EXECUTIVE SUMMARY:
- Similar to a prev post we chop the US48 up into latitude bands and
see whether storm activity seems to "interrupt" the travel of UFO's
from E to E or vv.
- We find relevant patterns. We also find windstorms don't seem to be
as important as hailstorms when looking for an effect on UFO
activity. Tornadoes also have an almost negligible impact on
E-W/W-E UFO travel.
- A possible "flight corridor" is revealed around long 85W. To the
west of that UFO's seem to "mostly" travel west. To the east of the
corridor they seem to mostly travel east.
- Combined with prev data posted is seems UFO's oftentimes come from
the N, get into the 85W corridor -- around the Ohio Valley and
Mississippi -- then split up east and west to do (whatever).


In a previous post we saw data on tornadoes used as a kind of "passive
radar" to track possible flight routes of UFO's.

The idea was to locate a pattern where tornadoes in a band of latitude
across the US seemed to have a positive corr with UFO sightings in a
band on one side of the tornado band, and a negative corr in a band on
the other side.

We could interpret this as evidence that more tornadoes in the central
band seemed to cause UFO's to "pile up" on one side of the band
because they were inhibited from travelling through the storms, and (consistently) "starved" on the other side.

In this way we could pinpoint in which direction UFO's might be flying
at various latitudes across the US.

It seemed from the data we looked at most UFO's seemed to be flying
from N to S because more tornadoes in various lat bands caused
sightings to increase to the N and decrease to the S of the tornado band.

We will now look at locating similar patterns using bands of
longitude. And we will also check the other types of storms the NOAA
has been keeping track of since 1950 -- i.e. hailstorms and
windstorms.

As before we chop up the US48 into 5-deg wide bands (~550 km), this
time running N/S. For each month since 1950 we compute the number of
tornadoes recorded in each band as well as the number of UFO sightings
recorded in the same longitude bands over the same period.

Then it's a matter of running a little program that will compute
time-series regressions using bands A, B, C where B is a band of storm
data and A and C are bands of UFO data -- B against A and B against C.
We look for the situation where both these TS regr are stat
significant and the signs of the betas are opposite.

The following table represents the summary results. The fine detail is
far too voluminous for mortals:

longitude band
stormtype A B C beta_ab beta_bc r2_ab r2_bc
hail -100 -95 -70 -5.65239 1.1156 0.58657744 0.30086143
hail -100 -95 -70 -5.65239 1.1156 0.58657744 0.30086143
hail -100 -95 -70 -5.65239 1.1156 0.58657744 0.30086143
hail -100 -95 -80 -5.65239 9.48006 0.58657744 0.28009550
hail -100 -95 -80 -5.65239 9.48006 0.58657744 0.28009550
hail -100 -95 -90 -5.65239 5.45152 0.58657744 0.24554827
hail -100 -95 -90 -5.65239 5.45152 0.58657744 0.24554827
hail -100 -85 -70 -4.11688 1.06516 0.48077574 0.33030840
hail -100 -85 -70 -4.11688 1.06516 0.48077574 0.33030840
hail -95 -80 -75 1.45559 -3.17667 0.47485289 0.44661260
hail -95 -80 -75 1.45559 -3.17667 0.47485289 0.44661260
hail -100 -90 -70 -5.72678 0.565266 0.46880559 0.06452935
hail -100 -90 -80 -5.72678 8.13944 0.46880559 0.22691812
hail -120 -80 -75 12.0204 -3.17667 0.32644105 0.44661260
torn -120 -80 -70 5.83422 -0.392899 0.26302291 0.12484681
torn -110 -80 -70 2.14279 -0.392899 0.25515163 0.12484681

As before, the "B" band is the band of storm data being used. In this
case with multiple types of storms in play all the storm data has been normalised (Z scores) so we can compare the effects of UFO activity on
each. The A band is the UFO activity (NUFORC sightings) to the west of
band B and the C band is the UFO activity to the east. The beta's
show the effect of B on A and C and the r2's show the "explanation
power" of the relevant TS regr.

I've sorted the lines by the first R2 field so we can see the "most
likely" relationships at the top and the "less likely" at the
bottom. But, remember, all these regression passed a test that said it
could occur by chance alone only 10% of the time.

So the first thing to note is the B longitude bands. They show storms
down the bands 95W-90W, 85W-80W, 80W-75W, and 90W-85W seem to show the
pattern we're looking for. When storms increase in that band UFO
sightings increase on one side of the band and decrease on the other.

For the lines that show a -ve beta_ab we have some evidence that UFO's
may be normally travelling east to west at that longitude and storms
curtail some of that travel -- causing sightings to pile up to the
east and drop off to the west.

Lines of the table that show a +ve beta_ab are the opposite. Normal
UFO travel seems to be from west to east and storms in the B band cut
that travel down.

Another thing to note -- even tho we calculated these statistics for 3
types of storm only hail and tornado turned up. It seems UFO's are not particularly fussed by windstorms. At least compared with the other types.

It seems hailstorms affect UFO travel patterns the most. Only 2 of
the lines in the table show tornadoes at some longitude affect UFO
sightings more than hail.

And, finally, we see there seems to be a "corridor" around longitude
85W. The evidence seems to point at UFO's "mostly" travelling west to
the west of that corridor, and mostly travelling to the east to the
east of that longitude.

Combined with other evidence it seems UFO's are oftentimes coming down
from the N, moving over to long 85W, and then dividing up east and
west to do whatever it is they are up to over the US48.

Of course I didn't check any of this with a Youtube psychic so take it
all with a grain of salt and do your own study.

--- SoupGate-Win32 v1.05
* Origin: www.darkrealms.ca (1:229/2)