Thread: increasing the FDSN Backbone

Dear FDSN WG1 members,

Here are some suggestions for discussion on increasing the size of the
FDSN Backbone.

Best regards,

Rhett

===

First, a very brief history. The FDSN Backbone was "originally" the
FDSN network. At the Potsdam meeting in 2004, FDSN members expressed
their interest that their networks actually become part of the FDSN
network. To include this concept and still to maintain a globally-
distributed, high-quality, openly-exchange subset of the FDSN network,
the former FDSN network was re-named the FDSN Backbone Network. This
FDSN Backbone has been largely unchanged in about 5 years, at about
200+ stations.

Therefore, there are actually two FDSN networks: The Network of all
FDSN stations, and a globally-distributed subset, the FDSN Backbone.

In principle, the "all-FDSN" network is contained in the FDSN
spreadsheet, and the Backbone is composed of those stations designated
by asterisks* in this list. The "all-FDSN" list is now increasingly
difficult to maintain, given the growth of broadband networks.

Before focusing on the Backbone, let me suggest that we need to define
and maintain the "all-FDSN" network with a working subset of those
stations which FDSN members want to exchange openly. Let's consider
this "virtual network" as _FDSN-all. This list would include for each
participating FDSN network, the network code, station code, data
center(s) where data are accessible, and real-time and/or archival
status. This _FDSN-all is implicitly a distributed network, and FDSN
data centers would need to work out bilateral ways to link to and
openly distribute the data. By necessity, for the data to be
exchangeable, it would need to conform to basic standards (SEED being
our key FDSN standard, but some FDSN data centers may possess the
means for more diverse translation of formats, e.g., CD 1.x, antelope,
etc).

By participating in this _FDSN-all, a member FDSN network could then
distribute its data more easily to the global community through the
existing infrastructure of FDSN data centers. The participation would
be voluntary. Some FDSN members might choose to distribute their data
only through its own data center request mechanisms. Others might want
to share some stations openly, but retain closer distribution of other
stations in their networks. Some FDSN members may wish all of their
data to be openly distributed. Note, in this latter case that there
are now many FDSN members effectively
doing this through mechanisms available at FDSN data centers.


At this time, the FDSN Backbone is about 200 stations. To effectively
halve the inter-station spacing, about 4 times as many stations are
needed. Therefore, the next order (octave) of global coverage is about
800-1000 stations. Even a 1000 stations does not pose a difficulty
archiving with current Internet and data storage capabilities. One
hard part is maintaining the meta-data. Our current Backbone is based
upon good meta-data, and this should continue to be a standard. If
real-time data are available, this sharing should be encouraged (an
FDSN goal is "Improving Access to data in real time"), but not be
required. The current Backbone emphasizes very-broadband stations (the
STS-1 was the original favorite), so that three channels of very-
broadband are favored. However, there are now many instances of
quality, very broadband regional sensors with good S/N performance at
longer periods. Future (near-term) planned stations should be
considered for the Backbone if there is a funding commitment (this
follows earlier FDSN precedent).

The Backbone should emphasize high-quality, low noise stations. This
is harder to define specifically. Many island stations are very-noisy,
but provide essential coverage.

Redundancy (in moderation) should be viewed as okay. Redundancy
provides for some improvement on regional data availability. E.g., if
two nearby stations each have about 80% uptime, their combined uptime
may be better than 95% (assuming independence of cause for downtime).

Currently, every FDSN member is encouraged to nominate at least 1
station to the Backbone. Many FDSN networks do not now have a Backbone
station. We should review this with FDSN members. Note also that many
FDSN networks not represented yet in the Backbone already share good
data (and meta-data) through one of the FDSN data centers. Perhaps,
those FDSN data centers could nominate a known quality station on
behalf of the FDSN member network donating the data (subject to their
approval).

I suggest extending this concept internationally (after all, the FDSN
is the International Federation ...). Part of the densification of the
Backbone could include FDSN stations in countries which are not are
not in FDSN. The GSN and GEOSCOPE (and others) have many such examples
(e.g., Solomon Islands and Senegal). There are an increasing number of
FDSN members installing stations internationally. As FDSN has a role
in GEO/GEOSS, extending international coverage of the Backbone (which
is designated as a GEOSS observing component) broadens the
participation of countries in global/regional seismology. Therefore,
designating Backbone stations could include an element of extending
the numbers of nations with a Backbone station(s).

Another approach for Backbone densification is to improve coverage in
seismogenic regions. This seems obvious to a seismologist. Many
seismogenic zones span beyond local/regional/national network
boundaries. Fostering international data exchange through the FDSN
Backbone in seismogenic zones is commensurate with FDSN's goal of
"Pursuing free and open access to data" and may help transcend other
obstacles to regional data sharing. Here, real-time data would be best
for earthquake response.

To summarize extending the Backbone:

. (very) Broadband
. Global Coverage
. Increase from current ~200 stations to (perhaps) 800-1000 stations
through decreased inter-station spacing
. Redundancy is good in moderation
. High-quality data and meta-data
. Real-time encouraged but not necessary
. Every FDSN member should designate at least 1 station from their
network
. Broaden international coverage by FDSN stations
. Enhance coverage in seismogenic zones (emphasizing real-time)