After a preliminary non-coherent system (a copy of the early one at
Saskatoon) which ran during and after the MAC/SINE campaign
(June 1987) on a good Tx antenna (4x4 dipoles O/X mode), and
transmitter, but very marginal receiver antennas (shielded loops),
in collaboration with Truls Lynne Hansen (Tromsø
Geophysics Laboratory, TGO, and University of Tromsø, UiT),
and Trond Trondsen was run into the 1990s (but none of these data are
shown here - a reference is JATP 52, No. 10/11,971-980,1990),
it was decided to upgrade to a coherent system in 1996 using some
money for new receivers (2.78MHz) and a receiver interface from
the Canadian Network for Space Research (CNSR) budget. Chris Hall
became the main contact and collaborator, and replaced the loops
with "inverted-V" dipoles, later changed to suspended dipoles.
He was assisted by Tor Holmboe. Later Bjørnar Hansen and Bjørn Ove
Husøy have joined Chris Hall in the care of the system.
About 2000 Satonori Nozawa (STELAB, Nagoya University) provided a
25KW transmitter and joined the collaboration. Another 25KW "slave"
transmitter was added a few years later. For part of 1999-2000
the system was run with 2 min records instead of the usual 5
(as illustrated by the increased number of values in the corresponding
In early 2013, extra noise appeared in one antenna - this is thought
to be generated by a poor/corroded metal-metal joint near the antenna,
but it could not be found. Because all antennas seem to have picked up
some of this noise - and thus it is coherent, with a phase change
depending on distance - the system finds "pattern" delays between
antennas very close to zero. For obscure not-well-understood reasons,
this translates into very low speeds in the full correlation
analysis (FCA). This effect usually occurs where reflected signals are
very small, e.g. 50-80 Km depending on season.
To reject these data an arbitrary post processing criterion, S/N > 3dB
is usually placed on the results before plotting.
This criterion has been applied to data years 2012-2015 in the
following plots. The rejection is not perfect, there still could
be some loss of good data, or added bad (very low speeds) at the