I'm flying over Spain right now, and every interpolated station has wrong winds aloft. Southerlies rather than Northerlies. Every non-interpolated station has the right weather. I think this looks like a simple mistake in the calculation.It looks like HiFi programmers are taking simple averages of the directions for the winds. So, if you have winds from 020, 040, 340, and 320, you'll get 180 as the average. I suspect these are weighted by distance, but you'll still get the massive error, like what I'm seeing in the program.This is a pretty common mathematical problem (vectors in trigonometry), with a very easy solution. But could this be the source of the error in interploated winds aloft?BrynUsing Build 642 with FS9.

I'm flying over Spain right now, and every interpolated station has wrong winds aloft. Southerlies rather than Northerlies. Every non-interpolated station has the right weather. I think this looks like a simple mistake in the calculation.It looks like HiFi programmers are taking simple averages of the directions for the winds. So, if you have winds from 020, 040, 340, and 320, you'll get 180 as the average. I suspect these are weighted by distance, but you'll still get the massive error, like what I'm seeing in the program.This is a pretty common mathematical problem (vectors in trigonometry), with a very easy solution. But could this be the source of the error in interploated winds aloft?BrynUsing Build 642 with FS9.

If only it were that simple... In addition to distance, you have to take pressure gradients, temperature differentials, humidity, and several other things into account to come up with even the simplest approximations and even then you are guaranteed to be wrong a great deal of the time. I spent several years on a NASA sponsored program back in the 70's in an attempt to model weather over a relatively small area and the conclusion was that, for many reasons, you really can't - the atmosphere looks more like a boiling pot of water than a smoothly flowing river. After many man-hours and a lot of big computer time we showed that any attempt to model weather at a point was going to fail dismally even if the inputs were coming from data generated relatively close to that point. Things smooth out a bit as you get higher into the atmosphere, but I think looking at wind trends along a flight path will give you better results for our purposes (even though not 'real') than attempting to do some sort of nearest reporting point matrix.DJ

I hope you are working on it. Because it can run well.But now it is really a nightmare to fly online with other pilots using AS6.5. Groundspeeds can bei 100kts different with same airspeed.

DJ, I understand what you're saying, but my point is that there is an error in the very simple approach that Hifi uses.They take four datapoints (one to the NW, SW, NE, and SE) and average those to get the interpolated weather (possibly weighting by distance). HiFi does use a nearest reporting matrix... but my hypothesis is they do the calculation wrong.If you take a simple average of the wind directions, you will get the wrong average wind direction. This is the key point... the mathematics here is wrong, despite it being a simple approach.I think the correct calculation is to convert to radians, take the sum of the sines of the angles and the sum of the cosines of the angles and take the inverse tan of the sines divided by the cosines. And then convert back to degrees.Or in excel: atan2(cosines, sines)Or in c++: atan2(sines, cosines)But the programmers would have to check this calculation. Frankly, though, if this is the source of the error that's plagued upper wind interpolation, someone should be having a word to them about what they're doing writing interpolation algorithms if they can't get the basic maths right. Meanwhile, someone should have a word to me about my spelling...

I just shutdown ASE because of this error and switched to my old AS6 version. Winds aloft in NorCal and SoCal are out of the north at all levels today; however, an interpolated station just outside KMRY led to the ASE winds becoming southerly at all levels except for the layers above 34000. This bug is very annoying, I had it occur another time in the Seattle area and every time an interpolated station came into play, the winds would start changing by 180 degrees--and it happened many times as it switched to "real" stations. If I remember correctly, once again the winds were supposed to be out of the north, but interpolated stations would have southerly winds. This problem, at least in my case, always seems to come in to play with northerly winds. I really hope this can be fixed as it detracts from ASE for me.Best,Graham

I think ASE can calculate. Because the problem is a new one and it worked really well for a time.I think its something with the servers. If I refresh the weather sometimes it works, sometimes not.What I cannot accept is that HifiSim is always looking for the problem at the user.Thay should say, we are sorry about the problems and try our best to solve it. That is not happening.

Hi,I totally agree with you, they must look into this, this is a bug....Christian

DJ, I understand what you're saying, but my point is that there is an error in the very simple approach that Hifi uses.They take four datapoints (one to the NW, SW, NE, and SE) and average those to get the interpolated weather (possibly weighting by distance). HiFi does use a nearest reporting matrix... but my hypothesis is they do the calculation wrong.If you take a simple average of the wind directions, you will get the wrong average wind direction. This is the key point... the mathematics here is wrong, despite it being a simple approach.I think the correct calculation is to convert to radians, take the sum of the sines of the angles and the sum of the cosines of the angles and take the inverse tan of the sines divided by the cosines. And then convert back to degrees.Or in excel: atan2(cosines, sines)Or in c++: atan2(sines, cosines)But the programmers would have to check this calculation. Frankly, though, if this is the source of the error that's plagued upper wind interpolation, someone should be having a word to them about what they're doing writing interpolation algorithms if they can't get the basic maths right. Meanwhile, someone should have a word to me about my spelling...

I realy think you have nailed the problem here Bryn.

Hi All,Can someone explain how then at times people in this very thread can go weeks without seeing this?If it was a calculation error wouldn't happen at every single ADV INT station, every day, with every weather report?FYI, we are in the process of looking into this.

Hi All,Can someone explain how then at times people in this very thread can go weeks without seeing this?If it was a calculation error wouldn't happen at every single ADV INT station, every day, with every weather report?FYI, we are in the process of looking into this.

I believe it only shows itself when winds are northerly with some stations reporting something like 020 and another 340, etc. I saw the behaviour again today in an area with northerly winds. My route Oakland to Jackson Hole--the western US is currently a good place to look for this problem as the winds are predominantly northerly right now, and are forecast to be that way for at least the next 48 hours.

Hi All,Can someone explain how then at times people in this very thread can go weeks without seeing this?If it was a calculation error wouldn't happen at every single ADV INT station, every day, with every weather report?FYI, we are in the process of looking into this.

Graham is right, this type of error is going to show up when northerlies are prevailing, because that is when you might get stations reporting 340 to 020.

Graham is right, this type of error is going to show up when northerlies are prevailing, because that is when you might get stations reporting 340 to 020.

That is a good finding. Checked my screenshots and agree to that.

Hi,Please note that we have made some changes and things should be better now. Please re-post if they are not.

Hi,You might have done some fixes. The problem is alive. See this navlog. Within several miles the wind should turn by 180 degrees ?!?And the bad message: ASV6.5 did similar things. Not at this place but in a similar condition 300 miles north of this place. At one point 40 nm from south and at the next with the same speed from the north. Very unrealistic behaviour.You definitely have not fixed it until today.RgdsReinhard

Hi,We have checked the server this afternoon and everything is working correctly. We would need log files in order to look into this further.