How is SpaceX Cracking the Formula with Crucial Atmo Data?
HTML-код
- Опубликовано: 12 дек 2023
- Check out the amazing Atmo AI meteorology tech: atmo.ai
There's one unsung hero that plays a crucial role in rocket launches. Weather predictions! You might be surprised how intricately the two are linked, and how SpaceX is revolutionizing weather forecasting at Starbase using artificial intelligence provided by Atmo (atmo.ai) , a new AI meteorology start-up that's building some of the most precise forecasts on Earth.
Atmo's social links:
LinkedIn - / atmo-ai
X/Twitter - / atmo_ai
Instagram - / atmo_ai
Facebook - / atmoai
👕Like this shirt? Pick it up on any product you like here.
marcus-house.myspreadshop.com...
and in reverse
marcus-house.myspreadshop.com...
🎁 Marcus House Merch - marcus-house.myspreadshop.com/
You can support me on:
Patreon - / marcushouse
Join my Discord - / discord
RUclips Shorts - / @marcushouseshorts
Follow/Subscribe on Twitter - / marcushouse
TikTok - / marcushouse
The production crew:
GameplayReviewUK, TiagoCruz, Mr Pleasant, Virtu, Orbitly
Support from the below is always massively appreciated:
📷 NASASpaceFlight - / @nasaspaceflight
📷 RGVAerialPhotography - / @rgvaerialphotography
📷 Randolph Visuals - / cosmicalchief
📷 Greg Scott - / gregscott_photo
📷 Starship Gazer - / starshipgazer
📷 Cosmic Perspective - / @cosmicperspective
📷 LabPadre - / @labpadre
📷 Epic Spaceflight - / @epicspaceflight
Set models:
😍 Mini venting Starship/SLS - stardesk.peachs.co/a/marcus-h...
😍 Starship, & Crew Dragon by - morethan3d.com/
😍 Moon/Mars Mova Globes - www.movaglobes.com/
😍 Saturn V - LEGO - www.lego.com/en-au/product/le...
😍 Space Shuttle - LEGO - www.lego.com/en-au/product/na...
3D artist magicians:
✨ Tony Bela - / infographictony
✨ Ryan Hansen Space - / ryanhansenspace
✨ Erc X - / ercxspace
✨ Corey - / c_bass3d
✨ Alex Svan - / alexsvanart
✨ DeepSpaceCourier - / ds_courier
✨ SpaceXvision - / spacexvision
✨ Stanley Creative - / caspar_stanley
✨ Matt Ryan - / mattr5226
✨ TijnM_3DAnimations - / m_tijn
✨ Christian Debney - / christiandebney
✨ Evan Karen - / @evankaren
✨ 3D Daniel - / 3ddaniel1 
Наука
I'm really thrilled to have had this opportunity to work with the Atmo AI meteorology team from atmo.ai ! It has been a wonderful experience getting to more deeply understand how SpaceX selects the best time for launch. This new technology is quite exciting I must say.
Brilliant as always, Marcus! I really can’t state enough how appreciative I am of the dedication and quality work you do. Thank you. You’re one in a million.
Fantastic video Marcus! This is the first I have heard of atmo.
Imagine if you AI model a highly accurate prediction system AND you run back data simulations on ground based and offshore based stations that can add or remove heat all across a wide region. Then find the best predictions and changes and control of those systems to lesson the intensity or divert from populous areas the future results of weather. You could use AI to manipulate the weather. Seems like the logical next step. Like diverting asteroids by tiny nudges far away to get bigger results much later on.
Maybe eventually adding satellites later on that can add heat from space as well. Like a giant space laser. 😕Might be a tough sell though. Ground trials would go a long way to help sell it.
Have "space weather" too, actually. We still don't even know what that even is in point of fact ("what constitutes to create space weather) but let there be no doubt the further "out there" you go the more you will interacting with what is not protected Earth/Moon electromagnetism.
The Challenger rocket explosion was not an unavoidable “oops” caused by weather. The company Morton Thiokol supplied the solid rocket boosters and their subcontractor, HydraPak, supplied the booster segment's o-rings and gave NASA a no-fly low temperature limit. Mistake 1 - NASA, knowing the air temperature was below the value recommended by the Morton Thiokol/HydraPak engineers, still called Morton Thiokol and ask them if they could fly anyway and the Morton Thiokol/HydraPak engineers said “no”. Mistake 2 - The Management at Morton Thiokol overruled the engineers and, well, you know what happened (o-rings leaked, boosters exploded). Humanity has been learning this lesson over and over and over again in all facets of life.
Yes... as I have posted....The challenger explosion was not caused by or effected by poor weather prediction. NASA knew the o-rings had problems in the cold a long time before, they were told days before hand that it was going to be too cold to safely launch. NASA pushed the launch any ways.
Greed, arrogance in leadership causing loss of life… Happens in every sizeable company to this day.
Correct. High wind sheer had nothing to do with the failure of the o-rings. 6:38
Seven dead due to a bureaucratic failure. What changes were made to stop this? We pay (at the point of a gun), TPTB play. Details of how our money is being spent? "You don't get no stinkin' details, peon. You choose to work/pay/die, ours is to control you, play, live longer better lives." The "aristocracy" lives on. Still vote?
Hubble's issues were also nasa's fault. Perkin Elmer told them that it needed calibrated, but nasa thought they could do it themselves and save money. PE should have been contracted for the calibration before the launch.
When you talk about the loss of the Shuttle, you needed to mention that one the engineers on the ground warned NASA repeatedly that the O-rings could fail due to the cold temperatures. NASA ignored the warnings, and the rest is history.
NASA is a government office so theres always a politician somewhere that might want his or her name attached to a flight.
@@Irobert1115HD I think you've hit the nail on the head here. Why should we assume that the government is going to do any better or be any less egregious with AI?
Ahh, bureaucracy.
That's right, and he expected the whole stack to explode on the pad. Indeed, a plume of black smoke erupted immediately from one of the SRBs at ignition, which could have indicated a much quicker and lower altitude explosion, but somehow it unexpectedly resealed after getting heated up. But the seal was still poor, which allowed the O-ring to be eroded, as it often was during other flights, by burning gases, only much faster, leading to a burn-through near booster separation, and the destruction of the entire stack.
By the way, if the SRBs are functioning properly, there should be *ZERO* erosion of the O-rings, ever. They should, by intention and design, never even be touched by the burning gases within (presumably protected by high-pressure air or inert gas that a proper seal would maintain). But there was significant erosion discovered on a number of previous flights. This should have *immediately* grounded the Space Shuttle until the problem could be fixed, because as I said, there should never be any erosion at all. Instead, some very "wise", "smart" people who know better than the lowly engineers decided that since the worst known case had "only" about 33% of an O-ring eroded (Jesus Christ! A disaster waiting to happen!), they claimed that they had a large 67% safety margin. 🤨🙄😬 That is *NOT* true in any valid sense! Not in a design for which zero erosion was specified and the exact cause of the problem was still unknown. GAH! 🤬 They're cold-blooded murderers.
@@Irobert1115HD There are always politicians and managers trying to push things, but the Space Shuttle program was notoriously bad in this regard.
I cannot blame you for including Challenger as weather was obviously *the* factor that made the ship unsafe to fly. However, we cannot blame the weather for the failure in any meaningful manner. If you drive your car into the sea, you cannot blame the tide and better forecasting would not have helped. Challenger *involved* weather, but the disaster was *caused* entirely by human stupidity.
I was a young teenager, but I remember that morning well. Having been too young to have experienced Vietnam, Challenger was my first major lesson in placing trust in authorities. *I* , sitting on the floor in front of our console TV, was worried about the flight as the network newscasters were telling us that the temperatures were below official no-go levels and they even pointed to the solid booster seals as potential problems!
Those astronauts trusted NASA, as did I. That trust was not warranted.
challenger accident wasn't caused by incorrect weather forecast. They *knew* what weather was on that day.
WHOA! Hold up there. The Challenger accident was NOT weather-related incident. Completely unrelated to the topic of weather prediction. It was a complete failure of the internal decision-making process. Engineers who said, "NO GO" and administrators who said, "Do it anyway".
Great to see a midweek update commentary from you Marcus. Once a week was great but twice is even better. Thanks 😊
Weather prediction helps by not wasting labor and resources preparing for a launch that eventually gets scrubbed, but it does not necessarily ensure safety. 06:24 The tragic loss of Challenger was not a weather caused failure, but rather a human failure. Clear boundaries were violated in the operating envelope of the SRB’s. Ample warnings were given and they were ignored.
AI forecasting has been used for NASA space launches since the '80s. They were rather rudimentary then. Baysian methods were used during the 70s ... all from paper tables. The Challenger accident was not due to a failure to forecast dangerous conditions but because the weather conditions were ignored. The upper air winds were being measured that day using an experimental wind shear radar. The question is, without the data, how accurate can an AI forecast be? How can the AI forecast be verified? What basis set is used to build the AI model?
The AI has an All Go Rhythm 😊
LOL, without data AI is nothing but code, and 🤔 can't they just wait and see after a prediction, would've they have had it running constantly and fine tuning it against reality?
An airline pilot radioed in to the control tower, to pass on to NASA, about a strong upper wind (as he knew about the shuttle launch) and as far as I recall, the information was passed on to NASA, but not taken seriously.
So that was just one more factor, to add to all the other factors, that still did not prevent the launch. Systemic failures by the management, the philosophy at the time, the manufacturers of all the integrated systems being restricted by, of course, a budget, limiting what could be designed.
At least currently, there seems to be more on board system computers that determine if the rocket launch's or aborts. Possibly why we are seeing very few failures in rocket launch's this year.
Models, not AI.
I'm a big fan of your videos. Thanks for making them.
The Challenger disaster wasn't caused by weather. The disaster was caused because the people in charge flew despite engineers telling them not to fly. The conditions were outside of tolerances for those gaskets the engineers told them it was outside of tolerances and the assholes decided to roll the dice. I know this was a tiny part of your video but I believe it's important to not forget the correct lesson that should have been learned from the Challenger disaster. It was people deciding to ignore safety warnings and put other people in harms way in order to not look bad. Safety should always be the first priority, it wasn't for Challenger just like it isn't for many things today and people are paying the price again.
After experiencing my first tornado warning here in Kentucky a couple days ago, I understand just how crucial it is to improve our meteorological predications and protect our assets and lives.
Awesome mid week video. .
Throughly interesting from start to finish.
Thank you MH + Team for this extra nugget of space happiness
Great introduction to a previously under-reported but critical part of a rocket launch. Congrats!
Awesome mid-week video! So glad to see this channel growing so quickly, great work Marcus & team
How is their correlation between their longer term forecast (say +5 days) to the actual conditions experienced.
They may have an awesome predictive tool that's faster than it's contemporaries, but if it's not as accurate, speed means nothing. ;)
Fab video as usual, always great to see Aussie creators on here. :)
The challenger explosion was not caused by or effected by poor weather prediction. NASA new the o-rings had problems in the cold a long time before, they were told days before hand that it was going to be too cold to safely launch. NASA pushed the launch any ways.
yeah, that was the final piece, but this part was not relevant to this topic and not the only cause, but just the biggest hole in the swiss cheese model.
There was the unusually cold weather played that a role in this too. if it were more moderate temperatures that day this might not have happened. also it seems the Shuttle expierienced immense windshear during ascent, which is possibly also a contributing factor (and again also weather related).
@@unitrader403 Weather predictions could not have *changed* the weather. They *knew* the weather the day of launch and should have scrubbed. I studied the failure charts for the SRBs that were provided to NASA by the engineers, and they had not even tested the SRBs at such low temperatures. Failures had consistently occurred at temperatures tens of degrees F higher.
Pretty sure the point wasn't "better weather prediction could have saved Challenger" but "this is how significantly adverse weather affects launch safety"
Thank you for your videos. I drive every night, and you always help me stay informed as well as helping me pass the time
Challenger wasn't a weather prediction issue. They knew there was an issue with the SRB O rings in the cold and they knew it was too cold to launch but chose to launch anyway.
That was a really interesting and informative advertisement for them, thank you.
Thanks Marcus for this" mid week" post..! You and your team have probably been working on this for quite some time.. Great research and production. As a tech geek, its great to see you report on the advantages of AI advancing the state of the art, and being hugely beneficial to humanity, not the mainstream medias threats of doom and gloom everyone hears.. Cheers...!
Big Wednesday with Marcus!
Mid week video!? The ambassador is really spoiling us!
Excellent video! Thank you and your Team, Marcus!
I always appreciate your updates about the "spice" missions, Marcus House! Hugs from Italy
Thank you! ❤
AI weather forecasting isn't a magic bullet. The weather remains a chaotic non-linear system. You can increase accuracy with more input measurements, and the AI techniques may allow the quick processing of all that data, but there will still be inherent uncertainty in the prediction, and more so the further out in time the prediction is.
Love and respect for everything you and your crew accomplish Marcus! You guys are so professional.
Stayed home sick today but an unexpected Marcus House video has me feeling better!
Great documentary Marcus. Really informative
Great to see your new content @MarcusHouse .
Crucial information!
On the space shuttle vertical wind shear had nothing to do with the failure just the cold breaking the o ring seal causing hot gasses to burn through
The cold didn’t break it. The cold made it stiff so it couldn’t change shape as needed. Two things made the boosters change shape. 1) the whole shuttle (orbiter, tank, boosters) were mounted on the launch platform with bolts at the bottom of the boosters which made them bend. Once the shuttle was launched, they “twanged” back. 2) wind shear (but likely not a big problem in the shuttle boosters - a real problem for the very long and thin Falcon 9).
@@peterfireflylund The challenger explosion was not caused by or effected by poor weather prediction. NASA new the o-rings had problems in the cold a long time before, they were told days before hand that it was going to be too cold to safely launch. NASA pushed the launch any ways.
Maybe some report says that wind sheer was a contributing factor. I don't know, but if that is what the report says then Marcus needs to mention it.
As far as I know, it was because NASA choose to proceed while the temperature of the O-rings was to low. Against warnings from lower level staff. But I'm not a credible source.
@mennovanlavieren3885 yeah that's what I thought as well said in that nasa video
@@Ezzell_ yes, but my point was that the seal didn’t *break* the o-ring (it didn’t crack or split in two), it just made it too stiff to compensate for the bending where the boosters were joined.
What a pleasant surprise! Great episode Marcus! Great work!!!!!
As a meteorologist, tip of the hat.
If their weather predictions are so accurate, shouldn't it just become the new industry standard?
I'm guessing they're focusing on a very small region at a time?
Also, you don't really need to know the wind 2 km up as a regular person.
There are dozens and dozens of weather models to choose from already, and they all have their strengths and weaknesses. Weather sites and channels and anywhere else that you might get weather from all use different models. If you look hard enough, you'll find one that gives you a sunny forecast when the rest are calling for rain. Just because one thing is better than the others doesn't necessarily imply it will be the most popular. In fact, the most popular models aren't even the most accurate, they're just easier for the average person to access. There is no regulating body like NOAA that determines which model is the industry standard that everyone uses and shuts down all the others. That would be like the DOT deciding that Ford makes the best truck and you're not allowed to use any others.
The AI & trad models are complementary, working in different ways. Trad models simulate actual atmospheric physics, with comparisons of forecast vs outcomes used to advance the underlying model. No empirical parametric overlays or heuristics added - just head scratching until sound theoretical basis for changes to the model found. The AI models are different: no underlying theory, just train in sample, judge model purely based on test out of sample. If done well, neural net can be much faster as it’s like a fast approximation to the trad model. It can also be more accurate for very local or short term forecasts as you can afford to run with much more detailed grid in space or time. Weakness of NN is you don’t know why it works, so less trustworthy in extreme or unusual cases (with sparse or no representation in training data). So, trad models will continue for research, but most commercial or consumer forecasts will use NN with year or 2. Check google paper on weather forecasting for details.
@@Skinflaps_Meatslapper That's a bit black and white though. You could mandate that all weather reports have to evaluate their accuracy in hindsight, with a standard methodology and provide that number to the public.
it probably will in time. maybe the current technology is kind of like a prototype and not really something ready for wider use?
your midweek videos are always a welcome surprise
Urban drone traffic management, both for cargo and for passengers, really needs something like this where the weather around the city at meter resolution can be forecast in near real time. That is NOT an easy task!
Yet another part of the space industry so many don't think about! Great video!!
Absolutely incredible. I again anxiously await the days when we have something like orbiting shipyards above Mars.
You mischaracterized the Challenger loss. NASA knew well beforehand that they should not launch in weather that cold... It wasn't some surprise to them.
You're not usually this dramatic. That was fun - fun enough for me to leave my first comment. Thanks for the great work (and I love my razor).
Glad you enjoyed it!
Videos are getting more and better
Thank you for the extra video.
this was SUPER INTERESTING!!! AMAZING TECH!!!
Thank You, Marcus!!
I always smile away every moment I watch. HEY HEY!
Thanks for including metric measurements 👍🏻👍🏻😎
Weather is yet another reason why launching in South Texas is better than Florida.
The weather gets pretty extreme there at Starbase too. I've learned a lot making this video.
In Florida, 2 starlink missions were scrubbed so I agree
Mind blowing!
So relevant and important and that you discussed the topic of weather, that shows your level of insight....well done !
Glad you liked it!
Falcon Heavy crying right now 😢
Marcus, you ROCK. You are my Saturday morning coffee go to. I love what you do so keep doing it.
Nice coverage of how weather affects launches. Could have been better without the built-in infomercial.
He gives us free content. Advertisement means he gets to keep doing that.
Awesome. Ironically just as us Aussies watch cyclone Jasper wipe out Cairns. No AI needed to predict that one!
The SpaceX provided top view of Starship 2 launching completely eclipses any previous footage from the Apollo or Shuttle missions. Unbelievable perspective with all that mist falling off the booster...
Marcus, I just received your Starship 2 launch polo shirt and love it. How about a version with an embroidered patch instead of silk screening. Would love to have one of those for durability.
Great video, as usual. Thanks
What an outstanding video. Well done! Weather is one of the things that humans can’t control, but this new technology is a giant leap forward in forecasting. This is yet another example of why Marcus has over a half a million subscribers. Thanks to you and your team for all your hard work.
I'm glad you enjoyed it. I know it is a little less related to the fiery end of a rocket, but I found it really fascinating.
I hope so. Because as they have increased the number of model runs, and real time models. The forecasting has not improved. I would hope this is different, however as someone who has studied models for several decades. There was a time when a forecast model out 5-7 days was better than they are 48 hours out now.
you wouldnt know that from all the people who insist HAARP is weather control. but then again, the earth is flat.
Excellent video as usual 👍👍
Great content, every day's a learning day...
Yes🎉🎉🎉🎉 2 episodes on a week
Heck yes. I am exhausted, but we are rapidly smashing out the video for Saturday as well.
I spent a morning with the met guys at RAF Valley and was fascinated with how they predicted weather fronts, especialy the rain that appeared during the visit. This technology takes it all to the next level and is amazing.
It is fascinating and I've learned a lot about about all this over the past few weeks.
Richard Feinman....
Figured it out.
Shuttle o-ring issue.
Administration push for the launch.
Another 21st century major advance in the science and engineering of launch operations technology. Congrats. Thanks, Marcus, for informing us of this new progress.
In the first 20 seconds I could have sworn Marcus said "that's the magic of rocket fart."
Took me three times to hear it correctly!
The weather might be why they are installing doors on the Mega Bays.
The best weather for the Launch us the Ground Hog 😊 😂 which is accurately set by the shadow 😅
Thanks!
Thanks Charles! Sorry for the delay. I missed this one at the time for some reason.
Hi Marcus, I often wonder with the high humidity and Boca Chica being so close to the sea, would rust ever be a problem both for the rocket and the surrounding infrastructure?
So, here’s the thing, Marcus, what part does AI play in the process. I want (need) to know how AI allows, aids, enhances, ???, the rapid iterations of the model and allows the model to provide finer detail.
Looking at those aerial shots of the SpaceX Texas launch and assembly areas, how will they cope with a hurricane storm surge in sea level?
I was interesting but it doesn't seem as well-received as usual. Maybe you could've dove deeper forward as well as backward. How much launching from an airplane reduces the effect of weather, and why Virgin Orbit went bankrupt anyway. How much more frequently does SpaceX actually launch because of this compared to going with traditional forecasts and being more cautious? Can we use AI to accurately model and then compensate for atmospheric distortions and eliminate the need for space telescopes altogether?! Just some ideas. Overall I learned new things and enjoyed the video as usual!
Nice job
Soooo... do Starlink Sats have Any weather tracking capacity?
Do any of the massive consellations?
Because if they do/could have real time wearher data...
You have to launch as close to the equator as possible. Tropical or Sub Tropical weather is built into the launch window.
Go team!
SpaceX Rocks !
That was really cool! I can see where this weather forecasting could be used for other industries as well as general forecasting. This is a great example that us space enthusiasts can point to when people ask why we are spending money on rockets. This is going to help us all
The weather network needs this as well! Starting to wonder why I still check them for the weather.... has been getting worse over atleast the past 2 years...
What are their accuracy curves wrt time-in-advance like?
It would make sense that future iterations of Starlink satellites carry weather sensing hardware for instant real time global weather information data to feed into AI.
I'll be back.lol😅 You can love me or hate, but I am going nowhere. Good show. Weather can be a killer. Slainte 👍
I'll hold you to that. 😍
Cheers mate.
Good morning❤
12 Houston can you go to poo? " Too late sir I already did that in my pants" 😅
Marcus, this is a great insight into how things are changing for the better using cutting edge technology. Give the way climate change is affecting the weather it will be pushing AI hard
I wonder when we'll get to the point we're weather won't matter. Similar to how ocean voyages can now largely ignore the weather.
There's a saying about Texas weather... "If you don't like the weather just wait a minute."
That’s all very well but can they forecast the weather here in the U.K.? Our weather is particularly fickle. If they can do ours then I might be impressed.
Next week we take off for a Viking Cruise in NZ and AU. We’ll be in Hobart the 28th. Can you get us a detailed forecast for then? ;-)
Deepmind also released their Graphcast model, that was just written about in Nature.
"we had an enjoyable visit to the rocket site. We don't go to all the sites because they aren't all glamorous. The warzones for instance."
What a super update! We all know how weather plays a role but to see the tech which goes into forecast in before a launch is mind numbing! Thanks again!
Weather has gotta influence landings too
100%
Holy crap! Is it Saturday?!?! How long did I sleep?
why doesn't star ship have two booster on the side , like falcon heavy?
Weather is "scary" ...OOooOOOoo
Soyuz: Hold my beer...
HAL9000 had a role in space travel too
RUclips explain the little bit more on how this are weather protection actually works. Because it's not like asking Chat GPT whats the weather gonna be like - But more like using individual nodes in the neuronal network to represent local weather at one point and the neurons being the interactions between them.
Also the real weather is the training data - because you could use yesterdays weather and prediction to reverse calculate form current weather
I always wondered who is in charge of long range weather forecasting at nasa !
When do you think we will see SpaceX start to "fit" nosecones etc with the internal structures and stuff that people will inhabit? Also we hear all the time on F9 and will do presumably for SH+SS "onboard computers taken control", where exactly are these computers located on the vehicles?
Did they ever say what all that soot was? Are they still using gas generators on the booster?
So in effect, the old adage "Know your enemy" applies here.
Current weather models are less science and more statistics. That is, they don't take into account water and energy going into or out of the atmosphere to create the mechanics of a working atmosphere and see what is going to happen. Rather, they take the readings for various points, and looking into historical records for a match withing a month or two of the date. Records that are close are compiled, throwing out any that end up on the edges of the probability curve. That is then presented as the "future model". It's like trying to drive down the road in a car while only looking in the rear view mirror.
100 times more detail DOES NOT equate to 100x more accurate.
In fact, exactly the opposite is true:
It actually means there is a 100x higher chance of the forecast being wrong.