Ryan MacDonald

Let me introduce you to WASP-12b. This flaming ball of super-heated gas 870 light years away is an ultra-hot Jupiter. From its distance and the type of star it orbits, its (equilibrium) temperature should be about 2500 K. But observations of this planet tell a different story.pic.twitter.com/ROJPrM2xzE

Astronomers looking at light passing through the atmosphere of WASP-12b (transmission spectroscopy) found a temperature ~1000 K colder than expected. But it gets weirder.pic.twitter.com/OFLyjmQVE7

We went through the exoplanet literature, and it turns out that *most* exoplanets are much colder than we would expect. Suspecting there might be something more to this pattern, we decided to dig a little deeper.pic.twitter.com/2bgbYemWcC

In our latest accepted paper (https://arxiv.org/abs/2003.11548 ), myself, @astro_jgoyal, and @NikoleKLewis identified a potential explanation for the cold retrieved temperatures of exoplanet atmospheres pic.twitter.com/CkqoONJHl0

Exoplanets are rich 3D objects that can have wildly varying temperatures and composition throughout their atmospheres. But the common way we infer their properties, called atmospheric retrieval, often assumes they are 1D (i.e. uniform spheres).pic.twitter.com/Nwz3saiO33

In our paper, we show that planets with different compositions around their day-night terminator, if subject to a 1D transmission spectrum retrieval, come out far colder than the planet actually is.pic.twitter.com/MIiqMkcHzh

We had fun deriving some pretty scary equations to analytically show why these cold temperatures arise.pic.twitter.com/Q40q3YAAIe

But the equation looks friendly enough if you give it a chance.pic.twitter.com/nfJsMyxvoH

We then generated spectra with non-uniform terminators for a range of exoplanet types, and applied a 1D retrieval code to them. The 1D models do a pretty decent job fitting simulated @NASAHubble observations...pic.twitter.com/AbBTNecSe3

But the measured temperatures can come out waaaaay colder than either side of the planet actually is. It's like these tricky exoplanets have been hiding their true temperatures from us all along.pic.twitter.com/cYuGpn4mO2

This non-uniform terminator effect can also trick 1D retrievals into giving molecular abundance measurements that are > 3 sigma away from their real (terminator-average) value. If this bias is widespread, we need to be careful taking 1D abundances too seriously.pic.twitter.com/1HD4dMIbXs

Note: the biases are worst for ultra-hot Jupiters, but not catastrophic for cooler planets. As a rule of thumb, we should be prepared for 1D retrieved abundances to be not much more accurate than the 2 sigma level. This is true even for H2O abundances.pic.twitter.com/FDNXKGHLPy