Tangled in the Tarantula Nebula

Breathless… that is the word that best encapsulates my state of mind during JWST’s Christmas Day 2021 launch, when we saw humanity’s last view of the telescope as it slowly separated from the upper stage rocket, extending its solar panel as if to wave goodbye.

JWST separates from the Arianne 5 rocket. Credit: ESA/Arianespace

This was a recurring sensation throughout JWST’s commissioning activities: extending and tensioning the five layers of the sunshield; lowering the secondary mirror; unfolding the primary mirror; focusing the telescope; and now, as the first color images come together on my screen, I am once again left breathless. I feel extremely fortunate not only to have been a part of the committee that chose these first targets the telescope would observe, but also to be assembling the data from those first observations into JWST’s first full-color views of the cosmos! The hidden universe has become somewhat less hidden recently, and will only reveal more of its secrets as JWST begins systematically exploring the infrared universe.

JWST’s First Observations

Of course, we knew that these first observations would be impactful. These objects were chosen specifically to demonstrate the power of the telescope’s observational and spectroscopic capabilities; we knew they would be both scientifically intriguing, as well as visually stunning. Even so, I am still left breathless, as I stare at my screen and see a complex tapestry of gas and dust come alive in vivid color in front of my eyes! I actually had to sit back and take it all in, staring in amazement that the moment had finally come! At this point, we’ve already seen the bulk of JWST’s first observations. We intentionally chose to observe more targets than we would release at the end of commissioning. This choice gave us some flexibility and redundancy in terms of covering all of the observatory’s main science goals in that first press release package, giving us backup options if there were any issues with a particular observation. Thankfully, there were none, and so now we’re able to release our “extra” targets as well! In today’s post, I would like to focus the discussion on the processing of our most recent image release of the complex star-forming region known as 30 Doradus in our satellite galaxy, the Large Magellanic Cloud (LMC).

Astronomers used the ESO 1-meter Schmidt telescope to capture this image of the nearby Large Magellanic Cloud (LMC) galaxy in 1986. The bright pink region of active star formation, 30 Doradus, sits near the middle left of this image. Credit: ESO

The Early Release Observations (ERO) program was planned and implemented by the ERO Production Team; a group of about 30 scientists, engineers, writers, and two image-processing specialists (Alyssa Pagan and myself), who met nearly every morning from the time these first observations began in late May up until the release of JWST’s first images on July 12th, 2022.

The first observation executed as part JWST’s ERO program was actually 30 Dor. After waiting decades for the telescope to launch, and making it through the excruciating excitement of deployment and initial commissioning activities, there was a palpable sense of nervous anticipation as our first observations executed. Seeing the very first frames transmitted by the telescope of the regions around 30 Dor, we knew we had something very special on our hands, and I could not wait to get to work on combining the data taken in multiple filters into JWST’s first full-color image!

A Tapestry of Star Birth and Death

One of the largest mosaics in the EROs program, 30 Dor was second only to Stephan’s Quintet. The pre-processing, calibration and alignment of the mosaic frames presented some challenges. Anton Koekemoer, an astronomer at STScI and instrument specialist for the NIRCam instrument, ran the raw data through JWST’s processing and calibration pipeline, meticulously adjusting settings where necessary, and aligning all frames to reference stars provided by the Gaia star catalog. Alignment between different frames of the mosaic presented the largest challenge in these first processing steps.

Here is an example of alignment offsets near the core of 30 Doradus which were a challenge in the early stages of processing. This image comes from JWST’s NIRCam instrument using the F090W filter. Credit: NASA/STScI/J. DePasquale

After three iterations of processing, we were satisfied that Anton had produced the cleanest possible version of the raw images. Given the time constraints of the EROs program, we decided to handle the remaining alignment issues and image artifacts in the image processing stage. The slight misalignments in the image were handled through use of the “Puppet Warp” tool in Photoshop. This tool allows you to warp and distort an image on small scales, in localized spaces to ensure that individual stars line up between the different color channels—essentially performing the calibration step of removing geometric distortions by hand.

This short animation demonstrates the subtle effect of using Photoshop’s “Puppet Warp” tool to align stars in a small section of the image. Credit: NASA/STScI/J. DePasquale

Assembly of a Color Image

JWST is, of course, an infrared observatory, meaning that all of the light it is sensitive to is beyond what the human eye is capable of detecting. However, the process of applying color to these images is remarkably similar to the approach used with Hubble Space Telescope (HST) and other astronomical data taken in the visible spectrum. The human eye contains photoreceptive cells sensitive to light at wavelengths corresponding roughly to the colors red, green, and blue. All of the colors that we can see are composed of those primary colors and any digital image that you view on a screen can be decomposed into red, green and blue color channels.

We can take advantage of this evolutionary quirk of vision to translate light from wavelengths our eyes are not sensitive to into the visible spectrum. The process is akin to shifting audio from inaudible frequencies into frequencies that we can hear. We cannot hear a dog whistle, but if you record it with a device that can capture it, and then shift it down a few octaves, we can hear it!

We chose to observe 30 Doradus in several filters with the NIRCam instrument on JWST covering a range of wavelengths from near infrared at 0.9 micron out to 4.4 micron. Not all of the data were used in the final composite as some of the narrow-band images did not contribute additional information to the final image. The image is composed of three wide-band filters (0.9, 2.0, and 4.4 micron) as well as a continuum subtracted medium-band filter at 3.35 micron. Continuum subtraction is a technique astronomers use when processing images to isolate a very specific wavelength of light. In this case, we are looking for the light of polycyclic aromatic hydrocarbons (PAHs – essentially warm dust except the dust grains are extremely tiny hydrocarbon rings) within the region of 30 Dor.

An image showing the electromagnetic spectrum from Gamma rays to Radio waves. Note JWST’s focus on infrared light. Credit: NASA/STScI/J. Olmsted
Here is a look at how continuum subtraction is used to produce an image that isolates very specific features in the nebula. Credit: NASA/STScI/J. DePasquale

The data were first “stretched” using the free software FITSLiberator. Stretching is the process of rescaling the dynamic range of an image so that we can reveal details hidden within the data. Astronomical data typically contains a much larger dynamic range (or range of brightness values) than can be displayed on a screen. A mathematical function is used to increase the brightness of the darkest pixels, while maintaining details within brighter pixels in the image.

An example using FITS Liberator to stretch the pixel values of the F090W filter data from NIRCam. This image was created with the help of the NOIRLab/IPAC/ESA/STScI/CfA FITS Liberator

After successfully stretching the data, color is applied chromatically such that the shortest wavelength 0.9 micron image is assigned blue, the 2.0 micron data is assigned green, and the 4.4 micron data is assigned red. The 3.35 micron image is then overlaid in a reddish orange color to highlight the light of hydrogen, but this step comes later in the processing. Assembling the color image from these data in this way gives us the starting point for further image processing.

Initial assembly of the color composite image. Credit: NASA/STScI/J. DePasquale

As we’ve often stated in previous posts, this process becomes more subjective at this point, following the visual principles of photography as we work to balance the color of the image, and bring out details that may be hiding the often “flat” and “dull” appearance of the initial color composite. There is an enormous amount of visual depth and dimension contained within these data and it is our job to present it in the best possible light while maintaining the integrity of the original data.

Our final color composite of 30 Doradus gives us a new and unique view into a familiar region of space. JWST, observing in infrared light, is able to reveal new details that were not available to us with HST observing in optical light.

This animation compares views of the core of 30 Doradus from Hubble to Webb’s NIRCam and MIRI instruments. Credit: NASA/STScI/A. Pagan

That is not to say that HST is no longer useful in the JWST era. I like to think of this using the analogy of diagnostic medical imaging. X-rays are useful for peering into the human body to see bones, but we also have sophisticated MRI machines which rely on an entirely different kind of light to provide another perspective on the inner workings of the human body. Both are essential to understanding our inner space, just as both JWST and HST are essential to our understanding of outer space!


Leave a Reply

Create a website or blog at WordPress.com

Up ↑