Eureka Meteor Radar Winds: monthly quick look plots
The Eureka (80N,86W) meteor radar was originally installed as part
of the Canadian Network for the Detection of Atmospheric Change ( CANDAC )
originally led by Jim Drummond, now at Dalhousie University,
and currently (2018)
by Kim Strong (U. of Toronto). By various nefarious
means it was kept in operation beyond the end of the funded 5 years,
until a less-well-funded collaboration, CCAR/PAHA (Climate Change
and Atmospheric Research / Probing of the Atmosphere of the High Arctic,
began.
Due to serious interference from a nearby Star Photometer installed
in the fall of 2010, which despite much effort could not be reduced,
there was a major loss of winter data while it ran. Since the first
incident of extra noise (not Star Photometer), a sparse archive of
interim meteor signal files has been maintained which is processed to
yield noise levels during each day at each receiving antenna. The bad
intervals have been cut out before plotting the data presented here. [Read on
to see why I have relented and replaced them with uncut figures.]
The radar is a semi-commercial SKiYMET (Wayne Hocking, University of
Western Ontario, and Genesis Software Pty. Ltd., Australia )
Unfortunately the Tx failed in fall 2013 - and is taking a long time
for Genesis to fix. Still not in operation April 2015! But it's now back
running in another location starting September 2015 (with some ambient
temperature control problems!), However the noise level in the new
location for the antenna nearest the "Cottage" was about as bad (~ 5dB)
as in the old location!
Wayne moved that antenna to a quieter location in July 2016. If we use
those current data, how can we refuse winters 2010-2012 noisy data?
Year to year and month to month natural variations make it very difficult
to measure the effect of extra noise. Maybe the weaker trails are
rejected instead being accepted (and contributing some sort of bias).
An off-line copy of the analysis could be used to estimate this effect
(or on-line, as used in the fix for a twisted antenns, see below) but a
realistic noise generator is necessary.
[ On Oct. 23,2016 one of the receiving antennas became twisted by ~40 deg.,
discovered by Wayne Hocking, who noticed the problem in monitor plots
just after it occurred
and joined a trip to Eureka in July 2017 to investigate (and also to
perform maintenance on his VHF radar).
The cause remains unknown. This interval covers 9 months of data.
Since raw data are not kept - though I do download a sparse
selection to keep track of external noise - it is not possible to
simply untwist the resulting meteor parameters. Several correction
methods are described:
Method 1,2
Final method .
Due to (my) lack of control over that archive, the "corrected" data cannot
be put into the regular
DNT data stream, but they have been posted to the Toronto data archive: names
MVYYYYYMMDD.eureka.mpd (instead of mpYYYYMMDD.eureka.mpd). All MV headers
were removed except " data " . (The uncorrected mp* data are there as well.) ]
A similar problem occured in September 2020- an antenna cable was gnawed and, because
of travel restrictions due to COVID-19, cannot be fixed yet. It was not a complete break,
and because of snow cover may be safe until spring. But there is a 13dB signal loss and
a phase shift. The phase shift was estimated and corrected to some extent in September
and Octber data. In November the phase correction was put into the on-site analysis.
The monthly plots below are marked. A description of the estimate amd correction methods
is stored in the "Technical Reports" page of this site: CableDamage.pdf (includes Fortran programs),
and CableDamage-short.pdf. (See main page for link.)
Just for pretty - here is an animation of real meteor postions, and their
horizontal components of speed, in a 90 minute fit interval
slid by 10 min, and the resulting wind fits
meteor animation
(takes a few seconds to load). The general clockwise rotation is caused by
atmospheric tides (due to solar heating, not gravity).
The chosen January interval covers the yearly, very active but
short duration, Quadrantids meteor shower, Jan 3/4.
And finally, usually the major components of the daily wind are background
mean, diurnal (24 hr), and semi-diurnal (12hr) tidal oscillations ,
with seasonally constant hours of maximum. That is, for intervals of days
we expect to see a virtually constant wind at each hour and height.
Well - reality is different. The tidal phases can get pushed around, and
there are multi-day large scale waves, particularly in winter. To
illustrate, we show an animation
of a day. Data have been stretched by interpolating the hourly
means between days to simulate a 0.2 day step, resulting in a view Franz
Mesmer might have appreciated.
UPDATE: The cable to one Rx antenna was chewed and caused some phase problems
which were corrected to some extent- but finally were too variable to control by
adjusting the phase correction parameter.
Just after that the Tx had VSRW problems which the radar deemed fatal. So the
last data are in Jun 2022. If CANDAC survives an almost total funding shortfall
this may get fixed - but don't hold your breath.
Update (Nov'24): Surprisingly We are in business again (for now) with 1/2 Tx power.
all Rx feeders seem to be ok. Thanks to Wayne Hocking for dealing with the hardware
problems at Eureka.
