Abstract: The first color images from the James Webb Space Telescope (JWST, Gardner et al. 2023) were designed with the understanding that a NASA/ESA/CSA flagship must explain its purpose and science to the tax- paying public, and that to earn the opportunity to explain and educate, we must first inspire. The feeling that many people have when seeing a well-composed color image of a deep space scene for the first time is visceral. It connects to a deeper part of the psyche than that of pure reason, like the effect of music or other visual arts. It is in eliciting this emotion that we have an opportunity to draw people in and educate; inevitably questions follow: What do I see? What do the colors and shapes mean? How large, distant, or old is this? And how does it relate to us, the Earth, the Sun, and my environment? It is in answering these questions we can educate, encourage children to take up careers in STEM, and hopefully convince the public that science is a fundamental human endeavor.
JWST Early Release Observations
Early Release Observations (or EROs, not to be confused with the Early Release Science – ERS – programs) are a common activity at the end of commissioning an astronomical observatory. EROs are intended to provide a first, dramatic, look at science-like data from an observatory, but with the public as the primary audience. Clearly, we owe the existence of JWST to the decades-long efforts of thousands of dedicated engineers and scientists at immense personal and monetary cost. Given an observatory that outperformed expectations (Rigby et al. 2023), the ERO team was keenly aware of their responsibility in the eyes of the public, lawmakers, scientists, and the JWST project. They should clearly demonstrate that the observatory works technically well, but should also meet a spectacular standard, capable of inspiration and awe.
For JWST, the EROs were the culmination of a years-long process by a small advisory committee, coupled with an implementation team at the Space Telescope Science Institute. Targets were kept strictly confidential, to not set expectations for any specifics. Indeed, the final EROs were not known until close to the end of JWST commissioning, as the exact timing of the observations determined which targets would be visible to the observatory. The process for selecting ERO targets is described in detail in Pontoppidan et al. (2022), including the implementation team. During the final 6 weeks of JWST commissioning, a team of more than 30 experts at STScI observed, processed, and delivered the JWST EROs (Figure 1). Despite the very short amount of time, the team knew that nothing would be good enough and worked non-stop, through weekends, through the night at times, to meet this standard. Ultimately, the first ERO product (a deep NIRCam image of the SMACS 0723 galaxy cluster) was presented by US President Biden on July 11, 2022. The JWST EROs resulted in 200 billion impressions worldwide (meaning that every person on earth had many opportunities to view them).
Requirements of cosmic art
There is something special about the imagery produced by Hubble. The most famous Hubble images seem more real than most astronomical images. They pioneered many techniques in modern astrophotography (Levay 2021). You get the feeling that, if you were there in person, the image would be what you might see with your eyes. There is a photorealistic quality to them. Of course, this is not the case. The collecting areas of our human eyes – the pupils – are even in the darkest places only 1 cm2 or 250,000 smaller than that of JWST. Under these circumstances, a person would only be able to see a ghostly, grey scene, if anything at all. But this does not matter – outreach images provide us with a superhuman capability to view the invisible.
First, some of the unique quality of a Hubble or JWST images comes from high signal-to-noise (S/N) in every pixel. For JWST public imaging, we required S/N>20 per pixel per filter, and set exposure times well beyond what is typically used to support science investigations of the same scene (typically 30 minutes per filter). The deep field on SMACS0723 remains one of the deepest images obtained with JWST with several hours per filter. The NIRCam outreach images use a custom dither pattern to ensure as uniform depth as possible, with a minimum of 5 dithers per pixel, implemented as a mosaic pattern with a 71.5% overlap (Pontoppidan et al. 2022).
Second, the images appear extraordinarily sharp, with great detail and structure, and the ability to support exploration. One may think of a “Find Waldo” or “Busytown” image – you can zoom in and find something special happening in the image on multiple scales. It turns out that scales of least 1:1e3-1:1e4 are needed. That is, the overall motif is >1000 times larger than the smallest resolved structure. Conversely, viewers of images with much larger scale range than this (better than 1:1e5) are not able to appreciate the smallest scales, and individual stars become invisible. An important point is that this requirement is not directly related to the absolute angular resolution of the telescope. Rather, it sets a requirement of the total angular size of the image, given a spatial resolution. Thus, there is no reason that a smaller telescope cannot produce spectacular outreach images, only that there must be at least 1000 meaningful resolution elements across the image. That said, a large telescope with a high angular resolution has more opportunities for excellent photographic compositions in the sky.
Photographic composition
A central requirement for composition of outreach images is a well-defined object or motif. Examples include an entire galaxy/protostellar outflow, compact HII region, all of which would include a well-defined, mostly empty, outer boundary. Examples of these include Stephan’s Quintet or the Oph core (Figure 3). Alternative motifs can be designed via a strong leading line, such as the Cosmic Cliffs (Figure 2). We avoided a section of a cloud without a dark boundary, or leading line, or empty deep fields, which are mutually indistinguishable to the non-expert.
We used best practices from photographic art to design compositions for outreach images. These include the use of leading lines, rule-of-three (collections of three items is more aesthetically pleasing than two or four), brightness/color contrast, golden ratio, as well as the S/N and scale requirements discussed above. For instance, the famous Cosmic Cliffs in Carina uses an obvious leading line across the “mountainous” crest, placed at the golden ratio (the ratio satisfying the relation (a+b)/a=a/b). This composition (Figure 2) also uses NIRCam filters designed to create a strong color contrast between the cloud (PAH emission) and the photodissociation flow (Paschen Alpha). Finally, outreach images are nearly always designed to be rectangular, also when composed of multiple mosaic tiles.
Optical versus infrared outreach images
Some use the term “False Color” to apply to infrared color images, or indeed color images from any part of the electromagnetic spectrum. This is an unfortunate misnomer that should be avoided (Hurt 2013). Granted, JWST operates in the infrared and is therefore sensitive to photons that are not visible to the human eye. However, this does not make the colors “false”, as opposed to real. Color is defined as a ratio in the energy received from photons of one wavelength relative to another. If most energy is represented by photons with wavelengths near 550 nanometers, our eyes generally interpret that as a green color. There is no difference in representing differences in energy received in different infrared wavelengths as color. In fact, there are many more colors available in the infrared than in the visible; the light available to JWST is over 20 times more colorful than what can be perceived by humans! To call this false color would be a great injustice to Nature, as if we had just painted the images arbitrarily. Rather, we have translated the infrared spectrum into a visible one, like one might translate from one human language to another. To aid in this translation, rules are applied, the most important of which is chromatic ordering. In chromatic ordering, increasing infrared wavelengths are assigned increasing colors representing increasing visible wavelengths.
Another less strictly applied principle, but also important, is to limit the total spectral range from the bluest to the reddest filter. There are several reasons for this: One is that, since the spatial resolution of a space telescope generally scales inversely with wavelength, the reddest band will appear blurrier than the bluest band. If the range is more than 2-3 times in wavelength, this will affect the sense of “photorealism” seen in the best Hubble and JWST color images and can result in red halos around stars and sharp features. Further, a wider wavelength separation of bands also generally means that the bluest bands see very different physical processes than red bands. In this case, the spatial structure seen in each band may not match, leading to clumps of primary color (red, green, blue), rather than smooth color blends. This also degrades photorealism.
Defining the first year of JWST outreach imagery
JWST ERO imagery were designed and obtained with the purpose of creating spectacular color images. These are clearly separate from competed science programs, which are designed to accomplish science goals in the minimum amount of observing time, with essentially no consideration for aesthetics. That is, the outreach quality of color images from science programs is incidental, and rarely, if ever, optimal for this purpose. Consequently, the STScI Director allocated additional observing time during Cycle 1 to obtain additional outreach images, following the same style as that used for the EROs. This resulted in a new set of observations, including the L1527 protostellar outflow, the JWST version of the famous Pillars of Creation, the first anniversary image of the nearest young star-forming region in Ophiuchus, and not least the ice giants, Uranus and Neptune (see Table 1). These last set of images were designed to do something new: to image a planet in the solar system on a background of the deep Universe, creating a composition of our own back yard reaching out to the beginning of time. This is a challenge as there is a vast difference between the brightness of the ice giants and distant galaxies, and because the planets and its moons move relative to the sidereal sky, as well as relative to each other. With JWST, it is possible to see the planets rotate and the moons to move in minutes. To solve these issues, the ice giant observations were designed to include multiple images with different exposure times. A short exposure to capture the bright planetary disk, and a much deeper exposure to image rings and the background galaxy field. By registering each component separately, it was possible to create the composite image seen in Figure 4, striking in its contrast between the blue, icy moons and the ancient red universe beyond.
JWST has now transitioned into the routine of normal science operations, but there are surely many more spectacular images in the future.
Acknowledgements
A portion of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004).
References
Gardner, J.P. et al. 2023, PASP, 135, 068001. doi:10.1088/1538-3873/acd1b5.
Levay, Z. 2021, Space Science and Public Engagement: 21st Century Perspectives and Opportunities, 67. doi:10.1016/B978-0-12-817390-9.00010-5.
Pontoppidan, K.M. et al. 2022, ApJL, 936, L14. doi:10.3847/2041-8213/ac8a4e.
Rigby, J. et al. 2023, PASP, 135, 048001. doi:10.1088/1538-3873/acb293.