Advanced search  
Pages: [1] 2 3 ... 10   Go Down

Author Topic: Odds of Spotting Survivors from the Air  (Read 124214 times)

Gary LaPook

  • T5
  • *****
  • Posts: 1624
Odds of Spotting Survivors from the Air
« on: November 04, 2011, 10:52:39 PM »

Why would she be deep in the bush at the other end of the island, as Ric claims, so that she couldn't get to the beach quick enough to wave to the planes. Wouldn't she be more apt to be camped on the beach near the ship wreck? Ric thinks she landed near that ship, what would compel them to walk to the other end of the island, did they hear loud music playing from that direction?

This is deja vu of the running-down-the-LOP discussion.  Show me where I have ever claimed that Earhart was at the Seven Site at the time of the Colorado search.

As shown in the post-loss radio signals catalog, the last credible message was heard at 20:18 local time on Gardner on July 7 - less than two days before the Colorado's airplanes appeared overhead. Radio messages could only be sent from the Electra so it's apparent that Earhart and Noonan stayed in the immediate vicinity of the airplane until at least that time.  The first campsite had to be somewhere close to the airplane.  We call this theoretical campsite Camp Zero. Even if the plane was washed over the reef edge and sank shortly after that, I would not expect them to abandon that area until after the Colorado's planes had come and gone (for the reasons you list).  Once that happened, they had to know that they were faced with the likelihood that they would need to survive on the island for a long time.  With immediate rescue no longer a realistic expectation, the logical thing to do would be to explore the island for an area that provided the best chance for survival. The area near the plane left much to be desired in that respect.  It's in the lee of the easterly trade winds so there are no cooling breezes. (I've been there. It's miserable.) There is also no access to the lagoon for fish and clams.

We know that the Seven Site was the castaway's LAST campsite.  We don't know how many other campsites there were but by the time she got to the Seven Site she was down to only a few durable items essential for survival and had figured out how to catch fish and birds and collect and purify rain water.  In other words, the castaway who died at the Seven Site was an experienced castaway.

So if Earhart and Noonan were in the vicinity of Camp Zero (about a quarter of a mile north of the shipwreck and inland under the buka trees for shade) when the planes came over, why weren't they seen?  Earlier this week we sent every TIGHAR member a DVD of the 2001 Aerial Tour of Nikumaroro which includes an excellent illustration of how hard it is to see people on the ground from the air at Nikumaroro. Take a look at the DVD and tell me that Earhart and Noonan "would have" been seen even if they were out on the beach waving their arms off.

Camp Zero might have featured items salvaged from the plane that were left behind when they moved on.  We plan to conduct a search for Camp Zero when we return to the island.
----------------------------------
Ric, on page 225 of your book you state that there was only a 10 to 20% chance of Lambrecht's search of spotting anyone on the ground. How is this calculated?

gl
Logged

Ric Gillespie

  • Executive Director
  • Administrator
  • *
  • Posts: 5898
  • "Do not try. Do or do not. There is no try" Yoda
Re: Odds of Spotting Survivors from the Air
« Reply #1 on: November 05, 2011, 08:12:18 AM »

Ric, on page 225 of your book you state that there was only a 10 to 20% chance of Lambrecht's search of spotting anyone on the ground. How is this calculated?

In Finding Amelia, Chapter 22 "Banquo's Ghost" (page 225), I wrote:
"In 1937, the techniques and standards of aerial searching were in their infancy. According to present- day Civil Air Patrol Probability of Detection (POD) tables, the chance of Colorado’s planes locating the aircraft in the course of a single inspection of each island was on the order of 10 to 20 percent.   In other words, if the Earhart plane was on one of the islands of the Phoenix Group, there is an 80 to 90 percent chance that Colorado’s search missed it."

Like everything else in the book, the source is cited. The statement is footnoted to: 
"Civil Air Patrol, Mission Aircraft Reference Text, Chapter 9: “Search
Planning and Coverage,” section 9.2.1: “Probability of Detection Table,”
Civil Air Patrol."

The current on-line version is called Mission Aircrew Reference Text
The on-line manual is organized a bit differently than the cited text but the same information is there.  It's pretty interesting to compare the current guidelines for organizing and conducting an aerial search to how the aerial search for Earhart was done.

I'd suggest starting with Section 9.2.2 Probability Areas.  You'll see that the first thing you have to do is establish the LKP (Last Known Position). There's a list of primary factors used to establish the LKP:
•      The aircraft disappearance point on radar. (Yeah, right.)
•   The bearing or fix provided by other ground stations. (The Pan Am and Coast Guard DF bearings on post-loss signals.)
•   Dead reckoning position based on the time of LKP. (Dead reckoning down the LOP.)
•   Reports of sightings-either ground or air. (None.)
•   Emergency locator transmitter (ELT) reports.
There are instances where the above information is not available to assist the planner. To establish a probable position in these instances, the planner must rely on less specific secondary sources of information including:
•   Flight plan. (Known.)
•   Weather information along the intended route or track. (Sketchy at best.)
•   Proximity of airfields along route. (No airfields, but the post-loss signals indicated that the plane landed some place so "Proximity of islands along route.")
•   Aircraft performance. (Known.)
•      Pilot's previous flying record. (Known.)
•      Radar coverage along the intended track. (Dream on.)
•      Nature of terrain along the intended track. (Known. Open ocean and coral atolls.)
•      Position and ground reports. (Scant and ambiguous.)

Section 9.2.3 covers Search Altitudes and Airspeeds.  There are recommendations based upon the type of terrain:
 - Open Flat
-  Moderate Tree Cover and/or Hilly
-  Heavy Tree Cover and/or Very Hilly

The manual gives this example:
"A red and white Cessna 172 has been reported missing and presumed down in eastern Arkansas, in open flat terrain. At the time of the search, flight visibility is forecast to be greater than 10 miles. The incident commander determines, based on available aircraft and crews, that the single probability of detection for this first search must be at least 50%."

Note that this is the estimated probability of seeing an airplane, not people, in flat, open terrain.

As shown in table 9-2 on page 156 of the manual, a big factor in the probability of finding what you're looking for is "track spacing." In other words, when you fly back and forth over your search area (aka "mowing the lawn"), how close together are your respective passes?  Or, if you have multiple aircraft flying line-abreast, how far apart are they? The Lexington used a plan that put aircraft line-abreast with a spacing of 2 miles between planes. We don't know what the Colorado planes did but there is no indication that they had any kind of plan.  The CAP figures assume a search for an airplane.  A ten to twenty percent chance of seeing people on the ground at Gardner in a single pass around the island is probably generous. 
« Last Edit: November 05, 2011, 08:18:03 AM by Ric Gillespie »
Logged

Monty Fowler

  • T5
  • *****
  • Posts: 1078
  • "The real answer is always the right answer."
Re: Odds of Spotting Survivors from the Air
« Reply #2 on: November 05, 2011, 05:22:19 PM »

Something that struck me after watching the TIGHAR DVD from the helicopter was the lack of birds - or maybe they just didn't show up because of the speed of that lil' guy? I remember the Colorado planes decided to fly at 400 feet because of the risk of collision with clouds of seabirds - which virtually eliminated any chance of seeing anyone on the ground.

LTM,
Monty Fowler,
TIGHAR No. 2189 CER
Ex-TIGHAR member No. 2189 E C R SP, 1998-2016
 
Logged

Ric Gillespie

  • Executive Director
  • Administrator
  • *
  • Posts: 5898
  • "Do not try. Do or do not. There is no try" Yoda
Re: Odds of Spotting Survivors from the Air
« Reply #3 on: November 05, 2011, 05:42:29 PM »

Something that struck me after watching the TIGHAR DVD from the helicopter was the lack of birds - or maybe they just didn't show up because of the speed of that lil' guy?

Our speed was similar to the Colorado Corsairs' - about 90 mph.  We didn't have any close encounters with birds. 
Logged

Gary LaPook

  • T5
  • *****
  • Posts: 1624
Re: Odds of Spotting Survivors from the Air
« Reply #4 on: November 07, 2011, 02:15:41 AM »

Ric, on page 225 of your book you state that there was only a 10 to 20% chance of Lambrecht's search of spotting anyone on the ground. How is this calculated?

In Finding Amelia, Chapter 22 "Banquo's Ghost" (page 225), I wrote:
"In 1937, the techniques and standards of aerial searching were in their infancy. According to present- day Civil Air Patrol Probability of Detection (POD) tables, the chance of Colorado’s planes locating the aircraft in the course of a single inspection of each island was on the order of 10 to 20 percent.  In other words, if the Earhart plane was on one of the islands of the Phoenix Group, there is an 80 to 90 percent chance that Colorado’s search missed it."

Like everything else in the book, the source is cited. The statement is footnoted to: 
"Civil Air Patrol, Mission Aircraft Reference Text, Chapter 9: “Search
Planning and Coverage,” section 9.2.1: “Probability of Detection Table,”
Civil Air Patrol."

The current on-line version is called Mission Aircrew Reference Text
The on-line manual is organized a bit differently than the cited text but the same information is there.  It's pretty interesting to compare the current guidelines for organizing and conducting an aerial search to how the aerial search for Earhart was done.

I'd suggest starting with Section 9.2.2 Probability Areas.  You'll see that the first thing you have to do is establish the LKP (Last Known Position). There's a list of primary factors used to establish the LKP:
•      The aircraft disappearance point on radar. (Yeah, right.)
•   The bearing or fix provided by other ground stations. (The Pan Am and Coast Guard DF bearings on post-loss signals.)
•   Dead reckoning position based on the time of LKP. (Dead reckoning down the LOP.)
•   Reports of sightings-either ground or air. (None.)
•   Emergency locator transmitter (ELT) reports.
There are instances where the above information is not available to assist the planner. To establish a probable position in these instances, the planner must rely on less specific secondary sources of information including:
•   Flight plan. (Known.)
•   Weather information along the intended route or track. (Sketchy at best.)
•   Proximity of airfields along route. (No airfields, but the post-loss signals indicated that the plane landed some place so "Proximity of islands along route.")
•   Aircraft performance. (Known.)
•      Pilot's previous flying record. (Known.)
•      Radar coverage along the intended track. (Dream on.)
•      Nature of terrain along the intended track. (Known. Open ocean and coral atolls.)
•      Position and ground reports. (Scant and ambiguous.)

Section 9.2.3 covers Search Altitudes and Airspeeds.  There are recommendations based upon the type of terrain:
 - Open Flat
-  Moderate Tree Cover and/or Hilly
-  Heavy Tree Cover and/or Very Hilly

The manual gives this example:
"A red and white Cessna 172 has been reported missing and presumed down in eastern Arkansas, in open flat terrain. At the time of the search, flight visibility is forecast to be greater than 10 miles. The incident commander determines, based on available aircraft and crews, that the single probability of detection for this first search must be at least 50%."

Note that this is the estimated probability of seeing an airplane, not people, in flat, open terrain.

As shown in table 9-2 on page 156 of the manual, a big factor in the probability of finding what you're looking for is "track spacing." In other words, when you fly back and forth over your search area (aka "mowing the lawn"), how close together are your respective passes?  Or, if you have multiple aircraft flying line-abreast, how far apart are they? The Lexington used a plan that put aircraft line-abreast with a spacing of 2 miles between planes. We don't know what the Colorado planes did but there is no indication that they had any kind of plan.  The CAP figures assume a search for an airplane.  A ten to twenty percent chance of seeing people on the ground at Gardner in a single pass around the island is probably generous.
-------------------------------------
The reason I asked, Ric, is that I just re-read your book and I was curious how you came up with your estimate of 10 to 20% probability of detection of Earhart by the Lambrecht flight. I looked at my National Search and Rescue Manual (see attached) Inland Probability Of Detection (POD) table and come up with a much higher probability that the Lambrecht flight would have spotted Earhart and Noonan if they had been on Gardner. I wondered if you were using a different POD table and now that I see what you have posted as the table that you used, I see that your table and mine are the same POD tables. Your POD table is on page 156 (172 of the PDF) of the Civil Air Patrol Mission Aircrew Reference Text:
 http://www.ndcap.us/dept/es/ref_aircrew.pdf

Mine is on page 8-2 of the SAR manual.

After reviewing your reference, I still don't understand how you came up with such a low probability of detection. Looking at your table or the one in my attached Search And Rescue Manual, the probability of detection had to be at least 30% even if Earhart and Noonan were hiding in the tall vegetation and at least 75% if they were standing on the beach or standing in the water on the reef flat.

We can go through the computation looking first at the worst case example of "heavy tree cover." Since Lambrecht was flying at about 400 feet we  use the 500 foot table. We also know that the visibility was at least 4 nautical miles since Lambrecht's report said it was 30 nautical miles so we use the visibility 4 mi. column. We also know that the track spacing could not have been greater than one-half mile because the island is too narrow for a greater spacing, see attached diagram of .5 nm spacing.

Using all of these entry values and entering the POD tables (either yours or mine) we find the probability of detection is 30% for the worst case, not the 10 to 20% that you stated. Using the same information and looking at the table for "open, flat terrain," such as the beach and the reef flat, we see that the probability of spotting Earhart and Noonan standing on the beach or on the reef is 75%, again much higher than the value that you stated.

But this is not the end of the computation, your must then go on to the Cumulative POD table on page 157 (173 of the PDF) of your manual, page 8-2 of my manual. Every additional pass over the same area increases the probability of detection. For Earhart, hiding among the trees, the cumulative probability of detection increases to 45% after a second pass; 50% after a third pass; 60% after a fourth pass; 65% after a fifth pass; 75% after a sixth pass; 80% after a seventh pass and 90% after 8 passes.

Looking at the case of Earhart standing on the beach, the probability increases to 95% after only 2 passes. Each pass around the island by each of the three planes in Lambrecht's flight counts as an additional search. Page 157 (173 of the PDF) of the manual that you used states:

"If you, or another aircraft and crew, fly the same pattern a second time, the POD increases significantly."

Because the land is only one half mile wide, the three planes in the Lambrecht flight would have flown either in trail or in line abreast close enough to each other so that they each covered the same half mile wide strip of land so each circle the island by the flight constituted three searches. (See attached diagram of .2nm spacing.) Lambrecht reported that he circled other islands three times but for his search of Gardner Lambrecht reported "repeated circling and zooming" which must mean more than three and certainly not less than three circles of the island. Since there were three planes observing the same one half-mile wide strip of land, three circles of the island by the flight of three planes is actually 9 searches for the purposes of the cumulative POD table. So if Earhart was standing on the beach, which is most likely, then the POD was at least 95% and even if she was trying to hide from the planes in the tall trees the POD was at least 90%.

You state that the POD table is for searches for a crashed aircraft, implying that the table does not apply to a search for individuals but that is because you looked at the example given in your manual, a search for a small plane, but, in fact, the tables are equally applicable to searches for personnel. This is clear since on the the same page as the POD table, the SAR manual states:

"C. Lost-person SAR is the second most common inland search mission."

Nor is there a different POD table given for searches for personnel nor is there some correction factor given to change the probability for searches for personnel from that given in the POD table so it is clear that the POD table is applicable to finding Earhart on Gardner.

You point out that the Lexington's planes used a track spacing of 2 miles but what those planes did in searching open water has no relevance to how the search was conducted over Gardner Island. This is clear since the POD table you used is actually the Inland Probability Of Detection table and is not applicable to ocean searches. A completely separate and different procedure and extensive set of tables are used for planning searches of open water. You may not have realized this since you were looking at the Civil Air Patrol Text and the CAP is involved in searches only over land so maritime searches are not covered in your manual. Open ocean searches are extensively covered in my SAR manual and I can post the tables that apply to the Lexington's search if you like.

Ric, can you explain to us how you arrived at your 10 to 20% figure?

gl
« Last Edit: November 07, 2011, 02:56:40 AM by Gary LaPook »
Logged

Ric Gillespie

  • Executive Director
  • Administrator
  • *
  • Posts: 5898
  • "Do not try. Do or do not. There is no try" Yoda
Re: Odds of Spotting Survivors from the Air
« Reply #5 on: November 07, 2011, 11:03:38 AM »

The reason I asked, Ric, is that I just re-read your book and I was curious how you came up with your estimate of 10 to 20% probability of detection of Earhart by the Lambrecht flight.

Applying the POD tables to the Earhart search is an admittedly fuzzy exercise.  The tables are generalized statistical compilations intended to help plan a SAR operation using modern techniques and trained personnel.  The results of numerous past searches are complied and probabilities are derived, but the probability of success for any single search is always either one or zero, yes or no.  The searches upon which the tables are based were conducted using established techniques and trained personnel. Whatever probabilities may be derived from tables, the expected results from a search carried out using unknown techniques and personal with no training in search operations can reasonably be expected to be significantly lower.

Looking at your table or the one in my attached Search And Rescue Manual, the probability of detection had to be at least 30% even if Earhart and Noonan were hiding in the tall vegetation and at least 75% if they were standing on the beach or standing in the water on the reef flat.

I can think of no better illustration of how different the reality of Nikumaroro is from the POD guidelines.  Look at the video and tell me that there is a 30% chance of seeing a person "hiding in the tall vegetation."   A 75% chance of being seen if they are standing in the open?  In 1989, by prior arrangement, a Royal New Zealand Air Force Lockheed P-3 Orion visited Nikumaroro while we were there.  They buzzed the island and our expedition ship at about 500 feet.  There were about a dozen of us standing out in the open, wearing brightly colored clothing, waving like mad.  They never saw us.

Because the land is only one half mile wide, the three planes in the Lambrecht flight would have flown either in trail or in line abreast close enough to each other so that they each covered the same half mile wide strip of land so each circle the island by the flight constituted three searches.

You're doing it again. "Would have" is a guess masquerading as a fact.  We have no idea how the planes conducted their search of the island.

Lambrecht reported that he circled other islands three times...

He did?  He said he made "one circle" around McKean, "a circle of the island" at Hull, "several circles" of Sydney, no mention of circling Phoenix, no mention of circling Enderbury, "two or three turns" around Birnie (the smallest of the group), no mention of circling Canton (the largest of the group).

...but for his search of Gardner Lambrecht reported "repeated circling and zooming" which must mean more than three and certainly not less than three circles of the island.

You're taking it out of context.  Lambrecht is not talking about repeatedly circling the island. Lambrecht wrote:
"Here signs of recent habitation were clearly visible but repeated circling and zooming failed to elicit any answering wave from possible inhabitants and it was finally taken for granted that none were there."
The "circling and zooming" was done over the "signs of recent habitation" to "elicit an answering wave" from "possible inhabitants."    (A "zoom" is a dive and sharp pull-up.  It's purpose, in this case, was probably to make a lot of noise.)

You state that the POD table is for searches for a crashed aircraft, implying that the table does not apply to a search for individuals but that is because you looked at the example given in your manual, a search for a small plane, but, in fact, the tables are equally applicable to searches for personnel. This is clear since on the the same page as the POD table, the SAR manual states:

"C. Lost-person SAR is the second most common inland search mission."

Nor is there a different POD table given for searches for personnel nor is there some correction factor given to change the probability for searches for personnel from that given in the POD table so it is clear that the POD table is applicable to finding Earhart on Gardner.

Another great illustration of how generalized the POD tables are.  As a rule, people are quite a bit smaller than airplanes, and some airplanes are quite a bit larger than other airplanes - and yet the POD tables make no distinction in the probability of detection. Are the chances of spotting a person really the same as the chances of spotting a B-52?

You point out that the Lexington's planes used a track spacing of 2 miles but what those planes did in searching open water has no relevance to how the search was conducted over Gardner Island.

That's what I said.

This is clear since the POD table you used is actually the Inland Probability Of Detection table and is not applicable to ocean searches.

No.  The POD tables didn't exist in 1937.  The reason that what Lexington's planes did is not relevant to what Colorado's planes did is because were no standard search procedures and both ships were making it up as they went along.   Lexington drew up its plan on the way out from San Diego. There's no indication that Colorado had a plan at all beyond "fly out and take a look."

Ten to twenty percentage probability for Colorado's planes seeing AE and FN if they were there is an estimate.  I think it's generous.
« Last Edit: November 07, 2011, 12:22:38 PM by Ric Gillespie »
Logged

Andrew M McKenna

  • Administrator
  • *
  • Posts: 677
  • Here I am during the Maid of Harlech Survey.
Re: Odds of Spotting Survivors from the Air
« Reply #6 on: November 07, 2011, 11:07:02 PM »

Gary

I know you've flown a myriad of aircraft, but have you ever been involved in any actual Search and Rescue Flying, or search mission management? 

The discrepancy between your use of the POD charts and Ric's has to do with understanding Search Visibility.  This is not the same as aviation visibility which might be defined as "the ability to see and identify prominent objects by day and night as determined by atmospheric conditions."

In your example, you have assumed Search Visibility to be 4 mi.  That may be the distance you think you might be able to see prominent objects such as an airport, or perhaps an Electra on the beach, but certainly not a single human being out in the open, never mind in the bush. 

Search visibility is defined as "the distance at which an object on the ground can be seen and recognized from a particular height".  Google "CAP Mission Pilot Course slides" and you can find it on page 8 of the Mission Aircrew Course- Chapter 9: Search Planning and Coverage (Feb 2005).

So, lets see what the Search Visibility might be for a human being on the edge of the beach at Niku?  Let's assume that we're flying at 500ft, how far away can we still see and recognize a human on the beach, or in the bush.  I can assure you it would be significantly less than 1 mile, not 4 miles as you have assumed in your POD example.  At .25 of a mile SV, the POD would be 2.5%, at .125 of a mile, the POD would be 1.25%. 

So in theory, with track spacing of .5 miles, flying at 500 ft AGL and IF you were able to see and recognize a person in the bush at a lateral distance of 1/8th of a mile (660 ft), you'd have a 1.25% POD for a single pass.

However, the aerial tour video (have you seen it?) demonstrates that it is extremely hard to see a person on the beach at Niku from an altitude of a mere 200 ft, even when you know they are there and they're wearing a white shirt. 

We might conclude therefore that the Search visibility for a human in the bush is significantly less than 1/8th of a mile (660 ft), therefore the POD will also be significantly less than 1.25% POD.

For the moment let's use 1/16th of a mile (330 ft) as the search visibility to calculate the POD.  If the POD with a .5 mile track spacing, 500 ft AGL, and a 1 Mile Search Visibility is 10%, the POD with a Search Visibility of 330 ft at 500 ft AGL might be 1/16th of the POD at 1 Mile SV, or a POD of 0.625 %, (6.25/10ths of 1% just to be sure everyone understands). 

At that POD, (interpolating the Cumulative POD chart at the low end) it would take two passes to get to 0.8% POD, three passes would give 0.9%, four passes 0.95% POD etc.  You can see that it would take many passes at 0.625% just to get to a POD of 10%, never mind a POD of 20%.  And, the POD goes down the lower the altitude, so at 400 ft AGL the POD would be less than 0.625% for each pass.

So you can see that the POD charts are not really created for finding missing persons, they were created for finding missing aircraft (or the smoking holes they create when colliding with a planet), an object that is considerably larger than a single human, and visible from much farther away, hence the minimum Search Visibility of 1 Mile on the chart. 

Overestimating POD is one of the most common mistakes searchers make, and search managers routinely take POD estimates from the field and cut them in half.  In the woods of Colo, we would never assign a POD from a single pass of one aircraft at more than 5%, it would simply be unrealistic. 

The Air Force Rescue Coordination Center (AFRCC) will typically abandon a search once the POD gets above 80%, so it is important not to overestimate the POD lest you leave your target abandoned in the search area.  Sound familiar?

Ric's 10% -20% POD is generous assuming that the aircraft is no longer visible, and we're looking for humans in the bush.

As an aside, for those of us who like to get lost, the single most effective tool that you can have to increase your chances of being found by searchers is a high quality signal mirror with sighting device (OK, OK, a GPS signalling device would probably be better).  Mirror flashes can be seen from great distances, like 15 to 20 miles (farther than smoke signals), effectively increasing the Search Visibility exponentially.  Simply sweeping the horizon with a signal mirror will get some attention, assuming someone is out there.  People have been found by flashing airliners at altitude.  Cheap at any military surplus, keep one with your flight / camping gear and learn how to use it before you really need it.

Andrew
« Last Edit: November 07, 2011, 11:24:21 PM by Andrew M McKenna »
Logged

Gary LaPook

  • T5
  • *****
  • Posts: 1624
Re: Odds of Spotting Survivors from the Air
« Reply #7 on: November 08, 2011, 01:36:47 AM »

Gary

I know you've flown a myriad of aircraft, but have you ever been involved in any actual Search and Rescue Flying, or search mission management? 

The discrepancy between your use of the POD charts and Ric's has to do with understanding Search Visibility.  This is not the same as aviation visibility which might be defined as "the ability to see and identify prominent objects by day and night as determined by atmospheric conditions."

In your example, you have assumed Search Visibility to be 4 mi.  That may be the distance you think you might be able to see prominent objects such as an airport, or perhaps an Electra on the beach, but certainly not a single human being out in the open, never mind in the bush. 

Search visibility is defined as "the distance at which an object on the ground can be seen and recognized from a particular height".  Google "CAP Mission Pilot Course slides" and you can find it on page 8 of the Mission Aircrew Course- Chapter 9: Search Planning and Coverage (Feb 2005).

So, lets see what the Search Visibility might be for a human being on the edge of the beach at Niku?  Let's assume that we're flying at 500ft, how far away can we still see and recognize a human on the beach, or in the bush. I can assure you it would be significantly less than 1 mile, not 4 miles as you have assumed in your POD example.  At .25 of a mile SV, the POD would be 2.5%, at .125 of a mile, the POD would be 1.25%. 

So in theory, with track spacing of .5 miles, flying at 500 ft AGL and IF you were able to see and recognize a person in the bush at a lateral distance of 1/8th of a mile (660 ft), you'd have a 1.25% POD for a single pass.

However, the aerial tour video (have you seen it?) demonstrates that it is extremely hard to see a person on the beach at Niku from an altitude of a mere 200 ft, even when you know they are there and they're wearing a white shirt. 

We might conclude therefore that the Search visibility for a human in the bush is significantly less than 1/8th of a mile (660 ft), therefore the POD will also be significantly less than 1.25% POD.

For the moment let's use 1/16th of a mile (330 ft) as the search visibility to calculate the POD.  If the POD with a .5 mile track spacing, 500 ft AGL, and a 1 Mile Search Visibility is 10%, the POD with a Search Visibility of 330 ft at 500 ft AGL might be 1/16th of the POD at 1 Mile SV, or a POD of 0.625 %, (6.25/10ths of 1% just to be sure everyone understands). 

At that POD, (interpolating the Cumulative POD chart at the low end) it would take two passes to get to 0.8% POD, three passes would give 0.9%, four passes 0.95% POD etc.  You can see that it would take many passes at 0.625% just to get to a POD of 10%, never mind a POD of 20%.  And, the POD goes down the lower the altitude, so at 400 ft AGL the POD would be less than 0.625% for each pass.

So you can see that the POD charts are not really created for finding missing persons, they were created for finding missing aircraft (or the smoking holes they create when colliding with a planet), an object that is considerably larger than a single human, and visible from much farther away, hence the minimum Search Visibility of 1 Mile on the chart. 

Overestimating POD is one of the most common mistakes searchers make, and search managers routinely take POD estimates from the field and cut them in half.  In the woods of Colo, we would never assign a POD from a single pass of one aircraft at more than 5%, it would simply be unrealistic. 

The Air Force Rescue Coordination Center (AFRCC) will typically abandon a search once the POD gets above 80%, so it is important not to overestimate the POD lest you leave your target abandoned in the search area.  Sound familiar?

Ric's 10% -20% POD is generous assuming that the aircraft is no longer visible, and we're looking for humans in the bush.

As an aside, for those of us who like to get lost, the single most effective tool that you can have to increase your chances of being found by searchers is a high quality signal mirror with sighting device (OK, OK, a GPS signalling device would probably be better).  Mirror flashes can be seen from great distances, like 15 to 20 miles (farther than smoke signals), effectively increasing the Search Visibility exponentially.  Simply sweeping the horizon with a signal mirror will get some attention, assuming someone is out there.  People have been found by flashing airliners at altitude.  Cheap at any military surplus, keep one with your flight / camping gear and learn how to use it before you really need it.

Andrew
-------------------------------

Go back and read that page again. You have apparently confused "search visibility" with "scanning range." You correctly quoted the definition of "search visibility" as "the distance at which an object on the ground can be seen and recognized from a particular height". This "an object" is not the same as the subject of the search such as a person. If you can identify an object, such as a house, at four miles then the "search visibility" is four miles even if you are searching for a person. Search visibility is not the distance you expect to be able to spot the subject of the search, I was not assuming that a person can be seen at four miles. Scanning range is "the distance that a scanner is expected to have a good chance at spotting the search objective."

The distance that you can spot a person is less than a mile but that is not an entry value for using the POD table, which is the search visibility. According to the SAR manual, a person in an orange flight suit is visible at .5 miles so a person not in an orange suit would be visible at something less, say half of that, at a quarter mile, twice your assumption. This is the scanning range. You can prove this to yourself. Go outside and look down the sidewalk and I'll bet you can see people walking a quarter mile (two city blocks) down the sidewalk.  The POD tables show a minimum track spacing of .5 miles meaning the observer observes one quarter mile on each side of the flight track so should be able to spot a person with half mile track spacing. If this were not possible then the POD tables would have to show much closer track spacing. The POD table shows that the minimum track spacing that is needed is a half mile and no need to space them closer than that.
Then you make an unwarranted attempt to calculate a much lower POD than the table is designed for and that it does not allow. Based on your manipulation, the POD per pass gets so low that there would never be an airborne search for a person because there are not enough planes on earth to achieve any useful cumulative POD. And the track spacing that you would require are not achievable in flight.

gl
« Last Edit: November 08, 2011, 01:39:55 AM by Gary LaPook »
Logged

Gary LaPook

  • T5
  • *****
  • Posts: 1624
Re: Odds of Spotting Survivors from the Air
« Reply #8 on: November 08, 2011, 02:22:55 AM »

The reason I asked, Ric, is that I just re-read your book and I was curious how you came up with your estimate of 10 to 20% probability of detection of Earhart by the Lambrecht flight.

Applying the POD tables to the Earhart search is an admittedly fuzzy exercise.  The tables are generalized statistical compilations intended to help plan a SAR operation using modern techniques and trained personnel.  The results of numerous past searches are complied and probabilities are derived, but the probability of success for any single search is always either one or zero, yes or no.  The searches upon which the tables are based were conducted using established techniques and trained personnel. Whatever probabilities may be derived from tables, the expected results from a search carried out using unknown techniques and personal with no training in search operations can reasonably be expected to be significantly lower.

Looking at your table or the one in my attached Search And Rescue Manual, the probability of detection had to be at least 30% even if Earhart and Noonan were hiding in the tall vegetation and at least 75% if they were standing on the beach or standing in the water on the reef flat.

I can think of no better illustration of how different the reality of Nikumaroro is from the POD guidelines.  Look at the video and tell me that there is a 30% chance of seeing a person "hiding in the tall vegetation."   A 75% chance of being seen if they are standing in the open?  In 1989, by prior arrangement, a Royal New Zealand Air Force Lockheed P-3 Orion visited Nikumaroro while we were there.  They buzzed the island and our expedition ship at about 500 feet.  There were about a dozen of us standing out in the open, wearing brightly colored clothing, waving like mad.  They never saw us.

Because the land is only one half mile wide, the three planes in the Lambrecht flight would have flown either in trail or in line abreast close enough to each other so that they each covered the same half mile wide strip of land so each circle the island by the flight constituted three searches.

You're doing it again. "Would have" is a guess masquerading as a fact.  We have no idea how the planes conducted their search of the island.

Lambrecht reported that he circled other islands three times...

He did?  He said he made "one circle" around McKean, "a circle of the island" at Hull, "several circles" of Sydney, no mention of circling Phoenix, no mention of circling Enderbury, "two or three turns" around Birnie (the smallest of the group), no mention of circling Canton (the largest of the group).

...but for his search of Gardner Lambrecht reported "repeated circling and zooming" which must mean more than three and certainly not less than three circles of the island.

You're taking it out of context.  Lambrecht is not talking about repeatedly circling the island. Lambrecht wrote:
"Here signs of recent habitation were clearly visible but repeated circling and zooming failed to elicit any answering wave from possible inhabitants and it was finally taken for granted that none were there."
The "circling and zooming" was done over the "signs of recent habitation" to "elicit an answering wave" from "possible inhabitants."    (A "zoom" is a dive and sharp pull-up.  It's purpose, in this case, was probably to make a lot of noise.)

You state that the POD table is for searches for a crashed aircraft, implying that the table does not apply to a search for individuals but that is because you looked at the example given in your manual, a search for a small plane, but, in fact, the tables are equally applicable to searches for personnel. This is clear since on the the same page as the POD table, the SAR manual states:

"C. Lost-person SAR is the second most common inland search mission."

Nor is there a different POD table given for searches for personnel nor is there some correction factor given to change the probability for searches for personnel from that given in the POD table so it is clear that the POD table is applicable to finding Earhart on Gardner.

Another great illustration of how generalized the POD tables are.  As a rule, people are quite a bit smaller than airplanes, and some airplanes are quite a bit larger than other airplanes - and yet the POD tables make no distinction in the probability of detection. Are the chances of spotting a person really the same as the chances of spotting a B-52?

You point out that the Lexington's planes used a track spacing of 2 miles but what those planes did in searching open water has no relevance to how the search was conducted over Gardner Island.

That's what I said.

This is clear since the POD table you used is actually the Inland Probability Of Detection table and is not applicable to ocean searches.

No.  The POD tables didn't exist in 1937.  The reason that what Lexington's planes did is not relevant to what Colorado's planes did is because were no standard search procedures and both ships were making it up as they went along.   Lexington drew up its plan on the way out from San Diego. There's no indication that Colorado had a plan at all beyond "fly out and take a look."

Ten to twenty percentage probability for Colorado's planes seeing AE and FN if they were there is an estimate.  I think it's generous.

------------------------
I think this is quite humorous Ric since you were the one who  cited to the POD table as authority for your 10-20% estimate for spotting Earhart. Then when I go through the actual computation and prove that the authority that you based your claim on actually predicts a much higher probability of detection than you like, your come back and say that the POD is not valid.

So now you now start throwing around claims based on your anecdotal observation (something that you always criticize) that you claim are more accurate than the statistical tables made by search and rescue professionals after they examined reams of data. So either you are right and the professionals in the Air Force, the Civil Air Patrol, the Navy, the Army and the Coast Guard are wrong or vice versa. My money is on the professionals, not on you, Ric. After all, those guys do not have a dog in this fight.

But let's say you were right that the POD for a single pass was 10-20%. Looking at the cumulative POD table it rises to 65% after 6 passes, two laps around the island by three planes for the wooded area and it must be much higher for the clear area on the beach.

And why do you assume that the pilots were not trained in how to conduct a search, after all that is what those float planes were on the Colorado for, to conduct searches, not to fly the admiral to the golf course.

Re-read Lambrecht's report about the way he conducted the searches on the other islands, you mis-characterized what he said. You don't know what kind of search he made at Gardner but it is unlikely that he ordered the other planes to stay well offshore so that he could get all the glory of finding Earhart so it is more reasonable to believe that all three planes took part in the search of this island (readers can decide for themselves which makes more sense) so many passes were made raising the cumulative probability of detection to a much higher level that you claim.

gl

« Last Edit: November 08, 2011, 02:46:33 AM by Gary LaPook »
Logged

Ric Gillespie

  • Executive Director
  • Administrator
  • *
  • Posts: 5898
  • "Do not try. Do or do not. There is no try" Yoda
Re: Odds of Spotting Survivors from the Air
« Reply #9 on: November 08, 2011, 08:03:41 AM »

I can't believe you're still trying to defend this trench.

I think this is quite humorous Ric since you were the one who  cited to the POD table as authority for your 10-20% estimate for spotting Earhart. Then when I go through the actual computation and prove that the authority that you based your claim on actually predicts a much higher probability of detection than you like, your come back and say that the POD is not valid.
 

I thought I 'splained that my 10 to 20% was not derived directly from the POD but was an estimate based on the fact that the 1937 search was not conducted with the techniques and training used to compute the probabilities in the POD. 

So now you now start throwing around claims based on your anecdotal observation (something that you always criticize) that you claim are more accurate than the statistical tables made by search and rescue professionals after they examined reams of data. So either you are right and the professionals in the Air Force, the Civil Air Patrol, the Navy, the Army and the Coast Guard are wrong or vice versa. My money is on the professionals, not on you, Ric. After all, those guys do not have a dog in this fight.
 

 I am not throwing around claims based on my anecdotal observation.  I have provided every TIGHAR member (including you) with a video re-enactment of the actual conditions under which the 1937 aerial search was conducted.  To assert that generalized statistical tables are more reliable than real life observation of a specific situation is,  frankly, ludicrous.  I think those Air Force, the Civil Air Patrol, the Navy, the Army and the Coast Guard professionals would agree.
Real life always trumps statistics.

And why do you assume that the pilots were not trained in how to conduct a search, after all that is what those float planes were on the Colorado for, to conduct searches, not to fly the admiral to the golf course.

I didn't realize that you didn't know this.  In the days before radar, the mission of float planes on cruisers and battleships was to scout for and identify enemy ships and then act as forward observers to adjust the fall of shot from the big guns.  You're a former artilleryman.  I don't know how they did it at Ft. Sill but at the Benning School for Boys (aka Infantry Officer School at Ft. Benning) SAR was not part of the curriculum for FOs.

Re-read Lambrecht's report about the way he conducted the searches on the other islands, you mis-characterized what he said. You don't know what kind of search he made at Gardner but it is unlikely that he ordered the other planes to stay well offshore so that he could get all the glory of finding Earhart so it is more reasonable to believe that all three planes took part in the search of this island (readers can decide for themselves which makes more sense) so many passes were made raising the cumulative probability of detection to a much higher level that you claim.

I don't think I mischaracterized anything Lambrecht said and I certainly didn't imply that he ordered the other planes to stay well off shore. As we have both said, we don't have any solid information about how the overflight of Gardner was conducted.  There are a few clues. 
- We can make a ballpark estimate of how long they spent there - 18 to 28 minutes. See the FAQ How long did the three aircraft from the battleship Colorado spend over Gardner island on July 9, 1937?
- We can compare Lambrecht's observations with those of another pilot on the same flight, Lt.(jg) William Short.

Lambrecht makes no estimate of the island's length. 
Short says it's "about 2 1/2 miles long by a mile wide."  His estimate of the width is right on but he's way off on the length. Gardner is actually 3 3/4 nautical miles long.

Lambrecht says "Most of this island is covered with tropical vegetation with, here and there, a grove of coconut palms."
Short is more specific. "Almost completely covered with short bushy trees including two small groves of coconut palms."  In 1937 the island was almost completely covered with Buka trees 60 to 90 feet tall. If Short thought they were "short bushy trees," a person on the ground would appear much smaller than he expected.  In 1937 there were five small groves of cocos.

Both Lambrecht and Short commented on Norwich City. Lambrecht estimates the ship at 4,000 tons.  Short's guess is 5,000 tons.  Norwich City's gross tonnage was 5,587.  Net tonnage was 3,513.  Both estimates were in the ballpark, but evaluating ships was what these guys were trained to do.

Curiously, Short does not mention the "signs of recent habitation" or the "repeated circling and zooming."  Either he didn't consider it worth mentioning or he was off looking at some other part of the island.  The planes could not communicate with each other except by hand signals.

 From what little information we have it would appear that the pilots were pretty good at what they were trained to do - evaluate ships - but not so good at estimating sizes and distances relating to terrain (hardly surprising).  There is also some indication that they didn't fly organized sweeps but split up and did their own thing.
Logged

Gary LaPook

  • T5
  • *****
  • Posts: 1624
Re: Odds of Spotting Survivors from the Air
« Reply #10 on: November 09, 2011, 12:33:57 AM »

I can't believe you're still trying to defend this trench.

I think this is quite humorous Ric since you were the one who  cited to the POD table as authority for your 10-20% estimate for spotting Earhart. Then when I go through the actual computation and prove that the authority that you based your claim on actually predicts a much higher probability of detection than you like, your come back and say that the POD is not valid.
 

I thought I 'splained that my 10 to 20% was not derived directly from the POD but was an estimate based on the fact that the 1937 search was not conducted with the techniques and training used to compute the probabilities in the POD. 

So now you now start throwing around claims based on your anecdotal observation (something that you always criticize) that you claim are more accurate than the statistical tables made by search and rescue professionals after they examined reams of data. So either you are right and the professionals in the Air Force, the Civil Air Patrol, the Navy, the Army and the Coast Guard are wrong or vice versa. My money is on the professionals, not on you, Ric. After all, those guys do not have a dog in this fight.
 

 I am not throwing around claims based on my anecdotal observation.  I have provided every TIGHAR member (including you) with a video re-enactment of the actual conditions under which the 1937 aerial search was conducted.  To assert that generalized statistical tables are more reliable than real life observation of a specific situation is,  frankly, ludicrous.  I think those Air Force, the Civil Air Patrol, the Navy, the Army and the Coast Guard professionals would agree.
Real life always trumps statistics.

And why do you assume that the pilots were not trained in how to conduct a search, after all that is what those float planes were on the Colorado for, to conduct searches, not to fly the admiral to the golf course.

I didn't realize that you didn't know this.  In the days before radar, the mission of float planes on cruisers and battleships was to scout for and identify enemy ships and then act as forward observers to adjust the fall of shot from the big guns.  You're a former artilleryman.  I don't know how they did it at Ft. Sill but at the Benning School for Boys (aka Infantry Officer School at Ft. Benning) SAR was not part of the curriculum for FOs.

Re-read Lambrecht's report about the way he conducted the searches on the other islands, you mis-characterized what he said. You don't know what kind of search he made at Gardner but it is unlikely that he ordered the other planes to stay well offshore so that he could get all the glory of finding Earhart so it is more reasonable to believe that all three planes took part in the search of this island (readers can decide for themselves which makes more sense) so many passes were made raising the cumulative probability of detection to a much higher level that you claim.

I don't think I mischaracterized anything Lambrecht said and I certainly didn't imply that he ordered the other planes to stay well off shore. As we have both said, we don't have any solid information about how the overflight of Gardner was conducted.  There are a few clues. 
- We can make a ballpark estimate of how long they spent there - 18 to 28 minutes. See the FAQ How long did the three aircraft from the battleship Colorado spend over Gardner island on July 9, 1937?
- We can compare Lambrecht's observations with those of another pilot on the same flight, Lt.(jg) William Short.

Lambrecht makes no estimate of the island's length. 
Short says it's "about 2 1/2 miles long by a mile wide."  His estimate of the width is right on but he's way off on the length. Gardner is actually 3 3/4 nautical miles long.

Lambrecht says "Most of this island is covered with tropical vegetation with, here and there, a grove of coconut palms."
Short is more specific. "Almost completely covered with short bushy trees including two small groves of coconut palms."  In 1937 the island was almost completely covered with Buka trees 60 to 90 feet tall. If Short thought they were "short bushy trees," a person on the ground would appear much smaller than he expected.  In 1937 there were five small groves of cocos.

Both Lambrecht and Short commented on Norwich City. Lambrecht estimates the ship at 4,000 tons.  Short's guess is 5,000 tons.  Norwich City's gross tonnage was 5,587.  Net tonnage was 3,513.  Both estimates were in the ballpark, but evaluating ships was what these guys were trained to do.

Curiously, Short does not mention the "signs of recent habitation" or the "repeated circling and zooming."  Either he didn't consider it worth mentioning or he was off looking at some other part of the island.  The planes could not communicate with each other except by hand signals.

 From what little information we have it would appear that the pilots were pretty good at what they were trained to do - evaluate ships - but not so good at estimating sizes and distances relating to terrain (hardly surprising).  There is also some indication that they didn't fly organized sweeps but split up and did their own thing.

-----------------------------------------------

What you wrote in your book was:

"According to present- day Civil Air Patrol Probability of Detection (POD) tables, the chance of Colorado’s planes locating the aircraft in the course of a single inspection of each island was on the order of 10 to 20 percent.   In other words, if the Earhart plane was on one of the islands of the Phoenix Group, there is an 80 to 90 percent chance that Colorado’s search missed it."

There was no indication in your book that your estimate did not come directly from the POD table. Only after I went through the calculation in the POD table did you change to now claim that your estimate was not directly derived from the POD table. So for those who only read your book and who do not participate in this forum you have left them with the impression that the 10 to 20% figure was gospel, with the blessing of the Civil Air Patrol. They haven't heard of your current distancing from that 10 to 20% claim, so it turns out your book statement is quite misleading.

Having acted as a forward observer for an 8 inch howitzer battery, I am well familiar with spotting the fall of shot. You are correct that the primary duty of the scout planes was to look for enemy ships and to adjust fire but they must have had some training in searching  for small objects since the Colorado planes flew three miles line abreast when searching the ocean for Earhart. If they were searching for a large ship, longer than 300 feet, with the prevailing 30 mile visibility reported by Lambrecht the plane spacing would have been 41 NM and 27 NM if searching for a ship 90-150 feet long. According to the current SAR manual 3 NM spacing is appropriate when searching for something the size of a large life raft so was also appropriate for a search for the assumed to be floating Electra. The Colorado pilots must have gotten that information from someplace such as search training.

You state that the planes could only communicate by hand signals but the second sentence in Lambrecht's report says that the radios worked fine.

I appears that your main point is that the vegetation on Gardner was much denser than that used in calculating the "heavy tree cover" column of the POD table so that the POD table overestimates the POD of finding someone among the trees on Gardner. O.K. for the sake of argument, I give you that point. However, you can not make the same argument for the "open, flat terrain" column of the POD table, open is open is open and is applicable to Earhart standing on the beach or on the reef flat. The POD predicts a 75% probability for one pass over open terrain increasing to 95% after only two passes. You state that at that time Earhart was at "camp Zero" a quarter mile north of the ship and inside the treeline for shade, so very close to the beach. Given the time the planes were over the island, Earhart had sufficient time to go out on the beach or reef flat and be found by the planes. You state that a P-3 flew over you standing on the beach and he didn't see you but you must admit that this is certainly an anecdote. If he didn't spot you after one pass so what, the table only predicted a 75% probability so one out of four times the plane would not be expected to spot you but this one anecdote doesn't disprove the accuracy of the POD table in predicting the probability in other searches such as the one conducted by the Lambrecht flight.

You also state that the planes flew over Gardner for only 18 to 28 minutes and that they flew at 90 knots which means that the planes flew between 27 and 42 NM while over the island. The circumference of Gardner is 8.5 NM so each plane could have completed 3 to 5 complete circles of the island making a total of 9 to 15 passes counting all the planes. No matter how they arranged the search of Gardner, each area got the equivalent of 9 to 15 passes unless all the airplanes did not take part in the search and there is no evidence that that was the case.  If they used some other search method, such as each plane covering its own one-third of the island, then each third of the island got the equivalent of 9 to 15 passes but, in this case, from just one plane. Since the strip of land is only one-half NM wide, if the three planes flew line abreast then they were spaced only 1/6th of a mile apart and the furthest any observer had to scan to each side was 1/12th of a nautical mile, only 506 feet, plenty close enough to spot someone on the ground unless the view was completely blocked by vegetation (your point) but no vegetation blocked the view out on the beach or reef.

So, based on the small amount of land to be searched and the time and miles covered in searching it the cumulative POD table predicts a high confidence level that they would have been found if they were on the island. I note your dispute about this if they were obscured by the trees but there is no way for you to dispute the applicability of the cumulative POD estimate for the beach area.

gl



« Last Edit: November 09, 2011, 12:56:48 PM by Ric Gillespie »
Logged

Ric Gillespie

  • Executive Director
  • Administrator
  • *
  • Posts: 5898
  • "Do not try. Do or do not. There is no try" Yoda
Re: Odds of Spotting Survivors from the Air
« Reply #11 on: November 09, 2011, 12:53:20 PM »

What you wrote in your book was:

"According to present- day Civil Air Patrol Probability of Detection (POD) tables, the chance of Colorado’s planes locating the aircraft in the course of a single inspection of each island was on the order of 10 to 20 percent.   In other words, if the Earhart plane was on one of the islands of the Phoenix Group, there is an 80 to 90 percent chance that Colorado’s search missed it."
There was no indication that your estimate did not come directly from the POD table. Only after I went through the calculation in the POD table did you change to now claim that your estimate was not directly derived from the POD table.

Are we now going to debate the definition of "derived?"  I used the POD tables exactly the way Andrew McKenna - an experienced CAP search pilot - says they are used in actual practice.  I stand by what I wrote.

...they must have had some training in searching  for small objects since the Colorado planes flew three miles line abreast when searching the ocean for Earhart. If they were searching for a large ship, longer than 300 feet, with the prevailing 30 mile visibility reported by Lambrecht the plane spacing would have been 41 NM and 27 NM if searching for a ship 90-150 feet long. According to the current SAR manual 3 NM spacing is appropriate when searching for something the size of a large life raft so was also appropriate for a search for the assumed to be floating Electra. The Colorado pilots must have gotten that information from someplace such as search training.

Do you suppose that this mythical 1937 search training was different for carrier pilots than for cruiser and battleship pilots?  If not, then why did Lexington's planes use 2 NM spacing?  (This was Lambrecht's first cruise as a battleship floatplane pilot.  His previous billet was aboard Lexington.)  The discrepancy between the spacing used by Colorado and Lexington to search for exactly same thing in the same environment would seem to be a pretty good indication that there was no accepted standard. Isn't is more likely that everyone was making it up as they went along? 

You state that the planes could only communicate by hand signals but the second sentence in Lambrecht's report says that the radios worked fine.

They worked fine for what they were for - communicating with the ship using Morse code.  VHF plane-to-plane voice didn't come along until later. 

You state that at that time Earhart was at "camp Zero" a quarter mile north of the ship and inside the treeline for shade, so very close to the beach.

Where did I state that?  I don't know where she was when Colorado's planes came over.  Do you?  We speculate that there was a "camp zero" a quarter mile north of the ship and inside the treeline for shade but I don't think you understand where the tree line is.  Between the beach and the tree line of "Buka" forest is 100 meters or more of dense, almost impenetrable "scaevola" bush.  If AE and FN had a "camp zero" in the Buka forest they must have made a path around and/or through the scaevola.  I once cut a straight 50 meter path through scaevola by myself using a good sharp, well balanced bushknife.  It took me two full work days.  I think it's more likely that AE and FN found a way around the denser patches of scaevola and ended up with a winding path from the beach back to the forest.

But since you bring it up, let's address the question: Where were our heros when Colorado's planes came over? - as long as we accept that any conclusion we draw is speculation.

Colorado's planes launched at 18:30Z and it should have taken them about an hour and a three quarters to fly to McKean, take a look around there, and fly to Gardner.

Let's say it's 20:15Z on Friday morning July 9 - that's 09:15 Gardner time using -11. And, just for the sake of argument, let's say that AE and FN (if he's still alive) have a camp back in the Buka forest.  The sun has been up for two and half hours so it's already warm - low to mid 90°s F on a typical July day based on our experience.  The last credible message was sent on Wednesday evening so the plane is gone.  High tide was two hours ago and there's a pretty good surf running, so ambient noise is high.  It will be at least another hour before the water level on the reef is low enough to look for fish trapped in tidal pools so there's no motivation to go to the reef for food.

What are our castaways most likely doing? Are they asleep? Nursing their injuries? Out near the beach watching the horizon? Back in the forest or over on the lagoon shore searching for water or food?

I think it's reasonable to eliminate one of those possibilities. If they are out near the beach watching the horizon they should see the planes coming, get out in the open, and do everything they can to make themselves seen.  Seems like they should have been seen, but that didn't happen, so they probably weren't out on the beach waving - either because they weren't on the island at all or because they were someplace else on the island.


 Given the time the planes were over the island, Earhart had sufficient time to go out on the beach or reef flat and be found by the planes.

At least you didn't say "would have" so you apparently know this to be a fact.  How do you know?  Let's stack the deck in your favor and put her at the camp in the Buka forest.  The race-to-the-beach clock starts ticking when she realizes there are airplanes over the island.  If you're back in the trees you don't see anything overhead. Been there, done that.  And in my experience you don't hear an aircraft on Niku unless it's directly overhead - too much ambient noise from wind and surf.  So how much time she had to get to where she could be seen depends upon at what point a plane passed over where she was.  How fast she could get out to the beach depends upon how winding the presumed trail was and how fast she could move.  If she had an injury that effected her mobility it might easily take several minutes. How long did the planes stay over one spot?  How long did Lambrecht circle and zoom?  You can construct a scenario in which she should have been spotted and you can just as easily construct a scenario that has her getting to the beach just in time to watch the planes fly away.


So, based on the small amount of land to be searched and the time and miles covered in searching it the cumulative POD table predicts a high confidence level that they would have been found if they were on the island. I note your dispute about this if they were obscured by the trees but there is no way for you to dispute the applicability of the cumulative POD estimate for the beach area.

So what are you saying? What conclusion do you draw from the fact that Lambrecht didn't report seeing anyone?

« Last Edit: November 09, 2011, 12:57:18 PM by Ric Gillespie »
Logged

Monty Fowler

  • T5
  • *****
  • Posts: 1078
  • "The real answer is always the right answer."
Re: Odds of Spotting Survivors from the Air
« Reply #12 on: November 10, 2011, 07:23:40 PM »

*scratches head* Can we at least all agree on the bottom-line fact that Amelia and Fred were not seen by the Colorado's planes?

LTM,
Monty Fowler
TIGHAR No 2189 CER
Ex-TIGHAR member No. 2189 E C R SP, 1998-2016
 
Logged

Ric Gillespie

  • Executive Director
  • Administrator
  • *
  • Posts: 5898
  • "Do not try. Do or do not. There is no try" Yoda
Re: Odds of Spotting Survivors from the Air
« Reply #13 on: November 10, 2011, 08:38:51 PM »

Can we at least all agree on the bottom-line fact that Amelia and Fred were not seen by the Colorado's planes?

There are only two conclusions that can be drawn from what we know about what they saw:
- They saw "signs of recent habitation" on an island that hadn't been inhabited since 1892
- They didn't see any people
Logged

Gary LaPook

  • T5
  • *****
  • Posts: 1624
Re: Odds of Spotting Survivors from the Air
« Reply #14 on: November 11, 2011, 02:33:33 AM »

What you wrote in your book was:

"According to present- day Civil Air Patrol Probability of Detection (POD) tables, the chance of Colorado’s planes locating the aircraft in the course of a single inspection of each island was on the order of 10 to 20 percent.   In other words, if the Earhart plane was on one of the islands of the Phoenix Group, there is an 80 to 90 percent chance that Colorado’s search missed it."
There was no indication that your estimate did not come directly from the POD table. Only after I went through the calculation in the POD table did you change to now claim that your estimate was not directly derived from the POD table.

Are we now going to debate the definition of "derived?"  I used the POD tables exactly the way Andrew McKenna - an experienced CAP search pilot - says they are used in actual practice. [1] I stand by what I wrote.

...they must have had some training in searching  for small objects since the Colorado planes flew three miles line abreast when searching the ocean for Earhart. If they were searching for a large ship, longer than 300 feet, with the prevailing 30 mile visibility reported by Lambrecht the plane spacing would have been 41 NM and 27 NM if searching for a ship 90-150 feet long. According to the current SAR manual 3 NM spacing is appropriate when searching for something the size of a large life raft so was also appropriate for a search for the assumed to be floating Electra. The Colorado pilots must have gotten that information from someplace such as search training.

Do you suppose that this mythical 1937 search training was different for carrier pilots than for cruiser and battleship pilots?  If not, then why did Lexington's planes use 2 NM spacing?  (This was Lambrecht's first cruise as a battleship floatplane pilot.  His previous billet was aboard Lexington.)  The discrepancy between the spacing used by Colorado and Lexington to search for exactly same thing in the same environment would seem to be a pretty good indication that there was no accepted standard. Isn't is more likely that everyone was making it up as they went along? 

You state that the planes could only communicate by hand signals but the second sentence in Lambrecht's report says that the radios worked fine.

They worked fine for what they were for - communicating with the ship using Morse code.  [2]VHF plane-to-plane voice didn't come along until later. 

You state that at that time Earhart was at "camp Zero" a quarter mile north of the ship and inside the treeline for shade, so very close to the beach.

[3]Where did I state that?  I don't know where she was when Colorado's planes came over.  Do you?  We speculate that there was a "camp zero" a quarter mile north of the ship and inside the treeline for shade but I don't think you understand where the tree line is.  Between the beach and the tree line of "Buka" forest is 100 meters or more of dense, almost impenetrable "scaevola" bush.  If AE and FN had a "camp zero" in the Buka forest they must have made a path around and/or through the scaevola.  I once cut a straight 50 meter path through scaevola by myself using a good sharp, well balanced bushknife.  It took me two full work days.  I think it's more likely that AE and FN found a way around the denser patches of scaevola and ended up with a winding path from the beach back to the forest.

But since you bring it up, let's address the question: Where were our heros when Colorado's planes came over? - as long as we accept that any conclusion we draw is speculation.

Colorado's planes launched at 18:30Z and it should have taken them about an hour and a three quarters to fly to McKean, take a look around there, and fly to Gardner.

Let's say it's 20:15Z on Friday morning July 9 - that's 09:15 Gardner time using -11. And, just for the sake of argument, let's say that AE and FN (if he's still alive) have a camp back in the Buka forest.  The sun has been up for two and half hours so it's already warm - low to mid 90°s F on a typical July day based on our experience.  The last credible message was sent on Wednesday evening so the plane is gone.  High tide was two hours ago and there's a pretty good surf running, so ambient noise is high.  It will be at least another hour before the water level on the reef is low enough to look for fish trapped in tidal pools so there's no motivation to go to the reef for food.

What are our castaways most likely doing? Are they asleep? Nursing their injuries? Out near the beach watching the horizon? Back in the forest or over on the lagoon shore searching for water or food?

I think it's reasonable to eliminate one of those possibilities. If they are out near the beach watching the horizon they should see the planes coming, get out in the open, and do everything they can to make themselves seen.  Seems like they should have been seen, but that didn't happen, so they probably weren't out on the beach waving - either because they weren't on the island at all or because they were someplace else on the island.


 Given the time the planes were over the island, Earhart had sufficient time to go out on the beach or reef flat and be found by the planes.

At least you didn't say "would have" so you apparently know this to be a fact.  How do you know?  Let's stack the deck in your favor and put her at the camp in the Buka forest.  The race-to-the-beach clock starts ticking when she realizes there are airplanes over the island.  [4]If you're back in the trees you don't see anything overhead. Been there, done that.  And in my experience you don't hear an aircraft on Niku unless it's directly overhead - too much ambient noise from wind and surf.  So how much time she had to get to where she could be seen depends upon at what point a plane passed over where she was.  How fast she could get out to the beach depends upon how winding the presumed trail was and how fast she could move.  If she had an injury that effected her mobility it might easily take several minutes. How long did the planes stay over one spot?  [5]How long did Lambrecht circle and zoom?  You can construct a scenario in which she should have been spotted and you can just as easily construct a scenario that has her getting to the beach just in time to watch the planes fly away.


So, based on the small amount of land to be searched and the time and miles covered in searching it the cumulative POD table predicts a high confidence level that they would have been found if they were on the island. I note your dispute about this if they were obscured by the trees but there is no way for you to dispute the applicability of the cumulative POD estimate for the beach area.

[6]So what are you saying? What conclusion do you draw from the fact that Lambrecht didn't report seeing anyone?
--------------------------------------------

1. We have both stated our positions regarding the probability that Earhart and Noonan would have been detected by Lambrecht and the other 5 aviators in his search formation if they had been on the island. Your position is that there was a 80 to 90% chance that they would have been missed by the searchers and my interpretation of the cumulative POD table is that the probability of their being missed was only 10% if they were in the treeline and only a 5% chance if they were standing out in the open, on the beach or on the reef flat. I posted the tables so the reader can do his own calculation and come to his own conclusion, it ain't rocket science. But I want to add a little additional information to aid the evaluation.

Table 4-4 of the National Search and Rescue Manual addresses searching open ocean.  It states that when searching for a person in the ocean wearing only a life preserver, that the track spacing should be 0.4 NM. Of course, searching the ocean is different than searching land, it is much more difficult to spot a head bobbing among the waves than it is to spot a person on dry land. Since the searchers look to each side of the flight path a distance of half the spacing, then when searching for a person in the water, the searchers look 0.2 NM each side.  Table 4-4 shows that there is a high probability of spotting just a head bobbing at this distance, 0.2 NM, (1215 feet). Figure 5-19 of the SAR manual shows this probability to be 80% for one pass, increasing to 95% after two passes and almost 100% after only three passes. Obviously, if you can spot just a head floating among waves at 1215 feet then you have a much higher probability of spotting an entire person on dry land at the same distance.

In my prior post, for simplification, I assumed the strip of land making up Gardner Island between the lagoon and the sea was half a nautical mile wide (3038 feet), but this was an overstatement. In fact, 39% of this donut is less than 700 feet wide and a further 45% is less than 1200 feet wide. Only the northern end of the island is a half nautical mile wide. This means that the search planes flying down the center of the strip of land would only have to search 350 feet either side of the plane (a little bit longer than a football field) for 39% of the circuit and 600 feet for 45% of the circuit. Only on the northern tip, constituting the remaining 16% of the island,  would they have to search a quarter mile either side, 1519 feet. You can see then that for fully 84% of the circuit the the distance they would have to look was significantly less than the distance that would allow spotting a bobbing head out on the ocean so should have had a very high probability of spotting an entire person on dry land. Only on the northern tip would the search distance be slightly greater, 1519 feet versus 1215 feet, than you would expect to spot a bobbing head among the waves so you would expect to be able to spot an entire person at this distance.

2. Regarding the radios in the planes, even if the planes did not have VHF or HF voice capability but only CW to communicate with the Colorado there was nothing to stop them from sending morse code messages from plane to plane, the radios work both ways.


3. You asked "where did I state that?"

Well, you wrote the following in your November 4th post:

"So if Earhart and Noonan were in the vicinity of Camp Zero (about a quarter of a mile north of the shipwreck and inland under the buka trees for shade) when the planes came over, why weren't they seen?"

4. Since your statement as to where you think camp zero was located is pure speculation then I have the right to also speculate. I think that if they were on the island that they made their camp on the beach near the NC so as to be near where they expected any searchers would look first and to have the benefit of any little breezes that might come along instead of going through all the tough scaevola brush to be in a windless and oppressively hot location under the trees. They built a shelter out of either vegetation or with materials salvaged from the plane or a combination of both.

5. You speculate that Lambrecht (and the other two planes) only did a cursory "circle and zoom" and then flew away but using your own estimate of the time they were over Gardner, 18 to 28 minutes, it is simple math to show that each of the three planes had enough time to make three to five complete circuits of the island, a total of 9 to 15 passes over each spot on the island for the flight of three planes. They were motivated to find Earhart (everyone wants to be a hero) so it is much more likely that they did a very thorough job in conducting this search than your speculation that they were off on a lark.

6. Looking just at the Lambrecht search, as interpreted through the use of the modern tables in the National Search and Rescue Manual and the CAP manual which show that there is a very high probability that they would have been spotted if they had been on the island, my conclusion is that it is highly probable that they were NOT on the island at the time the search was made.

gl

« Last Edit: November 12, 2011, 03:20:08 PM by Gary LaPook »
Logged
Pages: [1] 2 3 ... 10   Go Up
 

Copyright 2021 by TIGHAR, a non-profit foundation. No portion of the TIGHAR Website may be reproduced by xerographic, photographic, digital or any other means for any purpose. No portion of the TIGHAR Website may be stored in a retrieval system, copied, transmitted or transferred in any form or by any means, whether electronic, mechanical, digital, photographic, magnetic or otherwise, for any purpose without the express, written permission of TIGHAR. All rights reserved.

Contact us at: info@tighar.org • Phone: 610-467-1937 • Membership formwebmaster@tighar.org

Powered by MySQL SMF 2.0.18 | SMF © 2021, Simple Machines Powered by PHP