Recent Changes - Search:

EPOS-N - European Plate Observing System

Projects 2019

Seismometers / vault construction

Digitizers

Interconnections

Stations: Upgrade/install notes

Calibrations

SeismoLAB

Ideas for improvements

Installation / Inventory

Misc

edit SideBar

Title

    EVALUATION OF SEISMOMETER SITE
    CONSTRUCTION METHODS

By
O.M.
Date
07 April 2014
Version
0.1

1 AMBIENT NOISE - IMPORTANT DOCUMENTS


2 VAULT CONSTRUCTION METHODS


  CONSTRUCTION METHOD A : SKAR, FAUSKE  
Click to enlarge. Click to enlarge. Click to see construction drawing.
  CONSTRUCTION METHOD B : MOR8, KMY 
Click to enlarge. Click to enlarge. Click to enlarge.
  CONSTRUCTION METHOD C : KBS 
  CONSTRUCTION METHOD D : SSNN. Source all pictures: http://snsn.geofys.uu.se
We consider using this method in the HOPEN 2014 upgrade project.
 
Source: http://snsn.geofys.uu.se Source: http://snsn.geofys.uu.se

LINKS:

 http://csegrecorder.com/articles/view/alberta-telemetered-seismograph-network-atsn-real-time-monitoring


3 EVALUATION OF SEISMOMETER SITE CONSTRUCTION METHODS


3.1 Quality criteria


In order to evaluate the construction method of our recently built SKAR site, we wish to compare power density spectra (PDS) of stations furnished with similar broadband sensors. SKAR has a Nanometrics Trillium 120PA sensor, with 3 dB bandwidth from 50 Hz to 120 seconds (= 8.3 mHz). A power density spectrum will display most factors that characterize a good quality site:

  1. Low susceptibility to wind noise.
  2. Low presence of anthropogenic (i.e. cultural, man-made) noise sources.
  3. Low presence of diurnal (daily) or seasonal variations
  4. Absence of electric noise sources, in particular 50 Hz and sum/difference artifacts of this frequency.
  5. Good performance in the very low frequency band of the spectrum, lower then 0.1 Hz (10 seconds). Noise in this frequency range is primarily caused by two factors:
    1. The two horizontal channels (North/South and East/West) are particularly influenced by even the most minute tilting and/or twisting movements of the seismometer foundation, due to either atmospheric pressure changes affecting the landscape in this way, or tidal effects from nearby ocean waters or the earth itself.
      1. http://www.knmi.nl/~evers/infrasound/web-article/web-article.html
      2. http://www.bfo.geophys.uni-stuttgart.de/
    2. Poor thermal insulation of the sensor and its foundation.
  6. For stations located in high latitudes - above 60 deg (which applies to most of the sites within NNSN) - we should consider if and how seismometer signals are affected by disturbances in the geomagnetic field, as described in this article: Influence of high-latitude geomagnetic pulsations on recordings of broadband force-balanced seismic sensors, by E. Kozlovskaya and A. Kozlovsky, Sodankylä Geophysical Observatory of the University of Oulu, Finland (the observatory that operates OUL, MSF, SGF and RNF stations; of which PSD plots from OUL is included here).

3.2 Obtaining information from PSD plots - a catalog of errors


No. 1: Cultural noise creates a diurnal noise pattern. Traffic and building noise. Only remedy is to relocate the sensor.
No. 2: Electric noise through inductive and capacitive coupling - poor or missing grounding and/or shielding - remnants of sum/difference components of 50 Hz mains frequency and higher harmonics in particular. Alias frequencies due to non-ideal low-pass filter in front of analog-to-digital converter can perhaps also be seen. If digitizer is located directly adjacent to the sensor many of these problems can be avoided (and this is part of current designs).
No. 3: This site is suceptible to wind noise which covers a fairly large part of the high frequency part of the spectrum.
No. 4: Low frequency noise.

3.3 Viewing signals in Frequency or Time domain


Power density spectra show signals in the frequency domain. It it of course necessary to also study time domain data - ordinary seismograms - which clearly shows noise spikes and sudden, unnatural steps in the signal level. Such disturbances have a high frequency content and would therefore, in a spectrum averaging process, be distributed (or "smeared") over a broad frequency range, not easily detectable. So, examining time domain data is absolutely necessary. To illustrate this point, two time-domain examples are shown below, one normally impeccable (from OULU, Finland) and one with errors (from Hopen Island, SE of Svalbard):

Seismogram OULO, Finland, short period filtered. Click to enlarge.
Seismogram OULO, Finland, short period filtered. Source: GFZ Potsdam. Click to enlarge.
Seismogram HOPEN. Click to enlarge.
HOPEN 10-hour data sample, with steps (erroneously) in signal levels. Source: Norsar. Click to enlarge.



3.4 "Best practices"


"Best practices" - sources of inspiration.
USGS: Methods of Installing United States National Seismographic Network (USNSN) Stations— A Construction Manual. Click to open PDF document.
"Methods of Installing United States National Seismographic Network (USNSN) Stations— A Construction Manual" click to open PDF document.

IRIS: Portable Broadband Seismology - Part 7: Seismic Vaults for Temporary Installations. Click to open PowerPoint presentation. Note: 77 MByte.
Eskdalemuir facility.
An idea from Eskdalemuir facility is also interesting. Click to open description.


4 POWER SPECTRAL DENSITY COMPARISON OF BROAD BAND STATIONS


  • Data source: Norsar. Credits Dr. M. Roth at Norsar for these very useful plots.
  • A similar table for December 2013, but here some plots start on 10 December 2013. In that case there are vertical lines in the left side of the plots. These lines are artifacts from the spectrum computation, and are not present in subsequent plots.
  • A similar table for January 2014.
Station code/
sensor model/
location
Power spectrum density (PSD) plots of some BB stations.
Below each graph the station code is indicated, followed by three-letter channel codes,
which are described in Appendix A in the SEED Manual (p. 133++). The last letter provides orientation information:
E = East/West, N = North/South, Z = Vertical.
  SKAR  
Trillium 120PA
Ål, Norway
Click to enlarge
  HOMB  
Trillium 120P
Homborsund, Norway
Click to enlarge
  BLS5  
Trillium 120P
Blåsjø, Norway
Click to enlarge
  STAV  
Trillium 120P
Stavanger, Norway
Click to enlarge
  DOMB  
Trillium 120P
Dombås, Norway
Click to enlarge
  FOO  
Trillium 120P
Florø, Norway
Click to enlarge
  MOR8  
Trillium 120PA
Mo i Rana, Norway
Click to enlarge
  STEI  
Trillium 120P
Steigen, Norway
Click to enlarge
  LOF  
Trillium 120P
Lofoten, Norway
Click to enlarge
  TRO  
Güralp CMG3T
Tromsø, Norway
Click to enlarge
  HAMF  
Trillium 120P
Hammerfest, Norway
Click to enlarge
  BJO1  
Trillium 120PA
Bear Island, Norway
Click to enlarge
  HOPEN  
STS-2
Hopen Island (SE of Svalbard)
Click to enlarge
  BER  
STS-2
Bergen, Norway
Click to enlarge
  KONO  
STS-1
Kongsberg, Norway
Click to see site description from USGS.
Vertical scale to 20 Hz.
Click to enlarge
  KBS  
STS-1
Kings Bay, Ny-Ålesund, Svalbard
Click to enlarge.
Vertical scale to 20 Hz.
Click to enlarge
  KEV  
STS-2
Kevo, Finland1
KEV, Finland
Click to enlarge
  KIF  
STS-2
Kilpisjarvi, Finland
Click to enlarge
  OUL  
STS-2
Oulu, Finland
Vertical axis to 12.5 Hz.
Click to enlarge

Notes:

1 Discrepancy in meta data sources: USGS states that station codes HHZ, HH1 and HH2 are tied to sensor STS-2 sampling at 100 Hz, whereas station owner Univ of Helsinki, Dept of Seismology states that sensor is Guralp CMG-3T ....

4.1 Station locations


Clisk to enlarge
Norwegian National Seismic Network locations (SKAR not included). Click to enlarge.

Seismometer locations in Finland.
Source: http://www.seismo.helsinki.fi/english/observation/stations.html

5 BER PSD


Station code/
sensor model/
location
Power spectrum density (PSD) plots of BER station.
Below each graph the station code is indicated, followed by three-letter channel codes,
which are described in Appendix A in the SEED Manual (p. 133++). The last letter provides orientation information:
E = East/West, N = North/South, Z = Vertical.
  BER, JAN 2014  
STS-2
Bergen, Norway
Click to enlarge
  BER, FEB 2014  
STS-2
Bergen, Norway
Click to enlarge
  BER, MAR 2014  
STS-2
Bergen, Norway
Click to enlarge
  BER, APR 2014  
STS-2
Bergen, Norway
Click to enlarge
  BER, MAY 2014  
STS-2
Bergen, Norway
Click to enlarge

6 LINKS


Edit - History - Print - Search
Page last modified on June 07, 2017, at 12:27 PM
Electronics workshop
Department of Earth Science - University of Bergen
N O R W A Y