upon the Lightcurve Wizard to measure the frames photometrically. First I bring a dark and flat corrected frame from early in the session on screen, visually identify the asteroid (from having observed it at the start of the observing run) and select five comparison stars [#19]. These should be fairly bright but not bright enough for their image pixels to be saturated, fairly close to the asteroid, and distributed on all sides of the asteroid. Some tradeoffs are necessary as all of these conditions are rarely achieved in practice. I move the cursor to each of five comparison stars and to the target asteroid. A circular measurement aperture for each is now created. The yellow circles are for the first comparison star, whose pixel coordinates become the reference for all others. The red circles are for the other four comparison stars, and the green are for the moving target asteroid. The well counts in all pixels inside the central circle are added. The middle annulus is a blank zone in which nothing is measured. The median well count in the outer annulus is used to establish the sky background and is subtracted from the well counts in the central circle. Star images in the outer annulus can be ignored as they occupy less than half of all pixels in this annulus and will not affect the median. For each comparison star and for the target the software now displays successively the (x, y) coordinates of the image centroid, largest well count of any pixel within the aperture, full width half maximum of the image, and signal to noise ratio. The entire procedure is repeated for a dark and flat corrected frame from late in the session. The (x, y) offsets of the other four comparison stars from the reference star have not changed, provided that there is good polar alignment and negligible field rotation. But the asteroid has moved, and the change in offset must be noted. The software now calls to the screen each processed target frame in time sequence from the first. The cursor is clicked on the first comparison star The yellow circles appear around this star, red circles around the other comparison stars at the appropriate offsets from the first comparison star, and green circles at the target asteroid. Accept is clicked if an image is satisfactory. Otherwise one may double click to automatically advance to the next image, and the preceding is not measured. On this particular night there was a moderate wind, which buffeted the mount and smeared the images slightly as is noticeable in the first measured image. For astrophotography this would be ruinous. For photometry only the sum of all well counts, not the number of wells in which photons register, is significant. The measuring aperture is enlarged slightly to compensate, but good results are often obtained. From the two times and positions measured for the asteroid, the position of the asteroid on subsequent images is interpolated from the time of exposure. As succeeding frames are measured [20-23], one may be distracted by the slow drift to the left caused by the drive tracking at not quite the sidereal rate. I invite you instead to concentrate on the movement of the asteroid relative to the field stars. Several good images are shown. Later in the night the wind declined, as is seen from smaller and more round images, and it was useful to reduce the size of the measuring aperture. The measurement process has produced a well count sum for each comparison star and the target. These linear counts are converted by the software to a logarithmic magnitude system. The comparison star brightnesses are internally adjusted to variations caused by changing atmospheric extinction. Except for observational scatter they will be uniform through the night. The magnitudes thus computed are called instrumental magnitudes, not calibrated by standard stars for real magnitude. Variation in asteroid instrumental magnitude can now be established by comparison with these comparison star magnitudes with all atmospheric effects removed. Except for instrumental errors any nonvarying star should have the same instrumental magnitude for all images. [#24] Instrumental magnitudes are displayed for the first comparison star]. A few individual measures are discordant, and these are deleted from the data set [25]. This is repeated for each of the five comparison stars. The lightcurve of the asteroid for this one night only is now generated [26]. The operator selects a range of possible periods. Here I selected a minimum of 3.5 hours, maximum of 7.5 hours to include all periods listed in the Lightcurve Data Base. This meant selecting a minimum 3.5 hours, step size 0.004 hours, 999 steps. The software constructs a lightcurve phased to 3.500 hours, another phased to 3.504 hours, and so forth up to 7.496 hours. For each the Fourier coefficients of harmonics up to tenth order and their residuals are computed. The lightcurve displayed is the one plotted to the single period which has the minimum residual. Discordant points are deleted from the solution and the process is repeated [#27]. A period near 6 hours rather than the expected 4 hours is found, an indication that the "official" period may be wrong. Now this process is repeated to include data for all 5 nights [#28]. To reduce the large number of data points they are binned in sets of three, with interval not exceeding 5 minutes. Binning in this sense means that the average for three adjacent data points, provided they are all within a five minute interval, is represented as a single data point. A range of possible periods between 6.0 and 7.0 hours, step size 0.001 hours, 999 steps, is selected. The result shows that all data points for the most recent night lie well above those for the previous nights. This is because the difference between asteroid magnitude and mean for the several comparison stars is different due to a new set of comparison stars being selected with the asteroid in a completely different field. The instrumental magnitudes for the fifth night are systematically adjusted downward to produce the best fit, and the process is repeated [#29]. This is the aha instant! A beautiful lightcurve of 6.082 hours period, fit very nicely by data from all five nights, is produced. I have showed that the Lightcurve Database period of 4.036 hours is wrong and can be ruled out! Just to check I try producing another lightcurve with minimum set at 4.000 hours with step size 0.001 hour [30]. A complete misfit results. It is useful to draw a graph of the period spectrum, for which the rms error in the Fourier coefficients is plotted against the period for all periods within the stated range [#31]. A narrow minimum indicates a well defined period. The period spectrum should always be inspected over a large range of values to search for other possible periods (alias periods) any one of which might actually define the correct period. I have often claimed that "Amateur observers with lots of telescope time often obtain much better data sets than the professionals of 20 years ago." As examples I show several of the early lightcurves. The first ever made of 65 Cybele, [#32] by Schober, Scaltriti, Zappala, and Harris, showing a period of 6.07 hours, turned out to be correct. Several others [33-37] in my opinion have so few data points as not to show any well defined period. Somehow through the years people began to accept the 4.036 hour value, pass it on from one investigation to the next, without in my opinion ever having convincing data. It is a tribute to the senior editor of the Asteroid Lightcurve Data Base, Alan W. Harris, that he was suspicious of the 4.036 hour value and did not assign reliability 3. The final step is writing the paper and publishing in the Minor Planet Bulletin [38]. In this case I learned shortly after the events described above that Robert Stephens had also obtained a lightcurve of 65 Cybele. He accepted my invitation to publish collaboratively, and provided two more lightcurves. The submitted article is illustrated [#39]. It has now been published in Minor Planet Bulletin Volume 37, 2010 January-March, pp 8-9. The next update of the Asteroid Lightcurve Data Base listed my 6.082 hour period as the preferred one for 65 Cybele. At the time I submitted the paper I believed that I had solved the period. But the Lightcurve Data Base still assigned only reliability 2. I needed more lightcurves at the next opposition in October, 2010. The first night showed a variation of only 0.01 magnitude during more than 6 hours, covering both possible rotation periods [#40]. The explanation is that the line of sight to 65 Cybele was nearly directed at the pole. A reliable rotation period could not be achieved in late 2010 with this very small amplitude, and I obtained no more lightcurves at that opposition. But this lightcurve did show that the rotational pole was within a few degrees of the position of the asteroid on the night of observation. Previous data had suggested the pole was at least 30 degrees from the ecliptic. The session yield new information. I must try again at the next opposition in December, 2011, to obtain sufficient data to obtain a secure rotation period. I summarize by stating that asteroids are like the beautiful women for which most of the lower numbered ones are named. Some are easy lays, but many do not yield their secrets readily. I invite all of you to examine or download any or all of my lightcurves from the web site of the Astronomical Society of Las Cruces (ASLC) http://aslc-nm.org. Toward the foot of the ASLC home page on the left side click on The Science of ASLC - Asteroid Lightcurves. A list appears of all the lightcurves which I have posted to date. Click on any line in this list to view or download the relevant lightcurve. Digital versions of the Minor Planet Bulletin, and the Asteroid Lightcurve Data Base, may be freely downloaded from http://www.MinorPlanet.info. I now invite questions from the audience.