If you are creating the world in a particular location in space, the first step is generally mapping out the neighborhood. Sometimes, particularly if you are creating a fantasy-type world or some other situation where the locals aren’t going to be too concerned about their position in interstellar space, it is completely unnecessary to map out to this macro level. For those whose stories would involve interstellar travel FTL or otherwise, the neighborhood can be of great interest. Others might simply find something of curiosity in this subject.
If you are situating your world somewhere near our own little planet, a good place to start would be Winchell Chung’s 3-D Starmaps site. There you will find links to local star lists such as Hipparcos, Gliese, and David Nash’s very useful HYG Database which mashes up stellar data on up to 120,000 stars within about 50 parsecs of our Sun. Once you have that data, Mr. Chung gives very good instructions on how to convert the right ascension, declination and distance information into cartesian coordinates useful for mapping or even galactic coordinates.
But what if your world is situated somewhere beyond the reach of local catalogues? Even many relatively familiar and visible areas, like the space within a few dozen parsecs of Rigel are only sketchily known. I am dubious if even a somewhat brighter than average star like the Sun would be visible at that distance and many small M-class red dwarfs are probably missing from local catalogues. Realistically, even if we could see all the stars in that region, distance measurements would be somewhere between formidably difficult and flat out impossible. What then of worlds set on the other end of the galaxy, behind the central bulge or even a long time ago in a galaxy far, far away? Even in the case of the “Rigel Sector” map probably all of the stars except Rigel, itself, could be imaginary, depending on the volume covered. Even the exact position of Rigel may be a bit conjectural, as distance measurements that far out are less than certain, generally based on making an(increasingly well-educated) guess about the absolute magnitude of the star and comparing that to the apparent magnitude of the star as seen in our sky.
This has been the subject I have been haltingly researching for the last few weeks. The first step is determining just how many stars I can expect to find within a given volume of space. My first source was an article in a very old Astronomy magazine that I have since apparently lost track of(boo-hoo). From an old program I wrote using that data, I was able to reconstruct that near the core the density of stars would have been about one per 52.6 cubic light years, or an average distance of about 2.3 light years between stars. This would give a probability of a star being within a given one cubic light year volume of 0.019. For the space near our own Sun that would be about one star per 270 cubic light years, or about 4 light years between stars on average. The local probability of a star within a one cubic light year region came to 0.0037.
This was some pretty old information and given pretty offhandedly. That and I lost my reference… Therefore, having failed to find my copy of the offending Astronomy magazine(if I’d found it more quickly, I probably just shrugged and used it anyway), I needed to do some more research. In the excerpts I have from C.W. Allen’s Astrophysical Quantities(3rd Edition), I found a chart on page 249 titled Number density in spectral classes which gave a total density of about 0.068 stars per cubic parsec. That works out to about 0.00198 stars per cubic light year. Hmm. That would be about one star per 505 cubic light years or an average distance between stars of about 4.94 light years. Am I getting this right? So the average distance between stars has gone from a little less than the distance from Sol to the Alpha Centauris to just a bit more. I particularly appreciate that the chart lists numbers of stars according to spectral class(by letter, you know the number of Ks but not K7s) and luminosity class(crudely: supergiants[0, Ia and Ib] and subgiants[IV] are lumped with giants[II, III]; early stars and sub-dwarfs are lumped with main sequence[V, VI and some freaks]). I don’t know if I would use this information in practice, I’m favoring the use of setting mass based on a good model of the stellar mass function and then adapting the Padova stellar evolution tables to figure out the parameters of a given mass of star at an age that can be determined as a uniform random range. In this case, I’d just be figuring out where the stars are at and separately determining mass and age and finally going back and plugging in the mass and age to get the other stellar parameters. Conceptually, this just seems easier to work, anyway.
I’m thinking the old Allen reference might be undercounting the little M-class stars just a bit. My suspicion is that with a better accounting for the dim little stars, these number densities would go up a bit, but all of those added stars would be piddling little things unlikely to harbor habitable planets. Although in my own imaginary universe, I like to find excuses to put more or less habitable planets around as many stars as I possibly can. Admittedly a planet orbitting a 0.07 Solar mass red dwarf will definitely be tide locked to it’s primary to receive enough insolation(instaration? inastration?) to be reasonably warm, that’s not necessarily a killer, and there would definitely be time for life to evolve if the star wasn’t to prone to superflares.
Another thing I’d like to work out is the prevalence of binary, trinary, etc. stars. I have a few sources presently, but they all seem to be game related! They also don’t all agree!?!
So at present, I’m looking to automate the first step, placement of stars. I have an old java program that basically just walks through a 3-dimensional grid in 1 unit(presumably light year in this case or possibly parsec) steps. At each step it places a “star” with a probability somewhere in the vicinity of 0.0037 or 0.00198. Over a large volume there should be a lot of random values and the actual number of “stars” in a given volume should form a pretty reliable bell-curve around the average. After it places “stars” at integer coordinate corners it then jiggles their positions by up to 0.5 units up or down at each coordinate, basically just a random translation.
I’d probably pick the higher probability value for another reason. I want there to be some human influence over the placement of stars. Perhaps I want to set up specific choke-points and rifts in the map by removing stars here and there. If a range of one star in 270 cubic light years to one star in 550 cubic light years is reasonable given current knowledge, then I could start with the higher density and go so far as to edit out up to about half of the stars without appearing too unreasonable.
So I guess at this point my research should go into finding a reasonable mass function for stars in the local region and maybe get an idea how that would vary in other regions. The second element is star age. Is that actually a uniform distribution? How does that vary from place to place? What fraction of star systems are multiple? What fraction are binary? What fraction are trinary? What fraction look like the Nightfall system 😮 ?
Once we know all of this and can generate stellar parameters based upon the given information we can then distribute planetary bodies orbiting our star of interest. I already have some ideas for this, but that discussion will have to remain for a future article.
Thank you for your attention,