Summary of the Re-Navigation Research
This research was designed to conduct a detailed assessment of the body of foregoing World Flight research, critically review evidence in prior related works, validate or critique those works, and localize future search options. A reduction in the planned search area would enhance the project by reducing search time on station.
Approaching this task as a location of a lost aircraft, and definition of a probable search area, required understanding, to the fullest extent possible, the exact possible flight paths, profiles, speeds, flight times, fuel consumption, and pilot behaviors. Specifically in this case, these factors would determine flight time endurance remaining upon arrival in the Howland area - directly related to where the aircraft could be located. With little direct information on any of these factors, and the importance of accurate assessment, this information had to be created from widely disparate sources, research, and analysis.
A detailed review was conducted of at least ten authors writing directly about the World Flight, more than a dozen reports, and more than 8,000 pages of data associated with the World Flight attempt. Other documents examined included the entire Amelia Earhart Papers of the George Palmer Putnam Collection of 2,221 images from the Purdue University e-archives; Lockheed Electra and period aircraft operating manuals; meteorological and oceanographic data including the Lae-Howland geographical climatology; the effects of the Northern Equatorial Current, and Northern Equatorial Counter Current in the Howland area; and other authors/pieces with various theories about the disappearance of Amelia Earhart and Fred Noonan.
Aerodynamic engineering data and aircraft performance were examined in great detail, from many sources and authoritative records. Aircraft performance is a major, critically important variable in this analysis, and largely determines the vertical and lateral flight profile from Lae to Howland Island.
Amelia Earhart’s collection of flight notes, biography, life events, and her career in aviation were closely studied to gain insight into her motivations and beliefs. Perhaps most important, we wanted to understand Amelia’s behaviors — how she planned missions, flew aircraft, thought about flying them, and how she actually conducted her flights throughout her career in the air.
Recreating the Lae-Howland flight segment, using as much hard data and facts as were available, was critical to meeting research objectives. A faithful re-creation based on fact was the primary objective, and offered the best chance to accurately locate the aircraft.
Three possible flight paths were defined and evaluated, each terminating in a high confidence, End-of-Navigation point. Among the three paths, one path appears most likely (Path C), with a very high confidence End-of-Navigation point; one path is unlikely (Path A); and one is possible but with a lower confidence that it was executed (Path B).
The highest confidence Path C results from a rigorous path recalculation, aerodynamic performance and fuel consumption assessments, with significant cross-validation of results and conclusions. Error sensitivity analyses were performed on results for variable wind velocities, wind directions, and fuel consumption.
A Search Grid was constructed around this Path C End-of-Navigation point to accommodate terminal area maneuvering that was inferred from aviation experience, and application of the most likely behavior for Amelia Earhart and Fred Noonan, on July 2, 1937.
The search grid was initially oriented, and modified, as shown in Appendix 2, further refining Autonomous Underwater Vehicle search strategies. Several iterations of the grid with the Search Team resulted in the final search grid included in this report.
Previous estimates of position for Amelia Earhart and Fred Noonan’s Electra include
Northwest of Howland Island at 375nm +/-100nm (Safford)
425sm southeast of Howland at Gardner Island (Gillespie)
North of Howland Island at 52nm (Long)
On islands of New Britain, Mili Atoll (Marshall Islands), Saipan (Various)
Northwest of Howland Island within 30 miles (Nesbit)
The following are among higher confidence data that support analyses
The fuel load of the Electra leaving Lae was likely between 1080-1100 US gallons.
Chater reports the fuel load at 1100 gallons.
Collopy reports the fuel load at 1100 gallons.
Swenson and Culick calculate the fuel load at 1080 gallons.
Thunderstorms were forecast in at least two weather reports from Hawaii, at 250-300 miles east of Lae, and Amelia received one of these reports before leaving Lae. The second report was broadcast from Lae, to AE, during the first 7 hours of the mission.
The Electra departed Lae at 0000 GMT.
Of thirteen position reports made by Amelia Earhart from Lae-Howland, only two included a latitude and longitude position, and one of those is potentially in error in time and/or location.
This is unusual given Fred Noonan’s experience with making detailed position reports on South Pacific proving flights with Pan Am in 1935.
Before joining the World Flight, Fred wrote about the importance of complete position reports, including latitude and longitude, air and ground speeds, wind direction and speed, and outside air temperature, in a post-flight report following one of these trips.
The Lae-Howland reporting history is also unusual and unlike that accomplished on the Oakland to Honolulu, first leg attempt.
On this initial attempt, Amelia made 9 position reports, 4 with position latitude/longitude data, and on which Fred’s log shows approximately 35 celestial and/or navigation fix computations taken en route.
“…In all cases [Oakland-Honolulu initial World Flight Attempt] Earhart provided dead reckoning positions. Of the four documented positions, three were provided with times, but the wording provided by the USCG Hawaiian Sector leads to some ambiguity as to when Earhart stated these positions. Interestingly, all four messages indicate that the positions provided were well prior to the actual broadcast times. [The aircraft is beyond the waypoint reported]. Based upon this analysis, one can easily speculate that Noonan’s method was to project future positions via dead reckoning, and provide that information to the pilot sometime prior to the radio broadcasts. In no instance does Earhart provide timely information, nor does she provide an actual navigational/celestial fix and time of the fix to help constrain exactly where the plane was.”1
“In summary [Oakland-Honolulu initial World Flight Attempt] Noonan made use of seven radio bearings, 14 star/planet LOPs (of which nine were used for navigational fixes), and the plane made only four course corrections. Analysis of the flight path versus weather maps produced after this date show major concurrence with the winds aloft patterns. It is clear that the navigator’s major responsibility was to monitor the progress of the flight, and to suggest course corrections only when deviations from desired flight path became too extreme. Use of projected, future DR positions allowed Noonan to check his forecasts vs. later navigational fixes to update his speed and direction over the ground, and to offer approximate positions, when necessary.”2
Fred Noonan may have used this technique, if only partially reported by Amelia Earhart, on the Lae-Howland segment. There is no evidence to support that Fred functioned differently on this, his most difficult segment, than on prior segments. The lack of reporting integrity and consistency may be understandable in that throughout the World Flight, position reporting was infrequent, and accomplished mostly on the Lae to Howland segment.
From AE’s aircraft performance and re-calculated time of arrival at waypoints, compared with the time AE reported those waypoints, there is behavioral consistency in the technique outlined above.
This helps to characterize the reasonableness of these comparisons, understand the probability associated with each path, and assess the accuracy of navigation.
Of note is that at 1745 GMT, AE reported “about 200 miles out.” This was a position likely provided by FN using celestial fixes throughout the night of good visibility, made from excellent celestial bodies available, and therefore, an accurate position.
The aircraft’s distance from the 1937 Howland Island coordinates at the time of this report is 204 nautical miles, according to the Path C re-calculations.
The report and the position occur at AE’s typical reporting time of 15 and 45 minutes past the hour.
This appears that at 1745, the Electra was on track and on course to Howland, and FN calculated their position with good accuracy for 1937 equipment and methods.
This level of accuracy, while not routine in that period, was certainly possible.
This creates the possibility that something happened in the last 200 nautical miles distance to Howland Island.
After 0718 GMT, position reports were made in the blind.
Amelia had no pre-arranged communications between Lae and Itasca.
There were no arrangements for communicating with Ontario.
There were no arrangements for communicating with Nauru Island.
Aircraft aerodynamic performance was established with a high degree of confidence through data integration from many sources, and with consideration for pilot behavioral performance.
While radio strength is not entirely related to distance, strengths associated with the final few reports are the only indication of possible relative terminal area position.
Our research concludes for Path C, the most likely path, an End-of-Navigation point 35-28nm southwest of Howland Island, bearing 067 degrees to the 1937 position of Howland Island. A water entry area is shown for three fuel exhaustion scenarios (Swenson and Culick, Nutter, and Kelly Johnson) which plot theoretical points of fuel exhaustion following AE’s arrival at the End-of-Navigation point, as a function of fuel remaining at the End-of-Navigation point. These comprise theoretical position boundary limits, assuming AE conducted the search pattern depicted, throughout terminal maneuvering in search of Howland and Itasca. A high confidence water entry area is shown for the time 2013 GMT until 2100 GMT, likely from either fuel exhaustion, or from controlled flight into terrain, resulting from loss of situational awareness, fatigue, or abnormal mechanical circumstances. The maximum fuel remaining at 1912 GMT is computed at 123 gallons, and with a failure of the Cambridge Fuel Analyzer (discussed later in this report), the fuel remaining may have been 63 gallons, enough for approximately 90 minutes flying time.
Fuel consumption is discussed extensively later in this report. It is very likely that fuel exhaustion occurred between 2013 GMT and 2100 GMT.