Meteor detection: path from Scotland to GRAVES

[The Admiral]

56 minutes ago, Stub Mandrel said:

While traces that drop past the transmission frequency are not uncommon, I don't see many that are all below it (i.e. whole track heading away from the line joining me to Graves). I can't explain this.

I know what you mean Neil. It is strange that no meteor strikes seem to have an increasing frequency. You would think some would. I can't get my head around the geometries involved.

The overall Doppler shift will be the sum of shifts from Graves to target and target to receiver, using the component of velocities in your direction. For that reason, where the track is going perpendicular to the line from you to the target there will be no Doppler shift from the target to you (i.e cos 90° = 0). Either side of that point of course, the cosine is < or > 0, and there will be a shift. Similarly, where the target is moving at right angles to the line between it and Graves, the Doppler shift from Graves will be zero. These points will be at closest approach to either you or Graves.

Hmm, I'm not sure what I'm trying to explain here, or indeed, if you even need an explanation! Either way, I'm making a bit of a fist of it!

Ian

[Stub Mandrel]

20 minutes ago, The Admiral said:

I know what you mean Neil. It is strange that no meteor strikes seem to have an increasing frequency.

Yes, one moving away from both transmitter and receiver should be shifted down, but have a reducing doppler shift as it slows.

A thought - maybe one heading away has the doppler shifted 'head' signal shielded by the static 'trail' which produces a signal with no shift.

[IanL]

Here us how I see it, but still trying to think and read my way around it.

Start with one component, say meteor to receiver. The only way there can be zero doppler shift is when the meteor is neither moving towards or away from the receiver. In the case of a moving meteor, for there to be a zero shift continually it would have to follow a circular path with the receiver at the centre, i.e. staying at a constant distance.

Clearly that is not the case, as meteors follow straight paths. In the case of a meteor travelling directly towards the receiver there is a doppler shift to a higher frequency, and directly away a shift to a lower frequency proportional to its velocity. Where a meteor is traveling on a path perpendicular to the to the 'direct' (towards/away) path, there will be a shift to a higher frequency as the meteor approches the direct path and then to a lower frequency after it crosses it.

If you measure the radial distance from the receiver to the meteor at points along that perpendicular path, you will see they get shorter until minimum distance at crossing the direct path, then get longer after it. Only for the moment the meteor is crossing the direct path is it neither moving towards or away from the receiver (i .e. at minimum distance), and at that point there is zero shift. 

The same is true for the transmitter to meteor component. Basically treat the meteor as the stationary reference point and calculate the shift due to the 'motion' of the transmitter from the meteor's perspective.

Upshot is that neither the transmitter to meteor component nor the meteor to receiver component can ever be zero except momentarily at closest approach.  This is known as the radial velocity, and is proportional to both the velocity of the meteor and the angle of its path relative to the transmitter or receiver.

Of course you have to combine both radial velocity components to get the net shift at the receiver. I assume there must be paths where one component cancels the other, but whether real meteors can make those geometries relative to us and Graves I do not know.

[Stub Mandrel]

11 minutes ago, IanL said:

I assume there must be paths where one component cancels the other, but whether real meteors can make those geometries relative to us and Graves I do not know.

Only paths parallel to a line joining the receiver to Graves, unlikely, but many paths will be close enough for the shift to be hard to discern, given most traces seem to be a few hundred hertz wide.

[The Admiral]

19 minutes ago, Stub Mandrel said:

Only paths parallel to a line joining the receiver to Graves, unlikely, but many paths will be close enough for the shift to be hard to discern, given most traces seem to be a few hundred hertz wide.

I think this is generally only true when the target moves along the direct path between transmitter and receiver, where the components of velocities are identical. In practice, the target moves at altitude, and I think the cancellation of Doppler shifts will only be transient. That's my gut feeling anyway, because the transmitter, receiver and target form a triangle, and as the target moves, the component of velocities vary according to trig functions, even if the paths are parallel to the line between Tx snd Rx. I'd need to work through thd maths to be sure.

Ian

[Stub Mandrel]

10 hours ago, The Admiral said:

I think this is generally only true when the target moves along the direct path between transmitter and receiver, where the components of velocities are identical. In practice, the target moves at altitude, and I think the cancellation of Doppler shifts will only be transient. That's my gut feeling anyway, because the transmitter, receiver and target form a triangle, and as the target moves, the component of velocities vary according to trig functions, even if the paths are parallel to the line between Tx snd Rx. I'd need to work through thd maths to be sure.

Ian

I suspect you are right. The maths are beyond me these days...

[The Admiral]

6 hours ago, Stub Mandrel said:

I suspect you are right.

I don't like to be defeated .

I plotted the Doppler shift for the ISS at different distances from the receiver and transmitter (in plain geometry, not spherical geometry! The earth is flat after all ), passing overhead both the receiver and transmitter, and the total Doppler shift was calculated (shown in green below).

What does this mean? Imagine the ISS is approaching the receiver from > 5000km away, and 1000km beyond the receiver under the flight path is the transmitter. Any radio waves reflected off the ISS will be Doppler shifted; initially the frequency will be raised because the ISS is approaching, but the shift will continue to get less until it is overhead (or where the line of sight is perpendicular to its motion), at which point the frequency shift will be zero. As it progresses beyond the receiver, the Doppler shift will be negative. This is the blue line.

At the same time the ISS is approaching the transmitter, so the frequency the ISS sees (plotted in red), and that which will be subsequently reflected, will likewise be raised above the transmitted frequency until it passes overhead the transmitter, at which point the reflected signal will be at the same frequency as the transmitted frequency. Beyond the transmitter, it will be below the transmitted frequency.

What the receiver therefore perceives is the combination of these two effects, shown as the green line. Note that at some point between the receiver and transmitter the total Doppler shift is zero, where the positive Doppler shift of one origin is counteracted by a like shift but of opposite sign from the other origin.

So, if I've calculated it correctly (and I have had a number of stabs!), there you have it.

Ian

 

[Stub Mandrel]

2 hours ago, The Admiral said:

there you have it.

Chapeau!

[The Admiral]

4 minutes ago, Stub Mandrel said:

Chapeau!

Actually, I think there is something not quite right, but the principle still stands I think. The green line should pass through zero shift when the two shifts add up to zero. That clearly isn't the case here!

Ian

[BiggarDigger]

That's a really good graphical representation of my ramblings the other evening Ian.

The two components will constructively or destructively combine.

I think the "something not right" may be due to the orbital path being vectors towards or away from both transmitter and receiver that are not either perpendicular or in line with both sites.

In these cases (which occur more often than not) the velocity towards one site may be larger than the velocity away from the other, or vice versa.

 

How or why this appears to result in significantly more negative than positive shifted Doppler traces from head echoes remains beyond me at present.

Richard 

[Stub Mandrel]

41 minutes ago, BiggarDigger said:

How or why this appears to result in significantly more negative than positive shifted Doppler traces from head echoes remains beyond me at present.

I see more positively shifted echoes...

For the ISS It may partly depend if it's on the southbound or northbound leg.

I tend to see it dropping from high positive to low negative.

[The Admiral]

57 minutes ago, BiggarDigger said:

I think the "something not right" may be due to the orbital path being vectors towards or away from both transmitter and receiver that are not either perpendicular or in line with both sites.

I think you're reading more in to this than it deserves Richard. I started off wanting to illustrate the effects of the two Doppler shifts, and I thought that rather than trying to draw the curves it'd be easier to plot them with Excel.  I'm quite happy with the red and blue curves, but something is not right with the way I'm summing them to get the green curve. The zero net shift should occur mid-point between the stations, in this idealized example. My brain isn't working as I need it to! Age, perhaps .  Also, I've taken the track of the ISS to be over both the receiving and transmitting stations, for 'simplicity'.

Ian

[The Admiral]

12 minutes ago, Stub Mandrel said:

For the ISS It may partly depend if it's on the southbound or northbound leg.

I tend to see it dropping from high positive to low negative.

Surely it will do, as it always approaches from the West (i.e. NW, W, or SW), resulting in high shift where it can be observed (Max 7.4kHz), moving towards both us and the transmitter, and departing in an Easterly direction, moving away from both us and the transmitter.

The ISS is easy, as we always know its direction, but I find meteors confusing as the head return is always high to low, where it can be seen, even though it's not actually slowing down, and I can't visualize its track.

Ian

[BiggarDigger]

12 minutes ago, Stub Mandrel said:

I see more positively shifted echoes...

That's rather curious Neil.

Have I understood your description: By positively shifted, do you mean the frequency of the head echo is increasing before the long tail trails out in time?  If so that's opposite to the majority of my recordings.

I do see quite a few with very little or no discernable Doppler which appear to be similar to so called underdense reflections. I see a very small number of reflections where the head echo frequency is small in magnitude and increasing, but the majority of my recordings are where the head echo frequency is significant in magnitude and reducing at very high speed.

If I've not misunderstood your description of your positively shifted scatter, I wonder what causes that?

Off the top of my head, could it be because I appear to be receiving forward scatter off the rear lobes north of the radar....could you be receiving predominantly backscatter from South of the array, and if so, would that account for a reversal in Doppler shift compared to my reflections?

 

20 minutes ago, The Admiral said:

I think you're reading more in to this than it deserves Richard. I started off wanting to illustrate the effects of the two Doppler shifts, and I thought that rather than trying to draw the curves it'd be easier to plot them with Excel.  I'm quite happy with the red and blue curves, but something is not right with the way I'm summing them to get the green...

Ahh, sorry, I assumed it was a different model.  Yes summing the two curves should result in a symetrical zero point.  It looks like there's an offset in the green curve.

Richard

[Stub Mandrel]

I've shown this before, a meteor that passes through 'zero shift' with a stationary ion trail echo.

 

[Stub Mandrel]

2 minutes ago, BiggarDigger said:

Have I understood your description: By positively shifted, do you mean the frequency of the head echo is increasing before the long tail trails out in time?  If so that's opposite to the majority of my recordings.

No, I mean starting with a positive doppler shift that then  drops towards or (as in the above example) past no shift.

I don't quite understand what you mean by " I see a very small number of reflections where the head echo frequency is small in magnitude and increasing, " which suggests teh target is speeding up.

I suspect we we both know what we mean, but are at cross-purposes!

[BiggarDigger]

13 minutes ago, Stub Mandrel said:

I suspect we we both know what we mean, but are at cross-purposes!

Possibly, yes!

I'll see if I can find an example of my rare but increasing frequency traces.

Richard

 

/edit, so these are some very quick finds where there appears to be a very small increase in frequency before the tail.  Note these are all weak signals, not like to more normal ones:

A record of two strikes: the first with a very small increasing frequency head and the second with the more normal decreasing frequency:

and a very peculiar one that I have no explanation for:

and finally a more normal trace from today:

 

[wxsatuser]

On 24/09/2018 at 19:42, BiggarDigger said:

A quick update on this: I haven't been able to capture any of the ISS passes since the middle of last week.  On face value, most evening's there's been two suitable passes, one to the north of GRAVES, the other either passing roughly overhead, or slightly south of the site.  None of these have resulted in a positive identification, despite opening up the IF bandwidth on the SDR, looking for large Doppler shifts......

As I was typing this, explaining why I think my antenna geometries may be difficult to detect the ISS, I got a trace: but according to Heavens Above, it's not the ISS!

 

Any thoughts about which object that is?  It passed through zero Doppler shift (i.e. a straight line to GRAVES?) at 18:20:42.  I'm at approx 55.6N, 3.5W (happy to share by PM more accurate lat and lon) if anyone can do the maths and/or look up an ephemeris.

From the Doppler shift, the velocity should be obtainable I think, which should in turn lead to orbital altitude... and, if the zero Doppler shift position is indeed an indication of crossing a straight line between plotted between my site and GRAVES, it should give a triangulation point.

If this object can be identified, I'd be very interested to know its track and location relative to GRAVES and whether it was North or South of the site as that will illuminate the reflection path.

Richard

I would say ISS.
At you location at that time ISS would be about 4degs alt at Dijon it would be about 27degs alt approx South East.

[The Admiral]

I think this is it now!

Ian

[SteveNickolls]

I have been looking through my records of past detections as the phenomenon of initial streaks with increasing and decreasing frequencies was something the The Admiral and I were intrigued over in the past. The number of increasing initial frequency plots were always very much fewer than the decreasing frequency ones (I will see if I ever made a guesstimate of the occurrences). Here are just a few recorded from back in 2015.

 

I also attach excerpts from my Astronomy log for 2014-2015 with some figures-

Excerpt from 2014-2015 Astronomy Log.docx

Best Regards,
Steve

 

 

 

[The Admiral]

Indeed, Steve and I were confounded by the fact that for virtually all of the meteor trails which displayed an initial 'hook', the frequency change of the 'hook' was from a high positive value to more or less zero, or perhaps a little below. Almost never did we detect a 'hook' which increased in frequency, save for those Steve has shown above, or a 'hook' which fell to a high negative value from below zero.

It seems to me that to a first approximation the ISS Doppler shift calculations would be fairly representative of a meteor head reflection. After all, the ISS travels at a constant velocity (well, OK in a circular orbit, whereas my calculations do not use spherical geometry), as meteors are also purported to do. Using an average height of 90km and an average velocity of 16km/s, and plugging that into the calculations gives:

So shouldn't we expect the meteor head reflections to follow the same pattern? The distance between the knee and the point of inflexion at zero shift is some 500m, which at 16km/s would take ~ 30ms. Indeed a very sharp fall.

Now of course, this is an idealised situation where the meteor's path travels towards and over the receiver and thence over the transmitter, with the meteor being illuminated by radio waves throughout. For Graves, which only (supposedly) transmits towards the South, all we would see would be the reflections with the meteor beyond Graves, and we would then be into the domain where the frequency falls below the 'zero'. But we don't see any of these!

We were at a loss to understand this. One would have thought that, certainly outside of meteor showers, the directions of the meteors' impacts would have been fairly random. We tried to explain the overwhelming bias towards 'hooks' beginning with a high Doppler shift and falling to a low one on the basis of preferential meteor directions, but I'm not sure that this can be so.

Any ideas?

Ian

[IanL]

On my cursory reading of the paper surveying and summarising the published literature (post some way above), it seemed to me that the head echoes that we can detect involve some preferential geometry. I.e. We don't detect any and all meteors in the radar's beam, only those where the geometry is suitable.

For example, I have many detections with large signals (and smaller ones) from the static trail, but no head echo is discernible. I suppose that in these cases the composition or speed of the meteor might not produce the plasma head needed for the head echo, but it seems more credible that we didn't detect it due to the unfavourable geometry. 

So could it be possible that the geometry required to detect a head echo is only capable of producing the high to low shifts and not the reverse? For example, the paper seems suggest that the head echo is only detectable within 6km of the specular reflection point between transmitter and receiver.  One would assume that meteors travelling at grazing angles would be far less likely to pass through that specular reflection point, whereas meteors descending more steeply would be more likely to do so.

I'm not suggesting that is the cause of the one-sided distribution of detections, merely illustrating that there are clearly factors at play that mean the detectable subset of meteors doesn't have randomly distributed paths, unlike the total population of sporadics entering the atmosphere.

[BiggarDigger]

It is strange that the vast majority exhibit this falling frequency hook.  My traces above show a handful of increasing frequency traces, but they a far outweighed by the classic profile of a falling frequency "hook".

What I have noticed, though it is somewhat anecdotal is that the traces of reflections with increasing frequencies are not mirror images of the more classic decreasing frequency head echo.  The (rather poor) examples I posted yesterday are however an indication, at least from my site, that these anomalous reflections are far weaker on average than classic decreasing frequency reflections.  In my examples, you have to zoom in and almost squint to see the increase in frequency, though it may be easier if I were recording at higher speeds.  Is this an indication of the need to have a particular geometry in order to cleanly detect the head echo?

I'm going to speculate about a 3rd element that may be missing in your model Ian.  The 3rd element being the decrease in velocity as the meteor enters the atmosphere.  Your model and charts look good for constant velocity, with the reflecting media moving toward or away from the source/receiver.  But what effect might there be if you add in the reduction in velocity as the meteor follows that path?  If the velocity is reducing as the reflecting media is moving toward on end or the other, would that result in a net increase in the resultant frequency as a result of combining all the shifts?  In other words, is it a multi dimensional problem?

Lastly on this line, it's even more confusing that one of my traces looks similar in nature to the traces that Steve posted: where the start of the tail (the portion of the reflection generated by slowly dissipating ionization trail) increases in frequency and then stabilizes to a steady frequency.  I've read about high altitude winds moving the ionized gasses at 100's of metres per second but if that's the cause we should expect to see the evidence throughout the tail and on both these anomalous reflections and more classic reducing frequency reflections.

 

Sorry, I may just be more confused than at the start!

Richard

[IanL]

8 minutes ago, BiggarDigger said:

I'm going to speculate about a 3rd element that may be missing in your model Ian.  The 3rd element being the decrease in velocity as the meteor enters the atmosphere.  Your model and charts look good for constant velocity, with the reflecting media moving toward or away from the source/receiver.  But what effect might there be if you add in the reduction in velocity as the meteor follows that path?  If the velocity is reducing as the reflecting media is moving toward on end or the other, would that result in a net increase in the resultant frequency as a result of combining all the shifts?  In other words, is it a multi dimensional problem?

I think I got this link on this forum (if not this thread) not long ago, but worth sharing again (paper starts at the bottom of the page):

http://adsabs.harvard.edu/full/1998JIMO...26..117R

It explains the radial velocity issue clearly, unlike my poor attempt above. More importantly it presents some example traces (albeit taken from a detector sited in the same location as the transmitter I believe, so a simpler geometry than we have). What is clear is that the head echo doppler shift is caused to the radial velocity (i.e. the combination of the true speed plus the movement across the detector's line of sight). Any slowing of the meteor in the atmosphere will produce a curved head echo. The effect seems quite small compared to the overall radial velocity effect, and I suspect that for the purpose of this discussion it isn't a significant factor that we need to consider.

[The Admiral]

Thanks for your thoughts chaps.

48 minutes ago, IanL said:

So could it be possible that the geometry required to detect a head echo is only capable of producing the high to low shifts and not the reverse?

48 minutes ago, IanL said:

I'm not suggesting that is the cause of the one-sided distribution of detections, merely illustrating that there are clearly factors at play that mean the detectable subset of meteors doesn't have randomly distributed paths

Well you could be right here, but I'd be happier if it could be confirmed that other types of response do indeed take place. I've so far not read of anything other than the descending frequency 'hook'. I'm sure there are a number of factors here with which I am ignorant, and you could well be correct.

46 minutes ago, BiggarDigger said:

I'm going to speculate about a 3rd element that may be missing in your model Ian.  The 3rd element being the decrease in velocity as the meteor enters the atmosphere.

My understanding from one or more of the documents I've previously linked to is that the meteor doesn't lose speed particularly but just ablates as it goes. This came as a bit of a surprise as I'd initially thought that the decreasing frequency 'hook' was a reflection () of the meteor slowing down, but apparently not, but rather is a result of the change in radial velocity (which, in fact, is what the above traces also represent).

So far as the condition of 'specular reflection' is concerned, I believe I'm right in that it only relates to the static column of ionisation and not to the meteor head?  I can't remember what I've read and what I haven't now . Brain like a blancmange!

Ian