Had a great time at CosmoVerse in Istanbul. This is me on a panel with a couple random call-ins 😀

Amazing new CMB results, congrats SPT!.

It begins.

RT @scifri: Is the universe expanding too fast for… physics? Two cosmologists tell us how they measure the universe, and what happens when….

RT @StartsWithABang: Is the Hubble tension a real problem?. #AskEthan. Famed astronomer Wendy Freedman claims (including in an interview wi….

Going to be a guest on NPR @scifri today during the 3pm slot with Wendy Freedman to talk about the latest excitement in cosmology. I hope it’s fun!.

Is it doable? I think yes. I always think when statistical floor improves, people find ways to push down the systematic floor too. But man it's going to be hard. (end).

Lauren Aldoroty's paper looks at simulated images and can already see getting photometry to the couple mmag level will be really hard.

The net effect, imho, will be a supernova survey with amazing statistics, amazing redshift range, and the possibility of reaching amazing constraints on dark energy. But as the Rubin et al. paper shows (fig here), things go bad quickly if photometric calibration isnt nailed.

So how can Roman nail this? Basically it needs to get to 1 mmag for calibration, which has never been done before. Our Roman supernova Project Infrastructure Team (SN PIT) just put out four papers trying to pave a path to get there: Finding the optimal survey.

The reason is twofold. The first is unlike distance ladder where calibration effects can 'cancel out' due to similar redshift, for dark energy, don't have that, and there is a factor of 3x due to the sensitivity to color/dust. Then, what Brodie shows, is that one has to train a.

A paper out tonight (led by my former grad student) Brodie Popovic did his own big sweep of photometric calibration of past datasets. Overall, he found really good agreement in terms of calibration, but also saw that ~small 5 mmag effects can propagate across a large redshift to.

One thing I always say is that measuring the Hubble constant with supernovae is a lot easier than dark energy. The reason is that for H0, comparing SNe at z=0.005 to z=0.05 (same surveys, little evolution, redshift effects) and Hubble Tension signal is ~0.18 mag. For evolving.

A number of new supernova papers, both for current datasets and @NASARoman, out on arXiv last few nights. Interesting to see limitations of current datasets on measuring dark energy, and how Roman can help. A 🧵.

One last thing. We are in an awesome moment in cosmology–a `crisis’ for us really means we are about to learn something new– at the same time, telescopes that people have been building for decades coming online this year and next. Let’s not destroy U.S. astronomy funding now.

So where are we now? We have two new big, JWST studies of ‘local’ H0, from TDCOSMO and from TRGB-SBF, different teams using different tools than SH0ES (which are Cepheids+Supernovae, already >5 sigma above CMB+LCDM), and both with high H0 (together they are >3 sigma above.

Since their lenses are at moderate redshift, they need to know about changes in the recent expansion rate to translate H0 to z~0, so they use some different options. Here are their results. You can see TDCOSMO itself is 74-75, but big errors on OmM. If combined with Pantheon+

So TDCOSMO literally spent the last six years getting new data from JWST and Keck and VLT to measure the mass profiles of their specific lens galaxies and where did they end up? Pretty close to where they started! with a similar value of H0, albeit with a more.

About ~6 years ago, TDCOSMO (the previous version of team was called H0LICOW) was finding H0 around 73-74 and tight uncertainties (~2). This combined with SH0ES measurements really accelerated the ‘Hubble Tension’ story. But then TDCOSMO took a big pause–they were making pretty.

TDCOSMO measures the expansion rate of the universe called the Hubble constant H0 by analyzing time delays between multiple images of strongly lensed quasars (their 8 here), where the light paths are bent by a foreground galaxy’s mass. By modeling the lens mass distribution and