Jay Anderson: Astrophysicist, Observatory Scientist, & Astrometry Specialist

As an astrophysicist and observatory scientist at the Space Telescope Science Institute, Jay Anderson spends his time pushing instruments past what they were designed to do. His work on point spread function modeling, charge transfer efficiency correction, and astrometric calibration makes high-precision measurements possible on the Hubble Space Telescope and the James Webb Space Telescope. In this conversation, he turns to the structures most people never notice — the signatures, motions, and detector behavior that quietly decide what the science actually means.

---

When you’re modeling a point-spread function (PSF), most of the work lives in structure no human eye can see. What’s the part of that process that actually drives the result but stays invisible to everyone else?

PSF is the impression a point of light leaves on a detector. It has to be smooth because nothing in the image can be sharper than the telescope allows. The real fight is sharpness control. You need the model sharp enough to match how light actually falls on the pixels, but not so sharp that it invents structure the optics can’t produce. That tension is what decides whether the PSF holds up.

You work with undersampled images where the signal hides inside the pixel grid. What’s the kind of distortion or bias that still fools people even after years of looking at HST data?

Photons aren’t the only things that hit detectors. Cosmic rays are energetic particles that can strike the detector directly, without passing through the optics, and leave signal in a small cluster of pixels. These features can be sharper than the PSF, and since PSFs can look very sharp, people often mistake cosmic-ray hits for stars. With a solid PSF model, it’s easy to show a detection can’t be a star because it’s simply too sharp.

Your black hole work relies on motion that’s technically “too small to notice.” What’s the line between a real astrometric signature and a ghost that comes from the instrument?

It’s common in astronomy that the thing you want is buried under something brighter. Einstein predicted light would bend in a gravitational field, and Eddington confirmed it by waiting for an eclipse so he could measure a star’s tiny shift next to the Sun. Kailash Sahu used that same position–shift idea when a star brightened in a way that hinted at an unseen object focusing its light. In our case, a bright nearby star sat right on top of the event. With an accurate PSF model, we could strip out that star’s influence and measure the real shift cleanly.

Globular clusters look simple from far out. Up close, they’re a mess of populations and histories. What’s something you can see in their internal dynamics that most astronomers never realize is there?

The stars in a globular cluster are like bees around a hive. Each one follows its own orbit, but those orbits respond to the combined gravity of every other star. They almost never collide, but they constantly trade energy. We used to think all the stars formed together. New data shows multiple generations, with a second one likely forming from leftover gas near the center. You’d expect that second generation to stay concentrated, but energy exchange flings them outward fast, so the two populations look mixed. The giveaway is motion. Second–generation stars tend to move on more radial orbits, and that kinematic imprint is much harder for gravity to erase.

Calibration sounds procedural but it shapes everything downstream. What’s the invisible step or assumption that, if removed, would break half the science people think they’re doing with HST or JWST?

Space detectors don’t collect photons themselves. Photons always move at the speed of light, so what we actually collect are the electrons they generate through the photo-electric effect. Those electrons sit in pixels and get measured by complex electronic circuits. Because an electronic device is measuring electrons, the detector inevitably leaves its own imprint on the signal. One of the first steps in the pipeline is removing that detector signature. Before the correction, only the brightest sources show up. After it, even extremely faint structure becomes visible. Take that step out and a huge fraction of HST or JWST science falls apart.

---

When Jay talks about measurement, he isn’t describing a picture of the sky. He’s describing the machinery behind it. the corrections, motions, and constraints that hold the data together. For him, precision isn’t decoration. It’s the only way to see what’s actually there.

Learn more about Jay at:

• STScI

• Data Archive

• Publications