Apollo 15 Expedition Science Project: Stellar Density Mapping

[Satsuma]

Stellar Density Mapping - How To Guide

Background - warning, very long thread

As part of the Apollo 15 Expedition, we're going to be doing another round of stellar density data collection over the course of the trip. The route follows along the Sagittarius Gap for about 10 kly before crossing it, and I think that could provide some interesting information on how stellar density changes between arms in the Elite Dangerous.

This project is an extension of work done by Bald Eagle, Jackie Silver, marx, MattG, and Rollo Rhadium, measuring the change in stellar density (Scale Height) of the galaxy in different places. This can be used to generate a map of stellar density, and has implications with route finding and exploration.

The technique for measuring this is simple: Fly a column stretching from the galactic plane "up" about 1000 light years, measuring the stellar density ("rho") along the way. For a single column, this generates a graph like this:

Graph by Jackie Silver​

This project is especially suited for console players, because it only uses the galaxy map and the nav panel.

Important: I'm not sure if the nav panel under Odyssey always shows the same number of stars as the nav panel under Horizons, so for now let's make sure to do these scans in Horizons!

How to perform a density scan

Before you begin, you'll need some way to record the values. Google Spreadsheets is ideal. I use a single spreadsheet, with a sheet tab for each density scan "column".

Open the galaxy map. You'll see coordinates on it that looks like this: "-390 : -27 : 4,280". The important one for us is the middle one, which measures "height" above the plane of the galaxy, or "Z". Pick a star on the map where Z = -20 and fly there. IMPORTANT: CLEAR YOUR AUTOPILOT ONCE YOU ARRIVE! SETTING THE AUTOPILOT AFFECTS WHAT YOU SEE ON THE NAV PANEL, WHICH ALTERS THE RESULTS OF THE SCAN!

Now you'll look at your nav panel. You'll see that it shows systems nearby. It tries to show systems within 20 light years, but if there are more than 50, it only shows the 50 nearest, including the system you are in. So the first thing you want to do is count them. If there are 50, then you are in a dense area of space. If there are less than 50, you are in a sparse area of space. An easy way to get a system count is to set the nav panel filter to show only systems, adding 1 to the count of systems (to add the system you are in).

If you are in a dense area of space, you will want to note the distance shown in the last system in your nav panel, which we'll call "r". This lets us estimate the density as rho = 50 / ((4pi/3) * (r^3)). If you are in the FleetComm discord, you can use the bot Jaques to calculate rho with the command "@Jaques rho_dense <r>".

If you are in a sparse area of space, you will want to note the number of star systems in the nav panel (including the system you are in), which we'll call "n". This lets us estimate the density as rho = n / ((4pi/3) * (20^3)). If you are in the FleetComm discord, you can use Jaques again, with the command "@Jaques rho_sparse <n>".

Neither method is perfect, but in practice they're good enough for our purpose.

Record this number in your spreadsheet, along with the system name. In my spreadsheets, I also include coordinates, but those aren't strictly speaking necessary. If you would like to include coordinates, EDDiscovery can give you those for the system you are in.

Now fly 50 light years up, to Z = 30 and do the same thing again. Keep flying up, repeating this process of gathering data and calculating rho until you reach Z = 980. Congratulations! You've done a density scan column! It's best to do them as close to vertical as you can. If you're in "proc gen" space, it's easy to tell if you're flying vertically enough because all the systems will start with the same prefix ("Traikaae", for instance).

A vertical density scan column. FLY THIS MANUALLY, DO NOT USE THE AUTOPILOT!

A completed set of density scans​

It takes about an hour to do a full column, so it's not necessarily something you would want to do too often. I usually do one every day or two, whenever the whim strikes me. Once you've assembled a few weeks worth of scans, send me a link and I'll include it in the analysis. For Science!

 

[Satsuma]

We're nearly a week into the Apollo 15 expedition, and we've done seven density surveys. One thing that surprised me was how low the scale height is in the areas we surveyed. Blu Thua has a scale height of 80 light years, for instance, compared to ~ 1000 light years for the thin-disk scale height of the real Milky Way. A bunch of us did density scans during Distant Worlds 2 and determined that the scale height outside the core is typically around 300 light years, so this is very low even for the Elite Dangerous galaxy.

Most of these surveys happened very close to the bubble, which makes me wonder if the low scale height is an artifact of the area. Jackie Silver documented zones of suppression that run directly "north/south" and "east/west" of the bubble, where the formation of bright high mass stars is reduced. I wonder if the scale height of these zones is also compressed. We're headed out of the zone of suppression soon, so we'll see if we see a sudden jump in scale height.

An example: Blu Thua, with a scale height of about 80 light years.

A sample from Distant Worlds 2: Byeia Chrea, with a scale height of about 300 light years.

 

[Satsuma]

A few weeks into the expedition and we're trolling along the interarm region between the Orion spur and the Saggitarius-Carina arm. The density scans have continued to be puzzling, with an extremely small scale height and generally low densities. One thing that's shown up in all the scans is that density is highest at the -20 LY reading, which made me wonder if the density is actually peaking further to the left than we are scanning. I did a wider scan and learned that density in the area peaks around -40, with an almost cliff-like drop in density.

This led me to change my methodology to be a center-weighted one straddling the galactic plane. The drop in density from the peak is so rapid that I've been taking density readings every 10 light years out to about += 80.

Here is a scan of Swoidai by CMDR Getedoi, clearly showing the extremely steep density peak. One of the things that's interesting about this is that density peaks not in the plane of the galaxy, but at -40 LY. Both the low density and the skew to -40 puzzled me. Since Sol is at 0 LY, however, I wondered if this skew simply reflected the fact that Sol isn't actually in the plane of the real life milky way.

Unfortunately the data didn't bear this out. The density around Sol peaks at 0 LY, at around 0.0045 stars per cubic light year, which is pretty much what you would expect with the real life galaxy. The scale height is still very low (~100 LY) compared to the real milky way (~1000 LY), too, which is also puzzling. At the same time, all the stars around Sol are most likely hand-placed, so this doesn't necessarily reflect what things look like in procgen space in the arms.

I'm starting to think that the low density we're seeing in the inter-arm region reflects how Elite Dangerous models the galactic arms. Unlike the real galaxy, where the arms are thought to be a sort of standing wave in the interstellar medium, Elite Dangerous appears to be modeling the arms solely with stellar density.

The expedition is due to head into the Centaurus Reach, in the Sagittarius-Carina arm, so I'll be curious to see what things look like there.

 

[Orvidius]

Interesting. Yeah, I've been considering the actual plane to be somewhat below Sol, at around -25, since this is where the sector boundary is, and the star density often seems to reflect a similar plane altitude as well. Granted, the larger data set from EDSM/etc has an awful lot of selection bias baked into it. But my graphs tend to show the peaks a little below the Sol/zero line, such as this one:

 

[Satsuma]

It would be interesting to see if that varied by position in the galaxy. The real milky way warps...

I'm kind of hoping this models that warp...

 

[Satsuma]

I've been doing some number crunching all through the Apollo 15 Expedition, and now that it's over I'd like to present the data that we've collected.

During Distant Worlds 2, we collected a lot of stellar density data using the original methodology (running from -20 to +980 LY in 50 LY increments). The density plots from DW2 match the general shape of the density plots at the start of A15X, and in retrospect I should have changed the methodology back then. Even still, the DW2 data showed that the scale height of the galaxy is small compared to what you'd expect from the real Milky Way.

The Apollo 15 Expedition data started to show just how strange the Elite Dangerous galaxy was, structurally. Data collected from the interarm regions showed a characteristic knife-edge of peak density, with stellar density falling off extremely rapidly.

Swoide (courtesy of CMDR Getedoi) is a good example of this, with a scale height of about 40 LY. These interarm scans also showed that the density dropped down extremely low, remaining fairly flat from that point on. The inter-arm regions were clearly modeled by dialing the stellar density way down. I suspect the knife-edge peak is there to provide a "membrane" of stars which can easily be transited. This agrees with my memories of what it was like to cross the interarm gaps in a 22 LY range Cobra Mk III in the pre-Horizons days.

I'd done a density scan of the neighborhood of Sol (read up a few posts) and saw a much fatter density peak, agreeing more with how we view the galaxy. Since most of the stars there are hand placed, though, I wanted to compare them with proc gen regions inside the arms.

This scan was taken in Prooe Phio/Oofaih, in the Norma Expanse, by CMDR Cpeterk24, shows a typical stellar density graph inside an arm. The overall density is much higher, but it still shows the knife-edge peak and rapid fall off, with the scale height under 100 LY. One interesting thing from this is the second peak at +30 LY. This looks like it fits with CMDR Orvidius' data, and it makes me wonder if it's showing another boxel boundary.

But here are the money shots: The combined graphs for all the scans we've taken. These graphs are all arranged with the X axis showing distance from Sol. I'll start with scale height, with two graphs. The first simply shows distance from sol:

The second shows the same thing, but labeled with the names of the regions the data came from:

The "Interarm" and "Arm" labels indicate whether the data came from a region within an arm or between arms. There doesn't seem to be a strong correlation between scale height and the placement of the region within an arm.

Next comes peak stellar density:

There does seem to be a correlation between peak stellar density and the presence of the region within an arm.

Lastly, with the help of a bottle of bai jiu and a shaky memory of calculus, I tried to estimate the number of stars within the scan column*:

This also shows a correlation between star count and the presence of the region within an arm.

*The star count estimate was taken by finding the area under the curve (giving a total number of stars per cubic light year for the whole scan column) and multiplying it by the area of a 20 LY radius sphere (which is the maximum volume displayed in the nav panel). I might be totally off with this, but it did give me an interesting number.

One thing I've been noodling over is how there could be 400 billion stars in the Elite Dangerous galaxy. The real life Milky Way is about the same size as the Elite Dangerous galaxy, but it's very thin compared to the real Milky Way, so it didn't make sense that there could be the same number of stars. The stellar density near Sol is about 0.004 stars per cubic light year, with a scale height of ~1000 light years) and we're very close to the galactic plane, while the interarm values in the Elite Dangerous galaxy were nowhere near that. Looking at the hand-placed Sol scan column, I'm estimating about 1000 stars within it. It's clear, though, that there are regions in the Sagittarius-Carina arm that exceed (sometimes greatly) this star count, and which could explain what's going on.

I'm planning on heading towards the core and doing more of these scans, because I really want to figure out what's going on with this.

In any case, I would like to especially thank the following CMDRs for gathering data throughout this expedition and helping to shine a light on the weird nature of the Elite Dangerous Galaxy:

• bioxlapatsa
• Cpeterk24
• Enigmatic
• ewanspence
• Getedoi
• Gnauty Gnate
• Jaennics
• Kacper Fleszar
• Poutnik Santiago
• Skytoucher

The raw data for these scans is being kept in a Google Sheet and is available for anyone who would like to use it.

 

[Orvidius]

Man, this is pretty cool. Actually that's an interesting point about the "400 billion star systems" in the galaxy. It's something I've always wondered about. FDev uses that as the estimate themselves, but I've always wondered whether they assumed that the StellarForge was hitting that target, or if there's some sanity checking that enforces it, or whatever. Since the procedural generation is sort of an octree structure that inherits mass downward, and they use a mass-density image map to seed it, do we really know for sure that there are about 400 billion systems in the StellarForge? That's an interesting question.

 

[Yanick]

This is a genuinely excellent project, and one that very much surpassed my expecations.
A big thank you to CMDR Satsuma and all the contributers to the A15X Science Project!
o7

 

[Nax Wren]

Neat study! One thing I have also wondered about is how many systems are actually in Elite Dangerous. Currently my view is there probably is 400 billion stars in the game. That does not mean 400 billion systems. Theres uncountable amounts of systems with multiple stars. Then there are nursery systems that have dozens of young stars. All of these secondary stars could make up that 400 billion number.

 

[Maolagin]

Very impressive indeed! Jives well with the observation that in the outer ~1/3 or so of the galaxy, the neutron layer is tenuous or missing in the interarm gaps. With low scale heights and such depressed base density, it makes sense that there's just no mass left by the time you get to +/-1 kly.

I wonder, has anyone tried to do a similar density study focusing on distance from the galactic center? Seems to me the radial density profile is the other big ingredient you'd need to check the 400 gigastars figure.

 

[MiloGreenRanger]

Nice project and nice methodology!

I've thought about doing some sample-based estimates for the distribution of stellar density - e.g., pick a nice experimental design and sample 20-50 ly volumes throughout the galaxy. It would be direct sampling like you are doing, rather than the statistical inference from the density estimates that rely on reported system counts (which is already plenty clever).

Good luck, CMDR. o7