- Jennifer, I think we can invite Michelle to share her slides.

- Okay.

- Okay. And is that coming all through? Okay.

- Yeah.

- And can you unmute just so we can test your microphone?

- I hear her, Michael.

- Is my mic still not coming through for you?

- I do hear you, Michelle. Michael, can you hear me?

- I also heard.

- Okay, great. And Michelle, I'll give you a warning at 22 minutes and then another one at 24 minutes. So, It would be great if we could leave some time for questions at the end.

- Yeah, absolutely.

- Okay. Terrific. Well with that, I would like to introduce Michelle Kunimoto. Unfortunately, George Ricker couldn't be here today. And so, Michelle is gonna step in and give both halves of what would have been a two person presentation. So please, Michelle.

- Thank you. Yeah. So, I'll be kicking off this talk by giving some broad updates in the world of TESS from 2021. And then, for the second part of the talk, I'll be talking about some of the research projects I've been doing with TESS. So, for those that don't know me, I'm a Postdoctoral Associate working on the TESS mission as the lead of the Quick Look Pipeline team or QLP at MIT. And QLP is dedicated to the discovery and identification of TESS Objects of Interest or TOIs in the Full Frame Image data. I also help manage the TOI core team at MIT. And obviously, I'm just one part of a very large collaboration of people. So, the early updates that I'll be sharing are really the, really thanks to hundreds of researchers and developers and observers that I've had the pleasure to work with on TESS. I'm really honored to speak here today. And this is actually my first ever ExoPAG meeting, so, thank you for having me. So, in the past year, TESS has reached actually a major milestone in achieving its main goal of the Primary Mission, which was to detect 50 planets smaller than Neptune and measure their masses. Now, on the right-hand side, I've just plotted a mass radius diagram with the TESS planets highlighted in orange. And as you can see, below the dotted line, there's actually 67 small TESS planets with measured masses to date. And according to this plot from the TESS Science Support Center showing publications per year, TESS has also inspired an ever-growing number of publications. With 2021 alone, seeing more than 400. And does a testament to the broad appeal of the TESS mission and increasing fraction of these papers are actually non exoplanet fields with about 57% of all publications to date covering other astrophysics. The 2021 publication rate was more than one paper per day on average. And we have every reason to believe that this is going to increase even further into 2022. So, who knows we might reach to a thousand publications in 2022 alone. This past August, TESS also held its second TESS Science Conference which was hosted online due to the pandemic, but it covered a wide range of science topics from exoplanets to extragalactic astronomy. There were actually over 900 registered participants at the second TESS Science Conference, which is more than three times as many as there were at the first TESS Science Conference held just two years earlier. And I think, this goes a long way to show just how much interest in the mission has grown and continues to grow. As for some exoplanet specific highlights, just before the Christmas break, I'm happy to announce that TESS actually passed 5,000 TESS Objects of Interest. And it's been less than a month since then, we've already added another 164 since. After removing the false positives and false alarms from this list of TOIs, TESS is at 4,371 planets and planet candidates, including 1300 that are smaller than Neptune and 161 TOIs have been confirmed. This time last year, TESS was actually at only about, well, not only, but at about 2,400 TOIs. So, that number has grown by double for the last year. And some of you might be wondering what the heck happened right here at around mid 2021, when there is a huge increase in TOIs. And to that, I'll say, stay tuned, I'll be covering that later in this talk. And despite not being initially expected to find many multi-planet systems, because each star, most stars in the sky are observed for only 27 days at a time. TESS has actually found 125 planets and planet candidates in multi-planet systems up till now, including 45 that orbit stars brighter than 10th magnitude. And because TESS observed such a wide variety and diverse set of stars, millions of stars in the sky, the mission continues to find examples of a lot of exotic exoplanets. I was just browsing through the list of publications over the past year alone and was reading papers on circumbinary planets, hot Jupiters with close transiting companions, planets in the Neptune desert, where there should be a dearth of planets and also planets around some of the youngest stars planets have been found around so far. And last but not least, last June, TESS released its official Primary Mission TOI Catalog, which was primarily authored by Natalia Guerrero. This was the first major planet catalog paper from TESS, certainly won't be the last and it's well worth the read for those of you who want to delve deeper into the TOI identification process and also read about some more specific exoplanet highlights from TESS. Now, someone who's entirely focused on exoplanets, it can be easy for me to forget just how wide a reach TESS has in other areas of astrophysics. TESS has been able to contribute to solar system objects, variable stars, and explosive and variable extra galactic sources. And this is just a brief list of all the areas of astrophysics TESS has been able to contribute to. Now, looking to the future, TESS has a proposed plan for a three-year Extended Mission 2 or EM2, which would nominally, if accepted, start in sector 56, which would be around this September and spend three years to sector 96. There's quite a few updates that will happen. The main ones would be that the FFI cadence will improve from the current 10 minutes down to 200 seconds or 3 minutes, 20 seconds. And this is going to further broaden the applicability of TESS data to other areas of astrophysics that really benefit from the improved cadence even further. There's also a goal to make our calibrated FFIs available as soon as 4-5 days after downlink. This is gonna be particularly beneficial for transient observers to have their transient data on hand sooner before their events fade. As for the pointing scenario, which I've just plotted one proposed potential pointing scenario on the right, the tentative plan is going to be to revisit the north and the south ecliptic hemisphere and also, finish up the ecliptic survey. Through the full seven years, TESS should reach a close to a 100%, and not quite a 100%, but close to 100% overall sky coverage. So, this brings me to the next segment of my talk, where I'm going to switch gears and talk about one of the research projects I've worked on for the past year. Given that EM2 is proposed to begin later this year, it's on the horizon. I want to know what we in the exoplanet community, can expect from TESS over the next few years. What are the exponent yields going to look like with EM2? What can we expect? In particular, there were some motivating questions that I wanted to answer. For instance, how many planets should be detectable and what kinds of planets? Are they going to look like different typical planets than we've already found so far. How many of those planets can we expect to be in the habitable zone? Which will be an especially important question as the observing baseline from TESS continues to increase for a lot of stars. How many of these planets are going to be promising follow-up targets? For instance, for radial velocity observations or atmospheric characterization. And also, how well do these predictions actually reflect the TESS exoplanet yield. Now that we have several years worth of TOIs, this can be an important reality check to see if these predictions are matching reality. Now, these kinds of questions have been answered in the past through simulations the TESS exoplanet yield. The first of these was by Sullivan et al 2015. That was kind of a landmark simulation paper that was used, pretty widespread throughout the community. Perhaps, right now, the most commonly used one is from Barclay et al 2018. And my work is going to be drawing a lot of inspiration from these earlier works. But also taking advantage of updating pointed scenarios, newer stellar catalogs, actual TESS data instead of a proposed or simulated TESS data to realistically simulate the TESS observations. And they'll also be able to use updated occurrence rate from the Kepler mission. Very briefly, how these simulations work, is, I'm going to start out with 9.4 million AFGKM stars in the Candidate Target Lists or the CTL, the most recent version of the CTL. The CTL is a small subset of the broader TESS input catalog, which is going to reflect all likely dwarf stars brighter than a TESS magnitude of 13th magnitude, as well as fainter stars that are K dwarfs and M dwarfs that are members of specialty curated lists. I'm then going to simulate planetary systems around each of these stars. Based off of what we know about the planetary abundance from the Kepler mission. Here, it's important to take into account that different types of stars are gonna have different types of planet abundance around them. For instance, the occurrence rates of planets around A dwarf stars is not necessarily gonna be the same as the kinds of stars you'd expect to find around M dwarf, so, I take this into account when simulating planetary systems. I'm also going to be assigning random orbital properties to each planet, as well as physical properties, such as its radius, its orbital period, the orbital eccentricity and the orbital inclination. Now, even through EM2, not every star is going to be observed by TESS. And every star in the sky is going to be observed for a different amount of time, a unique number of sectors. So, given a pointing scenario, I can calculate the full list of sectors observed for each star. Also, well, every star that TESS observes will have observations at the FFI cadence, a small subset of stars every sector will also receive two-minute cadence observations, typically about 15,000 to 20,000 stars per sector. So, using various priorities, I can predict which stars would be more likely to be selected for two-minute cadence observations. Lastly, not every planet that is around a star that TESS observes, is necessarily going to be detected. Perhaps, it's because the planet doesn't transit its star. Perhaps, it does transit. But none of the transits happen to land in the time period when TESS is observing that star. And even if transits are in the TESS lightcurves, it might not necessarily have a high enough signal-to-noise ratio to be detected by a standard planet search pipeline. But this is probably the most complicated part of the simulations, because I'm taking into account a lot of different factors. The cadence of the observations, 30 minutes in the Primary Mission, 10 minutes in the Extended Mission, and 2 minutes for those stars that have two-minute cadence observations. That can affect the signal-to-noise ratio of your transits. Whether a star is observed with two-minute cadence observations, how long each star is observed for, what camera it's in, because different cameras could have, it could be affected by different amounts of scattered light. And this is tracking this for 9.4 million stars throughout seven years of TESS observations. So, after estimating a signal-to-noise ratio for a given planet, I then probabilistically determine which ones get detected or not. The final list of surviving planets essentially reflects a TESS planet catalog similar to the TOI planet catalog. Now, because there's a lot of randomness in these simulations, I repeated these a hundred times, so that I could get a good sense of the spread impossible yields. So, the very first question that I wanted to ask and answer, was how many planets should be detectable in TESS data? My simulations indicate that with the Primary Mission data alone. So, just the first two years of the TESS mission, TESS should actually be sensitive to about 4,700 planets. On the left, is a period radius diagram of those simulated detections for just one example simulation. Planets found only in the FFI data are shown in blue, planets found only in the two-minute cadence observations are in purple, and those found in both types of observations are shown in orange. As you can see, FFI detections are dominated by giant planets. Whereas, the smallest planets and those that long over the periods tend to be found in the two-minute cadence observations, because they really benefit from the improved shorter cadence. One thing I want to point out, is you'll see that these simulations look kind of blocky, especially for those giant planets up there. That's just a consequence of the exoplanet occurrence rates that I use to simulate planets around each star, because these were grid based and they were defined in terms of discrete period radius bins. Through EM1, I'm expecting EM1 one to add another 3,700 planets, bringing the total TESS yields to 8,400 and through EM2, EM2 should add another 4,100 planets, bringing the total TESS yields to 12,500 planets. So, for those of you that are worried that TESS is going to get have some diminishing returns or has exhausted a lot of the exoplanets to be found, have no fear, there's plenty of more planets to find in the future. This also means that the overall TESS discovery rate per year is going to remain well above a thousand planets per year. This is breakdown of new planet detections as a function of planet radius, with Primary Mission detections in blue, EM1 detections in orange and EM2 detections in red. Well, most giants will have been discovered from the Primary Mission alone. The extra data for a lot of these stars is really going to improve the yields of the smallest planets. In particular, I find that the number of planets smaller than Neptune is actually going to double from the end of EM1 to EM2. Part of this is because EM2 is going to be observing stars never before observed by TESS. But the biggest explanation for this is just that the stars that are re-observed will have more data available, which means more transits in the data, which in turn, means higher signal-to-noise ratios for planets and that makes them a lot more detectable. This is a similar breakdown as a function of orbital periods. And as the mission goes on, TESS is going to be sensitive to planets with longer and longer orbital periods, even those beyond the 500 days sensitivity of the Kepler mission. But most planet detections are still gonna be primarily short period planets. This is a breakdown with spectral type. This is actually quite a bit different from previous simulations, where previous simulations predicted that F type stars should be the most common TESS planet hosts. Whereas, I'm predicting that throughout the mission, it's fairly consistent, G dwarf stars are going to be the most common TESS planet hosts. The main difference is just that I've used much more updated Kepler occurrence rates than previous works and that resulted in this update. I also find that M dwarf stars are going to be the stars that benefit the most from the additional observations. So, in particular, M dwarf exoplanet yields will significantly increase in EM2. So, the question I wanted to answer next, was what can we expect for planets in the habitable zone? Which is of course, a very popular topic in the exoplanet community. In particular, what about the yields of terrestrials? So, those will be planets smaller than two times the size of the earth. In this plot, I'm showing the yields for one example simulation compared to the optimistic habitable zone in blue and the conservative habitable zone in green from Kopparapu et al 2013. Planets that are smaller than twice the size of earth are shown as stars at the bottom. But just for some context, up to this point in the mission so, after 45 sectors, my simulations predicted that TESS should find 6 ± 2 terrestrials in the optimistic habitable zone and TESS actually has found 5 TOIs after you remove false positives. Similarly, in the conservative habitable zone, I predicted TESS should find 3 ± 1 planets and it actually found 2 TOIs. So, in this regard, my simulations are very closely matching reality. Now through the end of EM2, I expect about 18 small planets in the optimistic habitable zone and 9 in the conservative habitable zone, all of which will orbit M and K2 stars and this is double the yield that we'll get after the end of EM1. So, there's a lot of really exciting targets to be found yet. If you're more interested in all kinds of planet sizes, I predict the TESS should be sensitive to about 200 planets in the habitable zone in general. The third question was about the expected yield for promising follow-up. And to answer this, I'm going to be defining two broad categories for what's considered a good follow-up target. The first one is going to be to identify targets that are good for radial velocity observations, inspired by TESS's Primary Mission goal of detecting planets smaller than Neptune for mass measurements. So, for every detected planet that was smaller than Neptune, I estimated what it's semi amplitude K might be after assuming a particular mass-radius relation. So, let's say, we consider a good follow-up target as one that orbits a star, right greater than 11th magnitude. So, it won't be too difficult to get a lot of RV observations in a little bit amount of time given the exposures. And let's also say that a planet with a semi amplitude of at least three meters per second is a good target. So, those that satisfy both criteria after EM2, there should be 300 such planets to choose from, which is a plethora of really good targets for follow-up. The second category is what about targets that are gonna be good for atmospheric characterization? And this is especially important, given that JWST successfully launched and TESS will be able to contribute some good follow-up targets for the James Webb Space Telescope. So, to do this, I adopted the transmission spectroscopy metric from Kempton et al in 2018, which essentially quantifies the expected signal-to-noise ratio in transmission spectroscopy. So, they gave some suggested thresholds, past which, you can consider a planet to be a good candidate for atmospheric characterization. And I understand there's a lot of numbers if they're not on the slide here. I won't go too much into detail. And these are small subsets of a much larger table in my full paper. But I do want to emphasize that TESS is really going to identify a lot of attractive targets for both types of follow-up in the future. The last thing I wanted to understand was, how well do these predictions actually reflect the TESS yield? In other words, should we trust these simulations? We have already seen that there's good agreement in habitable zone, but what about much wider types of planets? Now, since TESS is firmly in the first Extended Mission, there's a lot of TOIs that have been released and I'm able to compare my predicted yields with the actual TESS yields. And this is something that no previous simulation work has been able to do so far. And before I present these comparisons, I do want to bring up a few key notes. The first thing is a reminder of my simulations are only looking at AFGKM stars in the CTL and TESS is really finding planets around all kinds of stars that are much, much more than the CTL. So, in order to make it a fair comparison, I'm only going to be comparing to TOIs that were found around the same stars I'm simulating planets around. Similarly, the simulations only go up to 16 earth radius, I'm not simulating planets larger than this. So, I'm only going to be comparing to TOIs that fall within this simulated range. I'm also only gonna be comparing to TOIs from the Primary Mission catalog, since these TOIs were specifically compiled to reflect the TOI yield from a completed mission stage. They've also been known about for a while now. So, a lot of the false positives in that set have already been identified and removed. So, I'm not comparing simulated planets with false positives. Finally, it's worthwhile to remember that TOIs in reality are contributed by many teams, whether officially affiliated with TESS or teams in the broader exoplanet community. Different teams use their own detection criteria, their own detection processes, they look at different types, they look for different types of planets and they look at different types of stars. That means that the TOI identification process is definitely not complete and it's also not straight forward to simulate. Now thankfully, there's a few things about the Primary Mission catalog that makes it somewhat easier to compare to. The first is that the SPOC pipeline at NASA has identified almost every two-minute cadence TOI that has been found so far. That means that when I'm stimulating planet detections with two-minute cadence observations, that should pretty much be a good reflection of TOIs found by the SPOC team. On the other side, the Quick Look Pipeline at MIT has identified almost every TOI from FFIs in the Primary Mission catalog. So, as long as I can stimulate the QLP detection process, that can be a good representation of FFI TOI detections. Now, QLP analyzed all stars brighter than a TESS magnitude of 10.5. So, before I make my comparison, I'm gonna be doing the same cut on my simulated FFI detections. So, after I applied those cuts, I was left with 1,227 Primary Mission TOIs around CGL stars in blue. And that's compared to my predictions of 1,259 ± 58 from my simulations, which are in black.

- [Michael Meyer] Excuse me, Michelle, you're at 22 minutes.

- Okay, awesome. So, these plots show how the yields compare across spectral type planet radius over the period and TESS magnitude and I'm really happy to see that there's great agreement across all dimensions. This gives me a lot of confidence that the simulations are doing a good job of predicting the TESS exoplanet yield. What this also means, is that the Primary Mission TOIs likely represent only about a quarter of all planets that one could have found with the TESS Primary Mission alone. Remember, that my full yield prediction was 4,700 planets and the number of TOIs that I'm predicting is only about 1200. And that's really because a lot of those FFI planets haven't been found yet, because FFI detections have almost primarily been around bright stars. And this last point here is actually one of the motivators behind another of my projects known as the Faint Star Search. This was dedicated to finding those planets around fainter stars. And as a result, I've been able to add 1600 new TOIs from the Primary Mission with my collaborator Thomson Dalan. And that's the explanation for that big batch of new targets, that was what there happened in mid 2021. There's also this little notch here and that's another 412 targets from the TESS Faint Star Search from the Extended Mission 1 search and that's ongoing. I'm sure there's hundreds more to go. So, just as I finish up here, I'm gonna leave you with some takeaways. The first is that more than 8,000 planets should be detectable by the end of the current mission and more than 12,000 should be detectable by the end of EM2. New TESS detections are going to be progressively smaller planets with longer orbital periods, orbiting fainter stars and the TESS number of planets that are smaller than four earth radius should actually double between the ends of EM1 and EM2. So, I know a lot of us are really interested in small planets and this is really exciting to know. Also, the number of planets that are greater than 20 days in orbital period will also double between EM1 and EM2. And lastly, these predictions have shown that there are thousands more TESS planets, just waiting to be discovered even with the data at hand. And that brings me to the end of my presentation. Thank you.

- [Michael Meyer] Great. Thank you very much, Michelle. Virtual and auditory applause are for you. And we have some time for some questions. There are quite a few. So, let me turn to the Q&A tool. The Barkley et al simulations from a few years ago, predicted closer to 10 habitable zone rocky planets to be discovered by the TESS main mission. This far, there has only been TOI 700 D and there appears to be another unpolished. Any idea why the number of TESS discovered habitable zone small planets has been lower than that prediction?

- That's a good question. So, one of the key differences between my work and the works by previous simulations, all previous simulations actually, including the one by Barkley, is in the detection criteria. So, previous TESS simulations have assumed that any planet with a signal-to-noise ratio greater than 7.3 can be detected. In reality, planet search pipelines are not a 100% efficient, even for high signal-to-noise ratio planets. And the detection probability is more like a smooth function with a signal-to-noise ratio. And especially near that detection limit with low SNR, you do not have a 100% efficiency at finding planets. It's more like 50% or lower. So, in my simulations, when I'm probabilistically determining based off the signal-to-noise ratio of a planet, which one is detected, rather than say, it's detected if it has a signal-to-noise ratio greater than 7.3, I actually used the interaction recovery TESS results from the Kepler DR25 pipeline, which is more empirical estimate of a typical Kepler detection probability. Now obviously, this is not gonna be exactly the same as a TESS pipeline detection probability, the SPOC team or the QLP team haven't done that kind of test yet. So, I just assumed, well, I'll use Kepler for now and under the assumption that it's gotta at least be better than just assuming all planets above 7.3 are detected. So, when you do that, those kinds of low SNR planets, the like planets in the habitable zone decrease by quite a bit.

- All right, thank you. Is the large proportional increase in M dwarf planets in EM2 due to the increased cadence?

- It's not too much due to the increased cadence. The increased cadence does help a very small fraction of planets, only those that have extremely short transliterations on the order of 15 minutes and that's not too many of the ones that are in this in the simulations. The big increase in M dwarfs is primarily because M dwarfs tend to be fainter stars and fainter stars have worst precision. Precision is gonna get, it's gonna get noisier as you get to fainter and fainter magnitudes. So, that's gonna affect the signal-to-noise ratio of your candidates and that's why getting three observations of these types of planets is gonna be really important.

- And will upcoming data releases from TESS revisit the Primary Mission data or should we consider the archive data as final?

- Sorry, can you say that again?

- Will the upcoming data releases revisit Primary Mission data that's already been released or should we consider those archived data as final?

- It's definitely not final. An example of Post-Primary Mission data release was the Faint Star Search that I mentioned. So, the total number of TOIs in official Primary Mission catalog was about 2240, something around there, and the Faint Star Search added 1600. I intend to revisit that once I have an improved vetting pipeline that's gonna be able to do it. I want to be able to reprocess a lot of the primary mission data, essentially, and see if there's even more planets that might've been missed. I'm sure that I'm not the only one. I know there are independent groups that are looking at Primary Mission data for young planets or, you know, those in specific clusters. There's certainly going to be more TOIs to come from the Primary Mission.

- And finally, I'll ask one more. Do you see any areas where the planet occurrence rates from Kepler deviate from the TESS planet yields?

- Yes. So, one of the main areas is in the hot Neptune desert. So, Kepler found close to zero depending on your false positive rate. Essentially, it was, you know, why we thought that there was this total dearth of planets and depending on what type of star you're looking at, there were no Neptunes in the hot Neptune desert from Kepler. TESS has already found that that's not necessarily the case that it's empty of planets. And there have been confirmed TOIs, even those that have gone through atmospheric characterization in that presume. So, if there was, for instance, a more TESS specific occurrence rate, then that would help to improve these simulations. There's other areas like giant planets around M dwarfs. So, the Dressing and Charbonneau occurence rate that I used for M dwarfs, only went up to 4 earth radius and we've already found examples of TOIs, several TOIs that have radii larger than those. So, M dwarfs could actually host giant planets as well. And that's not something that Kepler really was able to contribute to because of the really small sample of M dwarfs. So, there's a lot of areas where these simulations, obviously that's a shortcoming given it's based entirely on Kepler and one of the amazing things about TESS is just because it's looking at such a large sample of stars. There's gonna be so many more unique planets that we've never saw before coming out of the TESS mission.

- All right. Thank you, Michelle. There's a few more questions on the Q&A tool that would benefit from you having a look and perhaps we can post some answers there. Let's give Michelle another round of applause.