September 30, 2004
One of the astronomers on the Hipparcos astrometric satellite team believes he has discovered why the satellite's data put the Pleiades 10-percent closer to us than it is according to other measurements.
The "Pleiades distance problem" has plagued the Hipparcos project since its catalog appeared in 1997. The catalog gave the positions of 118,000 bright stars to an unprecedented accuracy of one milli-arcsecond (another 2.4 million stars were measured with less precision). The stars shifted positions slightly as Hipparcos viewed them from different positions along Earth's orbit. With these shifts in hand, astronomers could determine the stars' distances. But Hipparchos data gave a mean distance to the much-studied Pleiades of only 118 parsecs (385 light-years), or 10-percent closer than the 135 parsecs (440 light-years) previously calculated.
A closer cluster implied its stars were strangely underluminous, emitting almost 20 percent less light than expected. Astronomers instead trusted their knowledge of stellar physics and cast doubt on all Hipparcos distances.
This year saw a veritable onslaught on the Hipparcos Pleiades results in the astronomical literature. In January, a paper in Nature on the distance to Atlas, a binary star in the cluster, suggested firmly that the satellite data could not be reconciled with ground-based observations. The astronomers derived the Atlas distance from observing the stellar pair in its orbit, which established the stars' masses. From this, they determined the true distance between the pair and, from the system's apparent size in the sky, its distance from Earth.
The Hipparcos team retorted on its web site that this distance differed from the one measured by Hipparcos by about the quoted error, and so it could just as reasonably be seen as a confirmation of the satellite's measurement of the distance to Atlas. And Atlas is just one of the Pleiades. For there to be a "Pleiades problem," said project scientist Michael Perryman, you would have to show that Atlas is in the center of the Pleiades. Only then would the star's distance have to be the same as the average distance Hipparcos gave for the whole cluster.
Recently, however, more studies have put the ball in the Hipparcos court. Especially awkward is a paper by Susan Percival of the Astrophysics Research Institute in Liverpool, England, and her colleagues, which has been accepted for publication in Astronomy & Astrophysics. Percival's team estimates the average distance to the Pleiades by a technique called main-sequence fitting. This works by looking at the colors of the Pleiades stars, which provide clues to the stars' surface termperatures and intrinsic luminosities. Comparing this information with the stars' observed brightness gives their distances. This technique is very imprecise for a single star, but for a cluster, containing stars of different sizes and colors, it permits a good estimate.
Percival and her colleagues had suggested earlier that this technique actually could give the same distance to the Pleiades as Hipparcos, if you assumed the stars were especially poor in "metals," anything that is not hydrogen or helium. Now, however, the team has reversed its decision after extending the study by using infrared radiation. The Hipparcos distance, they say, really is too short.
And Floor van Leeuwen of the University of Cambridge, England, now agrees. A member of the Hipparcos team, he has been looking at the data for the past 4 years with a view to improve the catalog and also to prepare for the next European astrometric mission, GAIA, to be launched in 2011. In June, he presented the outline of his solution to the Venus Transit conference in Preston in the United Kingdom.
In a way, Van Leeuwen explained, the Pleiades partially blinded Hipparcos by presenting the satellite with many bright stars in one of its two "eyes." Hipparcos simultaneously looked in two directions that were 58° apart. This double field of view allowed stars' positions to be measured not only in relation to their neighbors, but also to stars in a totally different part of the sky.
For any one observation, Hipparcos compared the position of each star to every other star being observed. So if the Pleiades were in one field of view, and the other field contained only one or two bright stars, most comparisons were made among the Pleiades themselves: The other field literally didn't count for much. Once astronomers know about this problem, they can rectify it by decreasing the statistical "weight" of areas containing concentrations of bright stars in the calculations. Van Leeuwen expects to finish a revised catalog before the year is out.
Is this the sort of mistake that makes a person kick himself for not having thought about it beforehand? Yes, Van Leeuwen freely admits. "But you must understand that this was a totally new technique, which only very locally failed. The vast majority of data in the catalog is correct. I am now recalculating all the data, taking account of a lot of other, smaller issues that we have discovered."
As a result, the positions in the new Hipparcos catalog will have better than 0.2 milli-arcsecond accuracy (mas) for stars of 4th magnitude and brighter, whereas the original catalog had 0.5 to 1.0 mas. And the GAIA mission, says Van Leeuwen, should get down to 0.02 or even 0.01 mas.