Sturrock/Vallee Sidereal Time Debate
From: J. Maynard Gelinas <j.maynard.gelinas.nul>
Date: Wed, 5 Jun 2013 17:11:18 +0800
Archived: Wed, 05 Jun 2013 17:20:43 -0400
Subject: Sturrock/Vallee Sidereal Time Debate
Peter Sturrock, Professor Emeritus of Physics at Stanford, wrote
an interesting paper statistically comparing just under 12,000
UFO events collected in the Hatch "U" UFO database to Sidereal
Time. This paper was published in 2004 by the Society for
Scientific Exploration and is available here:
http://www.scientificexploration.org/journal/jse_18_3_sturrock.pdf
To explain, Sidereal Time is a unified means for astronomers
across the world rationalize the Earth's daily rotation and
yearly orbit in relation to astronomical objects like stars.
This way, the astronomical community knows they are discussing
the same objects on disparate calendar days and times across the
globe. Here's Wikipedia, which is much more detailed:
http://en.wikipedia.org/wiki/Sidereal_time
Sturrock stated that his intent was to expand upon previous work
done by Claude Poher and Jacques Vallee in the 1970s, which had
apparently shown a correlation to Sidereal Time using smaller
datasets. The importance of which, as Vallee noted in rebuttal
paper, is:.
"If UFO sightings tended to occur at a particular value of the
local sidereal time this might mean that a specific celestial
source was involved in the phenomenon. Therefore searches for
extraterrestrial signals (such as SETI) might be conducted in
the part of the sky in question." (Pg 1 of the Vallee rebuttal
paper)
Just a point of opinion, Sturrock's statistical work is very
rigorous, showing what a trained scientist with good analytic
skills can do with a large dataset. But the paper is a bear to
read. Comprising eight sections in the paper, he uses a range of
methods to slice up the data. He finally concludes in section
four a significant correlation to Sidereal Time.
I'll break this post down into two sections, a discussion on
Sturrock's paper and then Valle's rebuttal.
Sturrock Paper:
The Hatch data is broken up into four blocks, taken from 1892
through to 1999. These are: 1892 to 1957; 1957 to 1973; 1973 to
1982; 1982 to 1999. His histograms reflect this breakdown by
using four blocks per period analyzed.
He starts by breaking down the data into number of events for
year (fig 2) and total events per year (fig 3). Fig 2 is
particularly interesting as it shows a dramatic increase right
in per-year sightings around the mid-1940s. Why is this? I think
most would probably argue it was due to media saturation
regarding the Kenneth Arnold sighting in 1947. (And it should be
noted that Vallee threw out all the pre-1947 data in his
rebuttal).
Sturrock then breaks the data down into day of week (figs 4 and
5) which show little correlation to UFO reports, the curve
fitting perfectly within standard deviation blocks. But in Fig
6, time of day, we see a HUGE curve below and above mean,
indicating significant positive activity between 8pm - 11pm and
dead time from about 3am to 8pm. EXCEPT for block A (1892-1957),
which is still at or below mean but which shows a spike in
activity from about 10am through to 7pm. His mean and standard
deviation graph in Fig 7 shows that this result is significant
and not just a statistical artifact.
He then performed an 'hour of year' analysis (fig 8), which
broke up events into approx 15 day increments for 24 segments a
year. This showed an interesting peak across all blocks from
between Jul through Nov, except for block A (1892-1957), which
also showed significant spikes from around May to Jul). Then he
broke it down to by mean to standard deviation (Figs 9). Note
though that Standard Deviation is all over the place. The peaks
are significant, but the margin of error is high in some cases.
The power spectrum graph in Fig 10 is important. There are two
peaks, one at .42 which goes to a count of 308, and one at 1,
which goes to a count of 297. He ignores the peak at .42 and
focuses on the peak at 1y. This result shows that the amplitude
of spikes every year is very high in relation to events across a
five year span. That is, the probability of events peaks EVERY
YEAR, not every two years or every three years or every five
years.
Now we get to Section 4, which contains the primary argument for
correlation to Sidereal Time.
"We know that if the frame rotates with respect to the Earth
with a one-day rotation period, there is a very strong
modulation, shown in Figures 6 and 7."
Figs 6 and 7 are the Time of Day correlations. Those are known
to spike at between 8pm and 11pm.
"If it were to turn out that many events are due to observations
of the planet Jupiter, then one would expect to find a
significant pattern in terms of a frame that has the same mean
rotation rate that Jupiter has with respect to the Earth."
What this means is that the frequency of events happening yearly
during peak months is at a very high amplitude every year
relative to a scale from 0 to 5 years. Sturrock wants to know If
UFO reports were due to a planetary occurrence like Jupiter
intersecting that patch of the sky (a celestial
misidentification). So he decided to compare synodic period to
local sidereal time period. See this Wikipedia entry on orbital
periods:
http://en.wikipedia.org/wiki/Orbital_period
I found this section very difficult to interpret. Please offer
corrections if I've made a mistake. But, what he appears to be
doing is comparing Hour of Day and Hour of Year clustering peaks
for UFO reports with sidereal time to the orbital period for
Jupiter.
There's a good deal of math where he draws out this comparison,
but fig.12 - the result - is very important. The spectral
analysis on the top of the graph in fig 12 compares the
amplitude of the frequency of events of Sidereal Time on the top
to amplitude of the frequency for the synodic period of Jupiter.
??? (corrections please)
Notice we see a peak at 1y which goes up to ~125. And at 2y
going up to ~75, yet on the bottom of the graph there are no
counter negative peaks shooting down to about the same
amplitudes. What he's arguing is that if the positive peak is
high and the negative peak (Jupiter) is not symmetrical in
amplitude, it rules out a synodic period correlation. (if I
understand this right)
Sturrock then computes two additional graphs in figs 13 and 14
to confirm this. Ultimately concluding:
"Since the power spectrum computed from Equation 4.3 is not
symmetric in positive and negative frequencies, it appears not
to be due to the interplay of the HOD [Hour of Day] and HOY
[Hour of Year] modulations. The asymmetry indicates that there
is a modulation in terms of LST [Local Sidereal Time]."
Meaning, whatever the cause of UFO reports at the same patch of
the sky at the same time across the globe, it's not due to
Jupiter.
The rest of the paper are a series of statistical tests to
confirm the result and then a final short concluding statement.
Suffice it to say, his additional tests confirm the initial
result found in Section 4.
So now we move on to a short discussion in Section 8, the most
interesting of which is the patch of sky in Local Sidereal Time
he argues is the main focus of reported UFO events:
"If the effect does indeed have an astronomical origin, it
points to a source with right ascension in the range 21.5
Plus/Minus 1.5. Based on the analysis carried out to date, the
declination is unknown."
And then concluding that his analysis supports the Poher and
Vallee analyses of the 1970s.
"The analysis reported in this article appears to support the
early findings of Poher (1973) and of Poher and Vallee (1975),
which were based on independent and comparatively small
catalogs."
Vallee Rebuttal:
In 2007, Jacques Vallee responded to Sturrock's paper. I note
that Valle's work is a significantly easier paper to read and
understand.
http://www.ufoskeptic.org/Vallee_LST.pdf
Here is his counterargument synthesized in the abstract:
"While a frequency distribution indicative of a correlation with
star positions was indeed detected, control with a separate
catalogue compiled in France discloses an important artefact:
multiple entries for a single, particularly remarkable UFO event
have resulted in massively duplicated records. This calls into
question the significance of the claimed pattern."
The first thing Vallee does is point to a Johnson and Saunders
(2002) study which used a different UFO database to determine
CE2 type events against sidereal time. What's interesting here
is that the Johnson-Saunders study determined a DIFFERENT
sidereal time (14:00) from the Sturrock analysis. See Fig 1.
Vallee then used the same Hatch "U" database as Sturrock, though
a slightly newer revision. He also tossed every UFO event before
1947, arguing that they should be considered an entirely
different class of events. However, he confirms the same result
as Sturrock (See Fig 2)
"This graph matches well with Sturrock’s published distribution.
In particular it exhibits the same increase in the frequency of
reported events in the range of 20 hours to 23 hours, with a
noticeable peak between 21.1 and 21.5 hours of local sidereal
time." (Pg. 4)
But then, upon further analysis, Vallee found that one
particular event (Nov 5th, 1990, in France) had many multiple
witness sightings. Vallee realized that Sturrock had counted
each sighting of one event as multiple events in his analysis.
"Among the UFO events in the Hatch catalog for which a value of
LST can be computed there are 56 entries for this single case of
November 5, 1990. Such a large block of entries, occurring in
such a small interval of LST values, introduces a potential
distortion in the statistical correlation. ... When this
duplication is removed the frequency distribution becomes that
of figure 4, where the rise in frequency around 21.5 hours is
less noticeable, although still present." (Pg. 5)
OK. So, this means that Vallee is challenging the significance
of Sturrock's finding by saying that there were duplicates in
the dataset that were erroneously counted. And when Vallee
performed a similar analysis with the duplicates removed, he
found the same correlation, but with a much small peak of
significance.
Vallee then went on to perform a similar analysis using a
smaller French SERPA database. He discovered a huge peak at the
same time as Sturrock (21.5hrs; or 8pm to 11pm). BUT, he then
factored out IFOs from the database - Identified objects - and
found absolutely no correlation to the Sturrock result of
sidereal time at 21.5hrs (or 8pm - 11pm) whatsoever This
presumably refuted his prior work with Poher in the 1970s. But
he did find a smaller statistical correlation to sidereal time
in the early morning hours of between 3am to 5am.
"It can be seen that the peak at about 21.5 hours disappears
when identified cases are removed from the Sepra catalog. In
contrast, the secondary peak between 3.2 and 5.2 hours of local
sidereal time on figure 6 actually stands out more sharply on
figure 7, with a maximum at 3.8 hours. This particular feature
in the distribution survives not only the elimination of IFOs
from the list but also a review of the individual cases for
possible duplication effects or other biases. At this point we
have to consider it as unexplained."
So. What does all this mean? First of all, there's the question
of whether the Hatch database separates out and classifies UFO
events based on relevant 'strangeness' factors. Valle confirmed
Sturrock's work with the Hatch database and yet challenged it
with the SERPA catalog. Then Vallee seems to argue that the
French SERPA catalog has better identification between
identified objects and the truly strange, and that this is the
result of the discrepancy between Hatch and SERPA. And,
ultimately, depending on database, you get different results
using the same - or similar - statistical tools.
I think the ultimate result here is to challenge the utility of
UFO databases for statistical work. Is that true? That is the
question to the List.
(and, of course, I welcome responses which correct any errors I
may have made digging through these papers. Attempting to
understand the Sturrock paper, in particular, was quite
difficult.)
-M
Listen to 'Strange Days... Indeed' - The PodCast
At:
http://www.virtuallystrange.net/ufo/sdi/program/
These contents above are copyright of the author and
UFO UpDates - Toronto. They may not be reproduced
without the express permission of both parties and
are intended for educational use only.