FINALLY. I am proud to announce that fnstraj.org is online! After a year and a half of abstract tweets, strange Stack Overflow questions, dealing with NOAA GrADS, and half-sleepingly typing on my screen, hoping something would happen, I am so happy to have finally put this gem on the internet. Is it done? no no no, fnstraj is alpha-quality software, but it does work (usually). There are some outstanding bugs that I’m still stuck on, but they aren’t blocking functionality enough to hold back anymore. I haven’t flied with fnstraj yet, but its release will certainly give me a reason to change that.
So what is fnstraj?
Good question. It’s trajectory prediction and analysis software for high-altitude ballooning. You’ve probably heard the term “near space” or “space balloon” to describe the expeditions of our newest citizen scientists. This tool is to help them figure out where their GoPros are (or broke at). fnstraj does this by getting some data about the balloon, getting some data about the launch site, and looking at weather information all over the sky (it exists, by the way). Once it’s all put together in the calculator, you get a nice map with the predicted flight path on it. Does it work? Is it accurate? Can’t say numbers just yet, but predictions are lining up with other known predictors well enough for me to trust it.
There are some other standout features; like multiple weather model support (not all at the same time, but still), planned SPOT GPS Tracker integration, nice design, works on your iPhone, etc.
So can I try it?
YES! Please do. I can’t promise it will work everytime just yet, but if it does, you get a good idea of if sending something to near-space is practical to you. Remember that the Jet Stream moves around a fair amount, so where things land is dependent on the day, daylight, and the season. Can you fly with it? I’d make sure to pack a proper GPS system and verify it works, but if you see this and get a report, please do and let me know your experience!
So, what is wrong with it then?
Right now fnstraj isn’t running on something called “real physics”, meaning ascent and decent rates are using a constant value of 5m/s instead of predicting the ascent and descent rates based on balloon and parachute parameters. This isn’t necessarily a bad thing, other trusted predictors do the exact same thing. But I want fnstraj to be different and take those values a little more seriously. I may need to do some more research on Reynolds Number to verify it’s not harming that however.
Another issue fnstraj is experiencing involves GFS and GFS HD not working in certain hours of the day. I haven’t run a test suite to get the exact hours of the day it doesn’t work, nor do I know what the fix is just yet. I really want it fixed, so as soon as I find a lead, I will have that done.
SPOT support is shown as working on the fnstraj create flight page, but that is not the case just yet. This is coming eventually (I have a personal need for it, so there’s that), but the queue needs to be refactored and I have to decide a friendly way to handle SPOT support on a singular queue (since I can’t afford to pay for more worker processes right now).
Oh. And I don’t know if it works on the west coast yet. If something looks wonky, please screenshot and email me.