Stefan Kürti´s Astro Web

How to discover an asteroid in SkyMorph archive

SkyMorph image archive

You can discover a new uncatalogized asteroid searching CCD images from SkyMorph database. In order you could find out that this is possible, I wrote this tutorial. As it works for me, I hope it would work for you too.

The search for new asteroid seems very complicated at first, but at first everything looks complicated ...

We can divide search of images into several steps:

  • download of images from SkyMorph archive
  • moving objects identification
  • search positions from other nights
  • report the discovery to Minor Planet Center.

  • Here are internet addresses you need for search:

    SkyMorph     images from NEAT archive
    Astrometrica     programm for astrometry work
    ASTPLOT     gives you positions of minor planets
    Find Orb     programm for orbital computation of asteroids
    MPEph     ephemerides of asteroids and comets
    MPEC     recent MPEC circular

    Download of images from Skymorph archive

    You can choose from 2 ways how to select images from archive - either by coordinates of specific part of sky you wish to search by (Position/Time), or by specific minor planet or comet (Moving Target).

    In the first choice SkyMorph returns all images according to selected coordinates Ra, Dec and time. This search works also in case you choose only one of them - position or date and time. Time interval is really broad - from 1 second to 1 year. On the picture bellow you can see the result of search for selected coordinates and date in 1 hour scale. It returned only 1 triplet of images. When you are searching for images of specific area of sky or specific dates you should use this tool.

    For our aim, to find new asteroid, is better to select images by Moving Target menu item. Click on it and write into the window Target any minor planet name, its provisional or permanent designation. SkyMorph returns list of all images, where required asteroid should be find. I usually select my search field from recent MPEC circular, where lately discovered minor planets are announced. My main criteria for selection are asteroid observation arc and its absolute brightness (H). Longer observation arc means more accurate orbital elements and so the computed positions of searched object on NEAT images are also more accurate. And the bigger H means also higher chance to see the minor planet on images.

    When I found archival images of newly discovered asteroid, I reckon that it was probably not detected in time of NEAT images aquisition in 2001-2002. So then it is more likely that beside this specific object, there could be more unnoticed objects. And that is, what I am looking for!

    When you choose recently discovered asteroid for the search, write its designation into empty window Target. In my example for this article I choosed minor planet 2009 BE175 (in short vesion as K09BH5E). Then in Search column you check from available databases only NEAT database and submit it.

    The programm produces list of images where asteroid 2009 BE175 should be found. Now you select three images taken on same starry field (so called triplet) which you will review. My first criteria which images to scan, is the brighness of object, you can see in the column Mag. I selected images from 2002 May 11, where the minor planet´s brightness was 20.2 magnitude. Depending on seeing at the time when images were taken, you can find on images objects up to 20,5 mag., but sometimes you can detect asteroids even at 21th magnitude!

    Now you check in left column images you wish to search. In the table Request parameters you choose the size of returned images in NEAT pixels box. I recommend size around 1000 pixels. Check in also a box of Show singlets, other parameters leave blank and submit it by Display images.

    In a while you will see requested images on your PC monitor. Bigger pictures means longer time for dowloading. Now you download images by "Save as ..." Images are so called Astro Image type, their endings is "fits". Your computer save them in this format, unless you have different type of files in your setup for saving.

    Moving objects identification

    When you have saved images, you can view them and search for moving objects. I use for this Astrometrica software. It is freeware and it is fully operational during 100 days. If you find out that it works well for you, register it. You can download it from this location Astrometrica


    It is necessary to adjust several specific parameters in Astrometrica according to you as user. First you will need to change the configuration file for those observatory and instruments which were used to aquire NEAT images. There are three different config files: 644. cfg, 608.cfg and 566.cfg.

    During the whole NEAT observing period there were taken images from 3 locations with different setups. To receive precise measurements you have to use correct config file according to viewed images!

    You can download config files from here:     566     608        644

    How can you recognize which images belongs to which observatory?

    566 - all images before 2000 December (Haleakala)
    608 - images after 2001 February (Haleakala)
    644 - all images ends with letters a,b and c (Palomar).

    In my example I choosed triplet from 2002 May 11 from the produced list with asteroid 2009 BE175. There you can see letter "c" in images ID, which means that those images are from Palomar Observatory (644) and accordingly we use config file 644.cfg.

    Exposures of images taken before 2000 December were usualy 20 seconds and there is limited magnitude about 19. So if searched object is 20th magn. or even dimmer, you could not find it. You should skip Images from Haleakala (608) taken after 2001 February is better to skip. You would hardly identify moving objects there, unless you are an experienced "hunter" as these images quality is poor. It requires a lot of practise. Exposures were also 20 second, so you can go down only to 19 mag., not deeper. The best for searching are pictures from Palomar observatory (644). In those images you can easily identify objects to 20.0-20.5 magnitude. as their exposures were 60 seconds.

    On opening of Astrometrica you will see this question first:

    Answer Yes at first time and than everytime you do not need your older measurements anymore. When you say Yes, software opens a new MPCReport file. This way you may lose all your previous measurements. So otherwise your answer should be No!

    First step after you open Astrometrica is select the right config file. In our example as we have images from observatory 644, we choose 644.cfg by click on Settings in the File menu or just simply click on first icon in the header menu and open this specific configuration file.

    As you will send astrometry of your discoveries to the MPC, you have to add your name and contact into the configuration file! Fill these in Observing Site, Details: write your name, contact, e-mail address and in the Measurer row write your name as follows: first name as a first letter and than your family name.

    You have to adjust some data in Enviroment section too. Specify the route in your PC where you store the Astrometrica programm, in order that Astrometrica can find the way to needed data during work. Save your corrected config file and use it in this form.

    After your settings were done, upload your triplet to be viewed - via File - Load images. Find your images saved from NEAT archive and click on them.

    When you upload an image, a new window opens on your screen. Fill the time and date when images were taken. Be careful - should write the middle of image aquisition´ time, not a start! In returned list of SkyMorph images you find this in column Time.

    Data Reduction

    After images were uploaded, next step is matching starry field from images with the star catalog. In this step Astrometrica create a "grid" for RA and Dec and find reference stars. This is very important step - only when you have correct adjustment, you can measure exactly. In astrometry work, this is essential!

    In second menu row Astronomy click on Data Reduction, or simply use the shortcut by CTRL+A and then you will see a new window for coordinates input (RA a Dec) of searched CCD images. In the list you will find these in rows Predicted Object Position. You fill in both coordinates and click OK. Astrometrica checks now reference stars. In case you filled wrong data, Astrometrica could not find correct reference stars, so it matches them incorrectly. After this step in table Data Reductions Results look at Fit order.

    You need Fit Order = 3! Sometimes coordinates are shifted and Astrometrica could not find reference stars. In such a case a table Reference Star Match Error will occur. You can solve this in 3 ways. When you use first - Manual Reference Star Match, you can move reference stars by your own in a correct direction and match stars on your images with reference stars mechanically. But before you do this, I recommend to take second choice first - Automatic Reference Star Match using. And only after that would not help, then try to match stars manualy. To accept the third possibilty means to accept erroneus match, which I do not recommend at all. You will have wrong reference stars and so your measurements are wrong too ...

    Let´s assume your data reduction process ran OK with required Fit Order value, so in a moment Data Reduction Result table with Fit Order = 3 appears. Now you close this window and finally start your search.

    Now you just click with mouse on selected object and a new window Object verification, opens where you can see information of selected object.

    Images review

    Select in the ASTROMETRICA menu bar in the upper row Tools item and click on Blink Images, or use the shortcut CTRL+B. You can change the speed of blinkage in range 1-10. You will find which speed allows you to detect moving object most conveniently. In second row of menu bar find Blink Frequency button, by which you can change the speed of blinking - just moving up or down the arrow.

    Watch images as they appear on your screen and look for moving object against the starry background. Astrometrica has an automatic identification of moving objects (6th icon in menu bar - Moving Object Detection), but don´t use it! Better to rely on your eyes. At the start it will not be easy to catch the moving objects, later, after some practise, it will be easy. Automatic identification is so "perfect", that on image where are just a few real moving object, he would "detect" tens of them ....

    When you catch a moving object, measure its positions by clicking on it on all three pictures. When you click on moving object and Object verification window opens, you have to name it on the bottom of opened window. You can assign your "discovery" by 6 digits code. I name my objects as a krtxxx, where last three characters is the sequence number of object I am searching. After you named your object, click on Accept button. This way review whole image and measure all detected objects. Positions of measured objects are stored in the MPCReport file, which you can open via menu bar in Files View MPC Report File.

    In the picture below you can see MPCReport with measured positions of object named as krt214.

    In next step you have to find out whether your discovery belongs to known or unknown objects. For this I use ASTPLOT, the on-line service of Lowell Observatory. When you enter coordinates of your object from MPCReport, ASTPLOT draws a plot of specified starry field with positions of all known asteroids and comet within specified area. When there is minor planet in the middle of the chart, that means you measured positions of already known asteroid. Otherwise (when nothing showed up in the centre) you probably found something new!

    The chart with asteroids in the field of our example of krt214 shows, that it could be a new object - in the centre of chart there is nothing!

    Now you have to prove that your "discovery" is a real object and you did not measured something which is already known or did not exist at all. It means to find more positions from other nights. This is the most difficult part of your whole work!

    Search of additional night positions

    It this part of your work you need to determine orbital elements of your object. I recommend for this purpose Find Orb. This programm compute orbital elements based on your measurements in Astrometrica. Open the programm, click on Open and upload your MPCReport.txt. It shows you all your data. In our example there you see positions of krt214.

    Now click on krt214 and all positions shows in a window below. As your positions cover just a very short observational arc 30-60 minutes, to receive preliminary orbit you have to use Vaisala method of determination. Click on Vaisala button and you will get the preliminary orbit. Save the computation result with the same name as you named your object - e.g. krt214.txt - by clicking on Save elements.

    How will you know whether your measurements are precise? The value of RMS error tells you more about this. So called mean error - RMS error - should be less than number 1! When it is higher, you did something wrong. It could be caused by incorrect reference stars, or you took for moving object something that was not exist at all. It could be a cosmic ray, electronic noise on images, or perhaps a star.

    Note: minor planets moves according to physical laws, which means in this case straight movement at constant speed.

    When you determined an orbital elements, return back to the SkyMorph and open Moving Target. This time you wil not write anything in Target box! As you are searching for a new object, it could not be in MPC database ... Instead of it you will input your computed Orbital elements. You fill item by item as follows:

    Epoch - epoch of observation in Julian data
    Eccentricity - e
    Perihelion Distance - q
    Perihelion Date - Perihelion - date in Julian date
    Long. Asc. Node - Node
    Arg. of Perihelion - Peri
    Inclination - Incl.
    H magnitude - H

    Be careful with sign in front of decimal numbers - it is dot, not a comma! When you input all data submit them. SkyMorph returns a new list. Now you can check whether your present step was correct. When on this list is also your "discovery" triplet, it verifies and proves that you are on good track. In our example it is OK. Images from 2002 Sept. 15 are on the list! I marked them with red circles in the picture. But you should check also exposures times too - they should be the same too! In other case, you are on the wrong way ...

    From newly received list choose another triplet for reviewing as close to the first date as possible. It does not depend whether selected images are after or before it. Choose the nearest one. Search among images from Palomar observatory (644) and of course keep your eyes also on the predicted brightness of the object. Select your triplet. In our example there are shown two triplets - 6 images from 2002 Aug. 29, from the same date, but different star fields. You have to select images from the same triplet! You learn out which three is from the same one by columns Observation Center. Here are stated coordinates of images centres, do not slip those with computed coordinates of searching object! I marked the triplet with blue circles.

    Now you check in all three boxes in rows of selected images on the list´s left side and download them. Again open ASTPLOT and input computed positions of searched object to draw a new chart. When I put krt214 positions to ASTPLOT, there is nothing in the centre of the field and that is a good sign.

    Next step you already know - data reduction. When it is done, you can review images for moving object near the computed position which are in columns Predicted Object Position of returned SkyMorph list.

    You drop out all known asteroids plotted on the chart, you can see there also their coordinates. Your object should be near the computed position and as bright as it is writen in SkyMorph list. Surely not brigher or dimmer of 1-2 magnitudes. Direction and velocity must be also similar to the first night observations. When there are more unknown moving objects near the assumed position of your "discovery", you need to examine all of them. Measure them and you will learn out the right one after new orbital determination. But be careful! You have to name the object with identical name as at first night. In case there are more new objects, as first, choose the most prospective from them - with the similar movement and brightness. Mark it again as "krt214". Other objects name e.g. as krt214a, krt214b, ... Now you have two night observations. Return to the Find Orb and compute new elements based on two night observations. In case you have not found the correct object, go back to MPCReport, rename next object krt214a to krt214 and repeat computation as long as you detect the object you are looking for.

    Let´s assume you found the right object and now you have its positions from 2 nights. Upload the actual MPCReport in Find Orb, where you can see 6 rows of measured positions now. Check in in Perturbers presets all objects and this time you will use for new computation Auto Solve button. If you measured positions of the same object from both nights, RMS error would change just a little, no matter which way. When it arise much, that means that your positions are not from the same object. Higher value of RMS error than 0.8 tells you that you are on the wrong way. Perhaps you found position of another uncatalogized object... Later it could be your next searched target.

    When new RMS is close to previous RMS rate, you found probably second night positions! Well done. Save your orbital elements and go back to SkyMorph.

    Just repeat already learned steps - input new elemnets and select another triplet from new list of images. Though on previous list there were more images to search, do your third night selection from a new computed list! In our example there are images from Mt. Palomar observatory (644)- besides of 2002 Aug. 29 and 20, use for selection based on 2 days orbital elements. Reason for doing this way is simple - orbital solution based on two nights is better and should returned more accurate positions of hunting object.

    Repeat these steps until you check every possible images. More observation nights, more precise orbital elements and so more accurate predicted positions in SkyMorph list. If there are no additional images you could search, it´s the time to send your discovery to Minor Planet Center.(MPC).

    Sending observations to Minor Planet Center

    In order to receive a provisional designation for your discovery from MPC, you should send observations at least from two nights, but prefer 3-nights! It is not two-three following nights, it means three different nights, although 3 following nights is OK. You can submit discovery from 2 nights too, you can even receive a provisional designation for such an object. When MPC has in their archive another positions for submitted object and those cover longer observed arc, you lost your discovery credit. So instead of discovery you just sent follow up observations for other observer.

    Be careful with your e-mail form. In case your filled the settings in Astrometrica correctly, you can use its form to send your report to MPC.

    Chcek your MPCreport and simple post it. After sending it, you will receive so called Acknowledgment, where you can read how many observations did you sent. In order to send a proper e-mail, read the Guide of MPC - linkage I mention on the buttom of this page.

    Usualy you receive answer about your observation from MPC in 24 hours . When it is a really new and uncatalogized asteroid, you will receive a message called "Designation", where privisional nuumber of your discovery is stated. The e-mail in our example shows that for object I sent to MPC as KRT118 was named as K02QD5E (short version), or as 2002 QE135 in other form.

    If you received message like that on the next picture, that means that you measured precovery positions of already known object... Left bracket means already known asteroid.

    Positions of KRT090 I reckoned as a new, but they were finally positions of minor planet 2002 TZ329. Its orbital elements were so inaccurate, that ASTPLOT did not show this asteroid on the position I detected it. In this case I submitted furhter data which extended observation arc and so this way I helped to improve orbital elements of minor planet.

    In following days your discovery is published in Daily Orbit Update (DOU) which you can see on MPC web page. Our example discovery of 2002 QE135 is in DOU listed between New-one opposition orbits.

    Thereafter a new asteroid is added to catalogue and you can check your discovery at Minor Planet & Comet Ephemeris Service.

    Write in the window the desigation of your asteroid. Then go down a little, check in the MPC-8 line and as a third you check in a box of Show residuals blocks.

    Now click on Get ephemeris/HTML and you will see soon orbital elements of your discovery together with all observations listed in MPC of this minor planet. This way you can follow your discovery from very first positions as far as it will be numbered and even beyond.

    In this moment your discovery process is at an end. Your minor planet now requires follow up observations to refine its orbital solution up to the stage when it is ready to be labeled with permanent number. This period, started with provisional designation and ended by permanent number would last usually several years. But for you as for a discoverer, this process is over, untill you have big enough telescope with sensitive CCD camera to follow up your discovery. Further observations will obtain professionals during their surveys. But to name your discovery is still a long way to go...

    As I mentioned at the beginning, work with SkyMorph archive is quite difficult, and if you carried your reading up to this point, you can still agree with my statement. But after you learned these several steps, you would work easily. The shortest time I made a 3 nights discovery was about 30 minutes, but usually it takes much longer. Sometimes I am unable to find additional positions. The misfortune is, that there are no suitable images in archive, otherwise I would find them ... I cannot choose but put "discovery" ad acta and start new search.

    Searching in archival data is a great fun and delight for me, especially when I can add another new asteroid to my discovery list. Likewise I enjoy following how new observations are added. Furhtermore after every lunation it´s a big thrill whether some of mine minor planet is assigned with permanent number or even named ...

    At the end I would like to mention that before you start your SkyMorph search, do a little homework first and read carefully Guide of IAU about astrometry work:
    "Guide to Minor Body Astrometry". And I can recommend you further reading of Marco Langbroek´s Guide to using SkyMorph, which was my handbook when I started this work.

    In case that such a complicated work does not discourage you and you are venture for your own hunt, I am ready to help you. Just contact me --->>> here!

    Good hunt and enjoy it!

    slovensky     English

    Summary of NEAT discoveries:
          SkyMorph/NEAT (644):       145
           Trojans:                                3
           Main Belt asteroids: 142

           Numbered:  49
           Multi opposition: 53
           Single opposition: 41
           2-nighters: 2

    Last update: 22.05. 2012   

    Stránka optimalizovaná pre rozlíšenie 1024x768 bodov Microsoft Internet Explorer verzia 7.0
    This page is optimalized for default 1024x768 pixels and for browser Microsoft Internet Explorer vs. 7.0

    Copyright Š 2008-2013 S. Kürti All right s reserved