Feature story: Physicists at Lawrence Livermore National Laboratory, led by Marius Millot, used the Omega Laser Facility at the University of Rochester to shock-compress diamond samples to about 1 terapascal — roughly three times Earth's core pressure — and temperatures hotter than the Sun's surface, recreating conditions found deep inside Neptune and Uranus. The results resolve a 20-year, ~1,000-degree disagreement between lab measurements and quantum simulations over diamond's actual melting point, confirming the simulations were right. The team also found diamond stays in its normal crystal structure right up until it melts — no intermediate phase — and confirmed solid diamond floats in liquid carbon, the same basic physics as ice floating on water. Published in Nature Physics, the corrected melting-point data could help triple energy gain in inertial confinement fusion reactors by allowing gentler, more efficient laser compression of diamond-shelled fuel capsules.
The rest of the news:
- Roman Space Telescope's strange origin: three days out from its Sunday, August 30 launch (7:26am ET, Falcon Heavy), we trace how Roman's core optics began life inside the National Reconnaissance Office's canceled "Future Imagery Architecture" spy-satellite program, donated to NASA in 2012 after the program's spectacular 2005 collapse. The telescope is currently being mated to its Falcon Heavy at LC-39A, with a Launch Readiness Review Friday, August 28.
- NOAA storm watch: the Space Weather Prediction Center has issued an official G2 (moderate) geomagnetic storm watch for Friday, August 28, following Tuesday's M6.9 solar flare — aurora chances improve for northern-tier US states, the UK and similar latitudes; minor storming isn't expected to reach much past Tasmania locally.
- AI solar storm detection: NJIT researchers have built a Transformer-based AI model, EarlyDetect, that spots hidden precursor signals of solar active regions forming roughly 9.24 hours before they're visible — not yet ready for real-time forecasting, but a promising extension of the warning window.
- SpaceX Starbase Louisiana: a $100 billion, five-complex, ten-pad second Starbase announced for Vermilion Parish, Louisiana, alongside Governor Jeff Landry — construction targeted for 2027, first launch aimed at 2029.
- Tonight's Sky: a 96%-partial lunar eclipse peaks at 4:13 UTC / 12:13am ET August 28 — spectacular from the Americas, broad daylight in Sydney (~2:13pm AEST). Venus and Saturn remain the reliable local targets.
Links & sources:
- ScienceDaily — Scientists crushed diamond beyond Neptune-like pressures — and solved a 20-year mystery
- Gizmodo — Scientists Recreate the Melting "Diamond Rain" of Neptune and Uranus. It May Help Fusion Power
- Space.com — From spy satellite to space telescope: the unlikely origins of NASA's Roman Space Telescope
- Space.com — Nancy Grace Roman Telescope live updates: NASA readies Roman for launch
- Watchers.news — M6.9 solar flare produces Earth-directed CME, G2 geomagnetic storm watch issued for August 28
- Universe Today — AI Spots Hidden Solar Storm Signs 9 Hours Early
- EurekAlert / NJIT — New AI model detects hidden signs of solar eruptions hours before they emerge
- Space.com — Starbase Louisiana: SpaceX announces enormous $100 billion Starbase launch site
- SpaceNews — SpaceX to develop Starship launch site in Louisiana
- EarthSky — Partial lunar eclipse of the August 27-28, 2026
Follow us: @AstroDailyPod
Become a supporter of this podcast: https://www.spreaker.com/podcast/astronomy-daily-latest-space-news--5648921/support.
Sponsor Details:
Ensure your online privacy by using NordVPN. To get our special listener deal and save a lot of money, visit www.astronomydaily.io/nordvpn. You'll be glad you did!
Get the best secure and private email on the planet. Stop your Government, google and who knows who else spying on every email you write. Do what we did and use ProtonMail. They beleive in privacy and there are no ads in their business model...yet they still provide a free forever service. Check them out and get out special deal at www.astronomydaily.io/protonmail
Become a supporter of Astronomy Daily by joining our Supporters Club. Commercial free episodes daily are only a click way... Click Here
This episode includes AI-generated content.
00:00:01 --> 00:00:04 Anna: Hey, everyone. Welcome to today's Astronomy
00:00:04 --> 00:00:05 AstroDailyPod. I'm Anna.
00:00:05 --> 00:00:08 Avery: And I'm avery. It's Thursday, August 27th.
00:00:08 --> 00:00:11 Series five, episode 178.
00:00:11 --> 00:00:14 Anna: Bigger show than usual today. Huw,
00:00:14 --> 00:00:16 our human producer, has given us the green
00:00:16 --> 00:00:19 light to run long because there's genuinely a
00:00:19 --> 00:00:22 lot going on. We've got a lab experiment that
00:00:22 --> 00:00:25 recreated the inside of an ice giant, one
00:00:25 --> 00:00:27 of the strangest origin storeys in modern
00:00:27 --> 00:00:30 astronomy hardware. A heads up on some actual
00:00:30 --> 00:00:33 space weather this weekend. An AI that can
00:00:33 --> 00:00:35 see a solar storm coming before it arrives.
00:00:36 --> 00:00:38 A quick word on SpaceX's enormous new
00:00:38 --> 00:00:41 plans for the Gulf coast, and a, uh, lunar
00:00:41 --> 00:00:43 eclipse that's happening literally tonight.
00:00:44 --> 00:00:46 Avery: That's a lot of universe for one episode.
00:00:46 --> 00:00:48 Anna: It is. Let's get into it.
00:00:48 --> 00:00:50 Avery: Okay, start us off. The headline is
00:00:50 --> 00:00:53 intriguing. Scientists recreated diamond rain
00:00:53 --> 00:00:56 solving a 20 year mystery. But what did they
00:00:56 --> 00:00:57 actually do here?
00:00:57 --> 00:01:00 Anna: Physicists at Lawrence Livermore National
00:01:00 --> 00:01:02 Laboratory just settled an argument that's
00:01:02 --> 00:01:05 been running in planetary science for about
00:01:05 --> 00:01:07 20 years. And they did it by essentially
00:01:07 --> 00:01:10 recreating the inside of Neptune in a
00:01:10 --> 00:01:13 lab for about a billionth of a second at a
00:01:13 --> 00:01:13 time.
00:01:14 --> 00:01:16 Avery: A, uh, billionth of a second doesn't sound
00:01:16 --> 00:01:17 like much to work with.
00:01:17 --> 00:01:20 Anna: It's not, but it's enough if you know what
00:01:20 --> 00:01:22 you're looking for. The team, led by
00:01:22 --> 00:01:25 LLNL physicist Marius Millett,
00:01:25 --> 00:01:28 took tiny diamond samples to the Omega Laser
00:01:28 --> 00:01:30 Facility that's at the University of
00:01:30 --> 00:01:33 Rochester's Laboratory for Laser Energetics,
00:01:33 --> 00:01:36 and used intense lasers to vaporise the outer
00:01:36 --> 00:01:38 layer of each diamond, which drives a
00:01:38 --> 00:01:40 shockwave straight through the rest of it.
00:01:40 --> 00:01:42 That shock wave crushes the diamond to
00:01:42 --> 00:01:45 pressures around 1 terapascal. That's
00:01:45 --> 00:01:47 roughly three times the pressure at the
00:01:47 --> 00:01:49 centre of the Earth and higher than what
00:01:49 --> 00:01:51 you'd find at the centre of Neptune or
00:01:51 --> 00:01:53 Uranus, while flash heating it to
00:01:53 --> 00:01:55 temperatures hotter than the surface of the
00:01:55 --> 00:01:55 Sun.
00:01:56 --> 00:01:58 Avery: Um, and that's meant to simulate what? The
00:01:58 --> 00:02:00 inside of an ice giant?
00:02:00 --> 00:02:03 Anna: Exactly. That. Neptune and Uranus are called
00:02:03 --> 00:02:06 ice giants because under the clouds, they're
00:02:06 --> 00:02:08 thought to have deep mantles of compressed
00:02:08 --> 00:02:11 water, methane and ammonia under pressures
00:02:11 --> 00:02:13 and temperatures so extreme that ordinary
00:02:13 --> 00:02:16 chemistry stops behaving the way it does up
00:02:16 --> 00:02:19 here. Back in 2017, an earlier
00:02:19 --> 00:02:22 LLNL led experiment first showed that
00:02:22 --> 00:02:24 carbon squeezed under those conditions
00:02:24 --> 00:02:27 crystallises into nano diamonds,
00:02:27 --> 00:02:29 literal diamond rain falling through the
00:02:29 --> 00:02:32 interior of these planets, possibly for
00:02:32 --> 00:02:34 billions of years, and possibly forming thick
00:02:34 --> 00:02:37 diamond layers around their rocky cores. That
00:02:37 --> 00:02:39 was the original headline. What this new
00:02:39 --> 00:02:42 study nails down is something narrower but
00:02:42 --> 00:02:45 more important. Exactly what temperature
00:02:45 --> 00:02:47 diamond itself melts at once. You're that
00:02:47 --> 00:02:48 deep.
00:02:48 --> 00:02:50 Avery: Why would that be uncertain? Diamond's
00:02:50 --> 00:02:50 diamond.
00:02:51 --> 00:02:53 Anna: Because at those pressures, you can't just
00:02:53 --> 00:02:56 stick a thermometer in it. You have to infer
00:02:56 --> 00:02:58 the melting point indirectly. And for two
00:02:58 --> 00:03:00 decades, lab measurements and quantum
00:03:00 --> 00:03:03 mechanical computer simulations disagreed
00:03:03 --> 00:03:05 with each other by close to a thousand
00:03:05 --> 00:03:07 degrees. Nobody could say for certain which
00:03:07 --> 00:03:10 one was right. Millet's team used much
00:03:10 --> 00:03:12 sharper x ray diffraction diagnostics than
00:03:12 --> 00:03:15 earlier experiments had access to,
00:03:15 --> 00:03:17 essentially getting a cleaner atomic scale
00:03:17 --> 00:03:20 snapshot of the diamond mid shock, and found
00:03:20 --> 00:03:23 that the real answer lines up almost exactly
00:03:23 --> 00:03:25 with what the quantum simulations predicted,
00:03:25 --> 00:03:27 not the older lab estimates.
00:03:28 --> 00:03:30 Avery: So the computers were right and the old
00:03:30 --> 00:03:31 experiments were off.
00:03:31 --> 00:03:34 Anna: That's the headline finding, yes. And there's
00:03:34 --> 00:03:36 a second result buried in there that's
00:03:36 --> 00:03:39 arguably just as interesting. The carbon
00:03:39 --> 00:03:41 atoms stayed locked in their normal diamond
00:03:41 --> 00:03:44 crystal structure right up until melting
00:03:44 --> 00:03:46 actually began. No weird in between
00:03:46 --> 00:03:49 phase, the kind some models had predicted. It
00:03:49 --> 00:03:52 goes solid diamond, then straight to liquid
00:03:52 --> 00:03:54 carbon cleanly. And they confirmed something
00:03:54 --> 00:03:57 poetic while they were at it. Just like ice
00:03:57 --> 00:04:00 floats on liquid water, solid diamond floats
00:04:00 --> 00:04:02 on liquid carbon. Under these conditions,
00:04:02 --> 00:04:05 same underlying physics, wildly different
00:04:05 --> 00:04:06 substance.
00:04:06 --> 00:04:09 Avery: Diamond icebergs floating in an ocean
00:04:09 --> 00:04:12 of molten carbon inside a
00:04:12 --> 00:04:13 planet.
00:04:13 --> 00:04:16 Anna: Which is a genuinely wild sentence to be able
00:04:16 --> 00:04:19 to say and have it be real science. But
00:04:19 --> 00:04:21 here's where it stops being just a curiosity
00:04:21 --> 00:04:24 about ice giants. This result actually
00:04:24 --> 00:04:26 matters for something happening right here on
00:04:26 --> 00:04:28 Earth. Fusion energy research.
00:04:29 --> 00:04:32 Avery: How does melting diamond connect to
00:04:32 --> 00:04:33 fusion?
00:04:33 --> 00:04:35 Anna: Inertial confinement fusion, the approach
00:04:36 --> 00:04:38 used at facilities like the National Ignition
00:04:38 --> 00:04:41 Facility, works by using powerful
00:04:41 --> 00:04:44 lasers to compress a small fuel capsule
00:04:44 --> 00:04:46 often built with a diamond shell, until the
00:04:46 --> 00:04:49 fuel inside gets hot and dense enough to
00:04:49 --> 00:04:52 fuse. Getting that compression right is
00:04:52 --> 00:04:55 incredibly delicate. Jock the capsule too
00:04:55 --> 00:04:57 hard, too fast, and you introduce
00:04:57 --> 00:05:00 instabilities that waste energy and can even
00:05:00 --> 00:05:03 ruin the implosion. Knowing the precise
00:05:03 --> 00:05:05 melting point of the diamond shell, the
00:05:05 --> 00:05:07 number this study just pinned down lets
00:05:07 --> 00:05:10 researchers use a gentler, slower initial
00:05:10 --> 00:05:13 shock while still guaranteeing the shell
00:05:13 --> 00:05:15 fully melts at exactly the right moment.
00:05:16 --> 00:05:18 Avery: And a, uh, gentler shock means what?
00:05:18 --> 00:05:21 Anna: In practical terms, a more compressible
00:05:21 --> 00:05:23 fuel capsule. And models suggest that
00:05:23 --> 00:05:26 alone could roughly triple the energy gain
00:05:26 --> 00:05:29 from these fusion implosions. More energy out
00:05:29 --> 00:05:32 for the same energy in without needing
00:05:32 --> 00:05:35 bigger, more expensive lasers to do it. It's
00:05:35 --> 00:05:38 a genuinely rare case of a planetary science
00:05:38 --> 00:05:40 result feeding directly and immediately
00:05:40 --> 00:05:43 into an entirely different field's
00:05:43 --> 00:05:44 engineering problem.
00:05:45 --> 00:05:47 Avery: So one experiment, two totally different
00:05:47 --> 00:05:50 payoffs, how ice giants actually work
00:05:50 --> 00:05:53 inside, and a Possible tune up for
00:05:53 --> 00:05:55 fusion reactors here on Earth.
00:05:56 --> 00:05:58 Anna: That's the shape of it. The findings are
00:05:58 --> 00:06:00 published in Nature Physics and the team is
00:06:00 --> 00:06:03 describing the new melting point measurements
00:06:03 --> 00:06:06 as atomic scale benchmarks. A reference
00:06:06 --> 00:06:08 point other researchers can now build their
00:06:08 --> 00:06:11 own simulations against for modelling extreme
00:06:11 --> 00:06:13 matter anywhere from planetary interiors
00:06:13 --> 00:06:16 to fusion capsules to eventually other
00:06:16 --> 00:06:19 worlds we haven't even looked at closely yet.
00:06:19 --> 00:06:22 Avery: A billionth of a second of lap time
00:06:22 --> 00:06:25 unlocking 20 years of disagreement.
00:06:25 --> 00:06:28 Anna: Sometimes that's all physics needs. The right
00:06:28 --> 00:06:30 billionth of a second pointed at the right
00:06:30 --> 00:06:31 question.
00:06:31 --> 00:06:34 Avery: Alright, next one's got a bit of everything.
00:06:34 --> 00:06:37 Cold War hardware, a, uh, cancelled spy
00:06:37 --> 00:06:39 programme and a telescope launching in three
00:06:39 --> 00:06:40 days.
00:06:40 --> 00:06:43 Anna: The Nancy Grace Roman Space Telescope is
00:06:43 --> 00:06:46 genuinely days away Now, Sunday morning,
00:06:46 --> 00:06:49 August 30, 7:26am, um,
00:06:49 --> 00:06:51 Eastern on a SpaceX Falcon Heavy from
00:06:51 --> 00:06:54 Kennedy Space Centre. We gave Roman its full
00:06:54 --> 00:06:57 feature treatment a couple of days ago, so
00:06:57 --> 00:06:59 today we wanted to do something a little
00:06:59 --> 00:07:01 different and tell you where the telescope's
00:07:01 --> 00:07:04 hardware actually came from because it's one
00:07:04 --> 00:07:06 of the stranger origin storeys in modern
00:07:06 --> 00:07:08 astronomy. And it's been getting fresh
00:07:08 --> 00:07:10 attention this week as launch gets close.
00:07:11 --> 00:07:13 Avery: I feel like I've heard this before, something
00:07:13 --> 00:07:15 about a, uh, spy satellite.
00:07:15 --> 00:07:18 Anna: You have, and it's true. Back in
00:07:18 --> 00:07:20 1999, the National Reconnaissance Office,
00:07:21 --> 00:07:23 the US intelligence agency that builds and
00:07:23 --> 00:07:26 operates spy satellites, kicked off a
00:07:26 --> 00:07:28 programme called Future Imagery Architecture,
00:07:28 --> 00:07:31 contracting Boeing to build a next generation
00:07:31 --> 00:07:34 family of optical and radar reconnaissance
00:07:34 --> 00:07:37 satellites. It expanded further after 911
00:07:37 --> 00:07:39 on the back of heightened national security
00:07:39 --> 00:07:42 spending. But by 2005 the whole thing
00:07:42 --> 00:07:45 had collapsed under billions of dollars in
00:07:45 --> 00:07:47 cost overruns. The New York Times at the time
00:07:47 --> 00:07:50 called it, and I'm quoting directly, perhaps
00:07:50 --> 00:07:53 the most spectacular and expensive
00:07:53 --> 00:07:55 failure in the 50 year history of American
00:07:55 --> 00:07:57 spy satellite projects.
00:07:58 --> 00:08:00 Avery: So a, uh, failed spy satellite programme
00:08:00 --> 00:08:03 Anna: just sat there for a few years?
00:08:03 --> 00:08:06 Yes. Then in 2010 the National
00:08:06 --> 00:08:09 Academy of Sciences decadal survey.
00:08:09 --> 00:08:12 Basically the astronomy communities ra wish
00:08:12 --> 00:08:14 list for the next decade of big missions
00:08:15 --> 00:08:17 named what would become the Roman Space
00:08:17 --> 00:08:20 Telescope as its absolute top priority.
00:08:21 --> 00:08:23 NASA announced in 2011 that it
00:08:23 --> 00:08:26 planned to repurpose leftover NRO
00:08:26 --> 00:08:29 hardware for the mission. And in 2012
00:08:29 --> 00:08:31 the NRO formally donated two
00:08:31 --> 00:08:34 complete unused telescopes from the
00:08:34 --> 00:08:37 cancelled programme to NASA. Each one
00:08:37 --> 00:08:40 had an optical telescope assembly, primary
00:08:40 --> 00:08:42 mirror, nine additional mirrors structure,
00:08:42 --> 00:08:45 and all roughly comparable to Hubble's
00:08:45 --> 00:08:47 own optics, and each valued at around
00:08:47 --> 00:08:50 $250 million.
00:08:50 --> 00:08:53 Avery: Free telescopes essentially sort of,
00:08:53 --> 00:08:54 though free
00:08:54 --> 00:08:56 Anna: undersells how much work it took. The
00:08:56 --> 00:08:59 electronics had to be entirely stripped out
00:08:59 --> 00:09:01 and replaced. Since a spy satellite's
00:09:01 --> 00:09:04 internals aren't built for open astrophysics
00:09:04 --> 00:09:07 and large sections of the original technical
00:09:07 --> 00:09:09 documentation stayed classified and
00:09:09 --> 00:09:12 redacted. So Roman's engineers had to
00:09:12 --> 00:09:15 reverse engineer parts of a system built by a
00:09:15 --> 00:09:17 completely different team for a completely
00:09:17 --> 00:09:20 different purpose. Experts still genuinely
00:09:20 --> 00:09:22 disagree on whether repurposing the hardware
00:09:22 --> 00:09:25 actually saved NASA money. Overall, once you
00:09:25 --> 00:09:28 count all that rework, NASA still holds on
00:09:28 --> 00:09:30 to the second donated telescope.
00:09:30 --> 00:09:33 Incidentally, no announced plans for it yet.
00:09:34 --> 00:09:36 Avery: That's a wild pedigree for a mission about to
00:09:36 --> 00:09:38 go hunt dark energy and exoplanets.
00:09:39 --> 00:09:42 Anna: It really is Cold War era spy
00:09:42 --> 00:09:44 satellite optics sitting unused for the
00:09:44 --> 00:09:47 better part of a decade now, three days from
00:09:47 --> 00:09:49 launch as one of the most capable wide field
00:09:49 --> 00:09:52 observatories ever built. As of today,
00:09:52 --> 00:09:55 the encapsulated telescope has moved into
00:09:55 --> 00:09:57 SpaceX's hangar at Launch Complex
00:09:57 --> 00:10:00 39A and is being mated to its
00:10:00 --> 00:10:02 Falcon Heavy this week with a launch
00:10:02 --> 00:10:04 readiness review scheduled for Tomorrow,
00:10:04 --> 00:10:07 Friday the 28th. To confirm everything's go
00:10:07 --> 00:10:10 for Sunday will be all over the actual
00:10:10 --> 00:10:11 launch when it happens.
00:10:12 --> 00:10:15 Avery: From reconnaissance to cosmology in one
00:10:15 --> 00:10:16 very unlikely career change.
00:10:17 --> 00:10:19 Anna: Not a bad way to spend a second life.
00:10:19 --> 00:10:21 Avery: Quick update on something we flagged as, uh,
00:10:21 --> 00:10:24 a maybe earlier this week. It's not a maybe
00:10:24 --> 00:10:25 anymore, right?
00:10:25 --> 00:10:28 Anna: Tuesday's M M6.9 flare out of
00:10:28 --> 00:10:31 Sunspot Region 4513 sent
00:10:31 --> 00:10:34 a coronal mass ejection our way, and at
00:10:34 --> 00:10:37 the time forecasters were only calling it an
00:10:37 --> 00:10:39 outside chance of minor geomagnetic
00:10:39 --> 00:10:42 storming that's firmed up. NOAA's Space
00:10:42 --> 00:10:45 Weather Prediction Centre has now issued an
00:10:45 --> 00:10:48 official G2 that's moderate on their
00:10:48 --> 00:10:51 five step storm scale geomagnetic storm
00:10:51 --> 00:10:53 watch for this Friday, August 28th.
00:10:54 --> 00:10:56 Avery: What chains between outside chance and an
00:10:56 --> 00:10:57 actual watch?
00:10:57 --> 00:11:00 Anna: Better tracking of the CME's trajectory and
00:11:00 --> 00:11:03 speed plus a second factor stacking on top
00:11:03 --> 00:11:06 of it there's a coronal hole high speed
00:11:06 --> 00:11:09 solar wind stream also forecast to hit
00:11:09 --> 00:11:11 Earth's Magnetosphere starting the 27th.
00:11:11 --> 00:11:14 Essential tonight with the CME's
00:11:14 --> 00:11:16 effects layering in on top of that starting
00:11:16 --> 00:11:19 the 28th. Noah's language is that these
00:11:19 --> 00:11:22 disturbances are anticipated to affect
00:11:22 --> 00:11:25 geospace across both days rather than
00:11:25 --> 00:11:26 a single glancing blow.
00:11:26 --> 00:11:29 Avery: Does a G2 watch mean anything for people on
00:11:29 --> 00:11:32 the ground, or is this purely a space weather
00:11:32 --> 00:11:33 nerd milestone?
00:11:33 --> 00:11:35 Anna: At AH G2 levels you can get some minor
00:11:35 --> 00:11:38 fluctuations in high latitude power grids
00:11:38 --> 00:11:41 and a bit of extra drag on satellites in low
00:11:41 --> 00:11:44 orbit, but the part most listeners will
00:11:44 --> 00:11:46 actually care about is Aurora, uh, D2
00:11:46 --> 00:11:49 storms can push the aurora oval down into the
00:11:49 --> 00:11:52 northern tier US States, southern Canada,
00:11:52 --> 00:11:54 the UK and similar latitudes in Europe.
00:11:55 --> 00:11:57 It's genuinely a, uh, get outside and look
00:11:57 --> 00:11:59 night if you're up there.
00:11:59 --> 00:12:02 Avery: And for us down here, same answer
00:12:02 --> 00:12:05 Anna: as earlier this week. G2 is still a modest
00:12:05 --> 00:12:08 storm. And modest storms don't typically push
00:12:08 --> 00:12:11 the Aurora australis much past Tasmania on
00:12:11 --> 00:12:13 a good night. We don't have anything in this
00:12:13 --> 00:12:15 forecast suggesting it goes further than
00:12:15 --> 00:12:18 that. Worth a glance at the southern horizon
00:12:18 --> 00:12:20 tonight and tomorrow if you're somewhere
00:12:20 --> 00:12:22 dark. But we wouldn't build plans around it
00:12:23 --> 00:12:26 Avery: from maybe to NOAA officially watching it
00:12:26 --> 00:12:27 in about 48 hours.
00:12:28 --> 00:12:30 Anna: That's space weather forecasting for you. It
00:12:30 --> 00:12:33 sharpens fast as the event actually gets
00:12:33 --> 00:12:33 close.
00:12:33 --> 00:12:36 Avery: Since we're already talking space weather and
00:12:36 --> 00:12:38 there's a genuinely clever piece of research
00:12:38 --> 00:12:39 that ties right into this.
00:12:39 --> 00:12:42 Anna: It does. And the timing's almost too neat.
00:12:42 --> 00:12:45 A team led by researchers at the New Jersey
00:12:45 --> 00:12:48 Institute of Technology has built an AI model
00:12:48 --> 00:12:50 nicknamed early detect that can spot the
00:12:50 --> 00:12:53 hidden precursor signs of a new solar active
00:12:53 --> 00:12:56 region forming before it's even visible on
00:12:56 --> 00:12:58 the sun's surface. An average of about
00:12:59 --> 00:13:01 9.24 hours ahead of time.
00:13:01 --> 00:13:04 Avery: Nine hours before a sunspot region even
00:13:04 --> 00:13:07 shows up. How do you predict something before
00:13:07 --> 00:13:08 it exists?
00:13:08 --> 00:13:11 Anna: You look underneath essentially active
00:13:11 --> 00:13:13 regions. The sunspot clusters that produce
00:13:13 --> 00:13:16 flares and CMEs like the one we just talked
00:13:16 --> 00:13:19 about don't just pop into existence. There
00:13:19 --> 00:13:21 are subtle acoustic signals and shifts in the
00:13:21 --> 00:13:24 sun's subsurface magnetic field that happen
00:13:24 --> 00:13:27 first as new magnetic flux rises up from
00:13:27 --> 00:13:29 deeper inside the sun towards the surface.
00:13:30 --> 00:13:32 Those signals are faint and easy to miss by
00:13:32 --> 00:13:35 eye. But the NJIT team trained a, uh,
00:13:35 --> 00:13:37 transformer based AI model, the same
00:13:37 --> 00:13:40 underlying architecture behind tools like
00:13:40 --> 00:13:42 ChatGPT on hourly acoustic
00:13:42 --> 00:13:45 power maps and magnetic field data from
00:13:45 --> 00:13:47 NASA's Solar Dynamics Observatory to pick
00:13:47 --> 00:13:48 them out.
00:13:48 --> 00:13:51 Avery: Transformer models reading the sun's insides
00:13:51 --> 00:13:52 like they'd read a sentence.
00:13:52 --> 00:13:55 Anna: Pretty much the same basic idea just applied
00:13:55 --> 00:13:58 to helioseismic data instead of language.
00:13:58 --> 00:14:00 One detail the researchers highlighted that
00:14:00 --> 00:14:03 genuinely surprised them. A, uh, standard
00:14:03 --> 00:14:05 filtering step that normally cleans up noisy
00:14:05 --> 00:14:08 data actually hurt the model's performance
00:14:08 --> 00:14:10 here because it was stripping out faint
00:14:10 --> 00:14:13 fluctuations. That turned out to be exactly
00:14:13 --> 00:14:15 the signal the AI needed to catch early.
00:14:15 --> 00:14:18 Leaving the noise in made the predictions
00:14:18 --> 00:14:18 better.
00:14:19 --> 00:14:21 Avery: So where does this actually go next? Is this
00:14:21 --> 00:14:24 feeding into real forecasts soon?
00:14:24 --> 00:14:27 Anna: Not quite yet, and the team's been upfront
00:14:27 --> 00:14:29 about that. They described early detect as
00:14:29 --> 00:14:32 not yet ready for Real time forecasting and
00:14:32 --> 00:14:34 it still needs validation against a lot more
00:14:34 --> 00:14:37 solar events before anyone could rely on it
00:14:37 --> 00:14:39 operationally. But the ceiling here is
00:14:39 --> 00:14:42 obvious. Today's space weather warnings like
00:14:42 --> 00:14:45 the G2 watch we just covered, mostly start
00:14:45 --> 00:14:47 once a CME is already on its way. A
00:14:47 --> 00:14:50 tool that can flag the storm producing region
00:14:50 --> 00:14:52 before it's even fully formed pushes that
00:14:52 --> 00:14:55 warning window back even further. The team's
00:14:55 --> 00:14:58 also released a public dataset called Solared
00:14:58 --> 00:15:00 and an interactive platform so other
00:15:00 --> 00:15:02 researchers can build on this directly.
00:15:03 --> 00:15:05 Avery: Getting ahead of the sun's mood swings before
00:15:05 --> 00:15:06 they start.
00:15:06 --> 00:15:08 Anna: That's the goal. We'll keep an eye on it as
00:15:08 --> 00:15:09 it develops.
00:15:09 --> 00:15:12 Avery: Last thing before Skywatch and it's a big
00:15:12 --> 00:15:14 one, even though we're keeping it brief
00:15:14 --> 00:15:14 today.
00:15:14 --> 00:15:17 Anna: Basaks and Louisiana Governor Jeff Landry
00:15:17 --> 00:15:19 announced this week that the company is
00:15:19 --> 00:15:22 building a second Starbase. This one on the
00:15:22 --> 00:15:25 Gulf coast in Vermilion Parish, Louisiana.
00:15:25 --> 00:15:27 With a uh, jaw dropping price tag,
00:15:28 --> 00:15:30 $100 billion. The plan
00:15:30 --> 00:15:33 is five separate launch complexes, two
00:15:33 --> 00:15:36 Starship towers each. So 10 pads total,
00:15:36 --> 00:15:39 each with its own propellant farm, plus on
00:15:39 --> 00:15:42 site propellant production, power generation
00:15:42 --> 00:15:44 and deep water shipping access.
00:15:44 --> 00:15:47 SpaceX President Gwynne Shotwell called it a
00:15:47 --> 00:15:50 uh, fully self sustaining spaceport.
00:15:50 --> 00:15:53 Avery: 10 pads is an enormous number
00:15:53 --> 00:15:55 compared to what they've got in Texas right
00:15:55 --> 00:15:55 now.
00:15:55 --> 00:15:58 Anna: It's built for scale. Musk has talked about
00:15:58 --> 00:16:01 Starship eventually flying more than 30 times
00:16:01 --> 00:16:03 a day by 2030, something like
00:16:03 --> 00:16:06 10 flights a year across the whole
00:16:06 --> 00:16:09 programme. And one site in Texas simply isn't
00:16:09 --> 00:16:11 built for that kind of cadence. Construction
00:16:11 --> 00:16:14 on Starbase Louisiana is targeted to start in
00:16:14 --> 00:16:17 2027 with the first launch aimed at
00:16:17 --> 00:16:19 2029. And the project's expected to create
00:16:19 --> 00:16:22 around 3 direct jobs over the next
00:16:22 --> 00:16:23 decade.
00:16:23 --> 00:16:26 Avery: Uh, a 10 year commitment before a single
00:16:26 --> 00:16:28 rocket flies off that particular stretch of
00:16:28 --> 00:16:28 coast.
00:16:29 --> 00:16:31 Anna: That's the scale SpaceX is planning around
00:16:31 --> 00:16:33 these days. We'll keep tracking it as it
00:16:33 --> 00:16:35 develops. For now, just worth having on your
00:16:35 --> 00:16:36 radar.
00:16:36 --> 00:16:38 Avery: Alright, Skywatch, and um, this is the one
00:16:38 --> 00:16:41 we've been previewing for days. It's actually
00:16:41 --> 00:16:42 happening tonight.
00:16:42 --> 00:16:44 Anna: It is tonight into tomorrow morning,
00:16:44 --> 00:16:46 depending which side of the planet. You're
00:16:46 --> 00:16:49 listening to this from. The 96% partial
00:16:49 --> 00:16:51 lunar eclipse we've mentioned a few times
00:16:51 --> 00:16:53 this week, gets underway with a partial phase
00:16:53 --> 00:16:55 starting at 2:34 Utah UTC,
00:16:55 --> 00:16:58 reaches maximum eclipse at 4:13
00:16:58 --> 00:17:01 UTC and wraps up its partial phase
00:17:01 --> 00:17:03 around 5:52 UTC, all
00:17:03 --> 00:17:06 early morning on August 28th. Universal Time.
00:17:06 --> 00:17:09 Avery: Translate that for the Americas, since it's
00:17:09 --> 00:17:11 Anna: their show tonight for the US East Coast.
00:17:11 --> 00:17:13 That's partial eclipse starting around
00:17:13 --> 00:17:16 10:34pm Eastern tonight the
00:17:16 --> 00:17:18 27th. Maximum coverage at
00:17:18 --> 00:17:21 12:13am Eastern just after
00:17:21 --> 00:17:24 midnight, technically the 28th and the
00:17:24 --> 00:17:27 partial phase wrapping up around 1:52am
00:17:27 --> 00:17:30 Eastern. Good views right across north and
00:17:30 --> 00:17:32 South America and low on the horizon for
00:17:32 --> 00:17:34 parts of Europe and Africa as well.
00:17:34 --> 00:17:37 Avery: And for us folks down under, one more time.
00:17:37 --> 00:17:39 Anna: Honestly, one more time. Broad daylight
00:17:39 --> 00:17:42 here. That 4:13 UTC peak
00:17:42 --> 00:17:45 lands around 2:13pm Thursday
00:17:45 --> 00:17:48 afternoon in Sydney. Moon nowhere near the
00:17:48 --> 00:17:50 horizon. This eclipse simply belongs to the
00:17:50 --> 00:17:52 other side of the planet. If you've got
00:17:52 --> 00:17:54 family, friends or listeners over in the
00:17:54 --> 00:17:57 Americas, tell them it's completely safe to
00:17:57 --> 00:17:59 watch with nothing more than their own eyes.
00:17:59 --> 00:18:01 No filters, no eclipse glasses needed like
00:18:01 --> 00:18:04 you'd want for a solar eclipse. Just find a
00:18:04 --> 00:18:06 clear view of the moon and watch it slide
00:18:06 --> 00:18:09 into a deep coppery red as it moves through
00:18:09 --> 00:18:09 Earth's shadow.
00:18:10 --> 00:18:12 Avery: And um, this is landing the same couple of
00:18:12 --> 00:18:14 nights as the geomagnetic storm watch we just
00:18:14 --> 00:18:15 covered.
00:18:15 --> 00:18:18 Anna: Theme General Window yes, tonight into
00:18:18 --> 00:18:20 tomorrow is genuinely the stretch to watch
00:18:20 --> 00:18:22 the sky whichever side of the planet you're
00:18:22 --> 00:18:24 onan eclipse for one hemisphere, a possible
00:18:25 --> 00:18:27 aurora mostly for the northern one, and
00:18:27 --> 00:18:30 neither strictly speaking, ours to claim down
00:18:30 --> 00:18:30 here.
00:18:30 --> 00:18:33 Avery: So what's actually worth stepping out for
00:18:33 --> 00:18:33 locally?
00:18:33 --> 00:18:36 Anna: Venus is still the reliable one, bright and
00:18:36 --> 00:18:39 unmistakable low in the west shortly after
00:18:39 --> 00:18:42 sunset. Saturn's well placed too, rising in
00:18:42 --> 00:18:44 the east as it gets dark. Worth finding with
00:18:44 --> 00:18:47 binoculars if you've got a pair handy. Not as
00:18:47 --> 00:18:49 dramatic as a blood red moon, but a solid
00:18:49 --> 00:18:51 pair to go find tonight, regardless of what
00:18:51 --> 00:18:53 the other side of the world is looking at.
00:18:53 --> 00:18:56 Avery: Venus at dusk, Saturn overnight, an eclipse
00:18:56 --> 00:18:59 for the Americas and a storm watch for the
00:18:59 --> 00:18:59 far north.
00:18:59 --> 00:19:02 Anna: A genuinely full sky this week, even for the
00:19:02 --> 00:19:04 parts of it we don't get to see directly.
00:19:04 --> 00:19:06 And that's it for today's episode.
00:19:06 --> 00:19:09 Avery: Quick recap. Lawrence Livermore Physicists
00:19:09 --> 00:19:11 recreated the crushing pressures inside
00:19:11 --> 00:19:13 Neptune and Uranus in the lab, solving a 20
00:19:13 --> 00:19:16 year disagreement over Diamond's melting
00:19:16 --> 00:19:18 point and pointing toward a possible tripling
00:19:18 --> 00:19:20 of fusion energy gains. Researchers have
00:19:20 --> 00:19:23 built an AI model that can spot solar storms
00:19:23 --> 00:19:26 forming roughly nine hours before their even
00:19:26 --> 00:19:28 visible. SpaceX unveiled a $100
00:19:28 --> 00:19:31 billion second starbase planned for the
00:19:31 --> 00:19:34 Louisiana coast, and a deep partial lunar
00:19:34 --> 00:19:35 eclipse is unfolding tonight for the
00:19:35 --> 00:19:38 Americas, daylight for the rest of us. But
00:19:38 --> 00:19:40 Venus and Saturn are still worth
00:19:40 --> 00:19:42 Anna: a look if you enjoyed the show. The best
00:19:42 --> 00:19:44 thing you can do is tell a friend, leave us a
00:19:44 --> 00:19:46 rating wherever you listen and follow us.
00:19:46 --> 00:19:49 Just search Astro daily pod on Facebook,
00:19:49 --> 00:19:52 Instagram, TikTok X, Tumblr and YouTube.
00:19:52 --> 00:19:53 Avery: And while you're there, head to
00:19:53 --> 00:19:56 astronomydaily.IO and sign up for
00:19:56 --> 00:19:58 our free daily newsletter. A summary of the
00:19:58 --> 00:20:00 latest space and astronomy news straight to
00:20:00 --> 00:20:03 your inbox, plus an email alert every time we
00:20:03 --> 00:20:04 post a new episode.
00:20:04 --> 00:20:06 Anna: We'll be back tomorrow with more from across
00:20:06 --> 00:20:08 the universe. Until then, keep looking up.
00:20:08 --> 00:20:10 See you next time, and wishing you. Clear
00:20:10 --> 00:20:12 Skies Astronomy Day
00:20:14 --> 00:20:15 Storeys.