Elsevier

Icarus

Volume 221, Issue 1, September–October 2012, Pages 226-235
Icarus

Looking a gift horse in the mouth: Evaluation of wide-field asteroid photometric surveys

https://doi.org/10.1016/j.icarus.2012.06.046Get rights and content

Abstract

It has recently become possible to do a photometric survey of many asteroids at once, rather than observing single asteroids one (or occasionally a couple) at a time. We evaluate two such surveys. Dermawan et al. (Dermawan et al. [2011]. Publ. Astron. Soc. Jpn. 63, S555–S576) observed one night on the Subaru 8.2 m telescope, and Masiero et al. (Masiero, J., Jedicke, R., Durech, J., Gwen, S., Denneau, L., Larsen, J. [2009]. Icarus 204, 145–171) observed six nights over 2 weeks with the 3.6 m CFHT. Dermawan claimed 83 rotation periods from 127 detected asteroids; Masiero et al. claimed 218 rotation periods from 828 detections. Both teams claim a number of super-fast rotators (P < 2.2 h) among main belt asteroids larger than 250 m diameter, some up to several km in diameter. This would imply that the spin rate distribution of main belt asteroids differs from like-sized NEAs, that there are larger super-fast rotators (monolithic asteroids) in the main belt than among NEAs. Here we evaluate these survey results, applying the same criteria for reliability of results that we apply to all results listed in our Lightcurve Database (Warner, B.D., Harris, A.W., Pravec, P. [2009a]. Icarus 202, 134–146). In doing so, we assigned reliability estimates judged sufficient for inclusion in statistical studies for only 27 out of 83 (33%) periods claimed by Dermawan, and only 87 out of 218 (40%) periods reported by Masiero et al.; none of the super-fast rotators larger than about 250 m diameter claimed by either survey received a reliability rating judged sufficient for analysis. We find no reliable basis for the claim of different rotation properties between main belt and near-Earth asteroids. Our analysis presents a cautionary message for future surveys.

Highlights

► We evaluate two wide-field photometric surveys of asteroid lightcurves. ► We find that only a fraction of reported periods meet our reliability criteria. ► Thus, such survey results are not reliable for spin rate distribution studies.

Introduction

With the advent of large format, wide-field imagers (e.g., Suprime-Cam on the Subaru telescope; MegaCam on the CFHT), it has become possible to acquire images that simultaneously include many asteroids. Thus, instead of targeting just one particular asteroid, one can in principle image a sky area during a night or even a few closely spaced nights and obtain lightcurves on multiple asteroids at once. In this paper we evaluate in some detail two such surveys. On 2001 October 21 UT, Dermawan (2004) obtained 42 images using the Subaru 8.2 m telescope during a single night to obtain photometric measures of about 100 asteroids down to a limiting Cousins R magnitude of about 24. In September 2006, Masiero et al. (2009) obtained about 1000 images over about a 2-week period using the 3.6 m CFHT on Mauna Kea, in what they called the Thousand Asteroid Light Curve Survey (TALCS). They covered a total of 12 square degrees, or about 12 separate fields, so any one field area was imaged about 80 times. Although this smaller telescope did not reach as deep as the Subaru, they did identify a total of 828 asteroids in the larger area covered, down to a Sloan g′ magnitude of about 22.5. Of these, they fit periods and amplitudes to 278 objects, generally brighter than g  21.

With advancing technology and new sky surveys coming on line (e.g., Pan-STARRS, LSST, Gaia), we can expect to see more experiments of this sort, as well as catalogs of accurate photometric measures of asteroids extending over long spans of time. These observations can in principle be analyzed to determine rotation rates, shapes, pole orientations, and phase relations of large numbers of asteroids. It is therefore of more than just passing interest to evaluate these first surveys, not just to validate the results reported, but also to gain insight into the design of future surveys. In a recent paper, Warner and Harris (2011) presented a study by computer simulation of putative sparse-data surveys. Here we evaluate in some detail the two surveys mentioned that have actually been done and reported.

Another such survey report is just published (Polishook et al., 2012), based on analysis of images taken for the “Palomar Transient Factory Survey.” This survey is indeed a “gift horse” in that this one uses data mined from images taken for completely different purposes. We have not yet analyzed this work in detail, but we discuss it briefly at the end of this paper.

The two observers (Dermawan and Masiero) have kindly provided us with much of their data, indeed anything we have requested and far more than we could take the time to re-analyze. Our starting point was to examine all of the derived composite lightcurves in order to make our own independent assessment of reliability of each one, consistent with the ratings we make for all of the lightcurve results we compile in our Lightcurve Database (LCDB, see Warner et al., 2009a). This much was necessary simply to include the survey results in our LCDB in a consistent manner. Both of the surveys claimed to find a number of asteroids with spin rates faster than ∼2 h period and estimated diameters larger than a few hundred meters. Such objects would be nearly unique among some hundreds of other asteroids in our database in the same size range, and therefore seem to call for special attention. In the following analyses we examine these objects especially in each of the two surveys, and point out some cautionary lessons that come to light in the process.

In the LCDB (latest version available at http://www.minorplanet.info/lightcurvedatabase.html), we endeavor to include every published lightcurve that we can find in the literature. In doing so, we evaluate each lightcurve and claimed rotation period and amplitude. In some cases, we make adjustments to the published values, for example we find some observers report as the amplitude of variation the difference between the brightest and the faintest individual measurement, instead of allowing for the noise level of the data so that the real amplitude is generally less than this difference. We also sometimes deduce a different period, for example a different harmonic of the reported value, maybe 1/2, 3/2 or twice the reported value, when we feel that another period is more probable. Finally, we assign a reliability code to the result, which reflects our somewhat subjective evaluation of how certain is the claimed period to be correct. The scale is as follows:

  • 0.

    A result later proven to be incorrect (used only when there is a subsequent, more reliable, result).

  • 1.

    Very unreliable, may be completely wrong; insufficient for statistical analysis.

  • 2.

    Not likely to be grossly wrong, may be wrong by a harmonic value or cycle ambiguity. Sufficient for statistical analyses.

  • 3.

    A secure result with complete lightcurve coverage; unlikely to have a harmonic or cycle error.

We have recently begun assigning “+” and “−” qualifiers to the reliability estimates, e.g. “2−” or “2+”, to refine the scale a bit. Nevertheless, in our statistical analyses of rotation data (e.g., Pravec et al., 2007), we do not include any result that does not score 2 or above (without qualifier). Thus, “2−” is the lowest score that would be included in any of our statistical studies, and “1+” is still rejected. In what follows, one should pay particular attention to the possibility of harmonic ambiguity of “2” or “2−” results, where the “best” solution is for a period <2.2 h, but other solutions with periods >2.2 h are possible.

Section snippets

Evaluation of Dermawan survey

Dermawan (2004), in his doctoral thesis at the University of Tokyo, presented results of a one-night survey (21 October, 2001), using the Subaru 8.2 m telescope on Mauna Kea with the Suprime-Cam wide-field camera, of faint main-belt asteroids. This study has recently appeared in the more accessible and peer-reviewed journal PASJ (Dermawan et al., 2011). The camera has about 0.25 square degrees field of view, and is capable of reaching to a limiting R magnitude of around 24.5–25.0 with a 2-min

Detailed analysis of Dermawan FRA cases

In Table 2 above there are 10 asteroids (listed in boldface) for which the period claimed is less than 2.0 h and the diameter is >0.3 km. If confirmed, any of these would be nearly unique among asteroids of that size in lying above the “spin barrier”. Dermawan has kindly provided us with his data of all of the “FRA” objects, so we have taken a closer look at these 10 objects. In his analysis, Dermawan (2004) used power spectral analysis, which he then followed with a “Window-CLEAN” method

Analysis of TALCS results

TALCS (Masiero et al., 2009) is a more thorough survey, covering more sky area and more objects with more exposures over more time. Indeed, the survey design was influenced to some degree by the shortcomings of earlier attempts such as the Dermawan survey. The analysis used was the Fourier noise spectral algorithm, so not surprisingly our re-analysis found mostly the same periods and amplitudes, but differed mainly in our critical evaluation of lower reliability results. Masiero kindly provided

Evaluation of spin rates including these surveys

Our latest release of the Lightcurve Database (LCDB) dated June, 2012 (http://www.minorplanet.info/lightcurvedatabase.html), includes the Dermawan and TALCS results, with our U-ratings. Fig. 19 is a plot of rotation rates vs. diameter from that database. For all of the results other than Dermawan and TALCS, we plot only results that are rated U  2−, a total of 4466 objects. For the Dermawan and TALCS entries, we plot with separate symbols those results that are U  2− (27 and 87, respectively),

A brief look at the recent survey by Polishook et al.

A paper just published by Polishook et al. (2012) reports a survey of asteroid lightcurves taken from four nights with the Palomar 1.2-m Oschin Schmidt telescope. Four fields were imaged on each of the four nights, each field imaged between 20 and 30 times each night, resulting in up to about 100 measurements asteroids that did not cross into or out of the observed area during the runs. Polishook et al. report that they were able to detect 624 asteroids in their fields. Of these, they claim

Acknowledgments

We thank Budi Dermawan and Joseph Masiero for generously providing us with copies of their data for our re-analysis. We also thank Joseph Masiero and an anonymous referee for careful and thorough reviews that substantially improved the paper. A.W.H. and B.D.W. were supported by NASA Grant NNX10AL35 and National Science Foundation Grant AST-1032896. P.P. was supported by the Grant Agency of the Czech Republic.

References (14)

There are more references available in the full text version of this article.

Cited by (14)

  • Main belt asteroid collision histories: Cratering, ejecta, erosion, catastrophic dispersions, spins, binaries, tops, and wobblers

    2022, Planetary and Space Science
    Citation Excerpt :

    Here I use the August 2019 compilation from the Light Curve Database (Warner et al., 2018) with over 17,000 asteroids in the main belt and NEA space with the quality rating of ‘2’ or better. I do not include Trojans, Hildas, Centaurs, or the somewhat controversial (e.g. Harris et al., 2012; Warner et al., 2020) main belt data from recent Palomar wide-field surveys. The code also makes available the earlier 2013 database.

  • Asteroid Lightcurves from the MOA-II Survey: A pilot study

    2022, Monthly Notices of the Royal Astronomical Society
View all citing articles on Scopus
View full text