Portable Broadband Seismic Stations - Checking Data Quality Cairo, Egypt November, 2009 Noel Barstow Bruce Beaudoin Marcos Alvarez.

Portable Broadband Seismic Stations -

Checking Data Quality

Cairo, EgyptNovember, 2009

Noel BarstowBruce BeaudoinMarcos Alvarez

Checking Data QualityKeep good records of everything

> Before going to the fieldTest all Instrumentation and Cables

> In the fieldInstall Station and Check: Power, Timing, Sensor Orientation, Signal

Visit stand-alone stations frequently to: Download Data and Review Station Status and Metadata on site

> In the labArchive, Analyze, and Document Data

Hardware Maintenance and QC: Dataloggers

Function test all Dataloggers prior to shipment:Bit weight, oscillator drift, data time stamps

Keep maintenance records in an electronic databaseTest and upgrade firmware when necessary

PASSCAL’s Hardware Maintenance: Seismometers

Feedback seismometers are Pier-Tested for 4-7 Days and compared to a Reference Sensor:- Amplitude and Phase- Polarity- Spectral Content- Noise Characteristics

Electronic Calibrations:- Corner Period- Damping

Do the signals look the same?

close up of verticals

overlay surface wave

Nightly Plots of Pier Tests

Electronic Step Calibration

White Noise Calibrations

Input White Noise

Pha

se

Am

plitu

deSensor Output

Sensitivity

Pier Data Comparing Sensor Noise of Different Sensor Types

Ve

rtic

al C

om

pon

en

ts

Streckeisen STS-2

€

X = − 2 /3( ) U + 1/6( ) V + 1/6( ) W

Y = 1/2( ) V − 1/2( ) W

Z = 1/3( ) U + 1/3( ) V + 1/3( ) W

STS-2 Element Contribution to Signal

East

North

Vertical

STS-2 Element V problem

Z

N

E Noisy Sensor

Z

N

E

Triaxial SystemsR

efer

ence

Unfiltered Earthquake Filtered 10 -30 sec period

Z

N

E

Z

N

E

Tes

t S

enso

r:

Drif

t &

DC

off

set

Orthogonal elements2 Horizontal1 Vertical

Guralp CMG-3T

Replace Bent 3T Pivot

Comparing Reference and Problem Sensor

Before

After

Time Scale = 3 hours

Eas

t /

Wes

t C

ompo

nent

Sudden Noise on Guralp 3T Vertical Component

Noise begins

Noisy Spectra

Days:

1

2

3

4

Ver

tical

Dat

a

Spe

ctra

Locking and Unlocking the 3T provides a temporary fix

Amplitude change at DC offset

Time = 8 Days

N/S

E/W

Z

Amplitude Problem: Loss of Differential Output

Das, Sensor, or interconnecting Cables could be the cause

•PASSCAL develops and maintains software to extract data from the dataloggers, perform in-field quality checks, and manipulate the waveforms into archive formats.

•http://www.passcal.nmt.edu•ftp.passcal.nmt.edu

Data QC & Software

Data Archiving Support PASSCAL acts as an intermediary between

PASSCAL Investigators and the DMC

• Check completeness of data submitted to PASSCAL

• Help evaluate or fix timing

• Verify SEED Format

• Ensure agreement between data and metadata

• Check completeness of data submitted to DMC

• PASSCAL provides email and phone support for our users before, during, and after experiments to analyze specific problems.

• PASSCAL staff members are available for the initial testing and installation phase of field campaigns.

• Typically there are over 500 broadband and short-period PASSCAL stations deployed at any given time.

• Contact us: [email protected]

Troubleshooting

After the power point presentation

Live demonstration oflogpeek - plotting of RT130 station log

messagespql - PASSCAL Quick Look - to view

waveforms, fast fourier spectrum, split view, etc.

pql will read mini-seed, sac, and segy formats

Review Station’s Status

Firmware

Volts

GPS

Timing

Time Scale = 1 WEEK

‘logpeek’Plot of StationLog MessagesFromRT130 DAS