Abstract

I will present surprising observational results on the 3D geometry of high-redshift star-forming galaxies from the JWST Cosmic Evolution Early Release Science (CEERS) Survey. Using a differentiable Bayesian model with Hamiltonian Monte Carlo, I will show that there are many more flat, elongated dwarf galaxies than there are round, circular dwarf galaxies seen in projection at high redshift. This puzzle can be explained if ~50-80% of dwarfs at z~2-8 with stellar masses ~10^9-10^10 Msun (including Milky Way progenitors) are not axisymmetric (oblate) disks or spheroids. Instead, they may be significantly flattened along two axes either as prolate ellipsoids (cigars) or as triaxial ellipsoids (oval, surfboard-shaped disks). Both prolate and triaxial ellipsoids trace out a "banana" on the projected b/a-log(a) diagram with an excess of low b/a (edge-on) and deficit of high b/a (round) objects, as observed. Empirical simulations show that the deficit of round dwarfs at high-redshift is real, i.e., we are complete to face-on disks over a reasonable range of sizes and magnitudes. I will argue that imaging alone may be insufficient for distinguishing prolate, oblate and triaxial systems. Instead, deep high-resolution spectroscopy will be needed to constrain the orbits of the stars which are expected to be highly non-circular in these prolate/triaxial candidates. This is suggested by high-resolution cosmological simulations in which the formation of these elongated galaxies is tied to mergers along cosmic web filaments. I will make the case for how JWST, Roman and GMT are poised to definitively probe the nature of this dominant yet enigmatic population of early elongated systems. Finally, I will discuss how this seemingly niche topic bridges together many different areas of astrophysics including Galactic archaeology, dynamics and cosmology, and how it addresses one of NASA's fundamental driving questions: How Did We Get Here?