Print

[Gary LaPook]

Ah - yes, now you've jogged my memory, Gary - but I don't think I've seen it in the site here either.

Jeff, I have not been able to find the graphic that Ric had at the symposium showing the correlation of the radio reception reports with the tides at Gardner which makes this correlation very clear. I wish Ric would make it available to everyone on the forum since it would answer the questions that have been raised by several people. The graphic looks a lot like the file I have attached, which is  a graph that I have created from memory of what Ric's graphic looked like. Like on Ric's graphic, I have drawn  vertical lines representing the radio signals plotted against the local time on Gardner (Zulu minus 11 hours), these come from the TIGHAR data base so should accurately replicate the lines on Ric's graph. I don't have the tide data (it is not available on TIGHAR either) so I just freehanded in what I remember of Ric's graphic which shows that the radio signals were sent out when the tide was low at Gardner and only one was sent out when the tide was higher, which Ric said was done by using the battery.                                                                                 
                                                                                                                     
Take a look, does this graph match your recollection too?  Anybody else that saw Ric's graphic at the symposium see any glaring errors in my graph with the line showing the height of the tide just drawn in by hand to represent an approximation of the same line on Ric's graphic?

gl

[Jeff Carter]

Here's a plot I made for my own interest.  AS IS.  No warranties or guarantees that I did all the conversions right in the Excel spreadsheet.  Colors are from the Post Loss report.   

http://tighar.org/Projects/Earhart/Archives/Research/ResearchPapers/Brandenburg/signalcatalog.html which says:
    Identifiers in BLUE are TRANSMISSIONS to Earhart.
    Identifiers in RED are reported receptions that are judged to be NOT CREDIBLE.
    Identifiers in BLACK are reported receptions the credibility of which are judged to be UNCERTAIN.
    Identifiers in GREEN are reported receptions that are judged to be CREDIBLE.


http://i.imgur.com/NCXnH.jpg

I think all it shows is that everyone was active listening in the morning Z-Time.

[Jeff Victor Hayden]

Nice work with the graphs Gary and Jeff.
One point that sticks out like a sore thumb from Garys graph. If the hoaxers were indeed the source of the transmissions then it is one hell of a coincidence that they transmitted at Gardner island low tide nearly every time, why would they do that (was their transmitter in Bakers surf line) and, what are the odds of that being pure luck?

[Gary LaPook]

Here's a plot I made for my own interest.  AS IS.  No warranties or guarantees that I did all the conversions right in the Excel spreadsheet.  Colors are from the Post Loss report.   

http://tighar.org/Projects/Earhart/Archives/Research/ResearchPapers/Brandenburg/signalcatalog.html which says:
    Identifiers in BLUE are TRANSMISSIONS to Earhart.
    Identifiers in RED are reported receptions that are judged to be NOT CREDIBLE.
    Identifiers in BLACK are reported receptions the credibility of which are judged to be UNCERTAIN.
    Identifiers in GREEN are reported receptions that are judged to be CREDIBLE.


http://i.imgur.com/NCXnH.jpg

I think all it shows is that everyone was active listening in the morning Z-Time.

That's a nice graph and it shows that the "not credible" transmissions do not fit the tide pattern and that a number of the "uncertain" ones do. Unfortunately it does not have the tide information that Ric's graphic had which makes this relationship much easier to see.

gl

[Doug Giese]

Jeff Carter, Gary,

Here is Gary's tide chart overlaid on Jeff's radio traffic chart.

I'm pretty sure I got the Z time offset (I aligned t=1100 on the radio traffic chart with t=0000 on the tide chart) right since the radio intercepts matched up quite well between the two charts.

I know I've seen a TIGHAR chart like this. If the TIGHAR tide chart is located I can duplicate a chart like this with the TIGHAR data.

Edit: Ignore this attachment. I've found raw tidal information on Hull Island (tides offset from Niku by about 10 minutes) from two different sources. I posted a revised version using updated tides here.

[Alan Harris]

Here is Gary's tide chart overlaid on Jeff's radio traffic chart.

I know I've seen a TIGHAR chart like this. If the TIGHAR tide chart is located I can duplicate a chart like this with the TIGHAR data.

Interesting idea, but I don't think Gary meant that to be a "real" tide chart, it was just an example of something or other.  For one thing, in reality there are 2 high tides and 2 low tides every 24 hours ("semi-diurnal").  I don't think we've seen current TIGHAR charts for the whole period.  There was one shown recently for 9 July, and IIRC 2 and 3 July (GMT) were posted a longer while ago.

[Doug Giese]

Alan,  yes. I had some time and thought I'd overlay the loss chart onto the tide tables. I had intended to use the tide tables I know I've seen somewhere but couldn't find them quickly so used Gary's tide tables knowing when the real ones turned up it would only take a few minutes to update the Photoshop file.

[Joe Cerniglia]

Here is something I thought might be of interest for building overlay plots, or for casual interest in viewing the radio messages reported to have been heard on July 2, 1937 and the days afterward.  A year ago, I transcribed all the post-loss signal data to an Excel spreadsheet. The spreadsheet provides an easy way to group data by category of hearer, time of day, credibility assessment, frequency, and others.  You can also see many of the signals at a glance to compare and contrast.  Use the arrow at the top of the "Record Link" field to group related subjects (e.g. Noonan injured) that crop up in more than one reported message.  I found this a handy way to analyze or extract the data quickly.  Maybe you will too.

Joe Cerniglia
TIGHAR #3078 ECR

[Doug Giese]

Joe,

Wow. That's an impressive spreadsheet. I didn't realize the credible intercepts were 47.5% (edit-fix) - it seemed lower just reading the raw logs.

The credible intercepts are something that's a comforting backstop. Every time I think maybe they did just crash and sink I remember the intercepts. They had to be on a shore. Maybe it's not Niku, but somewhere deserted along the LOP.

Thanks for posting the spreadsheet.

[Dan Kelly]

Here is something I thought might be of interest for building overlay plots, or for casual interest in viewing the radio messages reported to have been heard on July 2, 1937 and the days afterward.  A year ago, I transcribed all the post-loss signal data to an Excel spreadsheet. The spreadsheet provides an easy way to group data by category of hearer, time of day, credibility assessment, frequency, and others.  You can also see many of the signals at a glance to compare and contrast.  Use the arrow at the top of the "Record Link" field to group related subjects (e.g. Noonan injured) that crop up in more than one reported message.  I found this a handy way to analyze or extract the data quickly.  Maybe you will too.

Joe Cerniglia
TIGHAR #3078 ECR

That is impressive work Joe - are you saying that around a third of the post loss signals are genuinely from Earhart. That is a lot.

[tom howard]

Joe,

Wow. That's an impressive spreadsheet. I didn't realize the credible intercepts were 31.3% - it seemed lower just reading the raw logs.

The credible intercepts are something that's a comforting backstop. Every time I think maybe they did just crash and sink I remember the intercepts. They had to be on a shore. Maybe it's not Niku, but somewhere deserted along the LOP.

Thanks for posting the spreadsheet.

I think Joe's homemade spreadsheet lists 120 receptions logged from "earhart".
57 he rates as credible
So Joe calculates 47.5% are credible of the 120. WOW.
I don't believe even Bradenburg went there, so like beauty, "credible" is in the eye of the beholder.

[Joe Cerniglia]

I think Joe's homemade spreadsheet lists 120 receptions logged from "earhart".
57 he rates as credible
So Joe calculates 47.5% are credible of the 120. WOW.
I don't believe even Bradenburg went there, so like beauty, "credible" is in the eye of the beholder.

Brandenburg did go there. This is the same data TIGHAR has had on its website for months. I'm not aware if the respective databases have been updated independently of each other, but I've seen no inconsistencies.  Any slight errors or omissions are mine.  However, I did take some number of weeks to build this, it has been reviewed by the same people who worked, tirelessly I might add, on cataloguing the reception data, so I have very high confidence in its rigor.  The database includes receptions known to have been sent requesting NR16020 to respond. (You can filter these in or out by selecting 'n/a' in the credibility column.) Your 47.5% does filter these out, so based on the math, that figure is TIGHAR's own interpretation, not Joe's.  If TIGHAR disagrees, it should so state.  If you disagree, sweeping generalizations won't do.  Show us your research on why specific signals can or cannot be viewed as credible. Bear in mind that the only sum total of actual signals that needs to have come from the Electra for it to have been on land somewhere in July 1937 is precisely 1.

Joe Cerniglia
TIGHAR #3078 ECR

[tom howard]

I think Joe's homemade spreadsheet lists 120 receptions logged from "earhart".
57 he rates as credible
So Joe calculates 47.5% are credible of the 120. WOW.
I don't believe even Bradenburg went there, so like beauty, "credible" is in the eye of the beholder.

Brandenburg did go there. This is the same data TIGHAR has had on its website for months. I'm not aware if the respective databases have been updated independently of each other, but I've seen no inconsistencies.  Any slight errors or omissions are mine.  However, I did take some number of weeks to build this, it has been reviewed by the same people who worked, tirelessly I might add, on cataloguing the reception data, so I have very high confidence in its rigor.  The database includes receptions known to have been sent requesting NR16020 to respond. (You can filter these in or out by selecting 'n/a' in the credibility column.) Your 47.5% does filter these out, so based on the math, that figure is TIGHAR's own interpretation, not Joe's.  If TIGHAR disagrees, it should so state.  If you disagree, sweeping generalizations won't do.  Show us your research on why specific signals can or cannot be viewed as credible. Bear in mind that the only sum total of actual signals that needs to have come from the Electra for it to have been on land somewhere in July 1937 is precisely 1.

Joe Cerniglia
TIGHAR #3078 ECR

I will say that if you look at even Bradenburg's analysis he changes dependent on the year asked. Take for instance the "281" message supposedly sent by Amelia. First he rated not credible in his study of 4 selected important cases. Then the next study it's rated as credible. I believe he did the same for the Responses to the Hawaii radio broadcasts. First not credible. Then Credible. I have read his reports and some of it conflicts with other parts of his reports. Like the above mentioned 281 transmission.

Credible and not credible is not etched in stone, it's Brandenburg's and Tighar's interpretation.
I have heard of other forum member's concerns with Bradenburg's ratings, and the main theme of the error of Brandenburg is that practically anything received on frequency 3105KHZ, whether a hum of a generator, a carrier wave, a Microphone being keyed, he rated credible. Read the forum, some very good analysis is done right here by pilots, engineers ect.  That is what is nice about the Tighar forum. We get both sides of the coin.

 Brandenburg rated one signal as credible because a microphone was apparently keyed for 2 straight hours. I believed Gary Lapook debunked that one. I could be wrong on the poster. However, I looked at Bradenburg's charts and indeed His reasoning on most signals he rated credible- Amelia was the only licensed aircraft that could use that aircraft frequency per FCC rules, that was in the area.

That narrow view of the world at the time is problematic.
America is not the world in totality.
I have heard from others on this website and read myself that Howland and other sources picked up Japanese and Russian music on the same frequencies. That is right on this thread, and the OP did a good job of showing that problem. If Japanese music was being sent and heard on the same frequency, how can Brandenburg then pick out a carrier wave, or a microphone clicked a few times and claim that is Amelia?
Therefore FCC rules and regulations had nothing to do with the use of this channel, yet Bradenburg time and again rates signals as "credible" based on it simply being on 3105khz. I am repeating other's postings, but they make sense. These channel frequency business is not a valid reason to mark a transmission credible. I agree with that.

If you cannot see the total illogic in that, then I am sorry, I cannot explain that fallacy in the methodology any better than was already done in the previous pages. It reads like common sense to me. Other nations were using Amelia's frequency.
 There was also lot of confusion, people were hearing snippets of other boats or transmitters calling Amelia and thought it was Amelia calling back.

 I believe the Navy was correct on this one. Lot's of hoaxers, and misinterpreted signals is what the US Navy thought of these transmissions fairly quickly.

Now I hold out hope ONE signal is correct, but so far I haven't heard of it, and am losing some faith there is such a signal. I would think after a week she would have said "Here I am 350 miles south, come get me".
I have yet to to hear that transcription.

The Norwich was stranded, they called for help, and were heard right away and were rescued. Amelia has a week to call, different times to call, night and day, with the whole Pacific Navy listening. and never gets a clear message through.
Now the excuse of "broke antenna" seems weak in this case.
Lae heard her from 400 miles away. Howland is 350 miles.
So at no time during the entire week, did the atmosphere allow a full sentence saying "Here I am, come get me"
I find that odd to put it mildly.

[Greg Daspit]

The Norwich was stranded, they called for help, and were heard right away and were rescued. Amelia has a week to call, different times to call, night and day, with the whole Pacific Navy listening. and never gets a clear message through.
Now the excuse of "broke antenna" seems weak in this case.
Lae heard her from 400 miles away. Howland is 350 miles.
So at no time during the entire week, did the atmosphere allow a full sentence saying "Here I am, come get me"
I find that odd to put it mildly.

Tom and/or other radio experts, can you answer these questions:

Regarding Lae hearing her 400 miles away and Howland not hearing her 350 miles: Does transmitting 1000 feet up in a plane work better than transmitting 0 feet up on a reef?

Regarding the comparison to the Norwich City:
Is the Norwich City’s higher and different antenna set up going to work better than the  antenna of a plane on a reef?
Was the Norwich City’s call for help in code or by voice?  Doesn't code transmit better?

[Doug Giese]

Here's an updated composite tidal chart and radio transmission plot.

The radio transmission plot was posted by Jeff Carter here.

The tidal data is from EasyTide. It's for Hull Island. My understanding is Niku tides lag Hull by about 10 minutes. I didn't account for the lag in the plot because it doesn't really change the story and would complicate the plot. The chart also shows day/night.

I checked the transmit times against the radio database here and it appears to match.

I also added the height of the Electra's wheels (52") and the propeller clearance (76") over the tidal datum as reported in the Post-Loss Signal Statistics with Tide Information research paper. This study is six years old and there may be new hypothesis on the location of the aircraft on the reef. If it were at the edge of the reef, the wheels would be only 8" over tidal datum (and underwater much of the time), and the propeller clearance would be 32" over the tidal datum. This was added to the plot too even though not plausible.  A position on the edge would eliminate quite a few credible intercepts so the Electra had to be away from the edge.

It looks like AE/FN were in the Electra for at least the first three nights.