[IGSMAIL-6404] short note on status of GEONET soon after the M9.0 Earthquake, Japan

[Yuki Hatanaka]

Dear colleagues,

Recently I get an inquiry to share information of our
experience of troubles occurred on GEONET (the dense
GPS observation network and system in Japan) during and
after response to the 2011 off the Pacific Coast of Tohoku,
Japan Earthquake (March 11, 2011, M9.0).
I would like to copy the short note below since there
may be IGS members who are interested in the information:

- 3 sites were heavily damaged by the earthquake and tsunami.

- The electricity was down in Tohoku area soon after
  the event. The power was supplied by backup batteries
  for several hours to ~72 hour depending on sites.
  Within 72 hours, the observation was stopped at about
  40 sites because the batteries were out and the
  electricity was not restored. Number of stopped sites
  had decreased as the recovery of electricity progressed.
  GSI has been setting solar power for 5 sites.

- There are two communication means to GEONET sites,
  IP-VPN and cell phone.
  The realtime data transfer by IP-VPN connection had
  been working for a while after the first event.
  IP-VPN connection to 358 sites in northern area of Japan
  was cut about 30 minutes after the first event
   (probably due to the Tsunami). The IP-VPN connection to
  about 100 sites was recovered within a few hours.
  Cell phone line are used to retrieve the data (not in
  realtime) from the remaining sites.
  The cell phone line was also damaged for two sites.
  (One is Oshika site where the largest displacement
  is observed.) The data of those sites are collected
  when GSI's staffs visited them a week after the event.

- The routine analysis system worked without trouble.
  However, the movement of the fiducial sites (TSUKUBA)
  is so large (~60cm) that not only the translation
  errors but also scale and rotation errors of
  a few tens of ppb were expected for the quick solutions
  (Q3; every 3 hours with 6 hours data window) and the
  rapid solutions (R3; every 24hours) for which the nominal
  coordinates were applied. Initially, we used these
  results without correction to get the rough estimate
  of fault parameters. The routine solutions and fault
  parameters were re-analyzed a few days later by taking
  the movement of the fiducial site into account.
  (The final solutions (F3) are almost free from such
  errors since the coordinates of fiducial site given
  by analysis with IGS stations in surrounding Asian region.)

----------------------------------------------------
Yuki Hatanaka, Ph.D,
Geospatial Information Authority of Japan (GSI)
e-mail : hata at gsi.go.jp
----------------------------------------------------