Timeline of a Mars Settlement

[eacao]

Part I: Starships

It looks to be the case that a fully-reusable, superheavy launch vehicle is going to be operational this year or next. As everyone on this forum is aware, this spacecraft, which is to have 1,000 cubic metres of habitable volume and a throw-weight of 100-150MT, is fit for the purpose of establishing a permanent human settlement on Mars.

To envision how this capability might be employed, let's begin with two scaffolding schedules. The first schedule lists the dates of windows for Hohmann transfer orbits from Earth to Mars:

The approximate dates for a Hohmann transfer departing Earth for Mars:

21st November - 12th December, 2024
25th January - 14th February, 2027
27th March - 15th April, 2029
29th May - 16th June, 2031
30th July - 17th August, 2033
29th September - 17th October, 2035
28th November - 16th December, 2037
30th January - 17th February, 2040

The second, much more speculative list, has to deal with the production rate of Starships between now and 2040. To tackle this problem, we'll have to extrapolate based on Musk's stated aspirations and from proxy industries. In 2020, SpaceX iterated through 14 starship prototypes (SN3-17). Musk has stated that he aspirationally wishes SpaceX to produce a starship every 72 hours in the future -- around 120 per year. The primary bottleneck in Starship production is currently the rate at which raptor engines can be produced. The current Starship design is built on six raptors; future designs are intended to contain nine. Thus, building 120 Starships per year will require between 730 and 1095 raptors per year. Additional raptors will be required for the superheavy boost stages, though significantly fewer boosters will be required.

The 730 figure fits with the stated intention to build a raptor every 12 hours. On the surface, this may seem like an ambitious production rate, but let's consider the production rate of the Boeing 737 MAX and its CFM LEAP turbofan. In January of 2022, Boeing produced 27 aircraft for the month (a production rate of 324 aircraft per annum, requiring 648 turbofans). The 737 MAX production line was intended to grow to 47 aircraft per month (564 per annum) in Q1 2023, but fate intercepted that plan. For 2023, Airbus had set a production target of 65 A320s per month (780 per annum), to aspirationally grow to 75 aircraft per month (900 per annum) thereafter.

If the airliner industry is taken as a suitable proxy, then the production figures Musk has floated for the 50+ metre Starship are within normal. Let's also consider the production rate of Tesla vehicles as a far looser proxy for raptor production. Between 2013 and 2020, Tesla doubled its delivery of EVs every two years, now surpassing a million vehicles per annum. While a $230,000 Raptor 2 and an electric 4WD / sedan might not appear to be very analogous at the outset, this comparison is less related to the products themselves, as opposed to the factories and the Musk Group's experience with high-volume production of complex goods.

Thus, I will take three assumptions and plug them into a formula to estimate the number of Starships that might be available to service Mars by 2040.

(1) Every Orbital Window (OW), SpaceX will double Starship production each OW (every 26 months)
(2) Production will be arbitrarily capped at 1,040 vehicles per OW (equivalent to 40 per month).
(3) The 2024 production figure will be taken as one Starship per month (12 per annum). This is lower than the 2020 figure, but higher than the 2021 or 2022 figures.

We therefore arrive at the following figures for new-build starships:
2024: 12
2027: 24
2029: 48
2031: 96
2033: 192
2035: 384
2037: 768
2040: 1,040 (capped).

But this does not capture the entire fleet departing Earth during each orbital window. Starship is intended to operate on a circuit between the two planets so that the total fleet might accrete over time more quickly than simple new-build production figures would assume. So the question arises: "can Starships on Mars arrive back at Earth in time for the next Hohmann transfer orbit window, so that they can be added to the fleet that year?"

To answer this, we must list the Hohmann windows departing Mars for Earth.

3rd July - 22nd July, 2026
3rd August - 22nd August, 2028
4th September - 23rd September, 2030
4th October - 23rd October, 2032
4th November - 23rd November, 2034
5th December - 24th December, 2036
6th January - 25th January, 2039

We can see that the opening of a Hohmann window for a Mars-to-Earth journey comes roughly 7-8 months before the closure of an Earth-to-Mars window. This is cutting it close, but on account of the large ΔV budget available to a fully-fuelled Starship (which I am confident will be increased as the years roll on), I have decided to include returned Starships from the previous OW to the mix. However, early travellers to Mars are expected to hold onto their Starships for additional pressurised living space or for the steel in the hulls. Contrarily, the draw of large regolith sample returns might prove irresistible. So let's consider the very first manned mission.

I will select 2027 as the year for the first manned mission. This assumes that an unmanned expedition might be launched in the December 2024 window, which verifies a suitable landing site for the first visitors. In 2024, four starships depart for Mars -- two carrying an international payload of public and private scientific equipment, and two Starships carrying critical life support systems and spare parts (MOXIE reactors, a sabatier reactor, consumables, spare solar arrays, the raptors built into the vehicles themselves etc).

In 2027, four more starships launch -- two filled with important cargo necessary to host a crew for two years, and two carrying consumables and the crew itself. I won't specify a number of crew members for the 2027 mission, but we should assume the number will be small.

The reusability of Starship presents a funny quirk for the program. If the colonisation of Mars turns out to be a secondary market for SpaceX (and this is a dubious assumption considering the volume of public moneys that might be made available for a launch provider that can deliver an arbitrarily large payload to Mars), servicing Earth orbit and Artemis should require a distinct minority of new-build Starships. The rapid turnaround time for a Starship servicing Earthly customers means a small fleet might satisfy not only the demand generated by Starlink, but all launch customers on Earth (including Artemis). Thus, we may end up seeing the bulk of new-build starships being constructed for the express purpose of delivering payload to Mars. In the 2020's, it may be Government customers and private enterprises that foot the bill. As capacity grows, that may shift in favour of private citizens purchasing one-way-tickets to the tune of the average American net worth.

To generate a schedule of the total number of Starships available to launch in any given OW, I will assume that 90% of Starships return to Earth and join new-builds in the next departure. I will also assume that in any given OW, only 5 additional Starships will be added to the fleet reserved to service Earth orbit. Note that in the 2020's, these assumptions must be modified to account for ad hoc factors relevant to the first three missions. Thus:

2024:
(a) 12 New Builds
(b) 0 Returned
(c) 5 Starships servicing Earth
(d) 7 Available for launch (12 minus the 5 reserved for customers)
(e) 4 Actually launched (unmanned)
(f) 3 Non-reserved Starships not sent to Mars

2027:
(a) 24 New Builds
(b) 0 Returned
(c) 10 Starships servicing Earth
(d) 22 Available for launch (24 minus 5 reserved for customers, plus 3 that were not sent to Mars in 2024)
(e) 4 Actually launched
(f) 18 Non-reserved Starships not sent to Mars (15 Starships plus 3 from the previous OW)

2029:
(a) 48 New Builds
(b) 1 Returned (sample return)
(c) 15 Starships servicing Earth
(d) 62 Available for launch (48 minus 5 reserved for Earth, plus 18 that have not been sent to Mars)
(e) 20 Actually launched
(f) 42 Non-reserved Starships not sent to Mars (62 available for launch minus 20 launched)

2031:
(a) 96 New Builds
(b) 18 Returned
(c) 20 Starships servicing Earth
(d) 151 Available for launch (96 minus 5 reserves, plus 18 returned, plus 42 not previously launched).
(e) 151 Actually launched

2033:
(a) 192 New Builds
(b) 136 Returned
(c) 25 Starships servicing Earth
(d) 323 Available for launch (192 minus 5 reserves, plus 136 returned)
(e) 323 Actually launched

2035:
(a) 384 New Builds
(b) 291 Returned
(c) 30 Starships servicing Earth
(d) 670 Available for launch (384 minus 5 reserves, plus 291 returned)
(e) 670 Actually launched

2037:
(a) 768 New Builds
(b) 603 Returned
(c) 35 Starships servicing Earth
(d) 1,371 Available for launch (768 minus 5 reserves, plus 603 returned)
(e) 1,371 Actually Launched

Sum of Starships landed on Mars by 2038: 2,543.

This is clearly a rather dramatic figure. Let me explain some of the thinking behind my assumed figures for Actually Launched in any given year. In 2027, I assume that SpaceX permits only two Starships to carry crew aboard for the inaugural landing because I expect these to be a small and hand-selected astronauts drawn from a specialist SpaceX training program and likely astronauts drawn from NASA's (then ongoing) Artemis program. Having verified that this is possible, I imagine a substantially larger group may be hand-picked from the public for the follow-on mission in a vein similar to Dear Moon (crazy tourists, effectively) in conjunction with paying customers from space agencies, wealthy enthusiasts, and perhaps SpaceX-trained technicians tasked with the pragmatic job of establishing a base.

In 2031, I've assumed that tickets go on sale to private individuals who have watched the previous two landings over the past four years, and decide (for whatever reason) to go. From there, each consecutive window and the normalisation of space colonisation in the minds of the world's populations (it would be difficult over six years to not come to grips with the new reality) spurs on a growing (yet infinitesimal per capita) pool of applicants.

It is difficult to estimate properly the passenger-to-cargo ratio. In the past, Musk has floated the figure of 10 tonnes per passenger. For lack of a good proxy, I will stick to this. Thus, I'll assume an average passenger count per Starship of 13, accompanied by 130 tonnes of payload. That is not to say that every Starship carries precisely 13 crew and 130 tonnes of payload -- perhaps ≈half of starships will carry 30 passengers and the other ≈half are dedicated freight carriers (making better use of the un-crewed vehicle's cavernous fairing). After 2,543 launches, this would equate to a little over 33,000 settlers and 330,590 tonnes of supplies and equipment.

How is SpaceX expected to launch hundreds, or even a thousand Starships to orbit to take advantage of a two-week orbital window?
Answer: not all at once.
I am assuming that Starships will be parked in orbit over the course of the year / two-years leading up to the OW opening. These parked Starships will have no crew aboard. While in orbit, they will be tanked gradually by a steady (but manageable) cadence of refuelling missions. Only once the OW approaches will crewed launches ferry passengers up to the waiting (and fuelled) starships. Crew will become acquainted with life aboard their vessel, the life-support systems will be tested, and any issues that may arise can be resolved with a more relaxed schedule. When the OW opens, ships will perform their Mars-Transfer burns one-by-one so that a steady stream of Starships depart Earth orbit in sequence.

If 323 Starships depart Earth orbit between the 30th of July and the 17th of August 2033 (19 days / 456 hours), then that equates to less than one Starship per hour. I imagine the night's sky will look rather brilliant from the ground for those two weeks every two years.

[eacao]

Part II: The Colonies

Site selection is rather important. To colonise this new planet, the colonists must select locations with an abundance of accessible ice, building materials, volatiles, flat terrain, and mineral deposits. The two candidates that have risen to the top of the list are the Arcadia Planitia and the Utopia Planitia. These two sunken plains sit abreast one another in the northern hemisphere and are widely accepted to be the choicest landing sites for the first settlers.

Both sites sit atop great water deposits that may be dozens or hundreds of metres deep, cover an area the size of Texas, and hold concentrations of ice in the regolith of 50-90% by volume. Finding water on Mars is not likely to be challenging -- even for relatively large populations of tens-of-thousands to millions, along with their industry. Similarly, these sites show an abundance of ancient volcanic lava flows that might offer colonists plentiful access to basalt. Basalt can be melted and extruded into basalt fibres, woven into fabrics, and set with epoxies to form a composite material similar to fibreglass. Coupled with plenty of iron in the regolith and carbon in the air (together forming the all-important steel), building materials will be widespread and accessible provided sufficient power can be generated.

The combination of carbon and hydrogen permits the formation of hydrocarbons and thence plastics. Of particular interest is ultra-high molecular weight polyethylene (UHMWPE). Carbon can be sourced from the atmosphere through cryogenic distillation (freezing it). Hydrogen is obtained through the electrolysis of water. The Sabatier process produces methane from these inputs, which can be cracked into ethylene using steam. From here, a catalyst is used to polymerise ethylene into polyethylene or UHMWPE.

Why is this important? Because UHMWPE is an incredible tough and versatile industrial material that can be employed in any manner of necessary products from pressurised cabins, to wheels, to building-frames. When extruded and spun into Dyneema, it can be woven to manufacture spacesuits and other necessary equipment for colonists on Mars. Due to the high hydrogen content in UHMWPE, it also makes for an excellent radiation shield. While colonists are expected to reside beneath the regolith in subterranean tunnels -- dug by The Boring Company's TBMs (which are clearly designed for Mars), journeying the Martian surface for any extended period will require ample radiation shielding on the roofs of their rovers. UHMWPE is a great in-situ choice.

Things become a little bleaker when it comes to volatiles. Nitrogen and phosphorus in particular are unfortunately scarce on the planet. The wispy atmosphere contains around 2.7% nitrogen at least (making rapid expansion of the food supply a rather energy intensive endeavour), but phosphorus will be a cherished commodity. While phosphorus can be found generally in the Martian regolith, it is in lower concentrations than on Earth. Phosphorus is often found in sedimentary rock, particularly those containing clay minerals so while I'm not aware of any phosphorus deposits having been discovered and verified, evidence of past waterflows in the Arcadia and Utopian Plainitiae warrant further investigation and prospecting. Early missions (manned or unmanned) with a mind to colonisation will likely have to devote much energies to fixing the location of phosphorus deposits.

But site-selection should be a challenging albeit straight-forward process. The resources that a colony will need can be catalogued and from there, it's a matter of prospecting to find what they need.

Industry is imho a more interesting challenge. Below is an itinerary of some of the capital that must be shipped in to generate a self-sustained civilisation on Mars.

1. Resource development:
Excavators, regolith, sifters, smelters, dump trucks, compactors etc. Extracting the raw materials needed for industry will require the processing of great volumes of regolith. You should expect to see expansive mining operations on the surface, buzzing with automated collection and transportation equipment, moving regolith to and fro. Picture shallow open-pit mines on Earth, but on the Red planet.

2. Industrial production and manufacturing:
3D printers (fused filament for polyethylene prints or laser sintering for metal piece), CNC mills, lathes, injection moulding equipment, plasma, laser, and high-pressure cutters etc. A factory on Mars will look like a factory on Earth, only smaller, more generalisable, and optimised for low-production volumes of specialist parts. I'll be envisioning a hypothetical Martian factory later in this post.

3. Power generation:
Solar panels seem like the obvious choice for the earliest settlers. Some poor saps will have to head out to clean dust from acre-upon-acre of PV panels, but at least that keeps the oxygen flowing. Taking advantage of the abundant water below, I suspect hydrogen fuel cells will form the mainstay of power storage. But cargo weight is going to be at a precious premium and just as with Earth, turning to renewables plus batteries comes with an unviable burden in building overcapacity and storage. So for the colony to grow, it is bound to turn to nuclear power at its earliest convenience. Without a biosphere, and with the population already residing in radiation-insulated underground habitats, there are few if any downsides to nuclear power for a growing Martian colony. Other equipment, like heat exchangers, possibly Stirling engines, or even internal combustion engines that run on methalox might be brought-in from Earth initially, but with the establishment of factories and industry on Mars, these sorts of mechanisms can be built locally quite early on.

This post is TBC


/* plan

4. Life support and habitation:
TBMs, inflatable habs, electrolysis etc etc

5. Transportation & logistics
List vehicles

6. Scientific & med

- Factory and production process for Mars-specific materials.
- Ways in which the colony might pay for its imports from Earth
- Main colony & outposts
- Transportation between colonies
- Culture
- AI & robotics 2030's, 2040's.

Other remarks

edit (additional planning):
AI CAD will generate equipment designs on the spot. It will also figure out how the parts can be fabricated using available tooling, how it the parts can then be assembled, and will guide technicians through the production process if available robotics cannot handle the entire job (which they likely will be able to). Thus, 3D printers, as the foundational fabrication units will enable small workshops to (given a little time) build anything. Perhaps neuromorphic chips will be grown in a biologically-similar manner rather than fabricated via lithography. Perhaps by the late 2030's, equipment similar to nano-fabricators will allow colonists to produce microscopic high-quality chipsets, which can then be autonomously assembled into larger integrated circuits. Or, perhaps the plant & equipment for manufacturing quantum computers will turn out to be more transportable than today's SoC fab process.

[eacao]

Part I: Extended

While I initially intended to continue with Part II tonight, I instead decided to extend Part I a little.

I made a simple excel model of the hypothetical growth of the colony through the 2030's and 2040's.



Without describing every aspect of the model, I will highlight the key assumptions and figures.

I've already discussed in Part I the assumption that an inaugural (and largely experimental) crewed landing is attempted in 2027. I've also assumed a Dear Moon-style landing in 2029, which brings with it some of the most basic and foundational base-building assets in early preparation for a more orderly and industrious effort beginning in 2031. So I've assumed that the Starship design that we've just seen explode is going to be much-of-a-muchness with the vehicles that will be employed through the rest of this decade. Thus, I've assumed 100 tonnes of cargo (when fully refuelled in orbit) and 30 astronauts per vessel.

Starting in 2031, I've assumed that all new starships will be built to an upgraded varaint. This gen 2 starship can deliver 150 tonnes of freight to the surface or 40 colonists. I've broken the total Starship count into Crewed Starships and Freight Starships. The reasoning here is to strip heavy life support equipment and crew spaces from the freight variants so they might maximise their down mass and make the best use of the ≈1,000m^3 fairing volume. I have assumed that one-third of Starships are built to carry passengers and two-thirds to carry freight. Thus, in the early days, the 40 crew aboard a Crew Starship should be accompanied by 300 tonnes of freight aboard two Cargo Starships.

I've also assumed that a set number of Starship production will be reserved for the service of customers on Earth. I've assumed that this market grows at an average of 50% every two years. All other starship production can be freely devoted to the circuit to Mars. You can see "New Earth Fleet", which plots out the number of operating starships servicing Earth & cis-lunar orbit. Beginning in 2040, this fleet begins to decline as the Starship is phased out of production in favour of the MCT.

Beginning in 2033, I've introduced a larger spacecraft design, which I've notionally named the Mars Colonial Transporter (MCT). These are able to transport either 100 colonists or 350 tonnes of freight. I've taken these figures from the original Interplanetary Transport System (ITS) design of 2017. In 2019 (iirc), Musk discussed a future steel starship design with an 18m diameter, so I'm assuming ≈4-5,000m^3 of payload volume (though this does not appear anywhere in the model).

In this model, each MCT displaces the need for two starships. Thus, between 2037 and 2042, Starship production recedes to 0 while MCT production ramps up. Introducing the MCT in 2033 allows the maximum number of vessels (of both types) to peak at ≈1,000 in 2037, which already represents a stupendous logistical stunt wrt refuelling these vessels in orbit. However, this is the vehicle count that Musk has described so this is the peak number I've chosen to work with.

I've allowed the 'MCT' figure to continue growing exponentially through the 2040's. This does not necessarily mean that there will literally be 6,000 MCTs launched in the 2050 window. Rather, I'm using the MCT as a sort of proxy for all other launch systems operating by the end of the 5th decade. I have assumed that in the '48-'50 timeframe, there will be many interplanetary launch providers, a robust refuelling infrastructure on / around the Moon, perhaps nuclear thermal or other propulsion technologies, and an organic demand for transport to Mars from a wide and varied base of customers across the four corners of the globe. In these latter dates, it is the cumulative population ("Cum Population") and cumulative cargo ("Cum Cargo") that counts. It was a pleasant surprise to see 2050 mark the year in which the population of Mars surpasses 1 million, which is roughly the timeframe that Musk has floated iirc.

The last row, nestled down the very bottom, is the population of arrived colonists + natural population growth. I've assumed 1% growth every two years. In hindsight, this is low and should be adjusted upwards to 1-1.2% per annum. At any rate, it's a rounding error in the timeframe we're working with.

Lastly, I've also assumed that the cargo-to-passenger ratio will initially remain high (in the early 2030's), but with each consecutive window, can be weighted more heavily towards passengers as Mars' own industry begins picking up the slack.