Amelia Earhart Archaeology
A discussion of research into the Nikumaroro Hypothesis on the 1937 disappearance of Amelia Earhart and Fred Noonan, hosted by research archaeologist Tom King.
Friday, November 4, 2022
Machine Learning With Amelia Finds An Audience in the U.K.
Wednesday, March 23, 2022
Machine Learning with Amelia
by Joe Cerniglia
Incredibly, it has been more than two years since the last post was published here on Amelia Earhart Archaeology. It is time to revive the search for Amelia!
But before I do, I need to insert a little background on what I've been up to, and then tie that back to the Earhart search. Lately, I have been trying to learn new skills. One of those is Machine Learning. I have been studying this subject independently for about four months now, and I will be taking my first Machine Learning course with eCornell, the online division of Cornell University, next week. Next year I will be working on the Machine Learning certificate program offered by eCornell.
One of our assignments in the current course I am taking was to locate a data set of interest, load the data into a Jupyter notebook and carry out some analyses. Jupyter notebook is a tool that allows coders to integrate their code and the result of the code, plus explanatory text. While this may not seem all that revolutionary, research papers that transparently show all of the methodology behind their research are all too uncommon. Very often it becomes very difficult for those who wish to follow up on research to see the work behind it and exactly how a result was derived. The tools and languages used are often disclosed but the exact code and data remain elusive.
There is a name for this phenomenon, and I think it's a good one. It's called the reproducibility crisis.
The Alan Turing Institute in the U.K. has recognized the crisis and is urging the adoption of uniform standards of reproducibility in academic research. They identify one source of the problem as follows:
Issues in reproducible research predominantly stem from academic incentives that encourage competition between research teams working on similar questions. The system means that an individual research team - behaving rationally - is likely not to share their data, code, protocols nor experimental design expertise with researchers working in the same area as themselves. The outcomes will likely result in siloed knowledge and a lack of transparency in research methods.
Find their article here: Turing Response to Reproducibility and Research Integrity Inquiry
Predictably, and unfortunately, this phenomenon is not limited to academia. Obscuration seems to be popular everywhere, even in such things as product packaging.
In my own small way, then, I want to help address the reproducibility crisis in Amelia Earhart research by making available a research effort of my own.
While it has been two years since this blog had a post, it has been about 10 years since news of the famous freckle cream jar, which just possibly may have belonged to Amelia Earhart, first hit the airwaves.
Our joint paper on this topic, A Freckle in Time, first released in October 2013, stated that the jar had an unusual chemistry that was closer to that of window glass than that of cosmetic containers. We had some data in books that informed us of this fact, but we lacked real data that we could use to verify this for ourselves.
Casting about for topics to incorporate into my coursework, I came upon a glass dataset from the University of California at Irvine Machine Learning Repository:
https://archive.ics.uci.edu/ml/datasets/glass+identification
Recalling our work on glass chemistry in our paper, I thought that this would be an excellent opportunity to put our finding of 'unusual chemistry' to the test.
My goal was to use the UCI glass database to build a machine learning model, and then to apply that model to both the artifact jar found on Nikumaroro and to the clear facsimile jar in the same size and shape (but not color) found on eBay. The model would then determine which of several types of glass would best categorize these two samples.
What I learned was that a prudently tweaked ML model can indeed spot the clear facsimile correctly as a container. Also, as we had suspected nine years ago, the artifact jar of reputed freckle fame was identified by the model as a 'window non-float,' a variety of glass that is most often found in churches.
Usually in Machine Learning, a misidentification is considered a failure, but in this case, I would consider the misidentification a success. It showed, in a reproducible and much more rigorously scientific way than previously we had shown, that the artifact jar is indeed unusual and original in its chemistry. I further speculate that the inability of the model to characterize it is not the result of a less than fully robust sample dataset, but rather the result of a lack of real-word siblings to be found, even in the 1930s.
While admittedly some of this IS speculation, the absence of evidence I observe is one that is drawn from years of searching for the exact twin to the artifact jar, so far without result.
My research, transparently presented with the original Python source code and research result, may be found here on the Binder website.
A word of instruction on using Binder:
When you click on the link above, you will be brought to Binder, a website that hosts Jupyter notebooks for interactive sessions. The site may take 2 to 5 minutes to load and, if the site is busy, it may not load at all. Binder is a free service, paid for by a small number of corporations committed to open source work. The website states in its documentation:
We are still working on defining what the exact goals for uptime and reliability should be.
Understandably, they are still working out the kinks. I find, however, that when I shut down my browser completely, and then re-click on the link, the service then becomes available.
When you reach the workspace, in the left-hand pane double-click on the first document: DatasetPlayground_interactive_SMOTE.ipynb. When this document has opened click on the >> tool button at the top of the page to execute the code. Inside the document, you will be instructed in how to use the two interactive exhibits it contains, so that you may explore the data and verify the findings independently.
The conclusions I draw will probably not be controversial, nor will they be greeted with the fanfare of the initial reports of freckle creme, but they do constitute progress.
I have many more articles to post here, but first I need to write them. This will take time, but they are forthcoming. I appreciate the loyal readers and followers of this blog, and most of all, I appreciate your patience!
June 15, 2022 Update: I have written a much-expanded version without interactive links that may be viewed here:
Saturday, November 2, 2019
Making Room for a New Guess
by Joe Cerniglia
Robert Ballard and crew members of the Expedition Vessel Nautilus have concluded their mission to Nikumaroro. From August 9 to August 21, 2019, Ballard and his team scanned the ocean floor adjacent to Nikumaroro for signs of large pieces of Amelia Earhart’s plane. The Nautilus’ search area was more extensive than any search of the island’s deep water margins to date, reaching, at times, a distance more than a mile from shore. However, no Electra, fragmentary or otherwise, was found. In related work, a team supervised by Fred Hiebert, archaeologist-in-residence for National Geographic, searched human bone collections held by the Te Umwanibong Museum and Cultural Centre in Tarawa, Kiribati and found a fragmentary skull. Although details of the provenance of this skull are lacking, members of the team believe it may be the same skull that was discovered in 1940 on Nikumaroro and sent on to Fiji for analysis in 1941. If the skull, retrieved from Tarawa and now undergoing analysis in Florida, yields DNA that can be matched with that of relatives of Amelia Earhart, Ballard has promised to return to the island to resume the underwater search for Earhart’s vanished Electra 10e.
And if the skull does not display a clear biological link to Amelia Earhart, then, since the airplane has also not been found, must the Nikumaroro hypothesis finally and irrevocably now be said to have been disproven?
Many bloggers and screenwriters have already sharpened their pencils and begun to write that it has.
Those who wanted the Earhart mystery solved had pinned their hopes on a large piece of the Electra peering through the murky depths at the camera lights of Argus and Hercules, the remotely operated vehicles of the Nautilus. The failure to find the Electra, however, should not invite hasty conclusions.
The missing airplane may, however, be an opportunity to reexamine what we think we know or what we suppose must be true in order for the Nikumaroro hypothesis to be correct.
1) Pieces of the Electra are not underwater because the Electra was never at Nikumaroro. (Other elements of the hypothesis for Nikumaroro, however, would in that case continue to defy a unifyingly plausible explanation.[2])
2) Pieces of the Electra are underwater in areas that have been explored, but they are too encrusted with coral and ocean silt or other deposits to be noticed.
3) Pieces of the Electra are underwater but are hiding in shoreward areas not explored by divers and too shallow for exploration by autonomous and remotely-operated vehicles.
4) Pieces of the Electra are underwater but the plane floated many miles offshore and sank into abyssal depths.[3] (The farther one searches from shore, however, the more difficult it becomes to demonstrate that the Electra, if found, ever actually reached Nikumaroro.)
5) Pieces of the Electra are underwater but have by now been pulverized into smaller pieces that are not large enough to be noticed [4], wherever they may lie, or they appear insignificant when compared with other nearby identifiable objects such as S.S. Norwich City debris.
6) Pieces of the Electra were underwater in shallow areas close to land but have been gradually washing up onto the shoreline, driven by wind and storms, with little of it visible today.
There may be other possibilities, but these appear to be the most likely ones. Other than the null hypothesis (#1), which excludes all the others, more than one of these possibilities could overlap; that is, some could jointly be true.
Anecdotal accounts, over a long time frame, lend some support to this conjecture.
- Emily Sikuli, who lived on the island from 1939 to 1941, stated she saw a large piece of an airplane on the northern shore of Nikumaroro near the wreck of the S.S. Norwich City. In an interview, when asked how large the airplane was, Emily stated, holding her arms out, "About four arm spans."[7]
- Pulekai Songivalu, who lived on the island after World War II as resident schoolmaster, stated, "I saw pieces of airplane in the lagoon."[8]
- Tapania Taiki, daughter of Pulekai Songivalu, who lived on the island from 1958 or 1959 and left in the mid 1960s, stated, "I saw a piece of airplane wing on the reef."[9]
- Between 1944 and 1945, John Park Mims (1920-2018) made many trips to Nikumaroro in a PBY from Canton to resupply the Coast Guard LORAN station stationed on the island. He stated that he witnessed a large fish having been recently caught by the islanders. The hook in the fish's mouth was, according to the islanders, made from aluminum from a downed airplane. The line leader was aircraft control cable, recognizable to Mims and to his friend, Eyvind Wahlgren, as from a type of aircraft that was smaller than a PBY.[10]
TIGHAR, after more than 21 years of exhaustively researching the photo, ultimately dismissed it in 2009 as a depiction of the wreckage of a Japanese Tachikawa Ki-54 on Agrighan Island. It based its identification of where the photo had been taken on the logs of the HMS Adamant, whose only recorded port of call on an island seeming to match Carrington's description in 1946-47 had been Agrighan. It based the identification of the airplane in the photo on the way the engine cowlings had come apart, which were different from how an Electra’s engine should have, and on what it claimed was the presence of a “small round inspection plate” just forward and a little rightward of the center line from the windshield, on the nose section of a surviving Tachikawa Ki-54s housed in the Beijing Aviation Museum. Both the airplane in the wreck photo and the airplane in the Beijing Aviation Museum, it claimed, carried this same inspection plate in identical locations.[12] What TIGHAR claims to be an inspection plate in photos, however, seems difficult to distinguish from what may actually be a small hole, nor is it apparent from its research that all Tachikawa Ki-54s carried an identical inspection plate, nor even that the oval-shaped spot on the photo of the Tachikawa K-54 in Beijing is an inspection plate.
Indeed, the Electra itself had its propellers turned significantly forward when it crashed in March, 1937 at Luke Field in Hawaii, thus indicating the engines were under full power at the time the propellers impacted the runway.[14]
A Tachikawa Ki-54 recovered from Lake Towada in Japan has its propellers turned significantly rearward.[15]
However, if the airplane in the photo was not attacked, it would seem that for it to sustain this kind of damage, without damage to the propellers, it must have been landed successfully and been abandoned by or lost from the control of its owner. If such an abandonment or loss of control occurred, this would seem to have been a very unusual event, highly atypical of most aircraft incidents or crashes. It is, however, exactly the scenario envisioned by TIGHAR for Earhart’s Electra.
Photos of different models of the Japanese Tachikawa Ki-54 with propeller hubs in place.
This photo of a wrecked Tachikawa Ki-54 shows one propeller hub is present and one is missing. The port engine has lost its propellers and its propeller hub. Both engines show significant compression damage around the cowling, indicative of violent impacts. The surviving propeller blades on the starboard engine are bent. [17]
Most noticeable in the background of the wreck photo are the presence of several tattered and wasting coconut trees. Only a year after the photo was said to have been taken, several telegrams to the District Officer on Canton Island express concern regarding the condition of the coconut trees on Nikumaroro. These telegrams suggest ways of coping with the difficulty, including burning Buka groves in order to rid these areas of substances “poisonous for coconuts.”[18]
The problem of maintaining the health of the coconut trees on Nikumaroro appears to have been a perennial concern up until the time the colony was abandoned. As late as 1963, the Pacific Islands Monthly related how drought conditions had killed off many of the coconut trees of the colony, resulting in the necessity of the colony’s abandonment.[19]
"Agrighan Island, although of volcanic origin, is covered for the most part by rich, black, loamy soil conducive to a heavy growth of vegetation and plant life. It is true that the greater portion of the coastline and a large section of the eastern half of the island is made up of sharp, jagged, lava rock in the form of caves, ledges and cliffs. But excluding these portions of the island, most any kind of vegetables or fruits could be produced to maintain many times the number of natives found there. There are areas on the southwestern, southern and southeastern part of the island that could easily be turned into pasture land for the maintenance of live stock (sic)."[20]
No mention is made in McAfee’s report of the types of problems with dying coconut trees that appear to have very much troubled the Nikumaroro colony.
If the wreck photo truly depicts a Tachikawa Ki-54 that crashed on Agrighan Island, none of the three aircraft that crashed on Agrighan fit that description.
A B-29 was reported lost on the northeastern side of Agrighan on April 27, 1945 [21] and the Bureau of Aircraft Accident Archives (https://baaa-acro.com) shows two aircraft accidents during the 1950s on Agrighan, occurring within days of one another.[22] On December 17, 1953, a Consolidated PB4Y-2 Privateer, conducting weather observations on Typhoon Doris, crashed somewhere off Agrighan. Three days later, a Douglas C-47 Skytrain (DC-3) took off from Andersen AFB on Guam in search of the lost crew. At some point, the airplane lost visual reckoning of its altitude, and crashed into the volcano in the center of Agrighan Island. All crew of both airplanes were lost, as was the crew of the B-29 despite a vigorous search for one member thought to have parachuted to safety.[23]
The B-29 was a four-engine bomber whose engines powered four-bladed propellers,[24] easily distinguishable from the two-bladed propellers shown in the Wreck Photo. The Douglas C-47 Skytrain was powered by three-bladed propellers.[25] The Consolidated PB4Y-2 Privateer was also powered by three-bladed propellers.[26] None of the three recorded aircraft wrecks on Agrighan have been documented in recent archaeological surveys.[27] It is, of course, not impossible that a Tachikawa Ki-54 crashed on Agrighan during World War II and was not documented, but one must needs ask what a Japanese aircraft on wartime patrol would be doing reconnoitering an island that was of limited strategic importance and that they already controlled. In 1945, Agrighan was inhabited by about 120 natives, who farmed a coconut plantation managed by four Japanese overseers.[28]
Even if many small Electra 10e pieces may have been collected by TIGHAR in its various expeditions to the island since 1989, and even if a large piece here and there were transported from Nikumaroro to other islands, such as to Canton in the 1970s[29], this would still leave a large portion of the airplane, presumably waiting to be discovered in the scaevola of Nikumaroro or along the beach, intact.
So far, however, no large engines or wings or cockpit have emerged from the many expeditions TIGHAR members have made to the island. No smaller pieces that were both easily and uniquely identifiable to the Electra 10e have been found.[30]
According to Dr. Thomas King, there is much evidence in the old colonial village on Nikumaroro of colonists having been interested in quarrying aluminum for handicrafts and tools. He writes:
“Lots of aluminum, all in small pieces, usually more or less rectangular, cut from larger pieces with tin snips or maybe a machete. Some pieces were in an area more or less between the houses—marked by the remains of a stone-lined platform—and the dense charcoal and bone deposit left by the cookhouse. With one of the aluminum pieces were pieces of pearl shell and a glass bead. It looked like someone at Manybarrels had been making handcrafts with aluminum decorations.”[31]
On the other hand, this interest seems to have been greatly tempered and restrained by a kind of superstition that anything to do with the airplane was haunted and, thus, dangerous and forbidden. Tapania Taiki, daughter of the schoolmaster on Nikumaroro, who lived on Nikumaroro from the late 1950s to the early 1960s, stated in a 1997 interview:
“The older people said they saw the skeletons of a man and a woman, one each. The elders said ‘Do not go to where the plane is, there are ghosts there.’ They were trying to scare us to keep us away from there.”[32]
It is difficult to imagine or explain the reasons behind this combined fascination with and prohibition against airplane wreckage that colonists reported having seen on the island. Nevertheless, no matter the motivations for doing so, the elder colonists appear to have processed and transmuted significant amounts of aircraft aluminum, and other pieces of equipment from some complex machinery, for their own purposes.
We see possible evidence of this transmutation in the debris of the old carpenter shop. There, the remains of an old workbench or storage cabinet is piled high with circular bearing sleeves, cables of various gauges, and metal tubes. Some of these items have been marked with orange and pink archaeological flagging tape from previous expeditions to the island.
- Ballard and his team searched the waters around Nikumaroro extensively to a depth of some 900 meters and to a distance more than a mile from shore. No obvious signs of an aircraft were spotted.
- Pieces of aircraft wreckage were reported by colonists at various times near or on the northern part of the island.
- These reports were echoed by reports from two supply personnel for the U.S. Coast Guard station on the island.
- A British seaman reported, and took a photo of, his 1946/1947 discovery of a radial aircraft engine and other pieces of obvious aircraft wreckage on what he took to be a deserted island in the Pacific.
If these reports or even some of them actually happened on Nikumaroro, then the best and perhaps the only place to look for what small remnants of Earhart’s Electra 10e might remain is in the land areas of the island where wreckage was reported, both in the village and on Nutiran.
This essay should be construed as an exhortation to acknowledge that something different, perhaps very different, happened on the reef with the airplane than what those who have considered it have traditionally been willing to acknowledge.
That acknowledgement for some could mean an airplane that was never on Nikumaroro. But for those who still consider the evidence, the acknowledgement might productively involve investigating new approaches (water-penetrating LIDAR, for example, if such a thing has existed, or will exist), and - perhaps - revisiting some old approaches, by way of some old, discarded evidence, such as devoting some of the resources in time and money to the village and Nutiran that have up to now been poured into the ocean.
Such an effort would not be wasted. We would learn much more than we do about the life ways of a hardy and indomitable people, the colonial village denizens of Nikumaroro. In the process, we might also find an artifact, or two, that tells us more about the fate of a hardy and indomitable explorer who unwittingly may have found herself unrescued on this beautiful but forbidding place the islanders called home.
Remains of the workbench or cabinet from the carpenter's workshop. Photos by Joe Cerniglia from the 2017 Nikumaroro expedition.
=====================
I am indebted to Pacific historian Scott Russell for providing me with the U.S. Navy's 1945 report on the Agrighan search and rescue operation for the crashed B-29.
This paper is dedicated to the memory of Lt. Col. Northrop, USAAF, missing flyer from the B-29 that crashed on Agrighan on the morning of 27 April, 1945.
Accessed 2 November, 2019.