Estimation of instrumental noise through a corss-correlation of outputs of two instruments recording side-by-side Holcomb, G. L. (1989). A Direct Method for Calculating Instrument Noise Levels in Side-by-Side Seismometer Evaluations, USGS Open-File Report 89-214 Data processing steps 1. cd FPSD This directory has the code required for the coherency analysis and also codes for PSD noise estimateion This codes here are an implementation of the USGS Open File Report on noise measurements There are several subdirectories src - source code and Makefile bin - location of executables TESTPSD - location of a test data set a) cd src edit the Makefile to give the name of your compilers. The Makefile is set up for gfortran and gcc I also include the plot program xyplt2 which requires the graphics libraries of Computer Programs in FORTRAN You do not require this since your can plot the output using GMT The psd and coher programs are pure FORTRAN programs make all This will create two programs: psd for noise analysis and coher for the comparison of output of two sensors b) cd TESTPSD This directory contains the codes for performing a noise analysis A data set is provided for the GSN station ALQ The files ending with .cmp are the data files that I was given The files ending with .asac are Sac files in ASCII format. You can use the Computer Programs in Seismology program asctosac to covert from this format to the SAC binary format for your computer, e.g., asctosac anmbz_08.asac anmbz_08.sac The scripts DOITALQBIN and DOITALQASC are essentially the same except that the first uses the binary sac file and the other uses the corresponding ascii file. The program psd outputs the noise spectrum in acceleration The program is controlled from the command line: psd [-h] [-D -V -A] -R response_file sac_files where response_file is an ASCII file with the first column indicating frequency and the second indicating the response. The second column is in units of counts/meter for the -D flag, counts/m/sec for the -V flag and coutns/m/s/s for the -A flag The output of this program is redicted to a file, here called junk.alq which consists of two columns. The first column is period and the second is thepower spectral density of the noise in decibels or mathematically PSD 10*log10(m^2/s^4 .Hz) The program xyplt2 creates a CALPLOT graphics file plotXXXXX, where XXXXX is a unique number. The PSD from the program is plotted and is compared to the low and high noise models, nlnm.acc and nhnm.acc respectively The CALPLOT graphcis file is renamed PLOTALQ and converted to an Encapsulated PostScript file (EPS) using the Computer Program in Seismology routine plotnps 2. cd TEST.ANMO This directory has the scripts for performing the coherency analysis for the two LHZ sensors at ANMO * CHANNEL(NSCL)IUANMO LHZ00 * NETWORK IU * STATION ANMO * COMPONENT LHZ * LOCATION 00 * RATE (HZ) 1.0 * INSTRMNTTYPE Geotech KS-54000 Borehole Seismometer * CHANNEL(NSCL)IUANMO LHZ10 * NETWORK IU * STATION ANMO * COMPONENT LHZ * LOCATION 10 * RATE (HZ) 1.0 * INSTRMNTTYPE Guralp CMG3-T Seismometer (borehole) This example starts from a SEED volume obtained from IRIS The shell script DOALL a) unpacks the SEED volume to get the waveforms in a SAC format and the response in a RESPONSE file b) uses the resposne file to create an amplitude and phase table for the acceleration sensitivity YOU WILL NEED rdseed, evalresp, gsac I got the pdf_E2010.220_S2010.220_cBHZ_l10_nIU_sANMO.png indepdendent estimate form IRIS for this day so that I can check my code, The KS borehole is much quieter than the Guralp 3. cd KMI.TEST This uses KMI data from January 2010 and compares the STS-1 to STS-2 sensors. The STS-2 high gain is noisier but more sensitive