IDL Plays Key Role in Hubble Space Telescope Servicing Mission

In 1990, the Hubble Space Telescope (HST) was placed in orbit and pointed at deep space. Astronomers around the world anxiously waited for images that might bring new understanding about the beginning of the universe, the nature of distant galaxies and the properties of celestial phenomena. However, instead of brilliant images unaffected by the inherent distortion of the Earth's atmosphere, the HST delivered fuzzy, blurred pictures that revealed little. Investigation into the HST's construction soon revealed that the reflective lens had been ground incorrectly and that astronauts would have to service the HST to rescue the multi-billion dollar telescope.

Two critical factors helped achieve the success of the servicing mission: tenacious, innovative scientists and technicians who worked quickly and within extremely exact tolerances and precision tools that provided the utmost quality. Paul Scowen, from Arizona State University (ASU), was selected to be an associate member of the Investigation Definition Team because of his exceptional CCD data manipulation skills and interest in star formation and nebulae. The ease of writing and sharing applications, compiling data and visualizing results compelled Scowen and the team to use IDL, the Interactive Data Language, as the software language for testing, calibrating and analyzing data from the Wide Field/Planetary Camera-2 (WF/PC-2). IDL provided a "powerful, flexible language that is easy to use and customize," says Scowen. "As a matter of fact, a colleague of mine has been known to say that with IDL you can be productively working on your data after only a couple days' exposure to the environment."

IDL Indispensible for Critical Testing

At the Jet Propulsion Laboratory (JPL) in Pasadena, CA the Investigation Definition Team assembled. Extensive testing was conducted to learn how the WF/PC-2 performed and to make the first calibrations. "Data came in fast and furious and it had been anticipated that a powerful set of analysis tools was going to be needed for the job," said Scowen. "Among other things, I had put together a GUI that incorporated numerous IDL applications in a simple, point-and-click mode. As soon as the data was transferred from the readout computer, the resident analyst would check the image. IDL was indispensible since it offered fast, accurate assessment and revealed any anomalies that turned up."

Scowen's research group at ASU developed and collected a suite of routines (more than 22,000 lines of code) for interpreting HST data on Digital Alpha workstations running OSF/1 and has ported the applications to Digital 5000s running Ultrix. By using IDL's journal mode during development, the team was able to work interactively with real data until they achieved the desired results. "Then it was a simple matter to edit the journal file and produce in very short order a working, proven routine that did not need debugging," says Scowen. Scowen estimates they saved "more than a year of labor over the course of three years by adopting IDL over other alternatives." And, IDL's transparent cross-platform development architecture enabled the team to efficiently share their work.

"Being able to take procedures written on one machine and directly transplant them to another at the drop of a hat was particularly advantageous," Scowen says.

IDL GUIs Speed Task Implementation

"We made heavy use of IDL's GUI library. IDL is one of the best environments I've seen to design GUIs quickly and easily," says Scowen. The functions that the team most frequently used in the test were accessed through GUIs, but command line input also was available. By combining these modes, the ASU group has come up with a suite of more than 200 applications that makes repetitive tasks easy to implement but remains flexible. Users can employ functions from the command line, such as applying masking or colortable enhancement, and immediately see the results.

IDL Data Reduction and Processing for Clear, Accurate Imaging

"Many of IDL's array-handling logic routines are invaluable," Scowen says. "For example, the ability to multiply two arrays in memory as fast as IDL does cannot be matched by other software." IDL's WHERE function is frequently used to define masks in both pixel and data-number space. "For very large arrays we find that the ASSOC I/O offered by IDL is very convenient and allows some daunting tasks to be done in short order; other interfaces would be choking with the same physical resources," says Scowen.

"We can produce a stunning, three-color picture in just minutes using IDL's twenty-four bit mode," explains Scowen. "The team is sensitive to maintaining the original integrity of the imagery and makes an active attempt not to modify the content of any data," he stresses, explaining that processing and presenting the images in full color doesn't alter the true picture of the object.

Making Sense of What We See

Scowen's group at ASU has focused their study on two cosmic regions. The first is the Eagle Nebula (M16), a vast cloud of ionized gas called an HII region (because the region is comprised mostly of ionized hydrogen). By pointing HST at M16, the group has been able to make new observations about the process of star formation. According to Scowen, "M16 is fairly unique in that it has three fingers of cold, dark, dense gas that protrude into the main HII region, offering a very clear view. A better understanding of the interface between the ionized volume of an HII region and the surrounding interstellar medium will tell us much about how HII regions affect star formation in general across galaxies."

Learning about this relationship was the original goal of the M16 project. However, after studying the images, Scowen and his colleagues learned "something more fascinating: the first stars formed in the region affect how new stars are being formed nearby. This is the first direct observational evidence that primarily-formed stars, the ones responsible for making the HII region in the first place, are directly affecting how the next generation of stars develop by altering the environment in which they are forming."

The second project is a study of the Crab Nebula, which was first observed in 1054 A.D. by Chinese astronomers and is thought to be the remnant of a supernova. Using the WF/PC-2, the group at ASU took time-lapse sequences of this nebula. They were surprised to find that the nebula is much more active than originally thought. As a matter of fact, the Crab Nebula is now considered to be the most dynamic of any celestial object observed to date outside our solar system. To find out more about the Hubble Space Telescope and Paul Scowen's research, visit http://tycho.la.asu.edu/scowen.html.