Progress Report #01

September 2015

After more than two years since the start of the project, we would like to summarize what has been done.

As described on our main page, the aim of the project is to reconstruct shapes and rotation states of asteroids by analyzing their lightcurves - this is done by the lightcurve inversion method. To 'officially' introduce A@H, we have summarized the overall concept of the project and its technical description in a paper "Asteroids@home - a BOINC distributed computing project for asteroid shape reconstruction". The paper (currently under review) will be published in the journal Astronomy and Computing.

Between September 2012 and December 2014, we processed data for about 330,000 asteroids. The original idea was to use data from Pan-STARRS project, but it turned out that the amount of data from Pan-STARRS was too small for our purposes -- the number of measurement per object was only a few points on average and we would need at least a hundred. Therefore, instead of Pan-STARRS, we used the so-called Lowell Observatory photometric database, which consists of re-calibrated brightness measurements of asteroids observed from many observatories over the world and reported to the Minor Planet Center. These observations are numerous (hundreds of points per object) but suffer from large random and systematic errors. Oszkiewicz et al. (2011) took the MPC data, selected only the best observatories, corrected for some systematic trends and re-calibrated the data. The result is the Lowell Observatory photometric database that we can use in A@H project. Nevertheless, the photometry is still very noisy and the signal we are looking for (brightness variations caused by the changing shape projection) is in most cases drowned in the noise. That is why the number of models we were able to derive is still only a small fraction of the total number of processed asteroids.

In total, about ~330,000 asteroids with enough data were processed in the framework of A@H. The number of data points per asteroid and their photometric quality in general decrease with the order number of asteroid (asteroids with higher numbers were discovered later). Therefore, we have much more models for asteroids with numbers < 10,000 than for those with numbers > 10,000. We derived new models for 237 asteroids with numbers 1-100,000. There are also some models for numbers > 100,000, but because of large errors in the data we do not consider them reliable yet. We will published the models in a peer-reviewed scientific paper and then store them in the DAMIT database. Preliminary results are already posted on our Results page.

What next?

Since February 2015, we have been processing Lowell data together with data from WISE satellite. The data from WISE were obtained in thermal infrared wavelengths, which means that they are sensitive to the temperature of the surface, not only on the cross-section as the reflected visual data. The ultimate goal for A@H in the next years would be the full analysis of Lowell + WISE data using the model that can compute not only the reflected light, but also the temperature on the surface and then the infrared flux. We are now developing the code and hope for its implementation in BOINC.

Meanwhile, we are testing a simpler approach that the WISE data (or that part observed at wavelengths closer to visible light) can be used the same way as reflected light with the same "period_search" application. The results are promising - we can indeed get unique models form combined data in cases when Lowell nor WISE data are sufficient alone. On the other hand, sometimes we get a false positive solution (we know that the model is different from the "correct" one based on independent data). So the tests we are running now (different resolutions of the shape model, different weights between the Lowell and WISE data, etc.) should show us the optimum setup and limitations of this combined approach.

Palomar Transient Factory

Recently Waszczak et al.(2015) published data and rotation periods for about 10,000 asteroids. These data will be joined with Lowell photometry and processed by the "period_search".


The first intermediate data release of Gaia catalogue is scheduled for mid-2016. After the release, we will add Gaia photometry (expected for ~10,000 asteroids) to Lowell data and re-run the analysis with more data. Although the number of points per object from Gaia won't be high (tens of points), their accuracy will be much higher than anything we have now.

To conclude, the project is running, new data are coming and new models will follow. Thank you all for your participation and patience!