Home

User's Guide to HYPOINVERSE-2000, a Fortran Program to Solve

Contents

1. D E ATE EL longer station codes FMC OR XMC COMMANDS Read binary file of all stations tation file format Read ASCII file of all stations T 0 Read XMAG cors T 0 Read FMAG cors use all data use all data i CUSP phony station codes N 4 Hirshorn for Fmag2 Eaton s new magnitude constants Extra dist amp depth terms for use 2 06 71 1 35 for the 42db Z stations Eaton EW CODE D L Use lowgain verts for FMAG2 Zmag AG components amp label MI hetic WA components comps Eaton Fmags Eaton Xmags ST home calnet klein hypfiles allsta2 bin H71 11 3 use new s STA home calnet klein hypfiles all2 sta C home calnet klein hypfiles all2000 xmc 1 C home calnet klein hypfiles all2000 fmc 1 NK 6 IRG1 IRG2 IRG3 IRGE WWVB IRG MULTIPLE CRUSTAL MODELS RCR home calnet klein hypfiles multmod2 bin Read all model info in binary READ MODELS AND DELAYS IN ASCII home calnet klein hypfiles multmod2 hyp MAGNITUDE CHOICES MAG 1 T 3 1 Use Eaton for Fmagl DUR 81 2 22 0 0011 0 5 0 9999 1 Set DU2 005 40 0006 350 014 10 DUB 2 06 2 95 0 001 0 S 0 9999S 1 Zmag of Hirshorn 1 35 compensates for the 1 35 gain corr used FELI D 4 VHZ VHE VHN VLZ comps for FMAG1 ECZ 82 Lh VL XC1 X 7 VHZ VLZ VLE VLN VDZ VDN VDE X XC2 L 6 WLN WLE HHN HHE BHN
2. OUTPUT FILES T IF APPENDED TO F PRINTOUT test prt F ARCHIVE test arc ARCHIVE FORMAT CODE 1 JB aa 1 DEC 1999 0 08 SEQUENCE NO 1 ID NO 21069577 ADJUSTMENTS KM I ORIGIN IATN LON W Z NWR RMS DT DLAT DION DZ RR NF MOD 1 5 56 35 56 14 120 31 19 5 00 80 94 0 817 2 645 0 440 0 000 2 681 3 P FREE DEPTH 2 6 38 35 57 57 120 30 90 5 00 80 05 0 170 0 539 0 233 1 916 2 004 4 PB 3 6 55 35 57 86 120 31 05 3 08 80 07 0 003 0 156 0 177 0 585 0 631 4 P BEGIN DISTANCE WEIGHTING BEGIN RESIDUAL WEIGHTING 4 6 55 35 57 77 120 31 17 2 50 8 0 03 0 007 0 038 0 092 0 189 0 214 4 P 5 6 54 35 57 75 120 31 23 2 31 8 0 03 0 003 0 037 0 003 0 082 0 090 4 P 6 6 55 35 57 73 120 31 23 2 39 8 0 03 0 001 0 015 0 042 0 084 0 095 4 PI ERROR ELLIPSE lt SERR AZ DIP gt lt 3 28 86 69 gt lt 0 58 213 12 gt lt 0 21 308 15 gt YEAR MO DA ORIGIN LAT N LON W DEPTH RMS ERH ERZ XMAG FMAG PMAG 1999 12 01 0008 6 55 35 57 73 120 31 23 2 39 0 03 1 14 3 08 0 56 0 96 0 96D SOURCE NSTA NPHS DMIN MODEL GAP ITR NEM NWR NWS NVR REMRKS AVH N XMG XMMAD T N FMG FMMAD T LF X 7 9 2 1 PM 137 6 6 8 2 8 MID 2 00 0 14 X 1 00 0 00D NNN XMAG2 N XMG2 XMMAD T S FMAG2 N FMG2 FMYAD T S PREE MAG N PMAG PRMAD T 0 96 1 00 0 00 D REGION Middle Mountain MODELS USED PMV1 00 STA NET COM L CR DIST AZM AN P S WI SEC TOBS TCAL DLY RES WI SR INFO CAL DUR W FMAG T AVP U PER W XMAG T PM NC VHZ
3. 28 an empty shadow record will appear after every output record Use the 200 command to invoke either the pre Y2000 or Y2000 century compatible archive formats with formats 1 or 3 Because the 200 flag is a global one it is not possible to read pre Y2000 and write Y2000 output 1 Full archive format The default Corresponds to COP 3 input format 2 No longer used 3 Shadow format Corresponds to COP 5 input format Weight codes name codes and other parameters common to all phase input formats Each phase is accompanied by a weight code that states the weight the phase carries in the solution from full use to no use By default the codes are 0 or blank full weight 1 weight 2 half weight 3 1 4 weight 4 9 no weight The actual numerical weights of codes 0 3 can be reassigned with the WET command The weight value assigned to a P time S time duration or amplitude is the product of those specified in the station and phase files The station site code must agree exactly upper lower case position of blanks etc with the name in the station file The site code may not be all blank To avoid confusion with shadow lines that could cause some programs to fail the name should not begin with a The station component may either be specified in the 1 letter field or the 3 letter field The 3 letter component in memory is ultimately matched with that in the station file to identify the station us
4. Supply the format number as follows 1 Traditional USGS full phase format No Y2000 format 2 digit years only 2 Unused 3 Hypoinverse archive format 4 Traditional phase format with shadow records PHASEOUT HYPO71 shadow option 5 Archive format with shadow records 6 One CUSP event supply CUSP ID number with LOC command 7 Many CUSP events ID numbers read from a file see FID command For options 6 and 7 See the phase input section for discussion of CUSP input and MEM files CUSP input applies to VAX version only Also specify the level of output to the MEM file being read for example COP 6 0 0 No MEM file output 1 MEM output to data structures 2 MEM output to shared memory region while timing events 3 MEM output to disk file Example COP 1 the default Determine the format of the input phase file already set with the PHS command automatically RIGHT NOW The first 2 lines of the phase file are read when you issue the FIL command and the input and output format numbers and Y2000 format mode normally set with the 200 COP and CAR commands are reset Does not work for binary MEM files As a convenience FIL also determines the format of the various types of summary files Hypo71 2000 etc which are not valid Hypoinverse input files The results are announced without resetting the input format number Example FIL there are no parameters Set the archive file output format The f
5. Supply DISW2 Example DIS 4 50 1 3 the default or DIS 4 15 3 5 7 works well for NCSN Set the parameters that govern progressive down weighting of stations with larger travel time residuals RMS is the root mean square travel time residual in seconds The residual weight is 1 0 for stations with residuals less than RMS RMSW1 0 0 for stations with residuals larger than RMS RMSW2 and is cosine tapered between these residuals If the RMS is smaller than RMSCUT RMSCUT is used in place of the RMS in the above cutoff residuals See the section on weighting and figure 7 Supply ITRRES the iteration on which residual weighting is to begin Iteration continues at least until residual weighting begins Supply RMSCUT Supply RMSWI1 Supply RMSW2 Example RMS 4 16 1 5 3 the default Specify the weight factors for the P and S phase weight codes 0 1 2 and 3 Weight codes 4 9 always result in a weight factor of 0 0 Code 0 should get full weight 1 0 Specify the weight for code 0 Specify the weight for code 1 Specify the weight for code 2 Specify the weight for code 3 Example WET 1 0 0 75 0 5 0 25 this is the traditional scheme the default or WET 1 0 0 5 0 2 0 1 this is a statistically better weighting function for the actual errors that typical seismogram readers use as they assign weight codes Set the S phase weighting factor Use 1 0 to give all S readings full weight and 0 to not t
6. 50 PAR CS Li e2h PAV C 0 16 PBI C 0 33 PBM C 0 26 PBP 0 25 PBR C 0 2 2 PBW Ce 0 13 PBY C 0 32 PCA C 0 72 This is a sample of an attenuation history file read with the ATE command Calibration files are similar ATTENUATION HISTORY FILE FRAGMENT ABL CI VHZ 6 1984103123 12 1985040120 18 ARV CI VHZ 12 BCH CI VHZ 6 1985092400 12 BMT CI VHZ 6 CRG CI VHZ 12 ECF CI VHZ 18 ETG CI VHZ 18 LJB CI VLZ 48 MAR CI VHZ 12 PKM CI VHZ 6 1985040120 12 Here is the command dialog in a terminal window that located the sample event The user commands and responses that were typed at the keyboard are in bold COMMAND DIALOG 103 puna 45 hyp HYPOINVERSE 2000 STARTING COMMAND ini INITIALIZING WITH COMMAND FILE home calnet klein hypfiles cal2000 hyp 3473 STATIONS READ IN BINARY 540 STATION XMAG CORRECTIONS SE 511 STATION FMAG CORRECTIONS SE 35 CRUST ODELS READ IN BINARY 804 STATION ATTENUATIONS SET 53 STATION CAL FACTORS SET COMMAND phs PHASE FILENAME CR testin arce COMMAND fil FIND INPUT PHASE FILE TYPE amp SET PHS COP amp ARC CAR FORMATS INPUT IS A HYPOINVERSE ARCHIVE 2000 FILE NO SHADOWS SETTING FORMATS COP 3 CAR 1 COMMAND prt PRINTOUT FILE NONE FOR NON
7. An additional component correction may be set with the FCM command FMGN is normally 0 or and controls whether the gain correction is used The magnitude relation may be bi linear in log f p Hypoinverse uses the values FMA1 FMB1 FMF1 FMD1 and FMZ1 when f p is less than FABRK and FMA2 FMB2 FMF2 FMD2 and FMZ2 when f p is more than FMBRK If the relation is linear set FMBRK to 9000 so that FMA2 FMB2 FMF2 FMZ2 and FMD2 are never used Some examples of relations in use by the USGS Lee et al 1972 Mp f p 0 87 2 0 log fp 0 0035 D STACOR old Hawaii Mp f p 5 2 3 89 log fp 0 0037 D 0 013 Z STACOR f p lt 210 sec Mp f p 0 9 2 026 log f p 0 0037 D 0 013 Z STACOR f p gt 210 sec new Hawaii Mp f p 1 424 1 883 log f p 0 00418 f p STACOR f p lt 307 Mp f p 0 267 1 917 log f p STACOR f p gt 307 45 Hirshorn and Lindh low gain vertical stations Mp f p 0 71 2 95 log fp 0 001 Z STACOR G Eaton 1992 The additional terms are defined by the DU2 command Mp f p 0 81 2 22 log f p 0 0011 D Stacor G Component correction 0 005 D 40 if D lt 40 km 0 0006 D 350 if D gt 350 km 0 014 Z 10 ifZ gt 10km Lapse time tau magnitude expressions The supported form of the lapse time expression is M7 tau DMAO DMA1 log tau DMA2 log tau DMLIN tau DMZ Z DMGN G STACOR Component correction tau
8. Nanometrics HRD broadband net 1 275 1 084 BH Guralp Reftek broadband net discontinued 1 907 0 7252 BH Guralp Quanterra datalogger BDSN short period 2 5 0 5532 EPI L4C Quanterra datalogger BDSN broadband 2 5 0 5532 BH Streckheisen Kinemetrics K2 0 298 4 637 EHZ L4C Amplitude magnitude comments If the calibration factor CAL is found equal to 0 no magnitude will be computed for that station The useful range of XCOR is 2 4 If you want to compute a magnitude for a station but exclude the result from the event magnitude either give it a zero weight or use a value of XCOR equal to 5 0 plus the actual correction See the sections on station and phase input for more information If you are calculating amplitude magnitudes from vertical instruments you should correct magnitudes to account for the lower amplitudes on verticals than on horizontals for which the ML scale was originally defined The amplitude magnitude correction for vertical instruments in Central California is empirically about 0 25 You could add this to all magnitude corrections of vertical stations in the station file More easily if you have unique component codes for vertical stations a global correction applied to all vertical components may be set with the XCM command 55 Computing two amplitude magnitudes Like coda magnitudes the philosophy for calculating two magnitudes for amplitude is complicated Hypoinverse can calculate one amplitude m
9. s magnitude weight on and off with time change weight between 0 and 1 The XMC command reads the amplitude magnitude corrections and FMC reads the duration magnitude corrections for a specified date The XMC command can also assign instrument types to stations You must read your station file using the STA command before the magnitude correction files because Hypoinverse stores corrections only for the stations already read into memory This means that the magnitude correction file can contain the history of the whole network and only the data needed will consume space in Hypoinverse The handling of magnitude corrections is very similar to that for station attenuations and cal factors Using a separate magnitude correction file insures that the correct information will be used for each station on the date of the earthquake Each FMAG duration or XMAG amplitude correction has an associated expiration date When an event is being processed whose date is after the correction s expiration date Hypoinverse rereads the file and uses the updated correction Both FMAG and XMAG correction files have one station per line Unlike the delay file full 12 letter station codes are supported because the magnitude correction depends on the component and location codes The file consists of a station code followed by pairs of magnitude corrections and their expiration dates The last correction must have its expiration date left blank or set to zero to indi
10. DIST Distance in km at which ray reaches the surface If DIST 1 then the ray is trapped in a waveguide and does not reach the surface TIME Travel time in seconds REDUCED Reduced travel time in seconds given by TTIME DIST REDV where REDV is one divided by the reducing velocity L BOT The layer in which down going rays bottom Z BOT The depth at which down going rays bottom V BOT The velocity at which down going rays bottom DDIF The difference in distance between this and the preceding ray DDIF is negative on reverse branches BR Branch number It is incremented by 1 each time a new forward branch is encountered AMP The relative amplitude of the ray at the surface assuming an isotropic source and geometrical spreading It is just the ratio of the area of a ring on a unit sphere surrounding the source to the corresponding area at which rays emerge at the earth s surface AMP R 2 Amplitude times distance squared Used to estimate the differences between actual and ideal inverse square spreading REMK Remark such as RB reversed branch or WG ray in wave guide TTGEN also writes a series of files for plotting travel time curves such as figure 8 Each filename is composed of the model name the travel time table output filename input on line 1 of the file ttmod The file extension part of the name is a 3 digit number indicating the depth of the source in 0 1 km Thus hg3 050 is a possible name for data for the depth 5
11. F7 2 1X A1 F5 2 13 F4 0 F5 1 F5 2 F5 1 Data revised from original format Year month and day Must be blank Hour and minute Origin time seconds Latitude deg S for south blank otherwise Latitude min Longitude deg E for east blank otherwise Longitude min Depth km Preferred magnitude label code Preferred magnitude Number of P amp S times with weights greater than 0 1 Maximum azimuthal gap Distance to nearest station km RMS travel time residual Horizontal error km Vertical error km Remark assigned by analyst i e Q for quarry blast Quality code A D See note above Most common data source code i e W earthworm Auxiliary remark from program i e for depth fixed etc Event identification number Version of the information i e the stage of processing This can either be passed through or assigned by Hypoinverse with the LAB command 92 is the last filled column Pre Y2000 station archive format The archive format contains essentially all of the data about the event It has enough data to relocate the event and archive files can thus be both input and output Archive output can be reformatted by programs like ARCPRINT to look like the print file but the archive file does not contain the iteration history The archive file may be read by other programs to plot first motions 96 on the focal sphere FPFIT FPPLOT compile residual
12. Figure 3 Diagram of simple analog seismic system The seismogram amplitude in mm Ap is and similarly for Ar in digitizer counts Ap Sx Gx Dpx Hy 3 For the frequency band in which each system is linear between the natural period of the seismometer and the high cut filtering of the electronics about 3 10 hz the terms are constants e S is the motor constant of the seismometer in v cm sec For the typical L4 C seismometer S 1 0 v cm sec e Gis the dimensionless gain of the preamp VCO transmission and discriminator system For the USGS system it is 19460 x 10 or 1228 at 24 db attenuation The attenuator is in the preamp at the seismometer and is made in 6 db 2x steps Atten is the setting in db e Dis the recorder scale factor For the Develocorder viewer Dp is 40 mm Vvolt For the Menlo Park CUSP and earthworm digitizers also see below Dr is 4096 counts 5 0 volt 819 counts volt e Hy is the ground velocity in cm sec The gain of the electronics and recorder is also expressed as a calibration factor CAL originally defined as the amplitude in mm on the Develocorder viewer of a 10 microvolt RMS 28 3 48 microvolt peak to peak 5 hz signal applied to the VCO electronics in place of the seismometer Although defined historically CAL is the measure of system gain that Hypoinverse uses because the USGS keeps track of units equivalent to CAL factors as measures of gain It is useful to express CAL as
13. Mx log Ac x U 2 x CAL x R f x S F s F d XCORcomp XCORsra 7 where Ac is the peak to peak amplitude in digital counts Full digital systems with velocity seismometers This system is similar to that described above except that the amplifying electronics and the recorder can be visualized as combined into a digitizer 52 Seismometer Digitizer motor system factor constant S F Figure 5 Diagram of simple digital seismic system The response of this system is similar to equation 3 where Ac is the output amplitude in counts Ac S x 10 F x Hy 8 e S is the motor constant of the seismometer in v cm sec For the typical L4 C seismometer S 1 0 v cm sec e Fis the system factor of the amplifier and digitizer in microvolts per count The 10 factor converts from volts to microvolts e Hy is the ground velocity in cm sec Combining equations 3 and 8 and the relations between counts and mm of seismogram amplitude yields 10 F 819 G counts volt CAL 1 382 F Menlo Park and earthworm 9 CAL 0 3456 F HVO CUSP Thus when you know the system factor F of your system in microvolts per count you can assign a CAL to that station to compute magnitudes from equation 7 providing the code specifying the type of amplitude units is correct The relation 9 between F and CAL assumes that F is used to specify the entire electronic and recording system It works for a pu
14. and Hirshorn and Lindh s Mz relation for the low gain vertical seismometer network Step 1 Use Eaton s 1992 relation for coda magnitude 1 The 1 as the last argument indicates that gain corrections are to be used if they are available DUR 81 2 22 0 0011 0 5 0 9999 1 Step 2 Add the extra distance and depth terms to coda magnitude 1 These can only be added to coda magnitude 1 DUR not 2 DUB DU2 005 40 0006 350 014 10 Step 3 Add the component corrections to Eaton s coda magnitude The 3 indicates there are 3 pairs of component codes and their corrections that follow FCM 3 VLZ 06 VLE 30 VLN 30 Step 4 Use Hirshorn and Lindh s MZ relation for coda magnitude 2 DUB 2 06 2 95 0 001 0 5 0 9999 1 Step 5 Now assign the magnitude types to the two event coda magnitudes This assigns the Eaton relation DUR to FMAGI and the Hirshorn relation DUB to FMAG2 The four arguments of the MAG command are 1 Relation for FMAG1 1 coda 1 2 lapse time tau 3 coda 2 1 2 Whether to use assigned coda weights T for true 3 Relation for FMAG2 1 coda 1 2 lapse time tau 3 coda 2 3 4 The default log Ao relation to use in amplitude magnitude calculations 1 means the relation of Eaton 1992 is used for XMAG Presently there are 5 relations to choose from but these choices are treated in the section on amplitude magnitudes MAG 1 T 3 1 Step 6 Assign the 1 lette
15. is the homogeneous layer model which calculates travel times directly from the velocity structure The second model type uses layers with linear velocity gradients but requires that a travel time table be generated previously by the program TTGEN The table needs to be generated only once and Hypoinverse uses it very efficiently by merely interpolating from it to get all travel times and derivatives Tests on the VAX computer show that using a travel time table requires about 60 of the CPU time used with layer models Use the CRH command to read homogeneous layer models and the CRT command for reading gradient travel time table models The two model types may be used simultaneously Homogeneous layer models Each model may consist of up to 20 homogeneous layers including the half space Velocity must increase with depth Use the CRH command to specify the model number and the name of the file containing the homogeneous layer model The CRH command also reads the model into memory For example CRH 2 CRUST2 CRH The format of the homogeneous layer crust model file CRUST2 CRH in the example is Line format data Line 1 A30 Model name Lines 2 and later 2F5 2 Velocity of layer and depth to its top The first 3 letters of the 30 letter model name are used as the model code that appears in the print summary and archive outputs Use one line per layer top layer first The depth to the top of first layer must be 0 0 and the last l
16. origin time may be specified on the terminator line If a trial value is absent the standard value is used To specify a trial origin time you must supply hour minute and second To specify a trial latitude or longitude you must supply degrees and minutes A terminator which is a HYPO71 style instruction line and is blank except for columns 18 19 is a valid terminator but the instruction parameter will have no effect To fix the depth for one event only make the trial depth negative To fix the depth for all events set the default trial depth negative with the ZTR command To fix the entire location for this event only to its trial value put a non blank character in column 35 of the Hypoinverse terminator line This might be used for explosions for example 31 The origin time will be computed to minimize the average residual The non blank character will be copied to the output print file as for example the reason for fixing the location You may also use the hypocenter on the event header as a trial if you are reading any of the archive formats COP 3 4 or 5 You do this by choosing terminator format 3 with the H71 command COP formats 3 and 5 expect the header in Hypoinverse format and format 4 expects the header in HYPO71 format Using the previous earthquake location as a trial hypocenter may not reduce the number of iterations required or speed up the location run That is because several iterations may be required before di
17. 0 00 PAG C IASZ 35 43 9493 120 15 0209 4360 0 P 0 00 0 00 0 00 0 00 0 0 00 PAG C JAS 35 43 9493 120 15 0209 4360 0 P 0 00 0 00 0 00 0 00 0 0 00 PAG C KASE 35 43 9493 120 15 0209 4360 0 P 0 00 0 00 0 00 0 00 0 0 00 PAG C VVHZ 35 43 9493 120 15 0209 4360 0 P 0 00 0 00 0 00 0 00 1 0 00 PAN C VVHZ 35 46 8034 120 54 4279 4260 0 P 0 00 0 00 0 00 0 00 1 0 00 PAP C VVHZ 35 53 7514 121 22 0113 10440 0 P 0 00 0 00 0 00 0 00 1 0 00 PAR C VVHZ 36 14 9567 120 20 5662 4520 0 P 0 00 0 00 0 00 0 00 1 0 00 PAV C VVHZ 35 10 5500 120 37 9500 1330 0 P 0 00 0 00 0 00 0 00 1 0 00 A sample of the coda magnitude correction file read with the FMC command follows It could be a history file but in this case the corrections are assumed constant in time and their expiration dates are 0 blank indicating no expiration CODA MAGNITUDE CORRECTION FILE FRAGMENT HCA NC VHZ 0 04 HCB NC VHZ 0 24 HCO NC VHZ 0 53 HCO NC VLZ 0 53 HCP NC VHZ 0 09 HCR NC VHZ 0 28 HDL NC VHZ 0 22 102 This is a sample of a layered crust model file read with the CRH command LAYERED CRUST MODEL FILE PARKFIELD MODEL 1 42 0 00 3 24 0 25 4 82 1 50 536 250 5 60 3 50 5 87 6 00 6 15 9 00 6 60 15 00 8 00 25 00 A station delay file read with the DEL command in this case for one particular model looks like this STATION DELAY FILE FRAGMENT PAG C 0 10 PAN C 0 20 PAP C 0
18. 0 km Each line of the file contains four numbers DIST Distance in km at which ray reaches the surface If DIST 1 then the ray is trapped in a waveguide and does not reach the surface TIME Travel time in seconds REDUCED Reduced travel time in seconds given by TTIME DIST REDV where REDV is one divided by the reducing velocity INS ANG Incidence angle of ray at the source measured in degrees from nadir This is 180 EM ANG the emergence angle output to the print file Format of the travel time table generated by TTGEN and used by Hypoinverse Start Fortran Col Len Format Data Line 1 1 20 A20 Model title to appear on output 21 2 12 Number of velocity points n specified for the model 23 8 F8 4 One over the reducing velocity in sec km used to reduce travel times in table 118 Line 2 1 15 nF5 2 Depths of the n points of model Line 3 1 75 nF5 2 Velocities of the n points of model Line 4 1 7 F7 4 DD1 8 3 13 ND1 11 7 F7 4 DD2 16 3 13 ND2 Line 5 1 7 F7 4 DZ1 8 3 13 NZ1 11 7 F7 4 DZ2 16 3 13 NZ2 Travel time Blocks One block now follows for each of the NZ1 NZ2 1 depth grid points Line 1 of block 1 10 F10 4 Depth 11 10 F10 4 Velocity at this depth 21 10 110 Distance in units of 0 1 km at which a horizontally emergent ray reaches the surface This is used to resolve the ambiguity between upgoing and downgoing rays Line 2 etc of block 1 90 1516 Reduced travel times at each of
19. 1 1X Presently blank 14 4 14 Origin time year KYEAR2 18 8 412 Origin time month day hour and minute KMONTH 26 4 F4 2 Origin time seconds KQ 30 4 F4 1 Epicentral distance in km KTEMP 34 2 A2 P remark from phase line KPRK 36 1 11 P weight code LPWT 37 4 F4 2 Calculated P travel time for this station MTCAL 41 4 F4 2 P residual observed TT calculated TT residual KPRES 45 3 A3 3 letter event remark epicentral location code REMK 48 1 A1 Data source code for this station KSOU 49 1 A1 1 letter station remark KRMK 50 1 11 Station type 0 Wood Anderson etc from station file J TYPE 51 5 F5 3 Calibration factor from station or attenuation file JCAL 56 3 3X Presently blank Duration magnitude data for the STATION 59 4 14 Coda duration F P time for this station KFMP 63 1 11 Coda mag weight code for measurement 0 9 blank KFWT 64 1 11 Coda magnitude weight code for station 0 9 blank JJFWT 92 75 UUU F3 2 F3 2 F3 2 A1 2X Station duration magnitude correction JFCOR Component correction for coda mag from FCM command ICOMF Duration magnitude for the station KFMAG Type label for duration magnitude to tell which of the two possible relations was used LABF Presently blank Amplitude magnitude data for the STATION 77 83 N gt UUU F6 2 Peak to peak amplitude AMPK Units of amplitude M mm on Develocorder viewer or paper Wood
20. 40 reads delays for the model number specified by the DEL command The formats of the delay files are Delay file for one model DEL n filename nis the model number Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 1X Station net code Delays are for all components at this site 10 5 F5 2 1X P delay for model n 16 1 A1 Optional source code for delay not read by Hypoinverse Codes presently in use are blank delay derived for this model N delay from a nearby station A delay assumed from a slightly different crustal model B delay borrowed from a nearby region Delay file for all models DEL 0 filename Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 2X Station net code Delays are for all components at this site 11 1 A1 Put an A here to use alternate crust models with this station 12 40F4 2 P delays for each model in order of crust model number Alternate model file DEL 1 filename Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 2X Station net code Alternate status is for all components at this site 11 1 A1 Put an A here to use alternate crust models with this station You must read in the station list using the STA command before reading delays with the DEL command The station file contains all the stations you will use for locations and the STA command reads these into memory The delay fil
21. 81 TOPi non inna anoles a 74 TFTGEN V iena a sede ae daseteaae 13 TAA E E E 77 UNK eisoriaponoentasiianad ah 27 70 WE acho okie aaa dd chasse 29 58 85 WS Terron e A aE 14 64 D O EE AEEA 56 81 DOA E E iateyeduniecducee 56 82 DOCS TA sats seeak sass neti aes inen a in 55 81 XCM ares ei lianas ete 55 83 XMAG adana aeaa a 24 46 XMCare e a a aA 14 24 53 67 DEIN ee ra n linen A 56 82 ZR acsisvest Seuditeauasegiictadess eas 31 75 107 The old commands LES DLY ST5 and VER are no longer used but are recognized and produce an error message identifying the replacement command The p amplitude magnitude commands PAC PC1 PC2 PMA and PMC are functional but not fully implemented or documented 122
22. Anderson record C digital counts D digital counts HVO only CA1 Period at which amplitude was measured sec KPER Amp mag weight code for measurement 0 9 blank KXWT Amp magnitude weight code for station 0 9 blank JXWT Amplitude magnitude correction from station file JXCOR Component correction for amp mag from XCM command ICOMX Amplitude magnitude at this station KXMAG Amp magnitude type code from XC1 or XC2 command LABX Presently blank Duration magnitude data for the EARTHQUAKE 101 104 105 106 109 113 a KRW A230 KRW A320 Median event duration magnitude 1 FMAG1 Type label for duration magnitude 1 LABF 1 Most common FMAG1 data source code FMSOU Median absolute difference of coda magnitudes 1 KFMMAD Total of FMAG1 weights i e number of readings MTFMAG Presently blank Median event duration magnitude 2 FMAG2 Type label for duration magnitude 2 LABF2 Most common FMAGz2 data source code FMSOU2 Median absolute difference of coda mags 2 KFMMAD2 Total of FMAG2 weights i e number of readings MTFMAG2 Presently blank Amplitude magnitude data for the EARTHQUAKE gt BW 32 wW AUU Median event amplitude magnitude 1 XMAG1 Type label for amplitude magnitude 1 LABX1 Most common XMAG1 data source code KMSOU Median absolute difference of amp magnitudes 1 KXMMAD Total of XMAG1 weights i e number of readings MTXMAG Presently b
23. L Hanson R L 1974 Solving Least Squares Problems Prentice Hall 340 pp Michaelson C A 1987 Coda duration magnitudes in Central California U S G S Open File Report 87 588 120 INDEX OF COMMANDS AND SPECIAL NAMES Page numbers in bold are the complete command explanations in the command dictionary 76 76 6 7 76 QO iat seheee bite a a aa 5 67 PRIS a a iss Auden aoa 12 19 72 FAD Petes colons E E neta adel on soled aati 71 ARO aea e oN lic dosent aE 71 A TE erae a eaaa 14 17 22 65 ATN erna 16 21 44 65 83 BAS eondu ida sand a S AS 35 73 BUG cums aati 73 COAT eek ele Sets Maer si ati Sa cata 17 21 66 CAR aate hae T ANE 28 68 CON aa A A A TEN 85 108 COP ron N a As 28 67 CRA a Ey Mt A 11 13 63 CRA a E AT aS 11 14 63 CUS Pa adaa aa 28 32 53 DAM reanrannaa tae Ra 86 108 DE E 11 19 65 8 AEE ERE A E AT ERS 59 84 B 8 AE BAE EE E A 41 46 80 DUB aa a ese Soliant aes 40 80 PUIG rsaindchsstecrnneshisnenantsnemabanaccmbaiadstas 44 80 DUR ernaar n aE 40 79 Earthworm 5 40 52 53 B ea EE E rh E E A 75 112 ERP raan oun a O AR 71 ERR iaa o n eaa tee ania 75 112 EXOCE Ta in a e nes 35 CA AE EA AEEA T 42 78 OE EE S E A ET 42 79 FEM A A a 41 79 BLD AEEA E EEE E Mecdeatodeodes 33 69 j E i Ee EE E P EE A ss 27 68 FEMA Garena as a ta 24 EMAGI sich cs cenreit ann a 42 PMA G e a E tee teat aches 42 MO a a araa ae ales 14 24 66 ET ATE E Gand 14 15 32 33 69 HED ier deste A
24. M D H Format repeats 7 F7 2 1X 14 312 1X Use the CAL command to read the station calibration factor file The phase data file must be in chronological order to insure getting correct calibration factors if there is ever more than one value per station You must read your station file using the STA command before the calibration factor file with the CAL command because Hypoinverse stores calibration factors only for the stations already read into memory This means that the cal factor file can contain the history of the entire network and only the data needed will consume space in Hypoinverse For analog stations with cal factors supplied with the CAL command note that the instrument type must be 3 The CAL command and not the ATE command is appropriate for digital stations Station magnitude correction files Hypoinverse calculates independent coda and amplitude magnitudes and uses separate corrections for each There are two options for specifying station magnitude corrections 1 put the magnitude correction in the station file to use for the whole location run or 2 read the 23 magnitude corrections from a separate file and assign the corrections to stations already in memory Duration and amplitude magnitude corrections may both vary with time This might be because some changes are not reflected in the attenuation or calibration history The ability to change magnitude corrections with time can also be used to turn a station
25. a trial location 2 calculating the RMS at the current location 3 calculating an adjustment vector in the direction which minimizes the RMS 4 take that adjustment or a modification of it and 5 repeat steps 2 4 until the solution converges or meets some criteria Where to begin iterations In the absence of a specified trial origin time latitude longitude or depth on the terminator line a standard trial hypocenter is assumed Any one of the four trial hypocenter parameters may be specified independently however The trial origin time is two seconds before the first arrival and the trial epicenter is under the station with the first arrival Starting depth is at the trial depth ZTR specified with the ZTR command During the early iterations usually just the first depth is held fixed until the horizontal adjustment is less than DXFIX specified with the DAM command If the trial depth ZTR is negative all events in this run are held fixed at this depth at the positive value unless ZTR is temporarily overridden by a trial depth which is set for a particular event on its terminator line 107 How iterative steps may be modified Various parameters can be defined which damp the epicentral adjustments if the adjustment vector becomes large or unstable DAMP DAM command is the damping factor by which all hypocenter adjustments are always multiplied before an iterative step is taken Damping is automatically increased by cutting DAMP in
26. amp S times with final weights greater than 0 1 43 3 13 Maximum azimuthal gap degrees 46 3 F3 0 Distance to nearest station km 49 4 F4 2 RMS travel time residual 53 3 F3 0 Azimuth of largest principal error deg E of N 56 2 F2 0 Dip of largest principal error deg 58 4 F4 2 Size of largest principal error km 62 3 F3 0 Azimuth of intermediate principal error 65 2 F2 0 Dip of intermediate principal error 67 4 F4 2 Size of intermediate principal error km 71 3 F3 2 Coda duration magnitude 74 3 A3 Event location remark region derived from location 77 4 F4 2 Size of smallest principal error km 81 1 Al Auxiliary remark from analyst i e Q for quarry 82 1 Al Auxiliary remark from program i e for depth fixed etc 83 3 13 Number of S times with weights greater than 0 1 86 4 F4 2 Horizontal error km 90 4 F4 2 Vertical error km 94 3 13 Number of P first motions 97 4 F4 1 Total of amplitude mag weights number of readings 101 4 F4 1 Total of duration mag weights number of readings 105 3 F3 2 Median absolute difference of amplitude magnitudes 108 3 F3 2 Median absolute difference of duration magnitudes 111 3 A3 3 letter code of crust and delay model 114 1 Al Authority code i e what network furnished the information Hypoinverse passes this code through 115 1 Al Most common P amp S data source code See table 1 below 116 1 A1 Most common duration data source c
27. and returning to locate the corrected file EARTHQUAKE LOCATION METHODS Interactive earthquake processing Hypoinverse can locate a set of events interactively You can alter the input data and relocate several times until you are satisfied with an event Hypoinverse does this by stepping automatically through a preset list of events you wish to process The data for each event must be in separate files and therefore the computer file system does the necessary updating of files and retrieval of the correct events The two Hypoinverse commands that accomplish this process are BAS to establish the file naming you will use and PRO to actually process locate and edit the set of events The first step requires putting each event you wish to locate in a separate file The filenames must consist of a base name of 1 20 characters and a suffix of 1 8 characters for each input and output file type The same base name is used for each input and output file type associated with an event and each event must have a unique base name I suggest you use the date and time to make up the base name because files will then list in chronological order The base names are read as a text string from a file listing all events to be processed For example a base name might be 19840502133045 and a set of suffixes might be arc input arc output archive prt output print and sum output summary The input filename to read for this event wou
28. annulus in which the weight of the model relative to the surrounding models is decreasing from 1 0 Next several of the most important parameters set by the various commands are listed This includes the names if the input and output files 87 Earthquake iteration output The print output for each earthquake begins with a line with the event s date sequence number numbered sequentially within this run but not output in the archive or summary file and id number it is read and written and serves as a permanent and unique id number In the earthquake output one line of information per iteration is printed when the print control set with the KPR command is 2 or larger The final location data is always written to a print file and the station list data is written if the print control is 1 or larger I Iteration number ORIGIN Seconds part of origin time LAT N Latitude LON W Longitude Z Depth NWR Number of P amp S readings with final station weights larger than 0 1 This reflects all weight factors including the distance and residual weights which can vary from iteration to iteration RMS The root mean square travel time residual using station weights DT Origin time adjustment applied to ORIGIN to get to the next iteration location DLAT Latitude adjustment in km positive north DLON Longitude adjustment in km positive west DZ Depth adjustment positive down RR Length of adjustment vector in km NF Numb
29. are Eaton BSSA 1992 Bakun and Joyner BSSA 1984 Richter s 1958 approximation UC Berkeley Nordquist BSSA 1948 P amplitude presently unused A BWN Re Supply the number of components that will follow with their log Ao relation Supply pairs of 3 letter component codes and log Ao relationship numbers Example LAO 0 no special components the default or LAO 2 WLN 4 WLE 4 Use the Nordquist relation with 2 Wood Andersons 83 ATN Select whether to assume that station records have CALibration factors ATN F or attenuation settings ATN T The attenuations in db are entered in place of CAL factors on the station lines and must be a multiple of 6 See the amplitude magnitude section for the conversion between CAL factors and attenuation values If a value is entered that is not an integer multiple of 6 the CAL factor is left as zero which means no magnitudes will be computed This must be entered before the STA command to know how to interpret the numbers when reading station lines Set a flag F for Cal factors or T for attenuations Example ATN F MISCELLANEOUS COMMANDS NET Set the network number for assigning names to earthquakes based on their locations This option requires that prior definition of earthquake regions be coded into the KLAS subroutine If you do not have a net with defined names use a NET of 0 Present nets are 1 Hawaii 2 Northern California 3 new Hawaii See the QPLO
30. at 15 db where CAL 15 db 3 95 G 0 05 attenuation 0 753 Log CAL factor 0 05 attenuation 1 35 Note that G 0 for a station with 15 db attenuation CAL of 3 95 Not using a gain correction is equivalent to assuming an attenuation of 15 db which is close to the 12 or 18 db typical for most analog stations Either CAL factors or attentions may be supplied in the station file see the ATN command or time dependent attentions or CAL factors may be read from the appropriate history file see the ATE and CAL commands If you do not know the station attentions or CAL factors you should probably assume an average attenuation of 15 dB You may do this in several ways 1 Use a CAL factor of 3 95 and select the CAL factor option with the ATN command 2 Leave the CAL factor unknown by using 0 or blank on the station lines and select the CAL factor option with the ATN command or 3 Do not use a gain attenuation correction in the magnitude by setting DMGN 0 with the TAU command or FMGN 0 with the DUR command Do not directly specify an attenuation of 15 db because attenuations must be a multiple of 6 Note that duration gain corrections must be enabled both by setting FMGN 1 with the DUR command and by enabling corrections for some or all components with the DUG command The default setting of the DUG command is to apply duration magnitude corrections to all components and the default value of FMGN is 0 DUR command thus by default
31. average absolute value of discriminator output regardless of the channel s attenuation setting For digital stations no VCO and no attenuator setting the NCSN Earthworm systems are configured to terminate the coda at the average absolute value corresponding to a velocity of 1 729 e 5 cm s The Earthworm coda termination value in digital counts is determined separately for each sensor digitizer combination by using the sensor motor constant and the digitizer s sensitivity In terms of Earthworm parameters the coda termination value endcoda groundmotion of 1 729 e 5 cm s is EWcodaterm EWsensitivity 15dB L4Cmotorconstant 49 14 count 0 3519 uV count 0 000001 cm s uV 0 00001729 cm s Thus a digital velocity channel and a 15db analog velocity channel will report equivalent coda durations 15 db was chosen as the reference because this is the default gain for which no gain correction is made in Hypoinverse processing This 15 db assumption is also made for coda magnitudes from analog stations with unknown gain 40 The Earthworm end of coda amplitude is determined for all velocity channels analog or digital at any gain such that the resulting coda is equivalent to that measured by traditional Develocorder methods Thus as long as the Earthworm system is configured carefully as described above a coda magnitude relation developed for an analog network can work for an Earthworm system recording digital stations and a catalog wi
32. defaults by pressing just the RETURN key after the prompt This makes the program very easy to use providing you can remember the names of the commands Combining commands with and without their required parameters into a command file permits a variety of customized procedures such as automatic input of crustal model and station data but prompting for a different phase file each time All commands are 3 letters long and most require one or more parameters or file names If they appear on a line with a command character strings such as filenames must be enclosed in apostrophes single quotes Appendix 1 gives this and other free format rules for supplying parameters which are parsed in fortran When several parameters are required following a command any of them may be omitted by replacing them with null fields see appendix 1 A null field leaves that parameter unchanged from its current or default value When you start HYPOINVERSE default values are in effect for all parameters except file names Format of this document Throughout this document the file formats are given where appropriate Data files are fixed format column oriented The formats in this document use the fortran specification Thus A10 is an alpha field 10 characters long I5 is a five digit right justified integer F4 2 is a 4 column real number where the decimal point implied as 2 decimal places if the decimal point is not present etc Formats are given in Helvetica typefa
33. error ACTUAL TIMING ERROR y RDERR ERCOF RMSP Set ERCOF according to the influence you want the fit of your data the RMS to play in the location error calculation ERCOF should usually be in the range 0 1 inclusive The calculated location errors will be proportional to this actual timing error Supply ERCOF the error coefficient of RMS Example ERC 1 0 the default Used to skip events with too few reporting stations If fewer than MINSTA stations are present for an event no location is even attempted This is useful when small events are to be screened out See also the JUN command Supply MINSTA the minimum number of stations required to attempt a location Example MIN 4 the default CONVENIENCE AND CONTROL COMMANDS INI HEL MOR STO Initialize Hypoinverse by running a standard command file which the user has set up to set default variables read in station and crust files etc The unix version of the program executes the command file named in the environment variable HYPINITFILE Other versions have the file and pathname compiled into the program The file is set in hybeg f or hybeg for See the section Initializing with your defaults and input files Once set up this makes starting the program easy Example IN Typing HELP HEL or HELL in Hypoinverse gets a brief listing of the most important commands No detailed information is available through HELP Typing M
34. expiration 34 5 F5 2 1X Second magnitude correction 40 4 14 Second expiration year 44 6 312 1X Second expiration date M D H Format repeats 6 F5 2 1X 14 312 1X Other station comments The relationships used in handling arrival times and delays are as follows TOBS SEC CCOR OT 26 RES TOBS TCAL DLY where SEC observed arrival time CCOR clock correction only read from traditional format phase files OT origin time TOBS observed travel time TCAL calculated travel time DLY station delay RES travel time residual If a station is found in the phase but not the station file an error message will normally go to the terminal and print file This message may be suppressed for a certain list of fictitious stations such as those referring to digitized time code traces Use the UNK command to set the list of 5 letter site codes for which you want no error message All phase data for unknown stations is saved in a separate memory area and will be written at the end of the event in the output archive file but will not be listed in the print file PHASE DATA INPUT FORMATS The name of the input phase data file is specified with the PHS command The LOC command starts locating events For example PHS 1983 PHS LOC The phase file may contain any number of earthquakes Phase data may be in one of several formats see the COP and FIL commands All formats require a terminating line after e
35. found on the input archive header record and passed through to output are A is the authority code i e what network furnished the information Hypoinverse passes this code through V is the version of the information i e what stage of processing or completeness the event is in This can either be passed through or can be set by Hypoinverse to all events in the current run to the code set with the LAB command H is the version number of last human review Hypoinverse passes this code through The A V and H codes are in columns 114 163 and 164 of the Y2000 summary line The traditional USGS phase data input format not Y2000 compatible Some fields were added after the original HYPO71 phase format definition Start Fortran Col Len Format Data 1 4 A4 4 letter station site code Also see col 78 5 2 A2 P remark such as IP If blank any P time is ignored 7 1 A1 P first motion such as U D C D 8 1 11 Assigned P weight code 9 1 A1 Optional 1 letter station component 10 10 512 Year month day hour and minute 20 5 F5 2 Second of P arrival 30 25 1 1X Presently unused 26 6 6X Reserved remark field This field is not copied to output files 32 5 F5 2 Second of S arrival The S time will be used if this field is non blank 37 2 A2 1X S remark such as ES 40 1 11 Assigned weight code for S 41 1 A1 3X Data source code This is copied to the archive output 45 3 F3 0 Peak to peak a
36. half for the last 1 3 of the allowed number of iterations Thus if 15 iterations are allowed and convergence has not been reached after 10 iterations the remaining 5 iterations will be heavily damped Empirically this appears to improve convergence If an iterative step would place the hypocenter in the air the hypocenter is moved up to the fraction 1 DZAIR DAM command of its present depth Earthquakes above the model surface are not allowed because a flat earth model is assumed with all stations at its surface Thus the depth adjustment is DZAIR Z The depth adjustment may be independently damped if the adjustment is larger than DZMAX DAM command If it is the depth variation is damped by the factor DZMAX DZ DZMAX where DZ is the calculated depth adjustment If the value of RMS should increase by more than the amount RBACK DAM command after the last iteration the hypocenter is moved back by the fraction BACFAC DAM command toward the previous hypocenter This situation often occurs when a poorly constrained hypocenter iterates across a large velocity discontinuity in the crustal model The use of a generalized inverse scheme for finding the hypocenter adjustment allows great control over the adjustments actually taken For example we may choose not to make hypocenter adjustments in directions which are poorly constrained by the arrival time data and which are directions in which location errors are large The parameter EIGTOL D
37. in and passed through to output All of the shadow lines and the terminator line with any trial hypocenter are passed through to output without modification Pass through fields on the summary line include 29 e The first of the two 1 letter auxiliary remarks the one assigned by the user quarry blast etc e The externally determined magnitude its label and number of readings e The event identification number e The three 1 letter authority and version remarks see below Pass through remark fields on the station archive lines e P amp S remarks first motion weights etc e 1 letter station seismogram remark e Data source code who or where the data came from Many fields that are calculated by Hypoinverse are ignored on input for example 3 letter location remark Station distances and emergence angles Weight values actually used in the final location and importance values Travel time residuals If an archive file is read as phase input the 1 letter event remark assigned by the user is passed through to output Hypoinverse assigns the second 1 letter remark A pound sign indicates convergence problems with the final solution such as running out of iterations or failure of the hypocenter to reach a minimum in RMS A minus sign indicates the depth was held fixed either by the user or by insufficient data An X indicates the location was fixed to the trial hypocenter The three status remark codes for the event
38. is taken in that eigenvalue s direction Supply RBACK See BACFAC Supply BACFAC Ifthe RMS increases by more than RBACK from one iteration to the next move the hypocenter by the fraction BACFAC back toward the last location and continue iterating Supply DXMAX the maximum distance adjustment in km Supply D2FAR the maximum distance for the second closest station in km before iteration is terminated This prevents a distant earthquake from going out of control Example DAM 7 30 0 5 0 9 0 012 0 02 0 6 50 250 the default OUTPUT FILE FORMATS Print Output The amount of information in the printed output can be varied somewhat The LST command controls the quantity of station crust model and parameter data at the beginning of the print output file before any earthquakes are located The KPR command governs the level of data seen for each earthquake 86 User or data errors are marked with a specific message that begins with You can search for errors by searching for these 3 asterisks Errors can also be reported to the terminal as locations proceed see the ERF command The print file begins with a status line containing the date run and the run label set with the LAB command Station table NAME The 5 letter station site code NT The 2 letter station network code COM The 3 letter station component code C The 1 letter station component code normally treated as an abbreviation to the 3 letter compon
39. is the lapse time P travel time coda duration f p Z is the positive depth STACOR is the duration magnitude correction from the station line G is the gain correction A component correction may be set with the FCM command The coefficients DM are set by the TAU command The defaults are DMA0 1 312 DMA1 2 329 DMA2 0 DMLIN 0 00197 DMZ 0 and DMGN 1 AMPLITUDE LOCAL MAGNITUDES Local magnitudes from Wood Anderson seismometers The method for calculating local magnitudes is modeled after the reading of maximum peak to peak amplitudes from a standard Wood Anderson torsion seismograph If amplitude is read from an electromagnetic seismometer with velocity output it is corrected to an equivalent Wood Anderson response using Jerry Eaton s XMAG formulation 1970 1992 the seismometer motor constant and the response curve of the seismometer and recording system Digital amplitudes are handled also by using the appropriate system gain Richter s original magnitude formula is M log Awa 2 log Ao 1 where Awa is the maximum peak to peak amplitude in mm on the paper record and log Ao is an attenuation term and is a tabulated function of distance The division by 2 is because of the peak to peak reading This formula is applied to all stations of Hypoinverse instrument type 0 Wood Anderson The implementation of this formula in Hypoinverse is M log Awa 2 X CAL H F s F d XCORcomp XCORsta 2 wher
40. layer trace a circular ray path with infinite radius a straight line The use of linear gradients smoothes out the discontinuities in travel time derivatives which result from homogeneous layer models and gives a more realistic spread in emergence angles of down going rays than is possible with modeling rays as refracted from discontinuities One buried low velocity zone is permitted in the model This means that velocity may not decrease with depth except for one group of adjacent velocity points Hypocenters that occur within a low velocity zone may produce a shadow zone at the surface and rays in this distance range are calculated as if refracted along the layer above the low velocity zone TTGEN can handle models with homogeneous layers zero gradients but velocity discontinuities infinite gradients are not allowed Velocity gradients should assume reasonable values such as 0 0 or between 0 02 and 8 0 km sec km in the interest of numerical stability TTGEN operates by shooting rays out from the source and calculating time distance and other parameters where and if they emerge at the surface Layers with steep gradients such as might be used to model a Moho transition can produce reverse branches in the travel time curve and such layers should be at least 0 3 km thick to insure that enough rays will bottom in the layer to define the travel time curve and its reverse branch properly Errors can be introduced in the final travel time tab
41. list Send printer output to the file RUN1 PRT Don t list available stations or crust model in printout List only final solution amp station data on printout for each event Doesn t begin each new event at the top of a page Write summary data to SET1 SUM Read archive format input Define the phase file as OLD ARC Locate the events Example 4 Locate a set of events with then without S CRH 1 MOD1 CRH STA ALL STA PRT RUN1 PRT SUM WITHS SUM SWT 1 0 PHS SET1 PHS LOC SUM NOS SUM SWT 0 PRT RUN2 PRT LOC Read layers model 1 from the file MOD1 CRH Read station list from file ALL STA Send printer output to the file RUN1 PRT Put summary data with S in this file Set S weighting to 1 0 full weight Define the phase file as SET1 PHS Locate the events Put summary data without S in this file Set S weighting to 0 no weight Send printer output for the second run to this file Locate the same events as before 62 COMMANDS RECOGNIZED BY HYPOINVERSE The commands are grouped by function and thus all the commands dealing with amount of printout are listed together for example The required parameters are listed below each command The function of many commands is discussed above in the sections on various inputs and calculations The parameter defaults if any are listed in the examples If you do not supply parameters on the command line you will be prompted for parameters and
42. loads the earliest attenuation for each station This will require more updates from the attenuation file but will not require knowing the date of the first event This may 22 account for taking several extra seconds to locate your first earthquake Attenuations are converted directly into calibration factors when read in Calibration factor history file format pre Y2000 Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 2X Station net code 11 3 A3 1X Station component code 15 T7 F7 2 1X First calibration factor 23 8 412 3X Y M D H expiration date and hour of first calibration factor Minutes if supplied will not be used Leave the expiration of the last calibration factor blank to indicate no expiration 34 7 F7 2 1X Second calibration factor 42 8 412 3X Second expiration date Format repeats 7 F7 2 1X 412 3X Calibration factor history file format Y2000 century compatible Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 Station net code 9 2 A2 Station location code 11 3 A3 1X Station component code 15 7 F7 2 1X First calibration factor 23 4 14 Year of expiration of first calibration factor 27 6 3I12 1X M D H expiration date and hour of first cal factor Leave the expiration of the last cal factor blank to indicate no expiration 34 7 F7 2 1X Second calibration factor 42 4 14 Second expiration year 46 6 312 1X Second expiration date
43. magnitude relation is an extension of equation 2 Mx log Ap 2 x CAL x R f x S F s Fx d XCORcomp XCORsra 6 e Ap is the peak to peak amplitude on the Develocorder viewer in mm e CAL is the dimensionless calibration factor depending on system gain as described above e R f is the frequency dependent response curve of the USGS system relative to the Wood Anderson seismometer Hypoinverse interpolates this function from the table given below and plotted in figure 4 In the band 3 hz lt f lt 10 hz where the seismometer is flat to velocity the Wood Anderson is flat to displacement and no electronic filtering is in effect log R log f 1 e Sis the seismometer motor constant in volt cm sec e F s and F2 d are the log Ao distance correction terms discussed above They can be a function of epicentral distance d slant distance s or both e XCORcomp is the correction made globally to all components with a given component code See the XCM command 49 e XCORsgrza is the individual station correction This is specified for each site and component at that site Station corrections may be supplied in the station file or a magnitude correction history file may be specified with the XMC command Response function R f of NCSN analog system relative to Wood Anderson response log R 1 0 0 5 0 0 0 5 1 0 1 5 log frequency Figure 4 Response function R f of the standard NCSN analog sei
44. no duration gain corrections are used You can suppress gain corrections for individual stations by adding 10 0 to the coda magnitude correction This suppression can also vary with time See the FMC command One must be cautious when using both durations and amplitudes from digital stations If the gain correction is applied to durations from digital stations you should note that the same cal factor gain setting is used for both types of magnitudes and that the duration gain correction was only derived for analog stations This led to bogus duration magnitudes in Menlo Park s tests because the cal factors set for correct amplitude magnitudes did not give appropriate duration gain corrections Menlo Park decided to apply duration gain corrections to analog stations but not to digital stations Therefore gain corrections were enabled with the DUR and DUB commands The DUG command however is used to enable corrections only for the component codes corresponding to analog stations If the component codes for analog and digital stations are the same they are in the SEED convention for example EHZ you must selectively suppress corrections for individual stations by adding 10 0 to the coda magnitude correction This means that the durations from digital stations are assumed to come from a signal equivalent to an analog station running at an attenuator setting of 15 db or a cal factor of 3 95 In practice the coda termination level in counts for each
45. second duration relationship with the DUB command and the lapse time parameters with the TAU command Then use the MAG command to assign which of these three types to use for the primary coda magnitude and which for the secondary magnitude For example you could calculate only lapse time magnitudes as your primary coda magnitude or use the duration magnitude 1 as your primary and duration 2 as the secondary coda magnitude Step 2 Next choose which if any stations you want to assign to each of the two magnitudes You do this by selecting the component codes of stations you can choose no stations no event magnitude computed all stations or list up to 10 different 3 letter codes The FC1 command selects which components to use for the primary coda magnitude and the FC2 command selects which components to use for the secondary magnitude The default is to use all components for FMAGI1 and no stations for FMAG2 The component selection applies to all 3 of the magnitude types and is independent of it A magnitude is calculated for each station but if it is un weighted or if it is not on the list of components for FMAG1 or FMAG2 it is not used in that event magnitude A special case is when both coda duration magnitudes FMAG1 and FMAG2 are selected for a particular component i e a component code appears on both the FC1 and FC2 lists Because only one magnitude can be computed per station Hypoinverse chooses the secondary coda magnitude FMAG72
46. special case of reading many CUSP MEM files with the COP 7 command supply the file listing the CUSP ID numbers to locate 63 Phase filename Example PHS PHASES PHS there is no default BINARY FILES Binary files are designed for fast reading of station and crust data You would read in all the ASCII files once write them out in binary then read data in binary each time you use the program History data that changes with time such as calibration factors or station magnitude corrections are only read from the ASCII files WCR Write a snapshot of all crustal models and multiple model definitions currently in RCR WST RST memory to a file in binary form RIGHT NOW You should have issued all CRT CRH NOD ALT and MUL commands first Supply the filename to write to Example WCR home calnet klein hypfiles multmod bin Read a binary file of all crustal models and multiple model definitions previously written with a WCR command RIGHT NOW RCR replaces the CRT CRH NOD ALT and MUL commands Binary reads are several times faster than ASCII reads and worth the effort for frequently read files Supply the filename to read from Example RCR WE KLEIN MULT MULTMOD BIN Write a snapshot of all station data and delays currently in memory to a binary file RIGHT NOW You should have issued all STA and DEL commands first Note that only one calibration factor and one type of each magnitude correct
47. station is chosen to 44 correspond to 1 cm peak to peak Develocorder viewer of an analog station at 15 db attenuation This also means that codas are comparable to those for an old analog station before it became digital To work out the coda termination level in counts study the sections on system response in the amplitude magnitude chapter and keep in mind that duration magnitudes can also be adjusted for certain stations or components with the FMC and FCM commands Another specialized way of making gain corrections arose in Menlo Park in 2002 We will convert component codes to SEED compatibility and thus will loose the easy ability to distinguish analog from digital stations for purposes of deciding whether to make the coda gain correction We will implement a station by station choice of whether to use the gain correction by suppressing it for digital stations We will use the magnitude correction option of adding 10 0 to the correction as a signal to suppress the gain correction Duration magnitude expressions The complete form of the duration magnitude expression is Mp f p FMA FMB log f p FMF f p FMD D FMZ Z STACOR FMGN G The FM coefficients are set by the DUR and DUB commands f p is the end of coda F minus P time or duration D is the epicentral distance Z is the positive depth STACOR is the duration magnitude correction for the station G is the gain correction S is the slant distance S D Z
48. stations to be fully weighted for the first few iterations until the hypocenter approaches its final location and after residual weighting begins stations may be weighted in or out as convergence proceeds This generally allows about the largest 10 of the residuals to be weighted down if there are any bad readings If you have trouble converging I have found that manually weighting out the largest residuals then re introducing them if the residuals become small often helps in tracking down the bad readings To not use any residual weighting set RMSCUT to a large number like 1000 This forces full weighting of residuals less than thousands of seconds RMSWT gt RMSCUT POOR SOLUTION ABSOLUTE RESIDUAL RMSWT RMSW1 RMSWT RMSW2 RMSWT lt RMSCUT GOOD SOLUTION ABSOLUTE RESIDUAL RMSCUT RMSW1 RMSCUT RMSW2 Figure 7 The residual weighting function RMSCUT RMSW1 and RMSW2 are constants set with the RMS command RMSWT is the root mean square travel time residual computed before residual weighting is applied If RMSWT is larger than RMSCUT upper figure the function stretches out and scales with RMSWT such that most of the network stations receive weights somewhere in the tapering part of the function If RMSWT is smaller than RMSCUT lower figure the function is fixed Stations with residuals larger than RMSCUT x RMSW2 then receive no weight SOME SIMPLE COMMAND SEQUENCES Several examples of comma
49. summaries extract a subset of events SELECT EXTRACT etc The first line of each event in an archive file is identical to a HY POINVERSE summary line and acts as a header Next is one line per station and a terminator line terminates each event If the archive file has shadow records a shadow record beginning with a follows every header station and terminator line For historical reasons many fields are in the same position as the traditional phase input format In fact if you accidentally read an archive file expecting phase format COP 1 you will get a location but it will be incomplete and subtle errors may result Start Fortran Col Len Format Data 1 4 A4 4 letter station site code 5 2 A2 P remark such as IP 7 1 A1 P first motion 8 1 11 Assigned P weight code 0 9 or blank 9 1 A1 One letter station component code 10 10 512 Year month day hour and minute 20 5 F5 2 Second of P arrival 25 4 F4 2 P travel time residual 29 3 F3 2 P weight actually used 32 5 F5 2 Second of S arrival 37 2 3 A2 1X S remark such as ES 40 1 11 Assigned S weight code 41 4 F4 2 S travel time residual 45 3 F3 0 Peak to peak amplitude in Develocorder or WA mm 48 3 F3 2 S weight actually used 51 4 F4 2 P delay time 55 4 F4 2 S delay time 59 4 F4 1 Epicentral distance km 63 3 F3 0 Emergence angle at source 66 1 11 Amplitude magnitude weight code 67 1 11 Duration magnitude weight code 68 3 F3
50. that if one or more eigenvalues in S becomes small both the solution vector and error become large and unstable Each eigenvalue corresponds to one of the mutually orthogonal principal directions of the solution and if one eigenvalue becomes small both the adjustment DX and standard error VC in that principal direction become large in proportion to one over that eigenvalue The principal direction with the small eigenvalue will in general include components of origin time latitude longitude and depth Most often however the smallest eigenvalue has its largest component in depth If an eigenvalue should become smaller than the parameter EIGTOL set with the DAM command no adjustment is taken in that principal direction for which the error is also large The program does not add the term to DX originating from the small eigenvalue In other words solutions are prevented from becoming unstable and scattering out in the direction in which their error ellipsoids are very long In general the largest eigenvalue is of order 5 with its dominant component in origin time and the spatial eigenvalues are of order 0 3 to 0 7 The difference in size between origin time and 110 spatial eigenvalues arises because a change of several km in hypocenter location is required to produce the same change in an arrival time as a one second change in origin time Unstable or very poorly constrained situations occur when the smallest eigenvalue becomes less than ab
51. the ND1 ND2 1 distance grid points The reduced travel times are given as integers in units of 0005 sec minus 32000 Ths converts an integer range of 32000 into a time range of 0 0 to 32 seconds This encryption was done because Hypoinverse was originally programmed on a 16 bit computer with two byte integers to save space Appendix 6 Programmers notes The source code is in fortran77 with a few of the extensions The code is compatible with Sun computers running the sunos or solaris operating systems and DEC Compac VAX or Alpha computers running VMS System dependent fortran code like OPEN statements and system calls have been put in separate and small subroutines which can be linked into the appropriate executable For example the subroutine openr f opens files for reading on unix systems and OPENR FOR does the same on VMS systems When both a f and FOR version are present the former is for the unix system and the FOR is for the VMS system 119 REFERENCES Bakun W B and W Joiner 1984 The ML scale in central California Bull Seis Soc Am v 74 p1827 Eaton J P 1969 HYPOLAYR A computer program for determining hypocenters of local earthquakes in an earth consisting of uniform flat layers over a half space U S Geological Survey Open File Report Eaton J P 1970 and later Harmonic magnification of the complete telemetered seismic system from seismometer to film viewer screen U S Geological Su
52. the input phase file This is where you can make any change such as deleting stations changing data flagging quarry shots inserting a trial hypocenter etc When the event is relocated the input file is read first then the print file for new weights Note that weight changes in the print file will override those in the input file so do not attempt to change weights in the input file Then go to step 2 above When you finish processing a set of events delete all of the print files If you are on VMS purge the old versions of the input archive and summary files because they will not be needed If you interrupted the processing session without finishing you can edit the file listing the base names to either remove or comment out the events already processed before restarting Fixing depth or hypocenter Hypoinverse offers a few options for fixing hypocentral parameters One option is fixing the depth while solving for epicenter and origin time This can be done for all events in a run by making the trial depth set with the ZTR command negative Fixing depth can also be done on an individual event basis by using a negative trial depth on the terminator line To specify a trial depth on the terminator you must read an ASCII file You can t fix a single event depth while reading CUSP MEM files because there is no terminator You can fix the hypocenter in one of two ways The way to fix the entire location for this event only to its trial val
53. the station attenuation file Once an attenuation file has been read and a set of attenuations defined for a certain date the file is re read as needed during a location run to find a new attenuation value after the 65 CAL FMC expiration date of the previous attenuation value The new value is used until it expires This is why the earthquakes must be in chronological order when this feature is used See the section on specifying the station list for the file format Also see the event magnitudes and names section Supply the filename of the attenuation history file Supply the date and time year month day hour for which to load the initial station attenuations Hypoinverse will update the station attenuations as needed Use a year of zero to load the earliest attenuations Example there is no default ATE ALL ATN 1980 1 1 0 Read station calibration factors gains from a history file RIGHT NOW You must read your station file with the STA command before reading the station calibration factor file Once a calibration factor file has been read for a certain date the file is re read as needed during a location run to find a new calibration factor after the expiration date of the previous value The new value is used until it expires This is why the earthquakes must be in chronological order when this feature is used See the section on specifying the station list for the file format Also see the event magnitudes and na
54. to represent that station The component list also determines which of the station magnitudes are used in the two event magnitudes If a station s component is used for both the magnitude is used both in FMAGI and FMAG2 for the event This is true even when the two event magnitudes use different coda parameters Thus some station FMAG2 s can be used in event FMAGI s but this could be desirable See below for an example If all this seems confusing it is because Hypoinverse has been asked to serve different networks with different ways of calculating magnitudes simultaneously Coda magnitude examples A simple example is to use the original CALNET Lee et al relation for all stations to get the event magnitude The default settings use durations from all stations to compute coda magnitude 1 DUR command and compute the event s FMAG1 from them DUR 0 87 2 0 0 0 0035 0 5 0 9999 0 The first 5 arguments are coefficients for the constant log duration depth distance and linear terms in the magnitude vs duration relation for durations less than FMBRK 9999 seconds The second 5 terms are the unused terms for the relation for durations longer than FMBRK The last two arguments are FMBRK and a gain term coefficient 1 to use gain corrections and 0 not to use gains 42 A more complicated example is that of the current 2000 NCSN system NCSN uses Eaton s 1992 Mp relation for the high and low gain 1 hz seismometer network
55. 1 355 1 443 1 535 1 630 1 726 1 822 1 916 frequency 10 0 12 6 15 9 20 0 25 1 31 6 39 8 50 1 log frequency 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 log R 1 921 1 998 2 010 1 918 1 740 1 502 1 196 0 789 Relating Mx to M We can check our equations by equating Mx to M and seeing if the velocity seismometer response is reasonable Equating 1 and 6 yields Awa 2 Ap 2x CAL x R f x S Using 3 and the relation between Hp in mm and Hy in cm sec Hy 27f0 1 Hp and using CAL G 883 and using Dp 40 mm volt yields Awa 22200 f Hp R f If we restrict the frequency band to 3 hz and higher where the Wood Anderson is flat to displacement and its gain is 2080 Awa 2080 Hp yields R f 10 67 f This is close to the approximation given in the definition of R f above 51 Digital data recording and units of amplitude measurement Before the 2000 version Hypoinverse always assumed that the units of measurement were in millimeters as read from paper records or from the Develocorder viewer screen 100x magnification from the film The advent of digital seismographs and the reporting of amplitudes in digital counts meant that amplitudes had to be converted to mm before writing on the phase line for input to Hypoinverse Refer to equation 3 For the USGS analog recording the scale factor Dp 40 mm Volt applied to the Develocorder viewer For Menlo Park CUSP and Earthworm digitizing and recording systems using 12 bit digitizers 5 0 v
56. 1 4 14 5 4 212 9 1 1X 10 4 212 14 6 F6 2 20 3 F3 0 23 1 A1 24 5 F5 2 29 4 F4 0 33 1 A1 34 5 F5 2 39 7 F7 2 46 1 1X 47 1 A1 48 5 F5 2 53 3 13 56 4 F4 0 60 5 F5 1 65 5 F5 2 70 5 F5 1 75 5 F5 1 80 1 A1 81 1 A1 82 1 A1 83 1 A1 84 10 110 94 1 1X 95 1 A1 96 3 A3 Data revised from pre Y2000 format Year Month and day Must be blank Hour and minute Origin time seconds Latitude deg S for south blank otherwise Latitude min Longitude deg E for east blank otherwise Longitude min Depth km Blank Preferred magnitude type code Preferred magnitude Number of P amp S times with weights greater than 0 1 Maximum azimuthal gap Distance to nearest station km RMS travel time residual Horizontal error km Vertical error km Remark assigned by analyst i e Q for quarry blast Quality code A D See note above Most common data source code i e W earthworm Auxiliary remark from program i e for depth fixed etc Event identification number Blank Version of the information i e the stage of processing This can either be passed through or assigned by Hypoinverse with the LAB command 3 letter location remark 98 is the last filled column Y2000 station archive format Start Fortran Col Len Format 1 5 A5 Data revised from pre Y2000 format 5 letter station site code left justified 99 N gt hAWWWRAWBAAWRARARWONNRSBAANUTWADT
57. 137 7104 0 43 0 815 5504 21069577 19M 0108 0 42 612 0171 O O 829 6197 96 0 341 272ND ON ON 0259 0 69 860 199N 0285 0 0 80 0331 0 0 804 22343211 0 0 ON ON ONI 2 114 308 6 20 1014 OPMM NND 8X X X 106 APPENDIX Appendix 1 Rules for free format parameters following commands e Supply the parameters in free format following the command e The type and order of parameters is the same as in the command documentation e Free format values may be separated by either spaces commas or tabs e Character strings for filenames labels etc are delimited by apostrophes like MYFILE DAT e The form n A stands for n occurrences of the value A e A null field will leave the existing value unchanged A null field is specified by two consecutive commas by one leading comma or by two trailing commas Thus 2 MYFILE changes only the Ist and 3rd of 4 values e A slash at the end of a line means all later fields are null e The form n stands for n occurrences of a null field Appendix 2 Iteration and convergence The solution to the earthquake location problem is non linear It is solved by a series of iterations in linear steps to converge on the minimum RMS root mean square travel time residual which is the best estimate of the hypocenter The RMS is where rj are the individual station residuals and wi are their final weights The location method involves 1 guessing
58. 2 Amplitude magnitude correction If in the range 2 4 the correction is included by addition in the amplitude magnitude If you don t want a station s magnitude used in the event magnitude use a correction of 5 0 plus the actual correction You can also assign a zero weight see next Amplitude magnitude weight code Codes 0 9 and blank are used the same as the P amp S weight codes col 5 The actual magnitude weight used is the product of those on the station and phase lines See also col 48 Duration magnitude correction works the same as the amplitude magnitude correction Duration magnitude weight code works the same as the amplitude weight code Instrument type code Calibration factor Seismic network code 3 letter component or channel code 2 letter location code site component extension The HYPOINVERSE station data format 2 Start Col 16 19 NN Fortran Format A5 1X A2 1X A1 A3 1X A1 12 1X F7 4 Data Station site code The first character may not be the character Seismic network code Optional 1 letter station component code 3 letter component or channel code Station weight code in units of 0 1 by which the weights assigned each P amp S phase are to be multiplied Use the digits 0 9 for the weight in tenths or 0 for no weight or any other character including blank for full weight Latitude degrees Latitude minutes 26 1 A1 N or blank for north lati
59. 2 Period at which the amplitude was measured for this station 71 1 A1 1 letter station remark 72 4 F4 0 Coda duration in seconds 76 3 F3 0 Azimuth to station in degrees E of N 79 2 F2 1 Duration magnitude for this station 81 2 F2 1 Amplitude magnitude for this station 83 4 F4 3 Importance of P arrival 87 4 F4 3 Importance of S arrival 91 1 1X Presently blank 92 1 Al Data source code 93 1 Al Label code for duration magnitude from FC1 or FC2 command 94 1 Al Label code for amplitude magnitude from XC1 or XC2 command 95 1 A1 5th letter of station site code optional 96 3 A3 3 letter station component code 99 2 A2 2 letter seismic network code 101 2 A2 2 letter station location code component extension 102 is the last filled column 97 Y2000 OUTPUT FORMATS Recall that the 200 command selects between the old non Y2000 formats and the current Y2000 formats The Y2000 formats have 4 digit years Y2000 summary format also used as header in archive file Start Fortran Col Len Format Data revised from pre Y2000 format 1 4 14 Year 5 8 412 Month day hour and minute 13 4 F4 2 Origin time seconds 17 2 F2 0 Latitude deg First character must not be blank 19 1 A1 S for south blank otherwise 20 4 F4 2 Latitude min 24 3 F3 0 Longitude deg 27 1 A1 E for east blank otherwise 28 4 F4 2 Longitude min 32 5 F5 2 Depth km 37 3 F3 2 Amplitude magnitude 40 3 13 Number of P
60. 5 I5 NZ1 11 5 F5 2 DZ2 16 5 I5 NZ2 Line 4 Parameters for incrementing the grid spacing in distance see text 1 5 F5 2 DD1 6 5 I5 ND1 11 5 F5 2 DD2 16 5 I5 ND2 Line 5 1 20 A20 Name of model The first 3 letters of the 20 are the model code used throughout Hypoinverse and the output files as a model label prefer to make the model code the same as the root of the travel time table output file name Line 6 and later 1 5 F5 2 Velocity of first point km sec 6 5 F5 2 Depth of first point km This format is repeated for each velocity depth point of the model one line per point up to a total of 15 points The last point given sets the velocity and depth of the halfspace Here is a sample input file tttmod The moho is represented by a thin layer with high gradient between 13 7 and 15 5 km depth The halfspace under the model has an upper mantle velocity of 8 3 km sec The reducing velocity is 1 0 13 7 69 km sec hg3 PRT hg3 130 08 100 0 4 90 48 MAX 1 0 18 4 0 9 40 KM DEPTH MAX 1 0 28 4 0 13 135 KM DIST MAX hg32 19 0 0 3 0 1 4 O32 Bd 116 co N WN o W oO N A Reduced travel time sec Distance km DEPTH 2 00 0 0 40 80 Figure 8 A sample plot of the reduced travel time curve for the above input file for a source at 2 km depth The straight branch dashed is the Pn branch refracted from the halfspace The reverse branch is from rays which bottom in th
61. 7 command and specify the file containing the list with the PHS command The file may only have one CUSP ID per line but Hypoinverse will ignore any lines containing a blank field or spurious text The number should be right justified in its input field Use the FID command to specify the format for reading the list file the default is FID 110 If you want to locate all of the MEM files in a directory type the VMS command 33 DIR COL 1 OUT CUSPLIST MEM before running Hypoinverse In Hypoinverse use the commands COP 7 PHS CUSPLIST FID If all of the CUSP ID s are 8 digits use FID 1X I8 If you have a mixture of 7 and 8 digit numbers you will have to edit the file to get a field with only right justified numbers in it Locate all the events with the LOC command It is possible to write the new location back into the MEM file The COP command takes a second argument if the format choice the first argument is 6 or 7 The second argument controls the MEM output destination 0 none 1 data structures 2 shared memory region if working in a CUSP process and 3 MEM disk file Input of new phase data from the keyboard If you do not have files of phase data Hypoinverse has a utility to input arrival time data from the keyboard and write a traditional phase file that can subsequently be located This utility is entirely separate from the location functions of Hypoinverse The phase data may come from a read
62. 97 3 F3 2 100 3 F3 2 103 2 A3 106 1 A1 107 1 A1 108 1 A1 109 1 A1 110 1 A1 111 3 13 114 1 A1 115 1 A1 116 3 F3 2 119 3 F3 1 Data Year month day hour and minute Origin time seconds Latitude deg S for south blank otherwise Latitude min Longitude deg E for east blank otherwise Longitude min Depth km Amplitude magnitude Number of P amp S times with final weights greater than 0 1 Maximum azimuthal gap Distance to nearest station km RMS travel time residual sec Azimuth of largest principal error deg E of N Dip of largest principal error deg Size of largest principal error km Azimuth of intermediate principal error Dip of intermediate principal error Size of intermediate principal error km Coda duration magnitude Event location remark region derived from location Size of smallest principal error km Auxiliary remark from analyst i e Q for quarry Auxiliary remark from program i e for depth fixed etc Number of S times with weights greater than 0 1 Horizontal error km Vertical error km Number of P first motions Total of amplitude magnitude weights number of readings Total of duration magnitude weights number of readings Median absolute difference of max amplitude magnitudes Median absolute difference of duration magnitudes 3 letter code of crust and delay model Authority code i e what network furnished the in
63. AM command does exactly this and serves as a cut off below which eigenvalues of the inversion are deemed unstable and are suppressed See the appendix on the inversion scheme for a brief description of the matrix equations that might aid in using this parameter Two other parameters are useful in curbing large and generally erratic steps when the solution is poorly constrained Individual distance steps are limited to DXMAX DAM command When the distance to the second closest station exceeds D2FAR for example a large number like 250 km set with the DAM command stop iterating because the earthquake is outside the network and has no control Convergence and when to stop iterating Normally iteration can stop in any of 3 ways 1 when the number of iterations reaches the maximum allowed ITRLIM set with the CON command 2 When the change in the RMS residual between iterations becomes less than DRQT CON command or 3 When the hypocenter adjustment vector is less than DQUI T km set with the CON command The last two tests are only applied after a the depth has been freed from its trial value for at least one iteration and b after residual and distance weighting have been applied Appendix 3 Inversion scheme and use of eigenvalue cutoff In what follows bold variables are vectors or matrices and n or m are their dimensions This derivation follows treatments by Geiger 1912 Eaton 1969 and Lawson and Hanson 1974 Let s first co
64. ARBAANAWAaAD 2 letter seismic network code Blank One letter station component code 3 letter station component code Blank P remark such as IP P first motion Assigned P weight code Year Month day hour and minute Second of P arrival P travel time residual P weight actually used Second of S arrival S remark such as ES Blank Assigned S weight code S travel time residual Amplitude Normally peak to peak Amp units code 0 PP mm 1 0 to peak mm UCB 2 digital counts S weight actually used P delay time S delay time Epicentral distance km Emergence angle at source Amplitude magnitude weight code Duration magnitude weight code Period at which the amplitude was measured for this station 1 letter station remark Coda duration in seconds Azimuth to station in degrees E of N Duration magnitude for this station Amplitude magnitude for this station Importance of P arrival Importance of S arrival Data source code Label code for duration magnitude from FC1 or FC2 command Label code for amplitude magnitude from XC1 or XC2 command 2 letter station location code component extension 113 is the last filled column SAMPLE INPUT AND OUTPUT FILES The sample event is a small Parkfield earthquake detected and located in the regular NCSN processing Because NCSN processing has many large input files that are used regularly we use binary files and a startu
65. BHE LOCAL MAG Use BKY s Nordquist logAO relation for the WA amp synt LAO 6 WLN 4 WLE 4 BHN 4 BHE 4 HHN 4 HHE FCM 3 VLZ 06 VLE 30 VLN 30 Component corrections for XCM 2 VHZ 335 VEZ 20 Component corrections for PRE 6 3049 1109 2109 4449 3009 4 Preferred mag STANDARD CHOIC RMS 4 10 2 3 ERR 10 POS 1 78 REP F JUN T MIN 4 NET 2 ZTR 5 DIS 42309 3ST WET 2 25 62 min readings ES FOR CALNET Log events to terminal mag range In p ref order Residual weighting Standard timing error P to S ratio 0 0 9 Preferred mags don t print unweighted stations Force location of junk events Min number of stations Use California region codes Trial depth Distance weighting Weights for P weight codes 0 3 101 OUTPUT FORMAT ERE Send error messages to terminal TOP F No page ejects LST 1 1 0 No station list or models in printfile KPR 2 Medium print output each event PMAG SETUP 3 LETTER COMPONENT CODES USE PMA PMC PC1 PC2 AND PRE COMMANDS PMA T T 04 4 5 TURN ON PMAG CALC amp PRINTING CNT2MM FACTOR CLIPRATIO PMC 4 W 0488 P 04 R 04 0 04 DATA SOURCES W CNT2MM FACTORS PC1 TPY 1 418 1 760 1 PRIMARY PMAG FROM ALL COMPONENTS PC2 G 0 0 1 0 1 VLZ SECONDARY PMAG FROM L
66. CALN may represent 3 components at the same site Hypoinverse allows use of the full 12 letters or a subset of each of the 4 fields The LET command chooses how many letters from each field you want to use The newest station name field is the 2 letter location code This code discriminates between different channels that would otherwise have identical site net and component codes Examples include the same instrument recorded by different data loggers similar instruments at the site a station that was moved to different location but not given a new code or an experiment type code It can function like an extension of the channel or site code The LET command specifies how many letters of the 2 letter location code you want to test when matching stations in the station file with stations in other files In the LET command you separately state how many location letters to use 0 2 for matching station names in 1 phase files NSLOC or 2 other station files like magnitude correction or cal factor NSLOC2 For example say many stations were moved a few hundred feet to a better site but not given a new code at the time If you do not want to implement location codes use 0 as the number of location letters to compare and the location codes will be ignored Alternatively each site could get a different location code like 01 and 02 and comparing 2 letters with NSLOC 2 would use location codes and the appropriate station location to comp
67. E CR test prt COMMAND arc ARCHIVE FILE NONE FOR NONE CR test are COMMAND loc SEQ DATE TIME REMARK LAT LON DEPTH RMS PMAG NUM ERH RZ ID 1 1999 12 1 0 08 MID 35 58 120 31 2639 003 0D S Pel Coss 27069577 COMMAND sto puna 46 The input phase file set with the PHS command for this event is in the archive 2000 format The header line was not written by Hypoinverse because all fields are not filled in INPUT PHASE FILE 19991201 0 8 64735 5775120 3094 355 0 160 PST NC VVHZ IPD0199912010008 7 56 00 0 4 0 0 PMM NC VVHZ EPU2199912010008 7 71 700 0 4 19M 0 0 PVC NC VVHZ IPD1199912010008 7 81 8 98ESD2 00 0 0 6 0 PPC NC VVHZ IPU0199912010008 8 13 9 19ESU2 1600 0 4 0 0 PHP NC VVHZ EPU2199912010008 8 39 00 0 4 0 0 PSM NC VVHZ EPU2199912010008 9 98 00 0 4 0 0 BMS NC VVHZ P 419991201000899 99 00 0 4 22 0 21069577 The output print file set with the PRT command has some of the basic program test parameters listed before the earthquake output OUTPUT PRINT FILE HYPOINVERSE 2000 11 99 VERSION RUN ON Mon Apr 24 16 31 42 2000 RUN LABEL TEST PARAMETERS ITERATION AND CONVERGENCE WEIGHTING AND ERRORS MISCELLANEOUS 20 ITRLIM 0 900 DAMP 15 000 DISCUT 0 100 RMSCUT 4 MINSTA 0 040 DQUIT 0 001 DROT 3 500 DISW1 2 000 RMSW1 2 NET 7 000 DXFIX 0 012 EIGTOL 7 000 DISW2 3 000 RMSW2 30 000 DZMAX 0 020 RBACK 4 ITRDIS 4 ITRRES 5 000 ZTR 0 500 DZAIR 0 600 BAC
68. EARTHQUAKE LOCATION METHODS amp siacecssscassss sascces agecieosteacacke stesso tadecssaecdaros aiew eemiar et 35 Interactive earthquake processine si ciescrane Sh eiasieaee nee he ead ee 35 How to change weights in the print file cece ceccceseceecceeneeesseceecseeeeeseecsaeceseeeeeeeeaeeesaeen 37 Flow of steps in interactive processing yc Joc ekactehacdosctasaionadabiiaealagarudeelansaacieonal ease das 37 Fixing depth OF hypocentet osre ie e a a Ae a 38 Keeping or eliminating poor earthquakes wy cccsscc sssesMissieces vecontcsstsee Meedtccnsetianseiiesecteasecuodee tines 39 CODA MAGNITUDES iaoea re e RLE A aE A I ERTO EE EA RE AT E ah 40 Types of coda magnitudes and how the coda is measured ccccecsceesseestceceeceteeeeeeeeseeesaeees 40 Coda mag itude OPUS cao akare tet ctedans dscasuchate ca ecchnaeuaes iinn ekinini eidar sasise 41 Selecting coda magnitude types sssessssessesessressesstssresstsstesresseesrrsstesresetssressesstesresseeseeseessee 41 Coda m gnituderexamiplessironnino deinna ia a r Ea o i E a aia 42 Gain corrections to coda magnitudes ccisceaiskSssvsaessigdaidisesstaladasss sd eaeeaediaite an cusses aed 44 Duration magnitude expressiOns v csusshscusy sels ccs tots eadaeenesas cece ead eon 45 Lapse time tau magnitude ExpressiONns ci csssaccssssvcctesasessndessnesecssneesv ended cdlenssceeedaseeeanandnasendayes 46 AMPLITUDE LOCAL MAGNITUDES saisacasadpstivgecaiesssutcdshctedause gucadyatey a aaa i 46
69. F Duration magnitude X Amplitude magnitude The secondary magnitudes XMAG2 and FMAG2 and preferred magnitude PMAG will only appear in the printout if they were calculated XMAG2 N XMG2 XMMAD T S FMAG2 N FMG2 FMMAD T S PREF MAG N PMAG PRMAD T The secondary amplitude magnitude Number of secondary amplitude magnitudes used in the median XMAG2 It is the total of their weights which is the same as the number if all weights are 1 Weighted median absolute difference of the secondary amplitude magnitudes It is the error in the median magnitude calculated this way because it is a median The label code given XMAG2 by the XC2 command The principal data source code of XMAG2 The secondary coda magnitude Number of secondary duration magnitudes used in the median XMAG2 It is the total of their weights which is the same as the number if all weights are 1 Weighted median absolute difference of the secondary duration magnitudes It is the error in the median magnitude calculated this way because it is a median The label code given FMAG2 by the FC2 command The principal data source code of FMAG2 The preferred magnitude PMAG chosen from the magnitudes available Number of station magnitudes used in the median PMAG It is the total of their weights which is the same as the number if all weights are 1 Weighted median absolute difference of the magnitudes It is the error in the median magnitude calcula
70. FAC 1 000 SWT 0 100 RDERR 1 780 POS 1 000 ERCOF 104 DURATION MAG CONSTANTS DELAYS amp MISC STATIONS 0 810 FMA1 0 000 FMA2 1 0300 DMA0 T LMULT F ATIE 2 220 EMB1 0 000 FMB2 2 1000 DMA1 T CODAWT 5 SITE LET 0 000 FMZ1 0 000 FMZ2 0 0000 DMA2 T LJUNK 2 NET LET 0 001 FMD1 0 000 FMD2 0 0000 DMZ 3 COMP LET 0 000 FMF 1 0 000 FMF 2 1 0000 DMGN 1 000 FMGN 9999 000 FMBRK 0 0027 DMLIN 0 0050 DCOF1 40 000 DBRK1 1 LOGAO 0 0006 DCOF2 350 000 DBRK2 0 0140 ZCOF 10 000 ZBRK DUR MAG COMPON AMP MAG COMPON T T CORRECTIONS VLZ 06 VLE 30 VIN 30 T CORRECTIONS VHZ 0 33 VLZ 0 20 T MAGNITUDE LABELS amp COMPONENTS FMAG1 LABEI D COMPS VHZ VHE VHN VLZ FMAG2 IABEL Z COMPS VLZ XMAG1 LABEL X COMPS VHZ VLZ VLE VLN VDZ VDN VDE XMAG2 LABEI L COMPS WLN WLE HHN HHE BHN BHE FMAG1 MAGSEL 1 FMAG2 MAGSEL 3 INPUT FILES COMMANDS hnome calnet klein hypfiles cal2000 hyp BINARY STATION SNAPSHOT FILE home calnet klein hypfiles allsta2 bin DELAYS ATTENUATIONS home calnet klein hypfiles al12000 atn CAL FACTORS home calnet klein hypfiles al12000 cal FMAG CORRECT home calnet klein hypfiles al12000 fmc XMAG CORRECT home calnet klein hypfiles al12000 xmc PHASES testin arc PHASE FORMAT CODE 3 TERMINATOR FORMAT CODE 1 BINARY CRUST SNAPSHOT FILE home calnet klein hypfiles multmod2 bin
71. Local magnitudes from Wood Anderson Seismometers c ccccccesseeeseeesseceeceteceeeeeeseeeaeens 46 Amplitude magnitude distance COrrectiONns ccceccceseceseceeseeescecaeceeeeeeeeeseecaeccsaeeneeseeeensaes 47 Magnitudes XMAGs from electromagnetic velocity seismometePS cecceeeeeeeteeteees 48 Analog data transmission and recording cccceccceessceeseceeeceeseeeseeceaecnseeeeeeenaeecsaeeneeeeeeenaeees 48 Amplitude magnitude relation for velocity seismometers ceseseceeseeeeeeseceeceteceeeeeeenees 49 eG ait Mi tO M onient teh nena E ama E E ti aal yeti Aus Aled a et dine aa 51 Digital data recording and units of amplitude measurement ssssssesssssseseesesseseesessersessesee 52 Full digital systems with velocity seismometers ss snssseseesessessessresseesresresseesreserssresseseess 52 Seismometer instrument types sssessseseesseeseeseesstesesrtsstesesstessesstsstessesstestosseeseesresseesesressee 54 CAL factors for various digitizing systems ssessessesseeseessessresrosseesresrtessesrtssressessresresseese 55 Amplitude magnitude comments ee yess Sass cess de say als own sas cessandvanvoat oeteaidvecpacavieienlaevacdneeee 55 Computing two amplitude magnitudes nssssenseesesseessesesssresstssessresstestesresseesersstessessessressesse 56 PAIS TI OTE ES necne a n a a E A E haa a aidaa 56 Amplitude magnitude command example f 25 sessstsscrsaternauseelaceausisdscarpadiaaiaen
72. NOD 37 35 122 18 10 10 3 The transition regions surrounding the inner circles are where things get interesting as up to three models are used each with a weight less than 1 0 Hypoinverse uses the algorithm described above to find the weights for up to 3 nodes for earthquakes in the transition outer circle regions such that all weights total 1 0 Model 1 100 default model Weighted mixture of models 1 amp 2 Model 2 100 VVeighted mixture of models 1 2 amp 3 Weighted mixture of models 1 amp 3 Model 3 100 Node point the center of an inner and outer circle Figure 1 A example of nodes defining the areas for two crustal models and the transition regions between them Earthquakes in the dark gray areas within the inner circles use one model exclusively Earthquakes in the white area surrounding all of the circles use the default model exclusively Earthquakes in the lighter gray transition regions use a weighted average of models Numbers indicate locations where the weighting of different models is discussed in the text Consider these example cases for the numbered locations in figure 1 1 A simple location that is only in one of the node s outer circle regions Following the algorithm point 1 gets about 2 weight for model 2 Because the total weight is less than 1 0 the remaining weight to total 1 0 is made from the default model 1 Thus only models 1 and 2 are used in this area with a cosine taper to mak
73. ORE or MOR gets a listing of additional commands that do not fit on the basic HELP screen Files of commands can be executed as if they were typed at the keyboard by typing filename A command file may call another command file returning to where it left off and be nested up to 4 levels deep Any operating system command can be executed from within Hypoinverse by typing command This has no effect on current parameters and control returns to Hypoinverse when the command finishes This can be used to edit files check the directory for files etc Stop the program There are no parameters 76 SHO List the current input and output files on your terminal MAX List the current array maxima for stations phases crust models etc This will show the INP TYP maximum array capacity not the number of stations that have actually been read The number of stations read is announced after the see the STA and related commands have finished Input phases data from the keyboard with station prompting See the section on inputing phase data from the keyboard There are no parameters Type a line of text on the terminal This can announce something when running a Hypoinverse command file or give a hint about a command which will prompt for something No string apostrophes are needed Example TYP Now reading stations MAGNITUDES MAG Select the type of coda magnitude relation and weighting to use for the two coda magnitudes calc
74. OW GAIN COMPONENT EPRE Dp 28 po P20 94 6 AsO 9 Add pmags to preferred mag list xxxxxx x YOU SHOULD USE THE ATE COMMAND WITH THE DATE AND TIME OF YOUR FIRST EQ xxxxx x x USING A YEAR OF 0 THE ATE COMMAND BELOW NOT COMMENTED OUT WILL ALWAYS xxx k K WORK BUT IS INEFFICIENT xxx x AN EASY WAY IS TO USE THE ATE COMMAND BELOW WITH A 0 YEAR THEN xk KKK FOLLOW THIS COMMAND FILE WITH YOUR OWN ATE COMMAND kee ATE 1 1984 110 0 HIS REREADS THE STANDARD FILE BUT LOADS xxx x ATTENUATIONS FROM YOUR STARTING DATE ATE home calnet klein hypfiles all2000 atn 1984 1100 Put date of your first eq ATE home calnet klein hypfiles all2000 atn 0 Use this if first date is unknown CAL home calnet klein hypfiles all2000 cal 0 Load cal factors for digital stations The above control file cal2000 hyp reads binary station and crust files to save time because there are so many large files Here are a few lines of the station location file all2 sta that was read to make the binary file as read with the STA command It is in Hypoinverse station format 2 STATION FILE FRAGMENT PAB C VVHZ 35 9 4371 120 38 2559 1430 0 P 0 00 0 00 0 00 0 00 1 0 00 PABB C VVHZ 35 9 5352 120 38 3950 1700 0 P 0 00 0 00 0 00 0 00 1 0 00 PAD C VVHZ 35 38 4030 120 51 9729 4300 0 P 0 00 0 00 0 00 0 00 1
75. T documentation for a list of regions The 3 letter region codes are written to the summary archive and print files The full name is written to the print file Supply the net number Example NET 0 the default WEIGHTING OF ARRIVAL TIMES JUN Specify whether to force a solution of small and poor junk events If the number of stations remaining after distance and residual weighting are applied is less than the minimum number set with the MIN command the event will normally abort Setting the junk flag to TRUE cancels all distance and residual weighting for events that would otherwise abort Example JUN F the default DIS Set the parameters that govern progressive down weighting of more distant stations DMIN2 is the distance to the second closest seismic station in km The distance weight is 1 0 for stations closer than DMIN2 DISW1 0 0 for stations beyond DMIN2 DISW2 and is cosine tapered between those distances If DMIN2 is smaller closer than DISCUT DISCUT is used in place of DMIN2 in the above distances This means there is an expanding circle of stations if the earthquake is outside the network but the circle can t shrink too small for earthquakes under a dense part of the network See the section on weighting and figure 6 84 RMS WET SWT Supply ITRDIS the iteration on which distance weighting is to begin Iteration continues at least until distance weighting begins Supply DISCUT Supply DISW1
76. THQUAKE MAGNITUDES Hypoinverse potentially has available 5 different magnitudes 1 Primary duration magnitude 57 2 Primary amplitude magnitude 3 An externally derived sometimes called Berkeley magnitude 4 Secondary amplitude magnitude 5 Secondary duration magnitude 6 and 7 are reserved for un implemented P amplitude magnitudes Hypoinverse can select a preferred magnitude from those available The preferred magnitude is recorded along with all the other magnitudes but is generally the one that is reported or plotted The choice can require that certain criteria are met have a preference order and thus be somewhat complicated The procedure for choosing a preferred magnitude involves going sequentially through your list of choices and if a given magnitude matches your criteria it is chosen as the preferred value The tests made on each magnitude to evaluate it as the preferred magnitude are 1 Itis defined greater than 0 2 It have a minimum number of readings total of weights 3 It be in a certain magnitude range For example NCSN earthquake processing can have any of 4 magnitudes available The choices are 1 External magnitude Berkeley Wood Anderson any number of readings ie minimum 0 4 0 lt M lt 9 9 2 Primary duration magnitude at least 1 reading any M 0 0 lt M lt 9 9 3 Primary amplitude magnitude xmag at least 1 reading any M 0 0 lt M lt 9 9 4 Secondary amplitude magnitude rec
77. UL Indicate whether region dependent crustal models are to be used If true also give the NOD number of the default model to use outside the explicit regions Set a flag T to use multiple models or F to use one model Example MUL F the default or MUL T 1 Define a circle on a map by its center and radius within which all epicenters will use a particular crustal model and set of station delays Presently 124 nodes or circles are 71 SNO ALT allowed Each NOD command issued defines a new node you thus cannot reset or examine a node once defined The SNO command displays the current nodes on the terminal The transition width defines an annulus outside the inner circle within which the crustal model is used in combination with other models Set the latitude of the circle center degrees positive north Set the latitude of the circle center minutes Set the longitude of the circle center degrees positive west Set the longitude of the circle center minutes Set the radius km of a circle where the model is used exclusively Set the transition width outside the circle km for partial weighting Set the crust model number for this node Example there are no defaults NOD 37 10 122 5 4 30 10 2 This node for model 2 is a 30 km radius circle with a 10 km wide transition zone surrounding it Show the nodes centers and radii of circular regions for various crustal models which have been defined
78. UNITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY User s Guide to HY POINVERSE 2000 a Fortran Program to Solve for Earthquake Locations and Magnitudes 4 2002 version by Fred W Klein U S Geological Survey 345 Middlefield Rd MS 977 Menlo Park CA 94025 klein usgs gov Open File Report 02 171 Version 1 0 This report is preliminary and has not been reviewed for conformity with U S Geological Survey editorial standards or with the North American Stratigraphic Code Any use of trade firm or product names is for descriptive purposes only and does not imply endorsement by the U S Government 4 2002 program version TABLE OF CONTENTS TABLE OF FIGURES AND TABLES is xcnsseuenisvaaguinatietnyee son weap ions a R E R 4 INTRODUCTION airen A E E E E RE aaa me oe Oem 5 Initializing with your defaults and input TICS c c tocscediesdecdeseieeteeradendeandcceievssnsseasehlennacieds 7 Remark codes for earthquakes and station data ccecccccecsseceseceseceeeeeeseeceaeceeeeeeeenaeecsaeeeeenaes 7 CRUSTAL VELOCITY MODELS rerien ak n sae ihe Cubes Oana 8 M ltiplemodels miinor e REAR EEE E T A E T ARR ARSA 8 Alternate models ienien a a a aa ae 12 Model types homogeneous layer and linear gradient ns nnnsssnssesesseessessressessrssessessrssresseese 13 Homogeneous laveranodelsyi ish irate ccetcheneeahes eden A Saleatd sel boda 13 Linear gradient models using a travel time table eee cecceteceeeeceeeceeeceeaeec
79. VP 2 1 108 124 EPU2 7 71 1 16 0 89 0 29 0 02 0 46 NM 0 612 1 40 7 000M 19 0 42 X PST NC VHZ VP 3 5 171 102 IPD 7 56 1 01 1 16 0 15 0 00 2 29 N 0 829 PVC NC VHZ VP 4 7197 77 IPD1 7 81 1 26 1 39 0 09 0 04 1 14 N 0 341 1 40 6 0 96D ES 2 8 98 2 43 2 47 0 16 0 12 0 46S N 0 272 PPC NC VHZ O1VP 6 9 259 72 IPU 8 13 1 58 1 81 0 24 0 01 2 29 NM 0 860 1 40 16M 07 0 69 X ES 2 9 19 2 64 3 22 0 43 0 15 0 46S NM 0 199 PHP NC VHZ VP 8 0 285 72 EPU2 8 39 1 84 2 03 0 21 0 02 0 46 N 0 080 PSM NC VHZ VP 13 7 331 71 EPU2 9 98 3 43 3 09 0 37 0 03 0 46 N 0 804 1 UNWEIGHTED STATIONS NOT PRINTED 105 The first line of the archive file is the summary header The archive output file is difficult for people to read especially with the word wrapping of the summary line which occurs because the line is long OUTPUT ARCHIVE FILE 199912010008065535 5773120 3123 239 56 8137 2 0 OL 0 0 PM PST PVC PPC PHP PSM BMS NC VVHZ NC VVHZ NC VVHZ NC VVHZ NC VVHZ NC VVHZ NC VVHZ 21069577D 96 10Z 0 EPU2199912010008 IPD0199912010008 IPD1199912010008 IPU0199912010008 EPU2199912010008 EPU2199912010008 771 756 781 813 839 998 0 2 46 0229 4114 1229 2 46 3 46 P 4199912010008 99997826 0 O 0 07 00000 0 oO O 0 898ES 2 12 0 919ES 2 1516 0000 0 oO O 0 0 oO O 0 0 0 0 0 0 0 0 0 0 0 0 3 8669 32821312 58 96MID 21 0 29 0 2112404 0 15 0 3510204 46 9 16 47 7700 46 24 43 69 7204 0 21 O 80 7204 0 37 0
80. a file RIGHT NOW Also read in the stations to use with alternate crustal models The delay file if used must be read in after the station file because the station list must already be in memory The stations in the delay file need not correspond in number or order to those in the station file Note that the old DLY command no longer functions because DEL is a more flexible way of assigning delays Also see the multiple crustal models sections for the meaning of multiple delays See the section on station formats for the format of this file Specify a code selecting which type of delay file to read 1 36 model number to which delays apply specify one delay per station 0 read in the alternate status codes and delays for all models all models on same line minus one read in the alternate crust model code that states which stations use an alternate model Specify the file name with the station delays Example DEL 0 ALL DEL or DEL 1 MOD1 DEL There is no default Say whether the CAL FACTOR data field on the station record read with the STA command contains an attenuation or cal factor see amplitude magnitude section for explanation of these terms T The field is an attenuation setting F The field is a calibration factor Example ATN F Read station attenuations related to gain from a history file RIGHT NOW You must read your station file with the STA command before reading
81. a number related to gain G or to attenuation This assumes the values of D given above and removes them from the definition of CAL CAL thus does not depend on the recording system used Simple algebra yields CAL G 883 4 log CAL 1 35 0 05 x atten 5 The gain expressed as CAL may be specified for all seismometer types including the Wood Anderson and various digital systems It is only for the USGS NCSN analog transmission system and networks using that system that the concept of the attenuator setting makes sense Hypoinverse allows you to specify gain and gain history as attenuation only for the analog system using L4C seismometers and instrument type 1 see below If the CAL value is set to zero no amplitude magnitudes will be computed for this station Amplitude magnitude relation for velocity seismometers The X magnitude formula developed by Jerry Eaton for velocity seismometers is used in Hypoinverse Before the magnitude is calculated the amplitude is converted to an effective Wood Anderson amplitude using the period at which the amplitude is measured and the response curve for the seismograph type The magnitude calculation of course assumes that the location of the maximum amplitude on the seismogram is at the same time for both the velocity seismogram and a Wood Anderson displacement instrument This means the seismogram fits the definition of local magnitude where the W A seismogram has its maximum amplitude The
82. absolute travel time minus distance times REDV The values of reduced travel time read by Hypoinverse are limited to the range 0 to 32 seconds and the user is responsible for choosing a suitable reducing velocity to stay within these limits Using a reducing velocity equal to the halfspace velocity is a good choice The user specifies the amount by which the independent parameter Q is incremented to calculate distance and time for rays of various ray parameter and emergence angle Ray parameter P and emergence angle are functions of Q as follows 2 TAN Q Zy p SIN Va where Zy and Vy are depth and velocity at the hypocenter respectively Q is a better independent parameter than either P or since it gives a greater density of rays for deeper penetrations This also gives the distant travel time points a spacing in distance comparable to nearby points The parameter Q is incremented as follows It takes on the value 0 0 then NQ1 values at increments of DQ1 then NQ2 values at increments of DQ2 The largest value of Q is thus NQ1 DQ1 NQ2 DQ2 and the greatest number of rays maximum value of NQ1 NQ2 is 400 Ray calculation stops when down going rays begin to penetrate the halfspace Travel times appropriate to a refracted ray are used beyond this point Values of DQ1 0 04 NQ1 200 DQ2 0 2 and NQ2 200 are a good first try and generally insure that the entire travel time curve can be adequately defin
83. ach event The terminating record may be completely blank or may contain trial hypocenter data or an ID number Each station may report any or all of 1 P time 2 S time 3 amplitude or 4 coda duration An arrival time will not be recognized and processing will continue if any of the following are true 1 The remark field IP or ES is blank If the S remark is blank but the S time is non blank the reading will be used 2 The station is not in the station file 3 The phase line is incomprehensible or in a bad format 4 The P arrival time and date differs by more than 6 minutes from the first station S times are not checked If you are reading an archive format file and using the earthquake location from the header as a trial location the P phase times are checked against this reference time 5 The station site code is on the list of stations to ignore given by the UNK command Hypoinverse naturally reads the traditional phase format format 1 selected with the COP command which is compatible with HYPO71 phase lines It looks for other data in fields left blank or undefined by HYPO71 Hypoinverse archive output may optionally be read back in as 27 input Another option is the shadow format where every input line phase archive and terminator lines are followed by a line with additional data and which begins with a The input shadow record will be copied to the archive output if shadow records
84. acter fixes the entire hypocenter 63 72 10 110 Optional ID number HYP071 terminator format Start Fortran Col Len Format Data 1 4 4 A4 Must be blank 19 1 11 To fix depth either put a 1 here or make trial depth negative 20 24 5 F5 2 Trial depth 63 72 10 110 Optional ID number 32 Reading earthquakes directly from CUSP MEM files Hypoinverse can locate earthquakes directly from CUSP MEM files There is one MEM file per event whose name contains the CUSP ID number i e X100229 MEM You need not be in a CUSP environment or be on a computer running CUSP but you must be running VMS and define one name before running Hypoinverse on the Menlo Park VMS cluster for example type DEFINE EVENT DDOL PUBI CUSP HYPO EVENT DDL The subroutines to handle the MEM file input and output were written by Alan Walter Hypoinverse locates events given their CUSP ID numbers which are supplied in one of two ways 1 type in the CUSP ID numbers and locate events one at a time or 2 provide a file listing the CUSP ID numbers of the events to locate You may use the latter to locate all the MEM files in a directory by making a directory listing to a file before running Hypoinverse see below You can use the existing location in the MEM file as a trial hypocenter by setting the second argument of the H71 command to 3 If you want to examine the data that Hypoinverse is finding in a MEM file define an output pri
85. agnitude for each station and two amplitude magnitudes for each earthquake The Magnitude choice primary or secondary used for a station depends either on its component code or its seismometer instrument type The choice is made with the XCH command Several or all components could be used for each magnitude One could calculate a local magnitude for all Wood Anderson components and an X magnitude for all vertical components for example Similarly one could use the instrument type to select the two magnitudes The two event magnitudes the medians of a set of station magnitudes are independent but both are drawn from the set of station magnitudes according to the component list or instrument list chosen for each magnitude Choose which if any stations you want to assign to each of the two magnitudes The first way to assign a station to either a primary or secondary magnitude is by selecting the component codes of stations you can choose no stations no event magnitude computed all stations or list up to 10 different 3 letter component codes The XC1 command selects which components to use for the primary amplitude magnitude and the XC2 command selects which components to use for the secondary magnitude The default is to use all components for XMAGI and no stations for XMAG2 A magnitude is calculated for each station but if it is un weighted or if it is not on the list of components for XMAG1 or XMAG2 it is not used in that event ma
86. ance corrections to specific component codes is made with the LAO zero command The distance corrections available are where the item numbers are those you specify in the MAG and LAO commands 1 Eaton 1992 F 0 821 log S 0 00405 S 0 955 for S lt 185 3 km F 2 55 log S for S gt 185 3 F 0 09 sin 0 07 D 25 only if D lt 70 km 2 Bakun and Joiner 1984 F log S 0 003 S 0 7 F 0 3 Richter s 1958 approximation F 0 15 0 8 log S for S lt 200 F 3 38 1 5 log S for S gt 200 4 Nordquist s BSSA 1948 UC Berkeley relation This option computes F gt by interpolation within a table 47 5 P amplitude presently unused Compute F and F from option 1 Then add F as follows F 0 08 0 00942 D for D lt 52 F3 0 41 for 52 lt D lt 115 F 0 812 0 0035 D for 115 lt D lt 280 F3 0 168 D gt 280 Magnitudes XMAGs from electromagnetic velocity seismometers Analog data transmission and recording The method of calculating magnitudes was developed by Jerry Eaton 1969 1970 1992 and implemented in the Hypomag program for the USGS network It is easier to understand the procedure used in Hypoinverse if we discuss his method and the USGS network recording in simple terms Visualize the system from ground motion to seismogram like this with each box having an input and output Seismometer Electronics Recorder motor gain G scaling D constant S
87. ard errors calculated by this program has a 95 chance of containing the true hypocenter Hypoinverse also calculates the azimuths and dips of the principal axes of the error ellipsoid Note that the calculated error ellipse scales with the estimated error RDERR and the earthquake s RMS The error ellipse is thus also only an estimate though it contains all the geometry of the stations and ray paths A true calculation of the error ellipse would require knowing the uncertainties in every variable including the crustal model which is nearly impossible In addition the error ellipse is based on the partial derivatives at the hypocenter and knows nothing about real or modeled discontinuities such as layer boundaries or the Earth s surface The vertical error ERZ and horizontal error ERH are simplified errors derived from the lengths and directions of the principal axes of the error ellipsoid Each of the three principal axes whose lengths are the standard errors are projected onto a vertical line through the hypocenter and the largest value is ERZ ERH is simply the length of the longest of the principal axes when viewed from above and projected onto a horizontal plane Appendix 5 Generating travel time tables for linear gradient crustal models with program TTGEN Use of a travel time table Hypoinverse reads a travel time table which is generated independently of the location process by the program TTGEN Hypoinverse calculates travel t
88. are also selected for output The first line of each event summary headers with event information in archive files may have from 1 to 4 shadow lines The archive format comes in two flavors older pre Y2000 files with 2 digit years and newer Y2000 files with 4 digit years including the century CUSP input is not an ASCII format but invokes a series of database calls that read a binary MEM file CUSP input and output is only supported on VMS platforms One can choose the input format explicitly with the COP command or let Hypoinverse determine the format with the FIL command To determine the format set the input file name with the PHS command which must be an ASCII file not CUSP binary Then type FIL and Hypoinverse will tell you what kind of file it thinks it is and set the input format accordingly It will also invoke the Y2000 formats for all input and output if it finds a Y2000 format input archive file FIL will also set the appropriate archive output format in place of the CAR command For convenience FIL also detects the four different kinds of summary output formats which are illegal as input formats because they do not have phase times Some formats have shadow records which contain additional data that is not processed by Hypoinverse but just passed through Shadow records begin with a character and follow each header phase or terminator line In the archive formats the summary header lines may have from 1 to 4 s
89. ation factor 81 82 2 A2 2 letter location code site component extension Station delay file The Hypoinverse station file format holds two delays for models 1 and 2 the HYPO71 format holds one delay If you need more than the 1 or 2 delays for use with multiple crustal models see preceding section you must read station delays from a separate delay file using the DEL command Each line of the delay file has the station code and one or more delays The alternate model code for stations may either be in the station file or a file only containing codes for the alternate stations A station becomes an alternate if so designated in either file Only the station site and network codes are read in the delay files All components at a site are assumed to have the same delay and the delay will be assigned to all components at that site One could get around this assumption for a borehole station for example by using a different site code When matching the station codes in the station location file and the station delay file you can use less than the full 5 letter site code and full 2 letter net code See the LET command There are three types of delay file read with the DEL command The file type is determined by the first argument of the DEL command 1 reads a file with alternate codes designating which 19 stations use the alternate model set with the ALT commands 0 reads a file with delays for all models and the positive model number 1
90. ayer is the underlying half space Linear gradient models using a travel time table An alternative and more computationally efficient way to compute travel times is by interpolation within a table generated prior to running Hypoinverse Use of a table permits more complex travel time calculations such as linear velocity gradients within layers and capacity for a buried low velocity zone The travel time table must be calculated and written to a file using the program TTGEN appendix 5 prior to locating earthquakes A travel time table may be calculated for a velocity depth function consisting of from 2 to 20 points at which the velocity and depth are specified The velocity is then assumed to be linear 13 between points i e with a uniform gradient within layers Several restrictions apply to the possible models 1 No two velocity depth points may be at the same depth a sharp velocity discontinuity is not allowed Discontinuities may be modeled using thin layers with high gradients but the transition layer should be thick enough that one or two rays used to generate the travel time table will bottom within the layer and define a reverse branch of the travel time curve 2 The depth of the first point must be 0 0 and other points must be given in increasing order of depth 3 The last deepest point sets the velocity of the homogeneous half space assumed to underlie the model 4 The half space velocity must be the greatest of any speci
91. be recognized as unknown by a match of the first 4 letters of their 5 letter code but all 5 letters will be used when the data is written This is very useful for traces such as time codes where the data is to be archived but no true station or measured phase exists Example UNK 0 the default or UNK 3 ABCM IRIG WWVB Say whether to keep phase information from unrecognized stations and output them to the archive file Stations mentioned in the UNK command are also under control of the KEP command and can be saved or not Specify T to write unknown stations to the archive file or F to eliminate them Example KEP T Each event may have a 1 letter label code that marks its processing status or labels the run to distinguish it from other runs This is currently in col 153 col 163 of the Y2000 version of the Hypoinverse summary format This label can either be passed through from the input archive file or a new label assigned to all events in this location run Supply a 1 letter run label code blank is the default Set a flag T to pass the old 1 letter run label code from archive input to output files or F to insert the new label code into the output files Examples LAB T the default or LAB A F OUTPUT FILES PRT SUM Set the print output filename Supply the filename Use NONE or none to omit a printout file Example PRT NONE the default or PRT PRTFIL PRT Se
92. by NOD commands There are no parameters Example SNO Designate one crustal model as an alternate to another and use the alternate model for stations so indicated in the station file Ifa station is not specially marked it will use the primary models if it is marked as alternate it will use the alternate model if one is defined Alternate models allow using two different models with two different sets of stations for the same earthquake and may be used with or without multiple models for different regions Any number of models may have alternates An alternate model must be of the same type layer or gradient as the primary model Specify the primary model number to have an alternate Specify its alternate model number Example there are no defaults ALT 2 3 andALT 4 5 In this example an epicenter falling in the region defined by NOD commands for model 2 will use model 2 for normal stations and model 3 for stations marked as alternates An epicenter in region 4 will use model 4 for the normal stations and model 5 for the same set of alternate stations For example NCSN uses this feature for two model regions that straddle the San Andreas Fault The stations to the east of the fault are designated as alternate These alternate stations use the alternate models from every region where an alternate model is defined namely 3 and 5 Because the fault is a velocity discontinuity two different models are used for earthquakes on
93. cate that it applies into the future A station may have at most 6 corrections each for XMAG and FMAG If the magnitude correction for a station never changed only one correction is needed and no expiration date is needed Use the FMC command to read the FMAG correction file and the XMC command to read the XMAG correction file If there is more than one correction per station the phase data file must be in chronological order to insure getting correct magnitude corrections Normally the magnitude weight that governs whether the station magnitude is used in the event magnitude is taken from the station file Setting an optional flag in the FMC or XMC commands to false gives all stations not in the correction file a zero weight Setting that flag also gives all stations in the correction file full weight regardless of the weight in the station file Thus it is possible to ignore magnitudes from stations that don t have well determined corrections Most networks set the flag in the FMC or XMC commands to true and leave the station weights in the station blank to give full weight You can use magnitude corrections to give zero weight to magnitudes from individual stations when computing the event magnitude Just add 5 0 to the actual correction to suppress that station s magnitude contribution in the event magnitude This works for both amplitude XMC and coda FMC magnitudes If you want to suppress the gain correction coda magnitudes only fo
94. cd ccicad ccsaseescaddaecdases acvdncdiesdeasavadsers 53 Table 2 Seismometer instrument types ssssessssseessessresressetsersstesresrtsstessesstsstessessessresseeseesressee 54 Table 3 Gain factors of different seismic systems cccceceseesseceteceeeceeeeeeseecaeceeeeeeeeeaeeesaeees 55 Figure 6 The distance weighting function ssessessssssessessresressesstesrosseesesstesstesteseressessessressesse 60 Figure 7 The residual weighting function sessssesessssesseserssressessresessessreseesseeseesresseesresersseesee 61 Figure 8 A sample plot of the reduced travel time curve for the above input file for a source at 2 km Spats sc ees vatze Aspens dived ens Gavia ta ssca Ais eG E Ses as As Bc as A eee a A 117 INTRODUCTION Hypoinverse is a computer program that processes files of seismic station data for an earthquake like p wave arrival times and seismogram amplitudes and durations into earthquake locations and magnitudes It is one of a long line of similar USGS programs including HYPOLAYR Eaton 1969 HYPO71 Lee and Lahr 1972 and HYPOELLIPSE Lahr 1980 If you are new to Hypoinverse you may want to start by glancing at the section SOME SIMPLE COMMAND SEQUENCES to get a feel of some simpler sessions This document is essentially an advanced user s guide and reading it sequentially will probably plow the reader into more detail than he she needs Every user must have a crust model station list and p
95. ce In this document commands and certain other computer lines are given in the equally spaced Courier font This makes characters such as blanks and periods easier to spot and allows columns to line up as they do in computer files containing data where the column is important Hypoinverse is a complicated program with many features and options Many of these advanced or seldom used features are documented here but are more detailed than a typical user needs to read about when first starting with the program I have put some of this material in smaller type so that a first time user can concentrate on the more important information Hypoinverse commands that take parameters are in free format meaning that values are separated by spaces commas or tabs Other rules apply to free format regarding null fields in which values are not redefined but the existing value is kept Thus the command DUR 28 oO DA sets the second and sixth values but keeps the other values including the 7 and later as they exist Array sizes The current maxima array limits of various data types are Number of stations in station file 4000 Number of stations phase lines per event 1000 Number of phases P or S per event 820 Number of homogeneous layer crustal models 36 Number of linear gradient crustal models 36 Number of crustal models of any type 36 Number of layers per crustal model 20 Number of nodes for model regionalization 124 Chara
96. ching station entries regardless of location code of the two or more otherwise identical entries This means that the first location code in the station file will be used on all output archive and print files regardless of which location code was input This can generate errors It is thus best to fill only the station code fields that are actually being compared with the lengths given in the LET command 15 Station delays will be applied correctly to all stations with the correct site and net codes regardless of component or location codes The only restriction on naming stations is that the first four characters of the 5 letter site code must not be all blank This is to recognize the terminator record at the end of the event It is suggested that the component code be used for the component rather than combining the component with the site code as done previously This to take advantage of certain special uses of the component code such as applying a delay to all components at a site and using station magnitudes only from certain components to average in an event magnitude Hypoinverse requires a separate station input line for each component for each station This makes handling the database much simpler even though components at the same site generally share common information such as location or delay The station file is read in and each line is treated independently When a phase line is read Hypoinverse searches the station list from
97. cseesecseeseceeeeceaeeneeaeeaeaes 76 MAGNITUDES erlere e a cade aa ccuta coached tite ss atm deena haieas catia cama 77 CODA DURATION MAGNITUDES cimenta aata A E AA EER AEE roaten 78 AMPLITUDE MAGNITUDES peni r ia A EEE O Gav ea apes eas 81 MISCELLANEOUS COMMANDS aiclesss caccavssocentsseoaavertescastsauseicusioicers aocacds aero aeoaa ede 84 WEIGHTING OF ARRIVAL TIMES rrencia ae aia Meats aes 84 ITERATION AND CONVERGENCE PARAMETERS ceecceeccecscceessseeceeeenseeecessseeeeeeesaas 86 OUTPUT FILE FORMATS aten rea a E wana cy adpo E ae dg nao aed uae as Scam 86 Print OUMU oats Ges Rah tk Salers ai tie eek hat ia lee ach es tale ata ket acorns Na a els ht ok hil te 86 RADIO 0a WG secant eee ie ncaa osha ca loai din ale d aaa dl Gale eS matinee Medel A tenes ye ddes ca 87 Earthquake iteration output siei ierit itan aiae E EAA R ARE EEES 88 Fi al location data sea n a aa a e a eaei a a aeh 88 STATION LEST Aas pee hstracat cosy ecteans cence a e a aa a aan anes aE iaa aal 91 Ma gnit d data output fees erate teeta Shee aleaad ahaa ts tate aks Nt sa calaulas Dae toi fas lots tel oi eal onied a 92 Quality codes Hypo71 summary output ONLY cece eecceesceesceceteceeeeeeeeeeseeceaeceeeeeeeeeseeesaeenes 94 OLD PRE Y 2000 OUTPUT FORMAT S sisisvsssaccstinsiausietacendane lacusanrdaanuietaastatiacsdaniuozeauiande 95 2000 OUTPUT FORMA TS oi as E eaias Ge bess ae wee ee 98 SAMPLE INPUT AND OUTPUT FILES cjsccicpanict caveeideacantecicsacenitataaacetortea
98. cters in input shadow records 103 Summary shadow records in input archive format 4 Number of fictitious station codes UNK command 10 Number of unknown stations to copy to output 40 To find out the maximum sizes of the various arrays the program uses for stations phases etc use the MAX command If there are too many stations phase cards in an event and archive output is being generated the excess phase data is copied to the output file without any processing and with the calculated fields left blank These maximum array values are set by parameter statements in the source code in the common block include file and are easy to change for situations where more or less space is needed Initializing with your defaults and input files If a file called hypinst HYPINST with no extension in VMS is in your current directory the file is read on startup as a command file by Hypoinverse It may be used to set your own default values read station or crust model files that you always use etc You may then enter commands directly or transfer control to other command files to do specific jobs It is not necessary and may be a bad idea for a new user to use a hypinst file I never use one because it may execute at a later date without me realizing it Hypoinverse normally just issues a command prompt when there is no hypinst file Any command file may be run by typing filename Hypoinverse has an INI command to initialize the progra
99. date and time For the second and later events pressing RETURN to a prompt for year month day hour or minute gets the default which is that of the prior event 2 Prompts will be made for each station on the prompt list The P time is entered first followed by S time duration and amplitude if you requested prompts for them If you chose to enter full remarks and weights prompts for these will follow those for each P and S time 3 When the pre set station list is complete you will be asked if you want to enter another P or S time If yes you must enter the full station name as it exists in the station file time and full remark Pressing return and taking the default value of 0 skips that item Basic remarks such as P_0 and S_0 are provided as defaults 4 When the event is complete you will be asked whether you want to stop entering data and return to Hypoinverse command level If you answer f or false you will be asked for the date and time of another event which you must complete IMPORTANT NOTES 1 The case upper or lower of station names must agree with that in your station files 2 All numeric data may be entered in free format whole numbers may omit a decimal point leading zeros are optional etc Alphanumeric input such as station names and remarks however is used exactly as entered 3 Once you press RETURN the only way to correct mistakes is by leaving the Hypoinverse command level editing the phase file
100. e 46 CAL is a dimensionless scaling factor assigned to the station For example a low gain Wood Anderson running at a gain of 700 would use a CAL of 700 2080 0 3365 to correct the amplitude to the 2080 displacement gain of a normal Wood Anderson For a normal Wood Anderson instrument CAL should be 1 0 F s and F2 d are the log Ao distance correction terms that will be discussed below They can be a function of epicentral distance d slant distance s or both XCOR comp is the correction made globally to all components with a given component code See the XCM command XCORstz is the individual station correction This is specified for each site and component at that site Station corrections may be supplied in the station file or a magnitude correction history file may be specified with the XMC command Amplitude magnitude distance corrections The function F s F2 d where d is epicentral distance and s is the slant distance s d z represents the log Ao distance correction Several choices of this function are available in Hypoinverse Although all of these numerical functions are approximations to the same relationship in the earth often a seismic network uses a variety because different studies may have been used to determine different magnitude scales for different stations The default relationship choice is specified by the MAG command for components without specific distance corrections The assignment of dist
101. e correction Leave the expiration of the last magnitude correction blank to indicate no expiration Minutes if supplied is ignored 32 5 F5 2 1X Second magnitude correction 38 2 12 Second expiration year 40 6 312 3X Second expiration date M D H Format repeats 6 F5 2 1X 12 312 3X Duration magnitude correction file format Y2000 century compatible Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 Station net code 9 2 A2 Station location code 11 3 A3 1X Station component code 15 5 F5 2 1X First magnitude correction 21 4 14 Year of expiration of first magnitude correction 25 6 3I12 1X M D H expiration date and hour of first magnitude correction Leave the expiration of the last magnitude correction blank to indicate no expiration 32 5 F5 2 1X Second magnitude correction 38 4 14 Second expiration year 42 6 3I12 1X Second expiration date M D H Format repeats 6 F5 2 1X 14 312 1X The amplitude magnitude correction file may also contain the instrument type USGS analog short period or Wood Anderson for example Ifthe instrument type field is left blank the value set in the station file is used If the field in the correction file is non blank it over writes the station file value Use the XMC command to read the XMAG correction file The choice of which Y2000 format is used is governed by the 200 command For the instrument type codes 25 see the seismomet
102. e earthquake could then have both a duration magnitude and a lapse time magnitude However you can t calculate a lapse time magnitude for the EHZ and ELZ components and an F P duration magnitude for ELZ components because you may only compute one magnitude type for any individual component The two event magnitudes the medians of a set of station magnitudes are independent and each may be chosen from duration relation 1 duration relation 2 or the lapse time relation If both event magnitudes are lapse time for example the one set of lapse time constants are used to calculate magnitudes at all stations The two event magnitudes are medians of different sets of 41 stations defined by their components For example you could have FMAGI determined by all VHZ and VLZ components and FMAG2 determined by only the VLZ components This is probably confusing so consider the process in two steps 1 first choose the one or two magnitude types from the three available duration 1 duration 2 or lapse time you want to calculate for each event 2 Then choose which stations using their component codes you want to assign to each of the two magnitudes Remember that only one magnitude can be computed for each station but that two event magnitudes can be computed Step 1 First choose the two types of magnitude to be used from the three relationships available Define the parameters for the first duration relationship with the DUR command the
103. e may contain any set of stations delays for stations already in memory will be used and those for a station not in the station file will be ignored The delay file may thus be a master list containing all known delays and the stations 20 may be in any order Delays for a station in the station file but not the delay file will have delays left at 0 or as read in the station file for models and 2 Station attenuation and calibration factor files The station gain curve is specified by a standard response curve for that instrument type and by a calibration factor that fixes the level of the curve at a particular frequency See the section on amplitude magnitudes for a fuller discussion In the following discussion attenuation applies primarily to analog stations with a preamplifier and VCO located at the seismometer These are typical NCSN stations but are widely copied by other networks The pre amplifier located at the seismometer has a gain of 90db from which an attenuator lowers the net gain in increments of 6 db You can thus use either the calibration factor or the attenuator setting to specify gain There are five options for specifying station gain information for use in calculating amplitude magnitudes or correcting coda magnitudes 1 put the calibration factor in the station file to use for the whole location run 2 specify the pre amplifier attenuation in place of the station calibration factor on the station line see the ATN c
104. e set with the RMS command and are with sample values ITRRES the iteration number at which residual weighting is to begin 4 RMSCUT 0 10 sec RMSWI 1 0 and RMSW2 3 0 The residuals at which the taper begins and ends scale with RMS when it is larger than RMSCUT and are fixed when it is less than RMSCUT see figure 7 When RMS is larger than RMSCUT the residual at which the taper begins and ends scales with RMS and is full weight closer than RMS RMSW1 and no weight farther than RMS RMSW2 When RMS is less than RMSCUT the residual at which the taper begins and ends is fixed and is full weight closer than RMSCUT RMSW1 and no weight farther than RMSCUT RMSW2 Thus most stations will be fully weighted when RMS is large but as iteration and convergence proceed and RMS becomes smaller the weights of large residuals will lower RMSCUT prevents an inward spiral in which large residuals are discarded which then lowers RMS and which results in more large residuals being discarded etc There are two distinct values of RMS used by the program RMSWT is the RMS residual computed before residual weights are applied and on which residual weights are based The output variable labeled RMS is the value 60 of RMS computed after residual weights are applied and is used for convergence tests see below and as a final quality measure Residual weighting is re applied on every iteration starting with the iteration where it is to begin This allows
105. e the components F Use the instrument type from the station line or the amplitude magnitude correction file to choose which of the two amplitude magnitudes to apply the station to for the event Use the XTY command to choose the instrument types but you will also use the XC1 and XC2 commands to set the label codes for the two amplitude magnitudes Example XCH T Use components to choose amplitude magnitudes Define the station components to include in the primary amplitude magnitude XMAGI for an event A maximum of 10 different components may be used The components select which stations are used for the earthquake magnitude Also set the 1 letter label 81 XC2 XTY for XMAGI1 You would only be using this command except for setting the magnitude label if the XCH argument is T Supply the 1 letter label code for XMAGI Supply NCPX1 the number of components to use in calculating XMAGI Set NCPX1 1 to use all components in XMAGI1 Set NCPX1 0 to use no stations in XMAGI Set NCPX1 the number of components to use for XMAGI 1 10 Supply the NCPX1 1 letter component codes to use in XMAGI Examples XC1 1 all components used for XMAGI1 the default or XC1 X 4 V E N Z Use 4 components for X magnitude Here only the first of the 3 letters is used Define the station components to include in the secondary amplitude magnitude XMAG2 for an event A maximum of 10 different components may be
106. e the station importance bii will also be large Thus a large leverage through the partial derivative matrix A of a particular station on the solution is equivalent to a large station importance This can be seen intuitively from the relation A eV U e S mxn nxn mxn nxn where the matrices are as defined in the inversion section When the ith row of A is large corresponding to the ith station the ith row of this equation and hence of U will be large The ith diagonal element si of UeU will also be large An illustration of the relation between importance and partial derivatives is the fact that an S reading has a greater importance than a P reading from the same station The partial derivatives travel time space coordinate are larger for S arrivals than P arrivals at the same station by the factor Ts Tp and this means that rows of the U matrix and consequently the importance will be larger for the S arrivals This has an important consequence for assigning weights to arrivals When an S arrival cannot be read to the same precision as a P arrival it should be given less weight to compensate for its intrinsically larger importance The importance is a measure of the redundancy in the data and for example is individually small in distances and azimuths where there are many stations This can be seen from the following argument The inversion process for the over determined earthquake problem extracts n linearly independent combinati
107. e the transition smooth 2 This point is just outside the inner circles for two nodes both assigned to model 2 The weight for each node might be 90 to total 1 80 Because this number is greater than 1 0 it is normalized to 1 0 and model 2 gets full weight 50 from each node 3 This is near the outer boundary of two outer circles assigned to models 2 and 3 It might get 15 weight from each of models 2 and 3 to total 0 30 The remaining weight to reach 1 0 would be 70 and would be from the default model 1 4 This point is within the outer regions of 3 nodes and might get 30 40 weight from each If the cosine taper value is 30 the total weight from all 3 nodes is 0 90 Then model 1 contributes 10 model 2 60 and model 3 30 If the cosine taper value from each node is 40 the total is more than 1 0 and the normalized contributions are 67 from model 2 ad 33 from model 3 This example does not have a point where the outer regions from 4 different nodes all meet but if it did only the first three encountered in the list of nodes would be used to determine weights 10 Some simple guidelines in defining the nodes and their surrounding circles will help avoid location problems and sharp transitions between models that can form artificial discontinuities 1 Never have the inner circles for different models overlap This will make a sharp discontinuity between models For a given epicenter location each node is tested in the order it
108. e thin layer with a high gradient that represents the moho Because the final travel time curve is the earliest lowest of each of the branches the reverse branch is not present in the travel time table and the curve behaves like that of a refracting surface The thin high gradient layer also smoothes out the depths of hypocenters near it because there is no sudden discontinuity Outputs of TTGEN At each depth point specified in the grid a plot file is created of travel times and distance The actual travel time used in Hypoinverse will be the first arrival from each branch All reversed branches are indicated The condensed travel time table contains all the information necessary to identify itself and be used by Hypoinverse when read with the CRT command The format of the table is transparent to the user but is given below for completeness The print output of TTGEN contains one table for each depth grid point One line is printed for each ray calculation until the deepening rays reach the halfspace The tabulated data is as follows J The ray index an integer 1 2 3 Also used to reference rays defining the endpoints of a shadow zone or reversed branches Q The user defined parameter variable Equal increments of Q are designed to give a greater density of deeper rays where they are needed to define the travel time curve EM ANG Emergence angle of ray at the source measured in degrees from zenith 117 P Ray parameter in sec km
109. eaeceseeeeeeenseeeaeens 13 THE STATION LIST AND USE OF STATION DELAYS GAIN HISTORY AND MAGNITUDE CORRECTION Se nasennie cease feltanain a n R a a aa 14 Station locaton MeTs annae oee a i a a eon cones oo san Taa SEEE ASE EN E AESA 14 Station name codes and file formats ssessessessesseseesesseseesseseesesseseesesstssessessestesesseseesesseesessesee 15 Specifying instrument types and calibration factors ss sessssessesessseessessessressessresresseeseesressee 16 ETR el 2 WS DAE AE ERT E A E E 19 Station attenuation and calibration factor files s sessseessesseeseesseesesessseeseseesseessessrssressessessees 21 Station magnitude correction files c isssdcccessasededcstaciueddecerioied iceauaviauctesbocdoees cteasivietdesincge estes 23 Other staton Comments ie oeni ee a a a a a e e ae eaa EAR Te 26 PHASE DATA INPUT FORMATS Crepe u n A R R R R atetiate 27 PATCHY ES OULU ZON E 00e 1 PE AEE E A E 28 Weight codes name codes and other parameters common to all phase input formats 29 Archive files read as phase IM PUbiisiecs ides ssisstvdcsvenegeiadeci nati aepdoren tse eriela ava eeaanaead 29 The terminator line and trial locations all ASCII formats cccccccessccsscecesseecsssceesseeeeaes 31 Reading earthquakes directly from CUSP MEM files ccccceccceeseeeseeeseeceeceeeeeeeeenseeeeaeens 33 Input of new phase data from the keyboard cee cecceeseesseceseceeeceeceeeeeeeseeceaeceeeseaeenseeeaeees 34
110. ections to Eaton s amplitude magnitude The 2 indicates there are 2 pairs of component codes and their corrections to follow XCM 2 VHZ 33 VIZ 20 The weighting of duration and amplitude magnitudes The summary duration or amplitude magnitude for an event as reported in the event header and summary line is the weighted median of the station magnitudes The weighted median is the magnitude value for which half of the total weights are higher and half are lower Amplitude and duration magnitudes are calculated and reported separately and are never mixed Separate amplitude and duration weights may be specified on both the station and phase lines The weight used is the product of the two The weight codes are the same as the default weights for individual P and S phase times 0 or blank full weight 1 3 4 weight 2 1 2 weight 3 1 4 weight 4 9 no weight Specifying a magnitude correction larger than 2 5 will also give that station zero weight the correction used is the supplied value minus 5 0 Another way to avoid using a whole class of stations in the event magnitude is to omit their component code from those selected for use by the FC1 FC2 XC1 and XC2 commands The outputs of Hypoinverse report the weighted median magnitude the total of all weights essentially the number of magnitudes used in the average and the mean absolute difference MAD error for all four of the amplitude and duration magnitudes PREFERRED EAR
111. ed by less than 400 rays The grid points in distance and depth at which travel times are calculated for output to the final table are determined by eight parameters similar in concept to the Q parameters described above Travel times are calculated at depths of 0 0 then NZ1 values at increments of DZ1 then NZ2 values at increments of DZ2 This permits a fine grid spacing for shallow depths and a coarse spacing at greater depths where the travel time curve will be smoother Similarly travel times are calculated at distances of 0 0 then ND1 values at increments of DD1 then ND2 values at increments of DD2 Presently the maximum value of NZ1 NZ2 is 27 and ND1 ND2 may be as large as 41 These maximum array sizes can easily be increased by increasing the array dimensions and then recompiling Velocity model input format file ttmod Start Fortran Col Len Format Data Line 1 1 8 A8 Printed output filename 9 8 A8 Travel time table output filename The name must end in a period The extension crt will be added to indicate it should be read with the Hypoinverse CRT command 115 17 6 F6 1 REDV one divided by the reducing velocity used to condense the travel time tables Line 2 Parameters for inerementing the independent parameter Q governing ray spacing see text 1 5 F5 2 DQ1 6 5 I5 NQ1 11 5 F5 2 DQ2 16 5 I5 NQ2 Line 3 Parameters for inerementing the grid spacing in depth see text 1 5 F5 2 DZ1 6
112. ent code R The 1 letter station remark from the station line NCSN uses the first letter of the site code which designates the region LAT Degrees and minutes north LON Degrees and minutes west PDLY1 P delay for the first model Model code appears above the column A An A here designates stations for use with alternate models FCOR Duration magnitude correction FWT Duration magnitude weight factor for the station FMC EXPIRE Expiration date of the duration magnitude correction XCOR Amplitude magnitude correction XWT Amplitude magnitude weight factor PSWT Station weight factor for P and S phase times CAL Calibration factor related to gain CAL EXPIR Expiration date Y M D H of CAL factor 0 means no expiration date PER Default period at which peak amplitudes are measured TYP Station seismometer instrument or response type code PDLY2 P delay for the 2nd model Model code appears above the column Part two of the station table if requested lists the 4 letter station codes components and delays for all of the other crustal models Next the data listed for each crustal model includes the velocity at the top and depth to the top of each layer If regional crustal models are in use a listing of all geographic nodes assigned to the model is given Each node is indicated by its number center latitude and longitude and the inner and outer radii of the transition zone surrounding the node The transition zone is the
113. er instrument types part of the section on amplitude magnitudes Corrections can also suppress the magnitude weight of individual stations by adding 5 0 see text above Amplitude magnitude correction file format pre Y2000 Start Fortran 12 Year of expiration of first magnitude correction 312 3X M D H expiration date and hour of first magnitude correction Leave the expiration of the last magnitude correction blank to indicate no expiration Minutes if supplied is ignored 34 F5 2 1X Second magnitude correction 40 12 Second expiration year 42 6 312 3X Second expiration date M D H Format repeats 6 F5 2 1X 12 312 3X Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 2X Station net code 11 3 A3 1X Station component code 15 1 A1 1X Instrument type code Leave blank to use type from station file 17 5 F5 2 1X First magnitude correction 2 6 No Amplitude magnitude correction file format Y2000 century compatible Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 Station net code 9 2 A2 Station location code 11 3 A3 1X Station component code 15 1 A1 1X Instrument type code Leave blank to use type from station file 17 5 F5 2 1X First magnitude correction 23 4 14 Year of expiration of first magnitude correction 27 6 3I12 1X M D H expiration date and hour of first magnitude correction Leave the expiration of the last magnitude correction blank to indicate no
114. er of free hypocenter parameters solved for and adjusted Normally 4 but will be 3 on the first iteration since depth is held fixed If N is less than 4 it is because the solution is poorly constrained and one or more eigenvalues are less than EIGTOL see the DAM command MOD The primary largest weight crustal model code for this epicenter Next you will see the error ellipse data For each of the three principal errors that form the major axes of the error ellipse SERR is the one standard deviation error in km AZ is its azimuth in degrees east of north and DIP is its dip in degrees Final location data The final hypocenter data is below a printed row of dashes YEAR MODA Date ORIGIN Hour minute second LAT N Degrees and minutes LON W Degrees and minutes DEPTH In km RMS The root mean square travel time residual that uses all weights ERH The horizontal location error defined as the length of the largest projection of the three principal errors on a horizontal plane The principal errors are the major axes of the error ellipsoid and are mutually perpendicular ERH thus approximates the major axis of the epicenter s error ellipse 88 ERZ XMAG FMAG PMAG NSTA NPHS DMIN MODEL GAP ITR NFM NWR NWS NVR REMARKS AVH N XMG The depth error defined as the largest projection of the three principal errors on a vertical line Primary weighted median amplitude magnitude KMAGI1 Primary weighted
115. f different crustal models Travel time residual The residual may be flagged in the following column by an X if that station had a time but no assigned weight or an if the residual is larger than 0 50 sec Actual normalized weight used for this arrival including assigned weight distance weight residual weight station weight and global S weight S phases are flagged with an S following the weight The 1 letter data source code The source usually refers to the contributing network or processing system The 1 letter seismogram remark carried from the phase data The information or importance contribution of this arrival to the solution The total importance of all stations equals the number of unknowns usually 4 The calibration factor used for magnitude calculation Coda duration in seconds Duration magnitude weight code 0 4 Duration magnitude for this station The magnitude type code assigned this component by the FC1 or FC2 commands Peak to peak amplitude of the maximum part of the seismogram Amplitude units code derived from the units code on the input archive line M millimeters C digital counts D HVO digital counts Period in sec where amplitude was measured 91 W Amplitude magnitude weight code 0 9 XMAG Amplitude magnitude for this station T The magnitude type code assigned this component by the XC1 or XC2 commands Magnitude data output The magnitude data file contains some of the data i
116. fied to insure that rays can be refracted along the top of the half space 5 One buried low velocity zone is permitted in each model i e velocity may not decrease with depth except for one group of adjacent layers 6 Assigning the same velocity to two adjacent points may specify homogeneous layers Use the CRT command to specify the model number and the file name containing the travel time table to be assigned to that model The CRT command also reads the model into memory The first 3 letters of the model name as originally assigned before running TTGEN are used as the model code for labeling output For example CRT 1 MODEL1 CRT Reading crust model files is more efficient if they are read in binary instead of ASCII You may create a binary crust model file after reading in all crust model data including station delays and the multiple model parameters Use the WCR command to write a snapshot of the crust model arrays to a binary file Read the file back in with the RCR command The RCR command replaces the CRH CRT MUL ALT and NOD commands Tests show that binary reads using the RCR command are at least five times faster than equivalent ASCII reads and the speedup increases with more models THE STATION LIST AND USE OF STATION DELAYS GAIN HISTORY AND MAGNITUDE CORRECTIONS Station location file Specify the file containing names coordinates and other station data using the STA command For example STA 1984 STA The sta
117. file also controls whether S times amplitudes and durations are routinely requested for certain stations in addition to the P times that are requested for all stations One option is to not enter remarks such as IP and weights 0 4 for phases If you select this option all phases get full weight and a simple remark P or S remarks are 4 letters and include blanks such as IPU0 ES 2 or P__ where _ is blank Any numerical P or S time amplitude or duration assigned a zero value is automatically given zero weight weight code 4 Ifa P time is in fact 0 00 enter 0 01 instead Because INP writes a traditional format phase file a header line with a location is not required and a century is not possible Years must therefore be 2 digits When locating this file you will specify the century with the 200 command that will be used for all events in the file Station prompting file format Start Fortran Col Len Format Data 1 5 A5 1X Station site code A P time will be requested for all stations 34 7 2 A2 1X Station net code 10 1 A1 1 letter station component code 11 3 A3 1X 3 letter station component code 15 1 A1 If non blank prompt for an S time for this station 16 1 A1 If non blank prompt for an amplitude for this station 17 1 A1 If non blank prompt for a duration for this station The following prompts and entries are made for each earthquake event you wish to input 1 Enter the event
118. fix hypocenters the indicates that the remaining arguments if any are left unchanged and that a comment can follow CON 0 Set ITRLIM to 0 DIS 0 7 Begin distance weighting immediately RMS 0 Begin residual weighting immediately COP Sf Read shadow format with a summary header HIL 1 3 sl af Get the trial hypocenter from the header Keeping or eliminating poor earthquakes Most networks will encounter some poorly recorded earthquakes that do not have enough data to obtain a solution You may want Hypoinverse to pass all bad events through to output especially if the output archive file is your permanent catalog record Or you may want to screen out the poor events with fewer events output than are input Two commands control this behavior It is hard to predict how your earthquake will fare in Hypoinverse when you have few readings Four good P arrivals can solve for the four unknowns of the earthquake problem with an exact solution under favorable conditions In some cases 4 good P or S readings can produce a solution but of course there can be ambiguities in cases like good P and S from only 2 stations If there are less than 4 weighted readings Hypoinverse may attempt a solution with depth held fixed Using a trial location may help in cases with few readings The MIN command sets the minimum number of weighted phases P or S phases given non zero weights when they are input to require before a solution is written
119. formation Hypoinverse passes this code through Most common P amp S data source code Most common duration data source code Most common amplitude data source code Primary coda duration magnitude type code Number of valid P amp S readings assigned weight gt 0 Primary amplitude magnitude type code External magnitude label or type code Hypoinverse pass thru External magnitude Total of the external magnitude weights number of readings 95 154 WwWA ww wow A1 F3 2 F3 1 110 A1 F3 2 F3 1 A1 F3 2 F3 1 A1 A1 Alternate amplitude magnitude label or type code Alternate amplitude magnitude Total of the alternate amplitude mag weights Event identification number Preferred magnitude label code chosen from those available Preferred magnitude chosen by the Hypoinverse PRE command Total of the preferred mag weights number of readings Alternate coda duration magnitude label or type code Alternate coda duration magnitude Total of the alternate coda duration magnitude weights number of readings Version of the information i e the stage of processing This can either be passed through or assigned by Hypoinverse with the LAB command Version of last human review Hypoinverse passes this through Pre Y2000 Hypo71 summary output format Start Col m 2a AanNnNnAn RAWA aN RARAWA en lt Fortran Format 312 1X 212 F6 2 F3 0 A1 F5 2 F4 0 A1 F5 2
120. formation if specified as a history file is stored as CAL factor history 0 448 HS1 ie Nanometrics 8 0 Guralp ie broadband Reftek 24 0 Streckheisen STS 1 ie BDSN broadband NIN nn 15 0 Streckheisen STS 2 ie BDSN broadband Types 1 and 3 are for the same seismometer and have the same response In this case Hypoinverse also uses the instrument type to decide whether a station s gain history is in the attenuation history file type 1 or a calibration history file types 3 7 When processing encounters an expired gain value Hypoinverse then knows which of the two files to retrieve the new gain value from by checking the instrument type For types and 3 7 Hypoinverse uses interpolation within a digitized response curve to find R f This is the response curve of an L4C 1 sec seismometer with analog recording relative to the Wood Anderson The response R f approximates that of all velocity sensors in the frequency band where the response is linear i e 3 10 hz For instrument type 2 in the period range 0 1 to 1 9 second log R is approximately a linear function of the logarithm of period for type 2 log 1 R 0 41 0 56 log 0 2 PER 54 CAL factors for various digitizing systems Knowing the system factor F in microvolts per count will let a network operator assign a CAL factor to a station using equation 9 We give some typical values here for systems in use by the USGS in Me
121. ght 0 For a given earthquake location each node is tested in the order it was defined to see if the epicenter lies within its outer circle but testing stops when the epicenter is found inside 3 outer circles If the weights total more than 1 0 they are normalized to 1 0 If they total less than 1 0 the difference is made up using the default model model number 1 such that the weights always total 1 0 If the epicenter lies outside all circles the default model is used exclusively The mix of models is determined for each iteration and the epicenter may migrate from one model to another Figure 1 is a hypothetical example of two crustal models in addition to the default model used for all surrounding regions shown in white defined by 4 nodes These are the commands that might be used to define this model geometry MUL T 1 Enable multiple models assign model 1 as the default model NOD 37 38 122 24 10 10 2 NOD 37 35 122 24 10 10 2 NOD 37 32 122 24 10 10 2 These three NOD commands establish the centers in deg and min of latitude and longitude a 10 km radius of the inner circle dark gray a 10 km wide transition zone surrounding each inner circle light gray and assign the node to model number 2 The region assigned exclusively to model 2 is the union of the three inner circles hence the sausage shape to the dark gray area One node defines the circular area for model number 3 with the same 10 km radius
122. gnitude The XC1 and XC2 commands also assign a 1 letter label to each magnitude a practice that is strongly advised If you are using instrument types to choose the event magnitude use the XTY command You can t use both component and instrument to choose the magnitude The method you choose will depend on how your network is set up in terms of the naming of components or the mix of seismometers A special case is when both amplitude magnitudes XMAG1 and XMAG72 are selected for a particular component ie a component code appears on both the XC1 and XC2 lists Because only one magnitude can be computed per station Hypoinverse chooses the secondary amplitude magnitude XMAG72 to represent that station The component list also determines which of the station magnitudes are used in the two event magnitudes Ifa station s component is used for both the magnitude is used both in XMAG1 and XMAG for the event P amplitude magnitudes Hypoinverse presently does not support P amplitude magnitudes Theoretically P amplitudes offer a way to estimate magnitude very early in the seismogram before waiting for the coda to run out and determine the duration time P magnitudes may be an advantage only for networks with high dynamic range high frequency vertical seismometers that do not record S very well The increase in numbers of broad band seismometers make reliance on P magnitudes unlikely Most of the source code is present within the program but it
123. gnitudes are computed for each station and which magnitudes appear for the different event magnitudes Types of coda magnitudes and how the coda is measured Hypoinverse offers a choice between the traditional duration F P magnitude Mp set with the DUR and DUB commands and the lapse time tau magnitude Mr set with the TAU command Michaelson 1987 The choice of which of the duration magnitude types to use is made with the MAG command Both use the same duration coda or F P values on the phase line The lapse time magnitude code adds the P travel time to the duration to get tau Hypoinverse uses the calculated rather than observed P travel time so that durations can be specified without P times The USGS practice is to determine the end of the coda or F phase when the coda decays to10 mm peak to peak on the Develocorder viewer or an equivalent amplitude on an analog or digital signal This 10 mm Develocorder amplitude peak to peak corresponds to 0 25 volts of discriminator output peak to peak or 205 counts of CUSP 12 bit digitizer output Equivalently this coda amplitude corresponds to 0 12 volts 0 to peak or 0 06 volts average absolute value of discriminator output and to a velocity of 1 729 x 10 cm s See the amplitude magnitude section below for a description of the seismic system and where the amplitude voltage is measured For analog channels the NCSN Earthworm systems are configured to terminate the coda at 0 06 volts
124. hadow records If you stripped off all the shadow records you would have an ordinary archive file The shadow format in use by the NCSN is given in the output formats section as an example but Hypoinverse just treats shadow records as ASCII strings to be written to output The format choices with the COP command are as follows The numbers are the value to supply to the COP command Use the 200 command to invoke either the pre Y2000 or Y2000 century compatible archive formats with options 3 or 5 1 Traditional Hypo71 style phase format 2 digit years only The default 2 No longer used 3 Archive format generated in an earlier Hypoinverse run 4 Shadow phase format with a 1 line header and a shadow record after each line This format is obsolete and is PHASEOUT s HYPO71 shadow option 5 Archive format with shadow records This format is compatible with the PHASEOUT program s Hypoinverse option 6 Locate one CUSP event where the CUSP ID number is given with LOC command 7 Locate several CUSP events with ID numbers given in a file Archive output format The archive format is special because it is both an input and output format Many fields in the archive format are ignored on input because they are filled in with calculated results on output The archive output format command CAR has two choices Shadow records on input and output may be selected independently of each other If you select shadow output but not input
125. hase data input files and glancing at these sections is a good place to begin The program has many options because it has grown over the years to meet the needs of one the largest seismic networks in the world but small networks with just a few stations do use the program and can ignore most of the options and commands History and availability Hypoinverse was originally written for the Eclipse minicomputer in 1978 Klein 1978 A revised version for VAX and Pro 350 computers Klein 1985 was later expanded to include multiple crustal models and other capabilities Klein 1989 This current report documents the expanded Y2000 version and it supercedes the earlier documents It serves as a detailed user s guide to the current version running on unix and VAX alpha computers and to the version supplied with the Earthworm earthquake digitizing system Fortran 77 source code Sun and VAX compatible and copies of this documentation is available via anonymous ftp from computers in Menlo Park At present the computer is swave wr usgs gov and the directory is ftp pub outgoing klein hyp2000 If you are running Hypoinverse on one of the Menlo Park EHZ or NCSN unix computers the executable currently is klein hyp2000 hyp2000 New features The Y2000 version of Hypoinverse includes all of the previous capabilities but adds Y2000 formats to those defined earlier In most cases the new formats add 2 digits to the year field to accommodate the century Othe
126. hase or archive record has a 1 letter field for the data source contributing network timing system etc Multiple data records for the same station and component can co exist in the same event anyway but this source code can distinguish among them The processing for each event determines the most common data source code for each of P times codas and amplitudes allowing you where the data for different events comes from CRUSTAL VELOCITY MODELS All models are flat earth models with stations assumed to be at the earth s surface Station elevations are not used but the delaying effect of elevation can be mostly accounted for with station delays This flat earth assumption means that earthquake depths are relative to the average local surface defined by nearby seismic stations This is computationally very simple For each crustal model velocity can vary only with depth Lateral variations can only be accommodated with choosing a model based on location or by having two different models used for the same earthquake by different stations Multiple models The simplest case Hypoinverse handles is one crustal model and one set of station delays used for all epicenters and all stations Hypoinverse also allows considerable complexity by using multiple velocity models In any model velocity varies only with depth Many models may be used each assigned to epicenters in different areas Smooth transitions between adjacent models is accomplis
127. he CAL command These history files are read as necessary during a location run for updates to the calibration history with time To know which cal factor or attenuation of the history files to read for the analog stations the station is given an instrument type of 1 for attenuation history and 3 for cal factor history The HYPO 71 station data format Start Col 1 2 9 gt a 3 A 7 9 14 15 18 23 24 29 35 38 45 51 53 A gt OTH aH Ow ahd 58 63 6 Fortran Format A1 A1 A1 4X 1X F5 2 1X A1 2X F5 2 2X F5 2 1X 11 1X F4 2 1X F6 2 Data Optional 1 letter station component code Station weight code in units of 0 1 by which the weights assigned each P amp S phase are to be multiplied Use the digits 0 9 for the weight in tenths or 0 for no weight or any other character including blank for full weight Station name code The code may not be blank The first character may not be the character See also col 1 Latitude degrees Latitude minutes N or blank for north latitude S for south Longitude degrees Longitude minutes W or blank for west longitude E for east Reserved for elevation in m Not used by Hypoinverse P delay sec for delay set 1 Optional station remark field to copy to print output Duration magnitude correction Amplitude magnitude correction Instrument type code Default period in sec at which the maximum amplitude will be
128. hed by defining transition regions within which weighted averages of travel times travel time derivatives and station delays for 2 or 3 different models are used There is no fancy calculation of rays passing through different models only a simple numerical weighting of travel times and travel time derivatives Thus if the travel time for model 1 is 1 10 sec model 2 is 1 20 sec and the weight of model 1 is 20 and model 2 is 80 the travel time used is 1 12 sec It is that simple A weighted average of emergence angles is also calculated but is not used in the hypocenter solution The geometry of assigning models to different areas consists of a list nodes or points on a map Each node is assigned to a model along with the radius of a circle within which that model is used Several nodes may be assigned to the same model and if so the circles can and often do overlap It is thus possible to define an irregularly shaped region as the union of several circles Each node is defined by a NOD command This is algorithm for determining the crustal model s used If an epicenter lies within the inner circle surrounding any node that model is used exclusively An outer circle must also be defined for each node which describes how far out its influence extends If an epicenter lies between inner and outer circles it receives a partial weight for that model The weight is a smooth cosine taper between inner circle weight 1 and outer circle wei
129. ht codes as originally assigned and the final normalized weight used to two decimal places are preserved in the archive and print output files Note that the weights are normalized average value of the weight squared is 1 0 for the inversion and that a fully weighted station will often have a value greater than 1 0 Distance Weighting An ideal distance weighting scheme should allow for both reducing the weights of distant stations when an event is in the interior of the network and the maximum use of all stations when an event is outside the network The Hypoinverse distance weighting function is 1 0 for near stations 0 0 for far stations and follows a cosine taper in between The distance points at which weight tapering begin and end can be made to stretch out and scale with the distance to the second closest station DMIN2 for earthquakes outside the network The four parameters that govern distance weighting are set with the DIS command and are with some sample values ITRDIS the iteration number at which distance weighting is to begin 4 DISCUT 50 km DISW1 1 and DISW2 3 When the second closest station distance DMIN2 is larger than DISCUT the distance at which the taper begins and ends scales with DMIN2 and is full weight closer than DMIN2 DISW1 and no weight farther than DMIN2 DISW2 When the second closest station distance is less than DISCUT the distance at which the taper begins and ends is fixed and is full weight cl
130. ieaadesncessaedadioatade 56 The weighting of duration and amplitude magnitudes 20 0 0 ceecceeeeeeceeeeeeeeceeeeeeeeeseeeaeens 57 PREFERRED EARTHQUAKE MAGNITUDES c ccccsssccsstcccssncccesesessccessscceensccessnesessneeses 57 WEIGHTING OF P amp S TIMES ates reheat easels Sic sean i R A En S 58 Distance Weightin s ssrnarsncnkeri orrian a a a a Aa a daa 59 Re id ak weighting risie en orar E RE ER ATEA A TAA A Eai 60 SOME SIMPLE COMMAND SEQUENCES 2 245 s5 ceicsidy captensceheatethevcsteeditas wns eeenadacde tench doeshdeess 61 COMMANDS RECOGNIZED BY HYPOINVERSE cceseescsseeseeseeseeecseeeceeeeceseeeeeaeeaeeaeenees 63 INPUT FILES pieron vasa sence ttaeacsa RT N EA TA T A 63 BINARY FILES ieran a a iat anes Sa a E ew e meant seco 64 READING ADDITIONAL STATION DATA ssssssesssessssseesssesssrseressreesseesseesseesseressressseesseesse 65 FILE FORMATS AND RELATED CONTROLS 0 ccccccsssccessccseseccssssscessseseacsesnacsesseteesees 67 OUTPUT FILES peon Sods ipushetaavsssior cea sdecunytanduaiy sade coed E ton adandecaa shears ae a iA 70 MULTIPEE CRUSTAL MO DELS i5 dais adscruicsslisigin apaddindbasiases tant ueyu sae a R N ees 71 PROCESS EVENT SdN A PHASE FILE iceren n i a sa a 73 PRINTED OUTPUT FORMAT igre e teie ieee aTe vedas och gtesso vets goede Eea tosiaa eai 74 TRIAL DEPTH VELOCITY RATIO amp ERRORS 0000 ceccecesececeeeeeceeeeeceeeeecseeeeesteeeesaes 75 CONVENIENCE AND CONTROL COMMANDS cceeccesesscesesseese
131. if you will not use a MERGE program is to run EXOCET with none for a summary file This will make an individual file for every event Hypoinverse will reprocess only the selected events and all events both reprocessed and original can be reassembled back into a single file The Hypoinverse BAS command defines the filenames to be processed interactively First specify the file containing base file names of events to process This may either be a summary file or a list of files produced with a VMS command like DIR COL 1 0UT filename ARC Also supply the number of characters in the base name and the format like A12 for reading the base name Any blank spaces in the base names will be filled with zeros Any line beginning with or blank will be skipped Any lines with an invalid filename will generate an error message but will be skipped over The BAS command also requests the filename extensions for the input archive summary and print files If you don t want summary or archive files specify none for those names If your input phase and archive filenames are the same my preference Hypoinverse will read back the file it just wrote during interactive processing and all of your edits will be cumulative and saved If the names are different you might lose some changes unless they were explicitly made to the input file One of the processing steps is the examination of the print output file using an editor It is best to use a wind
132. ig iA E L to 76 MEPS Searhces calcd EEE Make trates odesiids 76 PY POT eaa Seta edb ae 32 HYPINIPEIC cs tcc naranai 7 PIV PUNIS Ese eoii say rate aoa T E aA 7 HY POT nnna 14 17 27 32 TID A EESE A ea ATTAT 7 76 INP a a naa a A 34 77 JUN assen u aaa ae a n 39 84 KEP aen a N EEES 70 KPR a A E 74 86 LAO fanana o Piatt bale 47 83 LAB a ee RT 30 70 89 95 96 98 99 EES eein ete Ne ead 29 CET a E A naa 15 29 69 LOC a a a 27 73 BS E E E ER T 74 86 MAG eini iana a 40 42 47 77 MAX ana an Aan ees 7 76 MEDernerinannenen a a a 71 MN a e RT a 39 76 MOR a E A 76 MU Mange T T a 11 71 INE EENE EE ENE 84 NOD anano a T a ER 8 13 71 PIAS eenn e hae as a ea NTs 27 63 POS A Ai 75 PRP tics EE E EO ak 58 77 PRO Saan costes e A 35 73 PR Enema a Aa N a 70 ROR aee Aa r OA 64 remark event cccccccee cece 29 30 31 89 remark locatiOn ccccccececeeeeesesseeees 30 89 POMAT KB ASE so 4s Seccasyiacaaeemoaacncyrerorkes 30 91 remark seisMOgramM cceeeees 30 31 91 remark station ccccccceceeseeeeeees 18 87 91 RE E gus teased et aacaadadsdueatatss 75 IRIS cs vinca sauna teu A i 60 85 TES NENT ETSE E E E 14 64 Shadow records ccccceeesesseseseeeeeeees 28 67 SHO aber Ree ee tense as he ote oy ne Pra iene e 76 SNO na teen at ate epee asta adetat at 72 DE Acie tence din eaa a 14 63 65 SE icra eds aise atiatend iat arene 76 SUMo ines tite a a 70 DWV cis catiegnisds ia aaa aN 58 85 TAU a ar E E T 40
133. ime travel time derivatives and emergence angles at the source by interpolation from this table Three point parabolic interpolation is used Linear extrapolation is used beyond the table which is exact for refractions from the underlying halfspace The table itself is a condensed grid of travel times as a function of distance and depth Two different grid point spacings are permitted for each of distance and depth so that travel times for shallow nearby sources may be accurately modeled with close spacing without wasting computer memory on deep or distant grid points where the travel time curve changes slowly The user may generate his own travel time table empirically 113 or with another program see later for table format or use the travel time generating program TTGEN to prepare a table from a given velocity depth function Allowable crustal models input to TTGEN Crustal models consist of from 2 to 15 points at which the user specifies velocity and depth Linear velocity gradients are assumed to connect the points The last point fixes the velocity and depth of the homogeneous half space underlying the model The halfspace velocity must be the greatest of any of the velocities specified to insure that rays can be refracted along the top of the halfspace Rays traveling through layers with a linear velocity gradient trace circular ray paths when viewed from the side Rays can thus bottom inside a layer with a gradient Rays ina homogeneous
134. ing sheet or other external source The phase data input utility is invoked with the INP command no arguments You must be prepared with a small file containing a list of up to 100 station names see below You will be prompted for data from these stations and will not have to enter station names unless you want to add to the station list as needed The input utility is mostly self explanatory and prompts for the data and the decisions it needs A description of its operation follows The input utility first asks for the filename to which phase data will be written If an existing file is named any data there is appended to Ifa file called stalist dat is present in the current directory it is read for the list of station names to prompt for on each event If the file stalist dat is not present another filename is requested and must be given Note that the first 4 fields of the station prompting file format are the same as the Hypoinverse station format 2 The prompting list of stations is a convenience data for other stations may be entered but the additional station name must be given along with each datum Also data need not be entered for every station in the prompting list Responding with just a RETURN either skips that item or takes the default value and goes on to the next piece if information in the list Stations reporting data for most events should therefore be in the prompt file for the most efficient operation The stalist dat
135. ing the number of letters 0 3 chosen with the LET command If you have chosen 1 letter components with the LES command the 1 letter component is transferred to the 3 letter component field before the single letter is matched to the station list in memory Thus you can use either the 1 or 3 letter fields If you use the long field for station matching you can think of the short field as an abbreviation that is easier for humans to recognize The traditional Hypo71 style phase format contains an optional 3 letter event remark For historical reasons this is abbreviated to 1 letter for output A 1 letter event remark may be derived from the first 3 letter remark encountered and output to the summary file The first 3 letter remark encountered in the input series of phases will be abbreviated to the 1 letter event remarks if they match one on a preset list Note that Hypoinverse assigns a second 1 letter event remark if there were problems with the solution such as depth held fixed In the following _ is a blank and is any character The pre set remark list now consists of FLT amp F__ becomes F felt TRM amp T_ becomes T tremor associated LP becomes L long period BLS becomes B quarry blast Q or becomes Q quarry blast or NTS becomes N NTS shot Archive files read as phase input Formats for the archive files are described later in the section on output files Many remark fields are simply read
136. ion are written no histories or expiration dates so the ATE CAL FMC and XMC commands must still be used Supply the filename to write to Example WST home klein hypfiles allcrust bin Read a binary file of all station data and delays previously written with a WST command RIGHT NOW RST replaces the STA and DEL commands Binary reads are several times faster than ASCII reads and worth the effort for frequently read files Supply the filename to read from Example WST home klein hypfiles allcrust bin 64 READING ADDITIONAL STATION DATA The following 5 commands can t be given until a station file is read with the STA command That is because these commands read station data and store it only for the stations already in memory All data read by these commands may also be supplied in the station file but some fields will be redefined by these commands Data management is much easier with station locations STA command delays DEL attenuation gain settings ATE or CAL duration magnitude corrections FMC and amplitude magnitude corrections XMC in separate files For example the network delays may all be stored in one place and need not be incorporated into every station file In addition one file can have delays for every conceivable station but only those matching stations in the location file will use memory in Hypoinverse DEL ATN ATE Read in the station delays in seconds for either one or all models from
137. is not fully developed and there is no place in the data formats for p amplitudes Amplitude magnitude command example An example is that of the current 2000 NCSN system NCSN uses Eaton s 1992 Mx relation for the high and low gain 1 hz seismometer network and local magnitudes for any of the UC Berkeley Wood Anderson or broad band with a WA response stations All of these commands are more fully explained below in the command dictionary 56 Assign the 1 letter abbreviation for the primary amplitude magnitude namely X for Mx The 7 means there are 7 component codes to follow These codes are for stations of instrument type 1 or 3 for which X magnitudes are to be calculated The codes represent high and low gain vertical 1 sec seismometers XC1 XK 7 VHZ VLZ VLE VLN VDZ VDN VDE Assign the component codes for local magnitudes M from Wood Anderson response stations These are of instrument type 0 L is the magnitude label and 6 means there are 6 component codes to follow XC2 L 6 WLN WLE HHN HHE BHN BHE NCSN sets the default log Ao relation to use in amplitude magnitude calculations to be that of Eaton 1992 This is indicated by 1 as the fourth argument in the MAG command MAG 1 T 3 1 NCSN uses Berkeley s Nordquist log Ao relation choice 4 for the 6 WA amp synthetic WA components LAO 6 WLN 4 WLE 4 BHN 4 BHE 4 HHN 4 HHE 4 Add the component corr
138. itude corrections Use a year of zero to load the earliest corrections and let Hypoinverse update the magnitude corrections as needed Example there is no default FMC ALL FMC T 80 1 1 0 66 XMC Read station instrument types and amplitude magnitude corrections from an XMC history file RIGHT NOW You must read your station file with the STA command before reading the magnitude correction file Once a file has been read it is re read as needed during a location run to find a new magnitude correction after the expiration date of the previous one See the section on specifying the station list for the format You can suppress the amplitude magnitude for a given station from being used in the event magnitude by adding 5 0 to the magnitude correction Supply the filename of the amplitude magnitude correction file Set the flag governing where the station xmag weights come from T to use the xmag weights on the station line the default F to give an xmag weight of 1 0 to all stations present in the correction file and 0 0 to all the stations that are not Supply the date and time year month day hour for which to load the initial station magnitude corrections Use a year of zero to load the earliest corrections and let Hypoinverse update the magnitude corrections as needed Example there is no default XMC ALL XMC T 80 1 1 0 FILE FORMATS AND RELATED CONTROLS 200 COP Choose whether to use Y2000 formats If you u
139. lank Median event amplitude magnitude 2 XMAG2 Type label for amplitude magnitude 2 LABX2 Most common XMAG2 data source code XKMSOU2 Median absolute difference of amp magnitudes 2 KXMMAD2 Total of XMAG2 weights i e number of readings MTXMAG2 Presently blank 93 Preferred magnitude data for the EARTHQUAKE 153 3 F3 2 Preferred magnitude PMAG 156 1 Al Type label for preferred magnitude LABPR 157 3 F3 2 Median absolute difference of preferred mag KPRMAD 160 4 F4 1 Total of preferred mag weights i e number of readings 164 1 1X Presently blank External magnitude data for the EARTHQUAKE 165 3 F3 2 External magnitude BMAG 168 1 A1 Type label for external magnitude BMTYP 169 3 F3 1 Total of preferred mag weights i e number of readings Quality codes Hypo71 summary output only A 1 letter quality code is output on Hypo71 summary lines for consistency with this older program Hypoinverse format has never calculated this code because it is an oversimplified parameter and is often misused Seismologists would be better to use the individual criteria number of stations location errors RMS etc rather than a simplified A D rating For example an epicenter outside the network might be well controlled with S arrivals but have a large DMIN or a large depth error and thus get a low rating even if epicenter control is what is most important Two quality ratings are computed and they are averaged t
140. ld then be 19840502133045 arc The minimum suffixes required are the input and print files Note that input and output files can both be arc files with the same name Hypoinverse reads the input file closes it then writes the output file which would replace the input file on most operating systems 35 A good way to build event files is as follows First locate an entire set a month say so that you have summary and archive files to work with Then select the events you want to reprocess You can use a program like SELECT or EQSELECT to get a summary file of events to reprocess or type the date and time fields in Hypoinverse summary format into a file Then run a program like EXOCET similar to EXTRACT to put each selected event into a separate file The filenames will be of the form YYYYMMDDHHMMSS ARC like the example above After reprocessing these files will be replaced by their revised versions and can be reassembled into one file in chronological order using the VMS COPY command or the unix cat command Note that if the origin time of the event changes from the original one during reprocessing the filenames will always be based on the original origin time There are two ways to handle the events that will not be reprocessed If you have a MERGE program get a file of rejected good events when you run EXOCET Later merge the good and repaired files together and there will be no duplicated events The simpler way
141. le by under sampling a too complicated or irregular velocity model with too few rays Using the program TTGEN TTGEN shoots rays with increasing ray parameter starting with vertically emergent rays and calculates distance travel time and other parameters for each ray see outputs of TTGEN section all at depth intervals specified by the user At each depth a printed listing of these results is produced noting any reverse branches or rays lost to a low velocity waveguide At the same time a file of distance travel time reduced travel time and emergence angle from nadir is made for all branches of the travel time curve which can be plotted later see figure 8 A separate file is produced for each depth with the model name forming the base filename and depth as the file extension The program then produces the final travel time table by interpolating travel times at regular distance intervals specified by the user The travel time curve used for earthquakes is the first arrival from among the various branches including refractions from the halfspace and top of any low velocity zone Input to TTGEN in the file TTMOD All model parameters including depth distance and ray intervals at which computations are to be performed are input to TTGEN in the file ttmod The program uses reduced travel times for 114 the table to save space You specify the inverse of the reducing velocity REDV in sec km The reduced travel time is the
142. lves R As DX m mxn n in the sense of minimizing R A gt DX This is done by pre multiplying by A transpose to get the least square condition A R A A DX nxm m nxn n 109 that now only requires inversion of the nxn symmetric matrix A A The solution can be sought in terms of the generalized inverse of A and in particular the singular value decomposition SVD of A This not only yields the usual least square solution but permits manipulation of the eigenvalues of A A calculation of the errors and evaluation of the information content of the data Hypoinverse uses the SVD subroutine of Lawson and Hanson 1974 and forms the above matrices from elements of the decomposition The decomposition of A is given by A U S v mxn mxn nxn nxn where U and V are eigenvector matrices and S is the diagonal matrix of eigenvalues of A A Also UU I and V V I I is the identity matrix and assuming that the number of linearly independent arrival time data exceeds the number of unknowns VV I When the resolution matrix VV equals the identity matrix the unknowns are perfectly resolved which is the usual case for the earthquake problem Then the least squares solution can be derived by substitution of A into the least squares condition and is given by DX V Si U R n nxn nxn nxm m The covariance matrix of the solution DX is given by nxn nxn nxn nxn where w is a constant We see at once
143. m by running a command file that applies to a group of people running it on a given computer The INI command does not require that you be in the directory where the command file is located It is for situations where a group needs to use the same parameters station and crust model files or where you rapidly want to initialize the program On unix computers Hypoinverse treats the environment variable HYPINITFILE as the command file to run when the INI command is typed On the swave computer in Menlo Park for example NCSN users currently put the command setenv HYPINITFILE we calnet klein hypfiles cal2000 hyp in their system startup file Remark codes for earthquakes and station data Many different kinds of remark fields are available for input and output We can t fully discuss all of them here but how to use the different remark types are explained throughout this document and the major discussions are indexed In creative hands these are very useful o Station remark i e region code letter o Event location remark 3 letters o Event type remark assigned by analyst 1 letter o Event location problem remark assigned by Hypoinverse 1 letter o Event processing and authority remarks three 1 letter codes o Phase remark impulsive etc including P first motion 2 letters o Seismogram or amplitude remark one per phase record letter o Data source codes are special remarks that are tabulated Each p
144. median coda magnitude FMAG1 The preferred magnitude chosen from the available magnitudes by the PRE command The number of stations phase lines read for this event The number of phases P and S read for this event Distance to the nearest station The dominant crustal model code for the event If some stations use an alternate model a follows the code The largest azimuthal gap between azimuthally adjacent stations Number of iterations required for the solution Number of P first motions reported for this event Final number of P amp S readings with weights larger than 0 1 Number of S readings with weights larger than 0 1 Number of valid P amp S readings assigned weights larger than 0 The first remark is the 3 letter region code based on location and depth Two auxiliary 1 letter event remarks follow and they are blank if there is nothing special about the event The first is assigned by the analyst and may include F felt Q or B quarry blast R or N explosion T tremor associated or L long period Hypoinverse assigns the second remark A pound sign indicates convergence problems with the final solution such as running out of iterations or failure of the hypocenter to reach a minimum in RMS A minus sign indicates the depth was held fixed either by the user or by insufficient data An X indicates the location was fixed to the trial hypocenter The three status remark codes for the event fo
145. mes section Supply the filename of the calibration factor history file Supply the date and time year month day hour for which to load the initial station calibration factor Use a year of zero to load the earliest calibration factor and let Hypoinverse update the station calibration factor from the file as needed Example there is no default CAL ALL CAL 1980 1 1 0 Read station duration magnitude corrections from an FMC history file RIGHT NOW You must read your station file with the STA command before reading the magnitude correction file Once a file has been read it is re read as needed during a location run to find a new magnitude correction after the expiration date of the previous one See the section on specifying the station list for the format You can suppress the coda magnitude for a given station from being used in the event magnitude by adding 5 0 to the magnitude correction You can suppress the coda gain correction for a given station from being used by adding 10 0 to the magnitude correction Supply the filename of the duration magnitude correction history file Set the flag governing where the station fmag weights come from T to use the fmag weights on the station line the default F to give an fmag weight of 1 0 to all stations present in the correction file and 0 0 to all the stations that are not Supply the date and time year month day hour for which to load the initial station magn
146. models are useful where different regions have locally determined models and the gradations between models are either smooth or unknown Keep in mind that station delays can also be adjusted to shift epicenters laterally or move epicenters on or off a fault The combination of alternate and multiple models each with their own sets of delays can become quite complex is very flexible but can possibly be misused To use the alternate model feature read in two models using any two different model numbers Then use the ALT command to designate the numbers of the primary and alternate models which then form a pair The station file or an alternate station delay file must designate which stations use the alternate model Alternate models may be used with or without the multiple model feature If you use both together be sure to designate the primary and not the alternate 12 model number in the NOD commands because the same set of nodes will be used for both models in the pair Several different models may have alternates but the same set of stations must use alternate models in all cases Thus a given station will be either primary or alternate for all of the primary and alternate models An alternate model must be of the same type layer or gradient as the primary model Model types homogeneous layer and linear gradient Models may be of two different types which are stored and calculated differently The simplest
147. mplitude in mm on Develocorder viewer screen or paper record 48 3 F3 2 Optional period in seconds of amplitude read on the seismogram If blank use the standard period from station file 51 1 11 Amplitude magnitude weight code Same codes as P amp S 52 3 3X Amplitude magnitude remark presently unused 55 4 14 Optional event sequence or ID number This number may bereplaced by an ID number on the terminator line 59 4 F4 1 Optional calibration factor to use for amplitude magnitudes If blank the standard cal factor from the station file is used 63 3 A3 Optional event remark Certain event remarks are translated into 1 letter codes to save in output 66 5 F5 2 Clock correction to be added to both P and S times 1 71 A1 Station seismogram remark Unused except as a label on output 72 4 F4 0 Coda duration in seconds 76 1 11 Duration magnitude weight code Same codes as P amp S 77 1 1X Reserved 78 1 A1 Optional 5 letter of station site code 79 3 A3 Station component code 82 2 A2 Station network code 84 85 2 A2 2 letter station location code component extension The terminator line and trial locations all ASCII formats One terminator line must follow each event If the line is blank a standard trial hypocenter is used it is at the standard trial depth beneath the station with the earliest time at an origin time two seconds before the earliest time Trial values for any or all of depth latitude longitude or
148. mponent code 15 2 12 1X First attenuation setting Must be a multiple of 6 db 18 4 14 Year of expiration of first attenuation 22 6 312 1X M D H expiration date and hour of first attenuation Leave the expiration of the last attenuation blank to indicate no expiration 29 2 12 1X Second attenuation 32 4 14 Second expiration year 36 6 312 1X Second expiration date M D H Format repeats 7 12 1X 14 312 1X Use the ATE command to read the station attenuation file The phase data file must be in chronological order to insure getting correct attenuations if there is ever more than one attenuation value per station You must read your station file using the STA command before the attenuation file with the ATE command because Hypoinverse stores attenuations only for the stations already read into memory This means that the attenuation file can contain the history of the entire network and only the data needed will consume space in Hypoinverse Note that the instrument type must be 1 for stations with attenuations supplied with the ATE command The filename given with the ATE command is read both when the ATE command is given and as necessary to update an attenuation The ATE command also asks for a date and time for which to load the initial attenuations If you use the date of the first earthquake you wish to locate Hypoinverse won t have to waste much time re reading the attenuation file If you specify a year of 0 Hypoinverse
149. n above Interactive Earthquake Processing Supply the filename listing events to be processed For each event the file must contain the base name text string used to form the I O filenames for the event The base file name is read from this file with the format specified with the FID command Normally base names are the dates YMDHMS of the earthquakes and the file is a summary file Number of characters in the base name Base names are fixed in length Format for reading base names from the file named above This is an ASCII string and must contain parentheses as required by fortran The following four parameters are the file suffixes for the file types read or written by Hypoinverse for each event The archive and summary file types are optional and may be suppressed by using NONE or none as a suffix string Input phase or archive file extension Archive output file extension Summary output file extension Print output file extension 73 Example BAS LISTFIL 12 A12 PHS ARC SUM PRT the default or BAS LISTFIL 8401 12 A12 2 ARC NONE PRT PRINTED OUTPUT FORMAT LST List stations crust and test parameters at the beginning of the print output file Set the print code 0 Print earthquakes only 1 Add the location parameters amp filenames to beginning of printout 2 Add a station list and all crust models If the print code is 2 add two more
150. n the archive file and more detailed data relevant to the magnitude calculations The print output file only has some of the data used for calculating magnitude All magnitudes in this file are written to two decimal places Enough information is written for example to recalculate magnitudes or determine magnitude residuals as a function of time and station Like the archive format the data for an event begins with a summary line and ends with a terminator line Manipulation programs such as EXTRACT may thus be used with either the magnitude or archive file types There is one line per station but only for stations reporting a non zero duration or amplitude The essential event data like date and event magnitude are written on every station line The file may thus be sorted by station and each line has enough information to calculate a magnitude residual for example Magnitude corrections are also written to the file so it should be possible to reconstruct the terms going into the final station magnitude The additional oddball correction terms defined with the DU2 and XCM commands however are not written to the file The format of the magnitude file is always Y2000 compatible Magnitude output format Start Fortran Col Len Format Data 1 5 A5 5 letter station site code STA 6 2 A2 2 letter station network code SNET 8 3 A3 3 letter station component code COMP3 11 2 A2 2 letter station location code component extension SLOCC 13
151. nd sequences will illustrate the flexibility of Hypoinverse The intent here is to point out some of the most useful commands and how they might be sequenced Example 1 The simplest possible run keeps all other defaults CRH 1 MOD1 CRH Read layer model 1 from the file MOD1 CRH STA ALL STA Read station list from file ALL STA PRT RUN1 PRT Send printer output to the file RUN1 PRT PHS ASETI PHS Define the phase file as SET1 PHS FIL Determine the phase file format from the file 61 LOC STO Locate the events Stop the program Example 2 Generates additional output files CRT 1 MOD1 CRT STA ALL STA PRT RUN1 PRT SUM SET1 SUM ARC ARC1 ARC PHS SET1 PHS LOC Read gradient model from the file MOD1 CRT Read station list from file ALL STA Send printer output to the file RUN1 PRT Write Hypoinverse summary data to the file SET1 SUM Write archive data to the file ARC1 ARC Define the phase file as SET1 PHS Locate the events Example 3 Read archive format phase data write a compact print file and use HYPO71 format summary output all with Y2000 formats 200 CRH H71 STA T 1900 0 1 MOD1 CRH 22 2 STAS HYPO71 PRT LST KPR TOR SUM COP PHS LOC RUN1 PRT 0 1 F SET1 SUM 3 OLD ARC Enables Y2000 formats sets some defaults Read layer model 1 from the file MOD1 CRH Use HYPO71 summary terminator amp station formats Read station
152. nents to which the duration magnitude gain correction is to be applied To apply the correction the parameter FMGN must also be set to 1 with the DUR command Supply NDUG the number of components to use with duration magnitude gain corrections 1 apply to all components 80 TAU 0 apply to no components n usen components 1 10 max Follow with n 1 3 letter components Examples DUG 1 the default or DUG 4 VHZ VLE VIN VLZ Define the parameters for the lapse time tau end of coda time minus origin time magnitude scale Mt This algorithm assumes that the duration F P time is entered on the phase line and adds the calculated P travel time to it to get tau The formula is Mt DMAO DMA 1 log tau DMA2 log tau DMLIN tau DMZ Z DMGN G station corrections where tau is the elapsed time P travel time coda duration Z is the positive depth G is the gain correction 0 05 atten 0 0375 atten 12 18 etc Also G 0 5 log CAL factor 0 3 Supply the 6 coefficients named DM in the above equation DMA0 DMA1 DMA2 DMLIN DMZ and DMGN Example TAU 1 03 2 10 0 0 00268 0 1 AMPLITUDE MAGNITUDES XCH XCl Choose the way to select the stations to use for each of the two different amplitude magnitudes Supply the flag to make the choice T Use the component code to identify stations for amplitude magnitudes Then use the XC1 and XC2 commands to choos
153. ngitude and depth in km with the other three variables held fixed They are the square roots of the diagonal elements of the covariance matrix The error ellipsoid consists of the lengths of the principal axes SERR and their azimuths AZ and dips DIP in degrees The principal axes are the standard errors in those directions in units of km The hypocenter statistically has a 32 chance of lying within the error ellipsoid given To obtain a 95 confidence ellipsoid multiply the standard errors by 2 4 See the sections on the inversion scheme and error calculations for more information Appendix 4 Error calculations The covariance or error matrix The covariance or error matrix is calculated from elements of the decomposition of the A matrix see section on inversion scheme as where S and V are matrices composed of eigenvalues and eigenvectors in the solution space of the hypocenter w is the variance standard error squared of the arrival time data Hypoinverse calculates w as w RDERR ERCOF RMS7 where RDERR is set by the ERR command ERCOF is set by the ERC command and RMS is the root mean square travel time residual RDERR represents the aggregate of all un modeled timing errors including estimated reading error in seconds of the arrival time data ERCOF is just a weighting factor for including the effects of a poor solution large RMS in the error calculations If you want the calculated errors in the hypocenter to reflec
154. nitude is the weighted median of the station magnitudes This means that outliers will have less effect than if an average was taken Coda magnitudes but not lapse time magnitudes may use a bi linear relation with different parameters for durations above and below some cutoff value In addition to several terms in the main duration relationship other specialized terms using component type see FCM command and non linear distance and depth terms see DU2 command can be invoked to calculate Jerry Eaton s 1992 Central California relationship Component corrections and non linear distance and depth terms are applied to coda magnitude 1 and the lapse time magnitude but not coda magnitude 2 Selecting coda magnitude types Hypoinverse can calculate one duration magnitude for each station and two duration magnitudes for each earthquake A station s magnitude may either be chosen from one of two possible duration F P relationships or from lapse time but only from one of these The availability of assigning three magnitude types to two event magnitudes can be complex but it is very flexible If more than one magnitude type is in use the magnitude type used for a station depends on its component code The component codes can be arbitrarily chosen Several or all components could be used for the magnitude For example one could calculate duration F P magnitude for all ELZ components and a lapse time magnitude for all EHZ components Th
155. nitude range See the section on preferred magnitudes The procedure is to run down your list of up to 10 choices until a magnitude is found that satisfies all criteria then it is chosen as preferred Note that a given magnitude type may appear more than once in your list if the criteria become more lax down your list The preferred magnitude and its 1 letter type code are output to all files The magnitude types with their numbers to choose from are 1 Primary duration magnitude 2 Primary amplitude magnitude 3 An externally derived sometimes called Berkeley magnitude 4 Secondary amplitude magnitude 5 Secondary duration magnitude 6 and 7 are reserved for un implemented P amplitude magnitudes Supply the number of possible magnitude types which follow to evaluate as the preferred magnitude For each of the magnitudes to evaluate supply Magnitude type 1 5 Minimum number of readings to qualify Minimum magnitude to qualify Maximum magnitude to qualify Example PRE 0 no preferred magnitude the default or in this example I use commas to visually separate each set of magnitude criteria but any separator would do PRE 3 4 449 3 00 9 4 0 0 9 magnitude type 4 if 4 or more readings and M gt 4 then any magnitude of type 3 then any magnitude of type 4 CODA DURATION MAGNITUDES FCI Define the station components to include in the primary coda magnitude FMAGI for an event A maximum of 10 diffe
156. nlo Park The USGS station codes are SEED Standard for earthquake information exchange compatible for broadband stations and were devised within the USGS for the variety of equipment for the short period stations The dash symbol in the station codes column stands for the direction letter Z E or N The relation 9 as used below between F and CAL assumes that F is used to specify the entire electronic and recording system It works for a purely digital system It also works for a hybrid analog digital system like earthworm or the Menlo Park CUSP system where the digitizer factor is Dr 819 counts volt and Ur 0 04883 counts mm If the system is like the Hawaiian Volcano Observatory s with a different digitizer factor the CAL values may not match those of an older analog system For systems with a different digitizer factor where CAL must match earlier values specify a different digital count units code in the input phase archive than the value 2 used for earthworm or the Menlo Park CUSP systems Table 3 Gain factors of different seismic systems Seismic system F CAL USGS Typical microvolts factor station seismometer count codes NCSN analog CUSP or Earthworm digital 0 9954 at 24 1 4 at 24 VH L4C recording 12 bit db atten db atten VL NCSN DST Digital seismic telemetry 12 6 865 0 2014 VD L4C bit Nanometrics HRD Hayward boreholes 1 037 1 333 VD EP HS1 L4C
157. nsider a simplified case If the solution to the earthquake location problem were 108 linear and if we had exactly as many independent data m arrivals times as hypocentral unknowns n the answer would be the solution of T A X G n nxn oon n where T is the n vector of arrival times X is the n vector of hypocenter coordinates and G is constant A is the n by n partial derivative matrix aX that may be directly calculated from an assumed velocity model But since the earthquake problem is nonlinear A is not constant we must seek successive linearized solutions and iterate toward the true solution until we have converged to the desired accuracy X and A must also be updated as iteration proceeds If To and Xo are the arrival time and hypocenter vectors calculated at the previous step or some initial guess on the first iteration which satisfy To A X G then subtracting the equations yields T TT As X Xo or R A gt DX n nxn n where R is the vector of travel time residuals observed times minus those calculated from the model at the previous step and DX is the hypocentral adjustment vector given in this simple case by DX A R Now we must consider the real case The number of observations m for the earthquake problem is often in the range 8 to 60 but the number of unknowns n is generally only 4 When m exceeds n however the true inverse A does not exist We seek the least squares solution which best so
158. nt file and set the print level to a number greater than 6 like KPR 7 This will generate much output but is useful for debugging There is a substantial list of limitations to the CUSP capability o Station data must be read from an external file as before using STA or RST o Hypoinverse cannot get instructions from the CUSP scheduler o Hypoinverse must be run in the directory containing the MEM files As before all other input and output files set in Hypoinverse may contain pathnames to other directories o Atpresent locating events from MEM files is noticeably slower than reading ASCII phase files There is an overhead in making database calls and sifting through unwanted information in the MEM file o CUSP MEM files are several times larger than an equivalent ASCII phase file because they contain seismogram data and entries for traces that were not picked CUSP data is treated like another input format selected with the COP command Use COP 6 to locate one event at a time by CUSP ID number In this mode the number is supplied with the LOC command i e LOC 100229 If you just type LOC you will be prompted for the ID number You must already have read in station and crust files and define other parameters as before In this single event mode any phase file specified with the PHS command will be ignored You may process a series of events by supplying a file listing their CUSP ID numbers Select this CUSP list option with the COP
159. o one for output Tests are based on several parameters RMS root mean square travel time residual ERH horizontal location error ERZ vertical location error NWR number of weighted station readings phases MAXGAP maximum angular gap in degrees between azimuthally adjacent stations DEPTH the earthquake depth and DMIN distance to closest station The first quality rating is based on errors and goodness of fit A RMS lt 0 15 sec and ERH lt 1 0 km and ERZ lt 2 0km B RMS lt 0 30 sec and ERH lt 2 5km and ERZ lt 5 0 km C RMS lt 0 50 sec and ERH lt 5 0 km D Worse than above The second quality rating is based on station geometry A NWR gt 6 and MAXGAP lt 90 and either DMIN lt DEPTH or DMIN lt 5 0 B NWR gt 6 and MAXGAP lt 135 and either DMIN lt 2 DEPTH or DMIN lt 10 C NWR gt 6 and MAXGAP lt 180 and DMIN lt 50 D Worse than above 94 OLD PRE Y2000 OUTPUT FORMATS Recall that the 200 command selects between the old non Y2000 formats and the current Y2000 formats with 4 digit years Pre Y2000 summary format also used as header in pre Y2000 archive file Start Fortran Col Len Format 1 10 512 11 4 F4 2 15 2 F2 0 17 1 A1 18 4 F4 2 22 3 F3 0 25 1 A1 26 4 F4 2 30 5 F5 2 35 2 F2 1 37 3 13 40 3 13 43 3 F3 0 46 4 F4 2 50 3 F3 0 53 2 F2 0 55 4 F4 2 59 3 F3 0 62 2 F2 0 64 4 F4 2 68 2 F2 1 70 3 A3 73 4 F4 2 77 1 A1 78 1 A1 79 2 12 81 4 F4 2 85 4 F4 2 89 2 12 91 3 F3 1 94 3 F3 1
160. o use any S readings Supply SWT the factor by which all S weights will be multiplied Example SWT 1 0 the default 85 ITERATION AND CONVERGENCE PARAMETERS CON DAM Set parameters governing tests for convergence of iterations to a final earthquake solution The solution is considered final as soon as either the hypocentral adjustment or RMS change falls below set limits or the maximum number of iterations is exceeded You may want to see the section on fixing hypocenters Supply ITRLIM the maximum number of iterations allowed Using a value of 0 is not the only requirement for fixing hypocenters Supply DQUIT If the hypocentral adjustment falls below DQUIT km iteration stops Supply DRQT If the change in RMS residual falls below DRQT seconds iteration stops Example CON 20 0 04 0 001 the default Set the iteration and damping parameters affecting hypocenter adjustments See the section on iteration and damping Supply DXFIX Keep the depth fixed until the epicentral adjustment is less than DXFIX km Supply DZMAX the maximum depth adjustment in km allowed without a forced damping of the adjustment vector Supply DZAIR If the depth adjustment would place the hypocenter in the air move the depth upward by this fraction toward the surface Supply DAMP the mandatory damping factor for all hypocenter adjustments Supply EIGTOL the minimum eigenvalue permitted before no hypocentral adjustment
161. ode See cols 68 69 117 1 A1 Most common amplitude data source code 118 1 A1 Primary coda duration magnitude type code 119 3 13 Number of valid P amp S readings assigned weight gt 0 122 1 A1 Primary amplitude magnitude type code 123 1 A1 External magnitude label or type code Typically L ML This information is not computed by Hypoinverse but passed along as computed by UCB 124 3 F3 2 External magnitude 127 3 F3 1 Total of external magnitude weights number of readings 98 130 1 Al 131 3 F3 2 134 3 F3 1 137 10 110 147 1 A1 148 3 F3 2 151 4 F4 1 155 1 A1 156 3 F3 2 159 4 F4 1 163 1 A1 164 1 A1 Alternate amplitude magnitude label or type code Alternate amplitude magnitude Total of the alternate amplitude mag weights no of readings Event identification number Preferred magnitude label code chosen from those available Preferred magnitude chosen by the Hypoinverse PRE command Total of the preferred mag weights number of readings Alternate coda duration magnitude label or type code Alternate coda duration magnitude Total of the alternate coda duration magnitude weights Version of the information i e the stage of processing This can either be passed through or assigned by Hypoinverse with the LAB command Version of last human review Hypoinverse passes this through 164 is the last filled column Y2000 Hypo71 summary output format Start Fortran Col Len Format
162. olts of discriminator output would produce 4096 digital counts Thus Dr 819 counts volt For the Hawaiian Volcano Observatory CUSP digitizing and recording using 14 bits of a 16 bit digitizer 5 0 volts of discriminator output would produce 16 384 digital counts Thus Dy 3276 8 counts volt The 2000 version of Hypoinverse supports amplitude measurement and input in digital counts in addition to mm The code for the type of amplitude units being used is located on the Y2000 archive phase format and is a 2 digit integer code The units code field is next to the amplitude field which has been expanded from 3 columns to 7 columns The print file lists the amplitude in the input units with a letter indicating the type of units Presently only four amplitude unit codes are supported Table 1 Codes for seismogram amplitude units Units Amplitude units U factor Units code code print file 0 peak to peak mm 1 0 M 1 0 to peak mm for Berkeley Wood Anderson data 2 M 2 peak to peak digital counts from Menlo Park CUSP or 0 04883 C earthworm digitizers 3 peak to peak digital counts from HVO CUSP digitizers 0 012207 D Hypoinverse converts amplitudes in digital counts to mm before applying equation 5 to calculate the magnitude The conversion factors for counts are Ur Dp Dr 0 04883 counts mm Un Dp Du 0 012207 counts mm The modified magnitude relation used by Hypoinverse for amplitudes in counts is then
163. ommand 3 put calibration factors on the phase line traditional phase format only to use only for that event and override that on the station line if present this will not preserve the calibration or write it out to the archive file This feature is for backward compatibility and is not recommended 4 read the attenuation history from a separate file and extract the attenuation for any given date see the ATE command and 5 read the calibration history from a separate file and extract the calibration factor for any given date see the CAL command Using a separate station attenuation or calibration file insures that the correct gain information will be used for each station on the date of the earthquake Each attenuation or calibration factor value has an associated expiration date When an event is being processed whose date is after the attenuation s expiration date Hypoinverse rereads the file and uses the updated value The cal or attenuation file has one station per line Unlike the delay file full 12 letter station codes are supported because the attenuation depends on the component The line consists of a station code followed by pairs of attenuation or cal values and their expiration dates The last attenuation on a line must have its expiration date left blank or set to zero to indicate that it applies into the future A station history may have at most 7 attenuations If the attenuation for a station never changed the line consists onl
164. omponents These corrections are in addition to all other corrections If you find that low gain horizontal stations typically give too large a magnitude for example you can apply a correction to all stations of that type without having to specify a correction for each station individually Supply the number of components that follow with individual duration magnitude corrections 0 4 Supply pairs of 3 letter component codes and duration magnitude corrections Examples FCM 0 no component corrections the default or FCM 2 VLE 0 3 VLN 0 3 negative corrections for two component codes Define the constants used to compute the first magnitude FMAG1 from coda duration F P time Two formulae may be used spanning different ranges of coda duration The formulae have the form Md FMA FMB log duration FMF duration FMZ depth FMD distance FMGN log CAL factor 0 6 station corrections Supply FMA1 FMB1 FMZ1 FMD1 and FMF for durations less than FMBRK Supply FMA2 FMB2 FMZ2 FMD2 and FMF2 for durations more than FMBRK Supply FMBRK the duration above which to use the second set of constants Set FMBRK 9999 to only use the first set of constants 79 DUB DU2 DUG Supply FMGN normally 0 or 1 to regulate the use of gain correction Also see the DUG command to regulate whether gain corrections are used Example DUR 5 2 3 89 013 0037 0 9 22026 2 013 20037 0 210 0 for Ha
165. omputed Wood Anderson at least 4 readings 4 0 lt M lt 9 9 5 External magnitude any number of readings any M 0 0 lt M lt 9 9 6 Secondary amplitude magnitude recomputed Wood Anderson any number of readings any M 0 0 lt M lt 9 9 Use the PRE command to specify the preference order For each magnitude you want to consider specify its type number 1 5 the minimum number of readings and the minimum and magnitude This NCSN example uses the PRE command PRE 6 3049 1109 2109 4449 300 9 40 0 9 WEIGHTING OF P amp S TIMES The actual weight given a P or S time is the product of several factors 1 The station weight A code on the station line results in a weight factor between and 0 in increments of 0 25 for that station for the entire run 58 2 The global S weight This number is set with the SWT command for an entire run and that factor multiplies all S weights For example setting SWT to 0 5 gives all S times half the weight they are individually assigned 3 The weight assigned each phase The weight codes 0 4 and blank yield weights from 1 0 to 0 0 The actual weight factors for codes 0 3 may be reset from their default 0 25 increment values using the WET command The codes 4 9 yield no weight 4 Distance weight Weight decreases from 1 0 to 0 0 with increasing distance 5 Residual weight Weight decreases from 1 0 to 0 0 with increasing absolute value of travel time residual The individual P amp S weig
166. ons of partial derivatives from the m combinations in the matrix A One unit of importance is attributed to each of these n independent combinations Hence the sum of the importances of all stations for a full earthquake solution is 4 If several data are redundant 111 i e linearly dependent or nearly so then the unit of importance must be distributed among them and the importance of each redundant datum goes down Eigenvalue and covariance error output If KPRINT is 3 or larger KPR command the four eigenvalues of the principal directions of the solution are listed in descending order These are useful in gauging the relative stability and error of the solution in the four principal directions Under each eigenvalue are the column eigenvectors corresponding to it The eigenvectors together make up the matrix V The elements of the column eigenvectors give the components of origin time latitude longitude and depth in the principal direction corresponding to that eigenvalue In other words the matrix of eigenvectors accomplishes the rotation between the principal and geographic coordinates The last eigenvector gives the mix of latitude longitude and depth that are most poorly determined and associated with the smallest eigenvalue The covariance matrix gives the variances diagonal elements and covariances of origin time latitude longitude and depth The errors listed are the standard errors of origin time in sec and latitude lo
167. ormat types correspond to some of the phase data input formats Supply the format number as follows 1 Archive format no shadow lines 2 Unused 3 Archive format write a shadow line after every line Example CAR 1 the default 68 FID H71 LET Set the fortran format for reading CUSP ID numbers to locate use with COP 7 This is determined by the format of the file identified with the PHS command for example a MEM file list Supply the format as a text string Examples FID files 10 the default or FID 1X 18 for current NCSN Choose between HYPOINVERSE and HYPO71 formats where they differ Set the summary output format 1 for HYPOINVERSE 2 for HYPO71 Set the terminator input format 1 Hypoinverse 2 HYPO71 3 Get the trial hypocenter from the Hypoinverse header record COP formats 3 or 5 only Set the station input format 1 Old Hypoinverse format 1 2 HYPO71 3 New Hypoinverse format 2 Example H71 1 1 1 the default Choose how many letters to match when finding correspondences between the station codes given in the various files such as station location phase and the various correction files Supply the number of letters for each of the site max 5 network max 2 and component max 3 codes A value of 0 means no characters are tested and any characters will match Normally you would specify the maximum number of letters your ne
168. oser than DISCUT DISW1 and no weight farther than DISCUT DISW2 See Figure 6 To keep the taper distance fixed at for example XNEAR and XFAR set DISCUT 1000 DISW1 XNEAR 1000 and DISW2 XFAR 1000 59 1 0 H DMIN2 gt DISCUT S EVENT OUTSIDE NET R DISTANCE 0 0 DMIN2 DISW1 DMIN2 DISW2 1 0 z DMIN2 lt DISCUT COSINE TAPER EVENT WITHIN NET Sb DISTANCE DISCUT DISW1 DISCUT DISW2 Figure 6 The distance weighting function DISCUT DISW1 and DISW2 are constants set with the DIS command DMIN2 is the distance to the second closest station If DMIN2 is larger than DISCUT upper figure the function stretches out and scales with DMIN2 such that most of the network stations receive weights somewhere in the tapering part of the function If DMIN2 is smaller than DISCUT lower figure the function is fixed Stations farther than DISCUT x DISW2 then receive no weight Residual weighting The purpose of residual weighting is to reduce the weight of arrivals with large residuals which may reflect large timing errors or travel paths for which the velocity model is very poor The residual weighting function is 1 0 for small residuals 0 0 for large residuals and follows a cosine taper in between In Hypoinverse the residual points at which weight tapering begin and end can be made to stretch out and scale with the root mean square residual RMS The four parameters which govern residual weighting ar
169. out 0 02 Looked at another way instability occurs when the condition number ratio of largest to smallest eigenvalue exceeds about 200 The value of EIGTOL should be chosen after attempting to solve for the most marginal events one wishes to locate with a given network and studying the eigenvalues and iteration history for these events The station importance and information density The station importance is a product of the generalized inverse approach and usually is not computed by other standard location programs It is a quantitative measure of the contribution a particular arrival makes to the hypocenter solution and includes the effect of the weights applied to the arrival data A result of the singular value decomposition of the partial derivative matrix A see section on inversion scheme is the information density matrix B UeU This is an m x m matrix where m is the number of arrival times reported Each diagonal element bj of B is thus associated with the ith phase alone and is the quantity printed and referred to as the importance of the phase arrival A feeling for what importance means quantitatively may come from realizing that the rows of U are linearly related to the rows of the partial derivative matrix A In other words when the partial derivatives of the travel time to the ith station with respect to the jth hypocentral coordinate OT OX are large for the ith station the ith row of A then the ith row of U and henc
170. ow with 132 columns to see all of the file You can browse through the file and change the P S and coda weights for the next location try You will actually edit the print file to change P S and coda weights Any other types of changes must be made to the input phase file directly If you choose to locate the event again the print file will be read after the input file so that any changed weights will override the original ones Note that weights changed in the print file will be passed to the output archive file If this archive file is then read as input input and archive file suffixes the same the weight changes will be carried along and not lost You have a choice of editors The BAS command asks for your editor choice The unix version of Hypoinverse will call either dtpad the solaris window editor vi or textedit the sunos window editor The command you choose should be in your search path The VMS version lets you choose either EDT or ED which you can define how you like 36 How to change weights in the print file Make the change on the line showing the station you wish to change In identifying stations marked in the print file Hypoinverse uses the number of letters in the site net and component codes you selected with the LET command It also uses the data source code in column 73 of the print file in case there are multiple sources for the same station If the station code is blank on the line you change as for multiple
171. p file identified by the environment variable HYPINITFILE under unix systems to be run by the INI command Samples of some of the original ascii files are listed here rather than supplying full sized files to do complete processing Files parameters and even program capabilities are always changing so these files are merely a sample The basic control file of Hypoinverse commands that is invoked by the INI command in Menlo Park is cal2000 hyp Ifa line starts with Hypoinverse interprets it as a comment Each Hypoinverse 3 letter command is followed by the necessary parameters The fortran parser 100 ignores everything after the on a line with free format parameters so these are also comments HYP COMMAND FILE cshre file puna swave ebird hollyhock STATION DATA 00 T 1900 0 D723 setenv HYPINITFILE home paths for hypfiles directory home calnet klein hypfiles homel calnet klein hypfiles ebird calnet klein hypfiles we calnet klein hypfiles i Enable yr 20 Use new 00 formats Hypoinverse startup file for NCSN for newer 2000 formats To run this file with the hyp command ini put a statement like this in your home calnet klein hypfiles cal2000 hyp kkk kk YOU MUST READ STATIONS 2 LE R X F U B EFORE USING TH
172. parameters Set the station detail code 0 List no stations 1 List station locations cal factors first delay etc 1 line per sta 2 Also list delays for all crust models adds 1 or 2 lines per station Set the crust model detail code 0 List no crust models 1 List the layers nodes and other data for each model Note that LST 1 and LST 2 0 0 produce the same result Example LST 1 the default or LST 2 1 0 KPR Control the amount of information in the print file for each event Supply KPRINT which controls print output Specifying a value also outputs all data output by lower values 0 Print final location only 2 lines 1 Add station list for final location 2 Add the location amp adjustments one line for each iteration 3 Add the eigenvalues covariance matrix amp error ellipse Also print a message each time an updated magnitude correction attenuation or cal factor is loaded from one of the station history files 4 Unused 5 Unused 6 Add the station list at each iteration Example KPR 3 the default TOP Start each earthquake at the top of a page in printout file 74 Set a logical flag T or F that controls whether to start each new event with a form feed Example TOP T the default Report each earthquake as it is located with a 1 line message on the terminal and control whether un weighted stations are listed in the print file Set a logical flag T or F that controls whethe
173. put in an editor editing the print output file You may examine the file and change only the P S and coda weights by putting codes in certain columns The history of successive location tries is also written to the print file 37 4 Ifyou are satisfied with the event you can QUIT the editor without saving changes to the print file If you change weights you must save changes EXIT in EDT you made in the editor to keep the changes 5 You will then get a prompt from the primary branch point within the processing loop in Hypoinverse The possibilities and their actions are e return This event is OK go on to next event step 1 above e T Relocate the current event Go to step 6 below e ZXZ Delete the current event including all versions of all input and output files Then return to step 5 e KS Killall S weights and relocate Go to step 6 below e KA Killall P amp S weights and continue to the next event This is useful to preserve the data but then merge the event with another set of picks for the same event e Any other response typed to this prompt will be interpreted as an operating system command You may do any special operation such as on VMS delete the most recent version of the archive file to cancel changes you just made DELETE 800101120030 ARC 0 If you want to stop processing issue a ctrl C unix or ctrl Y VMS at this point 6 Relocate the current event Hypoinverse first asks whether you want to edit
174. r abbreviation for coda magnitude type 1 namely D for Mp The 4 means there are 4 component codes to follow These codes are for stations for which coda magnitude is to be calculated The codes represent high gain 3 component and low gain vertical 1 sec seismometers ECI D 4 VHZ VHE VHN AVLAT Step 7 Assign the one component code for Mz The relation assigned by the MAG command to FMAG2 is the Hirshorn and Lindh relation Note that VLZ appears on both the FMAGI1 FC1 and FMAG2 FC2 lists See the confusing discussion above This means that VLZ stations are assigned FMAG2 as a station magnitude and that it is used in both the primary FMAGI and secondary FMAG2 event magnitudes NCSN feels that this makes FMAGI more stable as an event magnitude FC2 gt TAN 2 Via 43 Gain corrections to coda magnitudes The gain correction methodology for coda magnitudes was developed for the analog short period seismometers by Jerry Eaton 1970 1992 A station gain or attenuation correction is made to coda magnitudes if the flag on the DUR or DUB command is set to 1 Additional control over which components and stations are corrected can be made with the DUG and FMC commands Further control over applying coda gain corrections can be made to individual stations with the coda magnitude correction FMC command See the section above Station magnitude correction files The gain correction is G G log CAL factor CAL
175. r fields are sometimes rearranged or expanded to accommodate a better field order The Y2000 formats are invoked with the 200 command When the Y2000 flag is turned on all files are read and written in the new format and there is no mixing of format types in a single run Some formats without a date field like station files have not changed A separate program called 2000CONV has been written to convert old formats to new Other new features like expanded station names calculating amplitude magnitudes from a variety of digital seismometers station history files interactive earthquake processing and locations from CUSP Caltech USGS Seismic Processing binary files have been added General features Hypoinverse will locate any number of events in an input file which can be in one of several different formats Any or all of printout summary or archive output may be produced Hypoinverse is driven by user commands The various commands define input and output files set adjustable parameters and solve for locations of a file of earthquake data using the parameters and files currently set It is both interactive and batch in that commands may be executed either from the keyboard or from a file You execute the commands in a file by typing filename at the Hypoinverse prompt Users may either supply parameters on the command line or omit them and are prompted interactively The current parameter values are displayed and may be taken as
176. r individual stations add 10 0 to the coda magnitude correction These two flags for coda magnitudes function independently if you want to zero weight and not make the gain correction to a coda magnitude add 15 0 to the correction The filename given with the FMC or XMC command is read both when the command is given and as necessary to update a correction The FMC and XMC commands also ask for a date and hour for which to load the initial magnitude corrections If you use the date of the first earthquake you wish to locate Hypoinverse won t have to waste much time rereading the file If you specify a year of 0 Hypoinverse loads the earliest magnitude correction 24 for each station This will require more updates from the correction file but will not require knowing the date of the first event Use the FMC command to read the FMAG correction file The choice of which Y2000 format is used is governed by the 200 command Corrections can also suppress the weight of individual stations by adding 5 0 see text above FMAG corrections can also suppress the gain corrections by adding 10 0 Duration magnitude correction file format pre Y2000 Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 2X Station net code 11 3 A3 1X Station component code 15 5 F5 2 1X First magnitude correction 21 2 12 Year of expiration of first magnitude correction 23 6 312 3X M D H expiration date and hour of first magnitud
177. r to report basic data of each event as located to the terminal Data includes time location preferred magnitude and id number Set a logical flag T or F to control the printing of un weighted stations to the print file If the flag is T a station with zero weighted P S coda or amplitude will not print but will be written to the archive file Example REP T F the default TRIAL DEPTH VELOCITY RATIO amp ERRORS ZTR POS ERR Set the trial depth in km for the run which can be over ridden for individual events on their terminator lines If the trial depth is negative earthquakes are held fixed at this depth Supply the trial depth ZTR in km Example ZTR 7 the default Set the P to S velocity ratio POS All S travel times are calculated as POS times the model travel time P velocities assumed S station delays are derived from P delays by multiplying by POS Supply the P to S velocity ratio POS Example POS 1 75 the default Set the assumed reading and timing error in seconds This should be the total error from all sources including the reading error and all un modeled crust and delay time errors A good value to use is your typical RMS residual after you have adequate crust and station delays See the ERC command Supply the reading and timing error RDERR Example ERR 15 the default 75 ERC MIN Set the coefficient ERCOF of the RMS travel time residual in the expression for the actual timing
178. re given here in one place See the discussion of cal factors in the amplitude magnitude section and reading in the calibration and attenuation history files section for more information The calibration factor is used for amplitude magnitudes and is a measure of station gain relative to a standard seismograph The calibration factor was originally defined as the peak to peak amplitude of a 10 microvolt RMS signal at 5 hz applied to the VCO and measured in mm on the Develocorder film viewer For instrument types 0 and 2 the cal factor should generally be 1 0 A cal factor of 0 0 signifies an unknown response for which no amplitude magnitudes will be computed Ifa cal factor is given on a phase line traditional format only it overrides the value supplied for the station The instrument type specifies which frequency response curve and which seismometer motor constant applies to the instrument The response correction yields an equivalent Wood Anderson amplitude The seismometer types are listed in the instrument types table in the amplitude magnitude section below 16 The VCO preamplifier in the voltage controlled oscillator attenuation may be given in place of the cal factor see the ATN command An entire history of station attenuations with the dates of attenuation changes may be read from a separate file with the ATE command Equivalently an entire history of station cal factors with the dates of changes may be read from a separate file with t
179. read for this station Must be greater than 0 1 Default calibration factor for amplitude magnitudes The HYPOINVERSE station data format 1 Start Col Len 1 4 5 1 6 2 9 5 14 1 Fortran Format A4 A1 12 1X F5 2 A1 Data Station site code The first character may not be the character Station weight code in units of 0 1 by which the weights assigned each P amp S phase are to be multiplied Use the digits 0 9 for the weight in tenths or 0 for no weight or any other character including blank for full weight Latitude degrees Latitude minutes N or blank for north latitude S for south 15 19 24 29 32 34 35 36 42 48 53 woh OW an o 13 1X F5 2 A1 F3 1 A1 1X A1 A1 F5 2 1X F5 2 1X F5 2 A1 F5 2 A1 11 F6 2 A2 1X A3 A2 Longitude degrees Longitude minutes W or blank for west longitude E for east Reserved for elevation in m Not used by HYPOINVERSE Default period in sec at which the maximum amplitude will be read for this station Must be greater than 0 1 Optional 1 letter station component code Puta 2 or A here to designate this as an alternate crust model station Both alternate and primary crustal models must be in use Stations may also be tagged for use with an alternate model in the delay file Optional station remark field to copy to print output P delay sec for delay set 1 P delay sec for delay set
180. readings from the same station it will be inferred from the non blank line above o To change aP weight put a new weight code in column 1 before the station site code o To change a coda weight put a new weight code in column 9 between the net and component codes o To change an S weight put a new weight code in column 13 after the 3 letter component code This example shows the columns read for new weight codes in a typical station line 1111 column 1234567890123 Cc O weight d code P a S station V V V line BCS NC VHZ 01V 2 1 Site net component location amp 1l letter component codes These are the weight codes currently recognized e 0 9 New 1 digit weight code to replace old one e Remove reading by adding 5 to weight code e Restore reading by subtracting 5 from weight code Existing partial weight codes 1 3 will be converted to full weight 0 e Weight out this and all following P and S weight codes Flow of steps in interactive processing 1 Read the base name for a new event and form the input and output filenames When no base names remain in the file you are done 2 Open the event input and output files locate the event then close the files If you used the REP command to report events to the terminal you will see a list of the previous tries for this event with the most recent at the bottom so you can compare them and hopefully note improvements 3 You will then be
181. rely digital system It also works for a hybrid analog digital system like earthworm or the Menlo Park CUSP system where the digitizer factor is Dr 819 counts volt and Ur 0 04883 counts mm If the system is like the Hawaiian Volcano Observatory s with a different digitizer factor the CAL values may not match those of an older analog system For systems with a different digitizer factor where CAL must match earlier values specify a different digital count units code in the input phase archive than the value of 2 used for Earthworm or the Menlo Park CUSP systems Additional U factors can be programmed into Hypoinverse in the future If the Hypoinverse input is from a MEM file COP 6 or 7 the amplitude units code is sensed from the digitizer device code stored in the MEM file All amplitudes from AMF tuples amplitudes measured by analyst are in digital counts By default the units code is 2 unless the event is from the Hawaiian Volcano Observatory with digitizer device code OBI or XOB where the amplitude units code will be 3 Equations 7 and 9 mean that you can compute magnitudes from a variety of velocity sensor and recording systems The sensor output S in volt cm sec and system factor F in microvolts per count completely specify a system for the frequency band where response in linear ie 3 10 hz You set S by choosing a standard seismometer instrument type and set F by explicitly 53 entering a CAL factor In fact only the p
182. rent components may be used The components will select which of the station magnitudes are used for the earthquake magnitude Also set the 1 letter label for FMAGI1 Supply the 1 letter label code for FMAG1 Supply NCPF1 the number of components to use in calculating FMAGI Set NCPF1 1 to use all components in FMAGI Set NCPF1 0 to use no stations in FMAGI Set NCPF1 the number of components to use for FMAGI 1 10 Supply the NCPF1 3 letter component codes to use in FMAGI 78 FC2 FCM DUR Examples FC1 1 use all components used for FMAG1 which has no label the default or FC1 D 1 VHZ use VHZ component stations for the D magnitude Define the station components to include in the secondary coda magnitude FMAG2 for an event A maximum of 10 different components may be used The components select which stations are used for the earthquake magnitude Also set the 1 letter label for FMAG2 Supply the 1 letter label code for FMAG2 Supply NCPF2 the number of components to use in calculating FMAG2 Set NCPF2 1 to use all components in FMAG2 Set NCPF2 0 to use no stations in FMAG2 Set NCPF2 the number of components to use for FMAG2 1 10 Supply the NCPF2 3 letter component codes to use in FMAG2 Examples FC2 0 donot calculate FMAG2 0 components or FC2 Z 2 VHZ VLZ define 2 components for Z magnitude Define duration magnitude corrections for named station c
183. roduct CAL x S is used for scaling in the magnitude calculation Seismometer instrument types The instrument type code can be specified in the station file or in the amplitude magnitude correction file read with the XMC command If the type for a station and component is specified in the correction file with a non blank entry it overrides the one given in the station file Types 0 and 2 invoke the M magnitudes Types 1 and 3 7 invoke Mx magnitudes for velocity seismometers Choosing the instrument type also invokes the seismometer motor constant S in the Mx relation S is in volts cm sec Hypoinverse does not have a magnitude relationship for accelerometers Relating acceleration to displacement Wood Anderson response for an estimate of earthquake source energy is a bit of a stretch In any case the seismogram phase of maximum displacement typically does not coincide with the phase of maximum acceleration For non typical seismometers it is better to convolve the seismogram to Wood Anderson response before measuring the maximum displacement phase Table 2 Seismometer instrument types Type S Instrument 0 Wood Anderson torsion seismograph 1 1 0 1 hz L4C velocity transducer with 0 8 critical damping Gain information if specified as a history file is stored as attenuation history 2 Hawaii type Sprengnether seismometer with optical recording 3 1 0 1 hz L4C velocity transducer with 0 8 critical damping Gain in
184. rvey Open File Report 23 pp Eaton J P 1992 Determination of amplitude and duration magnitudes and site residuals from short period seismographs in Northern California Bull Seis Soc Am v 82 no 2 pp 533 579 Geiger L 1912 Probability method for the determination of earthquake epicenters from the arrival time only translated from Geiger s 1910 German article Bulletin of St Louis University v 8 no 1 pp 56 71 Lahr J C 1980 HYPOELLIPSE a computer program for determining local earthquake hypocentral parameters magnitude and first motion pattern U S Geological Survey Open File Report 80 59 59 pp Klein F W 1978 Hypocenter location program HYPOINVERSE U S Geological Survey Open File Report 78 694 113 pp Klein F W 1985 HYPOINVERSE a program for VAX and Pro 350 computers to solve for earthquake locations and magnitudes U S Geological Survey Open File Report 85 515 Klein F W 1989 HYPOINVERSE a program for VAX computers to solve for earthquake locations and magnitudes U S Geological Survey Open File Report 89 314 59 pp Lee W H K and Lahr J C 1972 HYPO71 A computer program for determining hypocenter magnitude and first motion pattern of local earthquakes U S Geological Survey Open File Report Lee W H K R E Bennet and K L Meagher 1972 A method of estimating magnitude of local earthquakes from signal duration U S Geological Survey Open File Report 28 pp Lamson C
185. se Y2000 formats you must do it for all input and output file types The only files which have different Y2000 formats are those with dates stations files for example are the same whether in Y2000 format mode or not The exception is the traditional USGS phase format which is always in its old format even when the program is in Y2000 format mode Thus if you have selected Y2000 formats with the 200 command and are reading traditional phase format a default century 1900 or 2000 is added to all the 2 digit years and full 4 digit years are used for all output files I have written a separate program called 2000CONV to convert old Hypoinverse files to their new formats Supply the format flag T for Y2000 formats F for old formats Supply the default century to use when reading the traditional USGS phase format Supply the default amplitude units code when reading the traditional USGS phase format The Y2000 archive format has a field that is not available in the old phase format describing the units mm counts etc of amplitude measurement 0 zero is the default code for the traditional mm units Example 200 F 1900 0 Set the input phase data format See the phase data input format section for the formats Shadow records contain additional data which is not processed by Hypoinverse but just passed through Shadow records begin with a character and follow each header phase or terminator line 67 FIL CAR
186. shown the current value If you want to keep the value just press the RETURN key Type HELP or for a listing of commands and a very brief description or type MORE for additional commands All of the commands are indexed The index thus contains an alphabetical command list with bold face page numbers referring to this command dictionary Unless indicated otherwise by RIGHT NOW commands set names or parameters without reading in data or doing any processing Normally one issues several commands some of which read files RIGHT NOW then ends with the LOC command to start processing INPUT FILES CRT Read a linear gradient crustal model expressed as a travel time table RIGHT NOW See the section on crustal models The table is generated by the program TTGEN from a series of velocity depth points Model number 1 36 File name containing the travel time table Example CRT 2 REGION2 CRT there is no default CRH Read a homogeneous layer crustal model RIGHT NOW See the section on crustal models Model number 1 36 File name containing the layer depths and velocities Example CRH 3 LAYMOD3 CRH there is no default STA Read a station data file into memory RIGHT NOW See the H71 command for station file formats Supply file name containing stations Example STA 1983 STA there is no default PHS Set the phase data input filename See the COP command for selecting phase format In the
187. smic system relative to the Wood Anderson displacement seismometer The analog system has a 1 sec velocity seismometer and high frequency cutoff filtering Dividing by this response curve converts velocity amplitudes to Wood Anderson amplitudes and produces x local magnitudes Hypoinverse uses the function R f for all velocity seismometers These are instrument types 1 and 3 7 The tabulated R f includes the response of a 1 hz 0 8 critically damped velocity seismometer relative to a Wood Anderson high cut and low cut electronic filtering and the Develocorder galvanometer response which attenuates high frequencies The function R f is 50 used in the magnitude formula for all velocity response seismometers whether used with analog or digital recording As long as you are in a frequency band 3 hz lt f lt 10 hz where the responses are linear and unfiltered R should be a good approximation to most systems Scaling the response to different gains can be done with the CAL factor The Hypoinverse user could code a new response table if modeling a specific system outside this band is necessary Here are the R values used by Hypoinverse frequency 0 16 0 20 0 25 0 32 0 40 0 50 0 63 0 79 1 00 log frequency 0 8 0 7 0 6 0 5 0 4 0 3 0 2 0 1 0 0 log R 0 288 0 432 0 561 0 680 0 786 0 891 0 983 1 066 1 138 frequency 1 26 1 59 2 00 2 51 3 16 3 98 5 01 6 31 7 94 log frequency 0 1 02 03 04 05 06 07 08 09 log R 1 205 1 276
188. sssssssessssessesessssessessessossessrssresseeseesresseeseess 114 Using the program TTGEN oa iee acne sausevetcvaucearassoucxswaeecea caeesetsbotaieasiwoae i seesetsbeanieenese 114 Input to TTGEN in the file ETM Ds seis ov sega y tals dea abst aaees eae enet Sechdveseiaaes natin 114 Outputs of TTGEN 55 cc sits ccetesadacey ae aaa a a Bd aate agua A 117 Appendix 6 Programmers notes yds spacd shea nae ooaca sia seat oats ea ecadeatea tag espsale aka 119 USS A E BAG Ba ls EEEE A E Pee DEO ES AS Poe OT UR Pee E E eee e error 120 INDEX OF COMMANDS AND SPECIAL NAMES cccccessessesecseeeceeeececeeseeaeeseeseeneeeeensens 121 TABLE OF FIGURES AND TABLES Figure 1 A example of nodes defining the areas for two crustal models and the transition regions between then sernai a a stacey talc aia area adel hate ahaa alia a ag i Ronee a alt 10 Figure 2 An example of regions for mutiple crustal models ceesceesceseeeeeteeeeeeeeeeeeeneeneeess 12 Figure 3 Diagram of simple analog Seismic system ccceecceecessseceseceeeeeecesseecsuecneeneeeenaees 48 Figure 4 Response function R f of the standard NCSN analog seismic system relative to the Wood Anderson displacement seismometet c ccsceceseceseceeseeescecsseceseeeeeeeeseecsaecnseeeeeeeneeeenaees 50 Table 1 Codes for seismogram amplitude Units cccccceccseessecsteceeeceeeeeeseeceaeceeeeeeeenseeeaeees 52 Figure 5 Diagram of simple digital seismic Sy Stems a s cccess lt
189. stance and residual weighting are invoked As iteration proceeds you may thus find that the hypocenter starts at its trial location iterates away then returns to the trial location after all weighting takes the same effect it had to produce the earlier location On the other hand a good trial location may help Hypoinverse find the right RMS minimum or put the earthquake in the correct neighborhood if it is unconstrained and thus produce a better solution An optional 10 digit ID number may be supplied in columns 63 72 a right justified integer of the terminator line It will appear in the print archive and summary outputs Columns 1 4 of the terminator line must be blank Use the H71 command to select either the Hypoinverse or HYPO71 terminator formats or to get the trial hypocenter from the header if reading an archive format If all terminator lines are blank it does not matter which format is used Either format may be used to enter a trial depth or to fix the depth but only the Hypoinverse format allows a trial epicenter or origin time on the terminator line Hypoinverse Terminator format Start Fortran Col Len Format Data 1 4 A4 2X Must be blank 7 4 212 Trial hour and minute 11 4 F4 2 Trial second 15 2 F2 0 1X Trial latitude deg 18 4 F4 2 Trial latitude min 22 3 F3 0 1X Trial longitude deg 26 4 F4 2 Trial longitude min 30 34 5 F5 2 Trial depth a negative value fixes depth 35 1 A1 A non blank char
190. strument types to use with the second amplitude magnitude XMAG2 0 3 or 1 for all types Supply the first instrument type for XMAGZ2 Supply the second instrument type for XMAG2 Supply the third instrument type for XMAG2 Examples XTY 0 3 0 0 3 1 do not use any instruments to choose amplitude magnitude the default or XTY 3 3 0 2 2 0 2 0 Use seismometer types 3 0 and 2 for XMAGI1 and types 0 and 2 for XMAG2 The comma is only used as a visual separator and is not really needed Define amplitude magnitude corrections for named station components These corrections are in addition to all other corrections If you find that low gain horizontal stations typically give too large a magnitude for example you can apply a correction to all stations of that type without having to specify a correction for each station individually Supply the number of components that follow with individual amplitude magnitude corrections 0 4 Supply pairs of 3 letter component codes and amplitude magnitude corrections Examples XCM 0 no component corrections the default or XCM 2 VLE 0 3 VLN 0 3 negative corrections for two component codes LA zero Choose the log Ao relation to use with specific components The log Ao relation is the distance term in the amplitude magnitude relation The MAG command chooses the default log Ao relation to use with all components not specifically mentioned here The relationships
191. t only the estimated errors set ERCOF 0 This will give errors that include primarily the effects of array geometry 112 If you want to include effects of poorly modeled travel times such as uncertainties in the crustal or delay models then set ERCOF 1 ERCOF can be set to any positive value or 0 The covariance matrix is a4 x 4 symmetric matrix whose diagonal elements are the variances standard errors squared of origin time in sec and latitude longitude and depth all in km The off diagonal elements are the covariances between these quantities This allows for example a quantitative estimate of origin time error and the tradeoff between origin time and depth The error ellipsoid is specified by the 3 x 3 sub matrix with origin time removed Error ellipsoid and vertical and horizontal errors The error ellipsoid is specified by the 3 x 3 sub matrix derived by removing origin time from the covariance matrix The 3 x 3 spatial covariance matrix must be rotated into the principal coordinates of the solution whose axes are the major axes of the error ellipsoid The three principal standard errors are calculated by taking square roots of the eigenvalues diagonal elements in diagonal form of the 3 x 3 covariance matrix The earthquake then has a statistical probability of 32 of lying inside an ellipsoid of error whose major axes are given by the three principal standard errors An error ellipsoid whose major axes are 2 4 times the stand
192. t the Hypoinverse summary output filename 70 ARC MFL ERF APP Supply the filename Use NONE or none to omit a Hypoinverse summary file Example SUM NONE the default or SUM OUT SUM Set the archive output filename This file contains all the data calculated for each station and can be read back in for a later relocation Supply the filename Use NONE or none to omit an archive file Example ARC NONE the default or ARC OUT ARC Set the magnitude data output filename This file contains precise magnitudes and other data necessary to recalculate magnitudes or evaluate magnitude statistics Supply the filename Use NONE or none to omit a magnitude file Example MFL NONE the default or MFL OUT MFL Error messages for bad data station names etc are always written to the print file if one is specified They may also be sent to the terminal to spot errors during a location run by turning error reporting on The most serious errors are always sent to the terminal Supply a T to send error messages to the terminal F otherwise Example ERF F the default Set the 3 logical flags that indicate whether an existing output file is appended to T or whether a new one is created F Supply 3 logical flags for 1 printout file 2 summary file 3 archive file Example APP F F F the default or APP F T F to append only to the summary file MULTIPLE CRUSTAL MODELS M
193. ted this way because it is a median and not an average The label code for the preferred magnitude If an external magnitude is present on the input header to the archive file the next line gives information about the external magnitude It might be labeled Berkeley mag REGION The full name of the geographic region 90 MODELS USED If you are using multiple crustal models the 3 letter codes of those actually used in the solution and their individual weights are given There may be 1 2 or 3 models used depending on the node geometry at the epicenter Station list data STA NET COM DIST AZM AN P S WT SEC TOBS TCAL DLY RES WT SR INFO CAL DUR FMAG AMP PER Station 5 letter site code name An asterisk after the station indicates that it uses the alternate crust model see the ALT command and column 34 of the station format The 2 letter network code The 3 letter station component The 1 letter station component code normally treated as an abbreviation to the 3 letter component code The 1 letter station remark from the station line Epicentral distance Azimuth to station in degrees east of north Angle if emergence at the hypocenter in degrees up from nadir P or S remark code and P first motion Assigned weight code and weight out symbol if any Observed arrival time Observed travel time Calculated travel time Station delay actually used which might be a mixture o
194. teing anew anaes 100 APPENDIX oeaan minded eas ested cca aa Nada asta E ose acd eae da hed tose 107 Appendix 1 Rules for free format parameters following command5s cceeeeseeereeeeeeee 107 Appendix 2 Iteration and Convergence scc 36 0520 05i5 cea ote ehsashaneontetr edocs Genet ahaa eae nae 107 Wh ere to begin iterations eusirinn esdeas vend dues a hag a a A a Whee aa 107 How iterative steps may be modified nnsnnnesessseesseseesressessresseeseserssesseserssressessessressesse 108 Convergence and when to stop iterating Sasa s ode eeucestelnund cy Stcaas va cdauis se auatees eens Nay deutes vate 108 Appendix 3 Inversion scheme and use of eigenvalue cutoff cece ccceceteceeeeeeeeeeeeenseenes 108 The station importance and information density 00 0 0 eececsceesceeeteceeeceeeeeeseecaeceeeneeeenaees 111 Eigenvalue and covariance error OUlPUtssiha5 casieisncdss seaoerbalinapcad iedela Sharon 112 Appendix 4 Error c lculatiohS is 01 s ciaet eden aise ato lenee Settee E E E a tas 112 THE Covariance OM GrrOr Matrik canst seed te ised ai ashe aa aaron ae eS 112 Error ellipsoid and vertical and horizontal errors cceseeseeseeeseeeeceeceaecneeeeeeeeeeeeeaeenaes 113 Appendix 5 Generating travel time tables for linear gradient crustal models with program IEKE A E E E N E S E EA E A E tty S AN E et tgde aaa etna 113 Use ofa traveltme tall 6 oie aiia o EE a E k 113 Allowable crustal models input to TTGEN ssss
195. th both types of stations can be processed together For NCSN Earthworm digital stations no VCO and no attenuator setting you do not want to make a gain correction to the coda magnitude You can suppress the gain correction 1 by using a cal factor of 0 0 unknown gain defaults to a 15 db attenuation equivalent 2 by suppressing gain correction for some or all components see the DUG command or 3 turn gain corrections off on a station by station basis by adding 10 0 to the coda magnitude correction set in the station file or read with the FMC command Note that using a cal factor of 0 0 will preclude you from determining amplitude magnitudes for that station Coda magnitude options Duration magnitudes can also use a station gain correction if the gain is known Gain may be expressed on the station line as either a calibration factor or as an attenuation in db positive numbers like 12 18 etc See the ATN command All magnitudes are calculated and output to each file to a precision of 0 01 The exception is that magnitudes output to the pre Y2000 format summary and archive files are to the nearest 0 1 Weights can be assigned to codas just like they are to arrival times Weights can also be ignored even when they are specified so that either all or named stations with a positive coda time will get a magnitude that will be weighted equally with the others in computing the event magnitude see the MAG and FMC commands The event coda mag
196. the delays for all models in one file If all delays are in one file the column a delay uses must correspond to the model number used for that model when the crustal files are read I find it easier to keep one model s delays in one file Models are labeled on output by 3 letter codes and not their number The MUL command is used to select either single model or multiple model modes and to define the default model 11 425 9 124 123 122 121 120 119 118 Figure 2 An example of regions for mutiple crustal models Model regions are defined by the union of circles Each group of circles is a different model The circles are defined by NOD commands Only the inner circles are plotted see text and the regions between model regions use a weighted combination of models perhaps including the default model Alternate models Hypoinverse also has an alternate model capability permitting the use of different models by different stations for the same hypocenter A typical use of alternate models is where the crust is different on each side of a fault like the San Andreas earthquakes occur on or near the fault and the rays to all stations on each side of the fault should use their own model Typically this alternate model feature is most useful only where two models and sets of delays have been derived for two station sets Alternate models are most useful where a single vertical discontinuity in the crust is known Multiple
197. the fault the normal model for stations and rays to the west and the alternate model for stations and rays to the east 72 PROCESS EVENTS IN A PHASE FILE LOC BUG PRO BAS Locate events RIGHT NOW This is the command that actually locates earthquakes using the files and parameters set by previous commands There are no parameters except when locating individual CUSP events COP 6 where you supply the CUSP id number as the parameter Example LOC Check phase file for format problems and station file for missing stations and write error messages to the print file RIGHT NOW Phase station and print files must have been specified before issuing the BUG command Works only with ASCII input formats The BUG command has not been tested with all program options so if you encounter a program bug try using the slower LOC command to debug your phase file Example BUG Interactively edit and relocate earthquakes in individual files one event per file RIGHT NOW See the discussion above under Interactive Earthquake Processing and the BAS command below There are no parameters Example PRO Set parameters needed for building the input and output filenames used in interactive processing started by the PRO command The filenames consist are a concatenation of a base name unique to the event and a suffix file extension for each file type The file extensions normally begin with a period See the discussio
198. the top until a match of the code is found This matching test uses the number of letters in each station code that are specified with the LET command Any later station with the same code is ignored You can choose how many letters including 0 of each of the 4 fields to compare when doing this search If your net field is blank on both station and phase lines for example it does not matter how many letters 0 1 or 2 you compare because you will always get a match See the LET command Note as with the location code that if a net code is supplied in the station file but not compared because the number of comparison characters is set to 0 by the LET command the first matching NET code will be output to the archive file regardless of input NET code The exception to the requirement that all 4 station fields match is as follows after you read in the station location file you can read in a separate delay file When matching a delay to a station in memory only the site and net codes are compared Searching continues through the station list to assign the delay to all components and location codes at that site These are the lengths of station name fields in different formats later extensions are shown as Site code Net code Component Location Separate 1 letter comp HYPO71 4 no HI 1 4 0 2 0 3 0 2 yes HI 2 5 2 3 2 yes Specifying instrument types and calibration factors These comments apply to all three station formats and a
199. tion data is read into memory as soon as this command is given and is kept until another STA command is issued The station file must contain one line per station Use the H71 command to select either Hypoinverse old style or new style or HYPO71 file format There is no separate Y2000 format because the files do not contain dates Reading station files is more efficient if they are read in binary instead of ASCII You may create a binary station file after reading in all station data including the multiple model delays Use the WST command to write a snapshot of the station arrays to a binary file Read the file back in with the RST command The RST command replaces the STA and DEL commands The RST command does not replace the XMC or FMC commands to read magnitude corrections If used the XMC and FMC commands must be given after RST because the magnitude corrections are applied to stations already in memory The RST command does not replace the ATE or CAL commands to manage gain histories Ifa calibration factor is read from the station line with the STA command it is written to the binary file The calibration history and expiration dates however are not written to the binary file If you are using attenuation histories issue the ATE command after RST to read and dynamically update the attenuation s 14 Similarly use the CAL command for cal factor histories Tests show that binary reads using the RST command are several times faster than equi
200. to the archive and summary files Phases that are later down weighted because of large distance or residual are still counted in this total of weighted phases The earthquake will always be written to the print file so you will see the output even if there are too few readings MINSTA is 4 by default but increasing this will screen out the poorly recorded events The JUN command can save the data from poor earthquakes by writing them to the archive and summary files even if the final number of phases was not enough to get a good solution If you set the junk flag to TRUE you will force a solution of small and poor junk events If the 39 number of stations remaining after distance and residual weighting are applied is less than the minimum number set with the MIN command the event would normally abort Setting the junk flag to TRUE cancels all distance and residual weighting for events that would otherwise abort CODA MAGNITUDES This section discusses coda duration magnitudes but maximum amplitude and external magnitudes calculated elsewhere and read in may also be reported for each earthquake You can select from two coda two max amplitude and the external magnitude to be the preferred magnitude for the earthquake Each of these magnitude types should have a single letter designator L for M etc It is strongly advised to use a letter for each magnitude type and to keep the letter unique This will avoid confusion about which ma
201. tude S for south 27 3 13 1X Longitude degrees 31 7 F7 4 Longitude minutes 38 1 A1 W or blank for west longitude E for east 39 4 4X Reserved for elevation in m Not used by HYPOINVERSE 43 3 F3 1 2X Default period in sec at which the maximum amplitude will be read for this station Must be greater than 0 1 48 1 A1 Put a 2 or A here to designate this as an alternate crust model station Both alternate and primary crustal models must be in use Stations may also be tagged for use with an alternate model in the delay file 49 1 A1 Optional station remark field to copy to print output 50 5 F5 2 1X P delay sec for delay set 1 56 5 F5 2 1X P delay sec for delay set 2 62 5 F5 2 Amplitude magnitude correction If in the range 2 4 the correction is included by addition in the amplitude magnitude If you don t want a station s magnitude used in the event magnitude use a correction of 5 0 plus the actual correction You can also assign a zero weight see next 67 1 A1 Amplitude magnitude weight code Codes 0 9 and blank are used the same as the P amp S weight codes col 15 The actual magnitude weight used is the product of those on the station and phase lines See also col 62 68 5 F5 2 Duration magnitude correction works the same as the amplitude magnitude correction 73 1 A1 Duration magnitude weight code works the same as the amplitude weight code 74 1 11 Instrument type code 75 6 F6 2 Calibr
202. twork actually uses In what follows is just a visible separator Thus for a code like CALB NC VHZ you would use LET 5 2 3 Ifyouused LET 5 0 3 you would also match codes like CALB VHZ or CALB VHZ The full code from the station location file is the one used on output files Thus if CALB NC VHZ were in the station file and CALB VHZ were in the phase file and LET 5 0 3 were used the station would match and CALB NC VHZ would appear in the archive and print files Note that station delays from the DEL files are always matched to all components of the station site and network as if 0 were used for the comparison length of components Supply the number of station site code letters to test in making matches 2 5 Supply the number of station net code letters to test in making matches 0 2 Supply the number of station component code letters to test in making matches 0 3 69 UNK LAB Example LET 4 0 0 Set a list of 4 letter station codes or the first 4 of 5 letters that you expect to be in the phase but not station file Stations in this list will not produce an unknown station error message but can be archived in the ARC file see KEP command Other unknown station codes not on the UNR list can be output to the archive file but will see an error message each time they are encountered Specify the number of stations maximum 10 and that many 4 letter codes Stations will
203. ue is to put a non blank character in column 35 of the Hypoinverse terminator line This might be used for explosions for example This flag to fix the hypocenter will be copied to the output archive file and thus will remain in effect until it is deleted from the file 38 The origin time will be computed to minimize the average residual The non blank character will be copied to the output print file as for example the reason for fixing the location To fix the hypocenters for all events in a location run you prevent any iterations away from the trial hypocenter You do this by setting the maximum number of iterations ITRLIM to zero with the CON command Be sure to specify a trial hypocenter and origin time when using 0 iterations Also be sure to invoke the distance and residual weighting on iteration 0 DIS and RMS commands because Hypoinverse will always iterate until the distance and residual weighting begin All calculations including travel times will apply to your trial hypocenter The origin time may shift slightly from your trial value Hypoinverse always removes the weighted average station residual from the trial origin time This minimizes the RMS residual solves for an origin time and is useful for finding the origin times of known quarry shots for example Thus you can never actually fix the origin time in HYPOINVERSE The calculated travel times are of course independent of the origin time Here is a command file fragment to
204. ulated for each event Set the magnitude type number MAGSEL for the primary coda mag FMAGI 1 Traditional duration F P magnitude Md Lee et al 1972 DUR command 2 Elapsed Time tau magnitude Mt Michaelson 1987 TAU command 3 Second duration F P magnitude DUB command Say whether to use the coda weight on the phase line T or to ignore any coda weight F o T Use the coda weight on the phase line 0 or blank on the phase line is full weight 1 3 are partial weights and 4 9 is no weight o F Ignore any coda weight All codas are given full weight except those with a weight code of X or N Set the magnitude type number MAGSE2 for the secondary coda mag FMAG2 1 Traditional duration F P magnitude Md Lee et al 1972 DUR command 2 Elapsed Time tau magnitude Mt Michaelson 1987 TAU command 3 Second duration F P magnitude DUB command Choose the log Ao relation see amplitude magnitude section 1 Eaton BSSA 1992 Bakun and Joyner BSSA 1984 Richter s 1958 approximation UC Berkeley Nordquist BSSA 1948 P amplitude presently unused AUN Example MAG 1 T 1 1 the default 77 PRE Establish the criteria for choosing the preferred magnitude for each earthquake from the five magnitudes potentially available If a given magnitude is 0 0 it has not been computed and is never chosen as preferred The criteria for choosing a magnitude depend on its availability number of stations used and mag
205. und on the input archive header record and passed through to output o Ais the authority code i e what network furnished the information Hypoinverse passes this code through o Vis the version of the information i e what stage of processing or completeness the event is in This can either be passed through or can be set by Hypoinverse to all events in the current run to the code set with the LAB command o His the version number of last human review Hypoinverse passes this code through These 3 codes are in columns 114 163 and 164 of the Y2000 summary line Number of primary amplitude magnitudes used in the event magnitude XMAGI1 It is the total of their weights which is the same as the number if all weights are 1 89 XMMAD N FMG FMMAD T SOURCE Weighted median absolute difference of the primary amplitude magnitudes It is the error in the median magnitude calculated this way because it is a median and not an average The label code given XMAGI by the XC1 command Number of primary coda magnitudes used in the median FMAG1 It is the total of their weights which is the same as the number if all weights are 1 Weighted median absolute difference of the primary coda magnitudes It is the error in the median magnitude calculated this way because it is a median The label code given FMAGI by the FC1 command The code for the most commonly used data source L Location P amp S times
206. used The components select which stations are used for the earthquake magnitude Also set the 1 letter label for XMAG2 You would only be using this command except for setting the magnitude label if the XCH argument is T Supply the 1 letter label code for XMAG2 Supply NCPX2 the number of components to use in calculating FMAG2 Set NCPX2 1 to use all components in XMAG2 Set NCPX2 0 to use no stations in XMAG2 Set NCPX2 the number of components to use for XMAG2 1 7 Supply the NCPX2 1 letter component codes to use in XMAG2 Examples XC2 0 do not calculate XMAG2 no components or XC2 L 2 WLE WLN use 2 components for L magnitude Set the seismometer instrument types to use with the two different amplitude magnitudes Use the standard type numbers You must supply all parameters to this command even if you are not using some of the instrument types for a magnitude i e there must always be 8 arguments to this command You would only be using this command if the XCH argument is F See the section on seismometer instrument types in the amplitude magnitude chapter for the type numbers available Supply the number of instrument types to use with the first amplitude magnitude XMAG 0 3 or 1 for all types Supply the first instrument type for XMAGI Supply the second instrument type for XMAGI Supply the third instrument type for XMAGI 82 XCM LAO Supply the number of in
207. ute the earthquake solution This is because 2 letters are used to match location codes in station and phase files and station ABC 01 will not match station ABC 02 Because the station was moved slightly and thus both location codes did not operate at the same time you can use the same gain history and magnitude correction files for both locations by setting NSLOC2 0 Thus station ABC in the magnitude correction file matches both ABC 01 and ABC 02 in the station file and the correction is used for both stations If you put magnitude corrections for each station in the station file matching is not an issue and it does not matter what NSLOC2 is Note that if you read station delays from delay files and not in the station file a delay will always match and be assigned to all components and all location codes In other words the delay for ABC will be assigned to ABC V 01 ABC V 02 and ABC E If a non blank location code is used on input in the station file it will be passed to the output archive file This is true even if the location code is not used to match stations in the station and phase files Note this warning If two stations have identical codes except for different location codes their data can be treated independently and output to the archive file properly only if location codes are compared with NSLOC gt 0 If you do not check location codes NSLOC 0 or NSLOC2 0 station picks or other data will be applied to the first of the mat
208. valent ASCII reads Station name codes and file formats There are 3 different station file formats currently supported They differ in the amount of station information supplied and in the number of letters allowed in the station name code The HYPO71 format is supported for backwards compatibility HYPO71 allowed only 4 letters Hypoinverse format 1 maintained the 4 letters but added other letters on in various places as needed Hypoinverse format 2 supports the full 12 letters and thus supports the full IRIS and SEED specifications and stores locations to 0 0001 minute The H71 command chooses the format to be used The complete code consists of a 5 letter site code assigned by each network a 2 letter net code assigned each network by Tim Ahern at IRIS a 3 letter component or channel code for different records at each site and a 2 letter location code to further discriminate channels 12 letters with the 2 letter location code extension are now available because of the proliferation of stations and the need to avoid duplication Hypoinverse also stores a 1 letter component field as an abbreviation to the full 3 letter field because this has been a common historical practice and because 3 letter components are often confusing to people Output files carry both the 1 letter and 3 letter components Although conceptually different we have often mixed letters representing these 3 types in the same field for practicality Thus CAL CALE and
209. waii the default or DUR 87 2 0 0035 0 5 0 9999 O traditional CALNET or DUR 2 06 2 95 0 001 0 5 0 9999 1 CALNET low gain DUR 81 2 22 0 0011 0 5 0 9999 1 Eaton 1992 Define the constants used to compute the second magnitude FMAG2 from coda duration F P time The parameters are exactly the same as the DUR command but are used for the second duration magnitude The parameters may have different values than the DUR command or may be used with another set of stations see the MAG and FC2 commands No component corrections FCM command or additional distance corrections DU2 command are used with FMAG2 Set additional constants for Eaton s special terms in the Md relation Recall Eaton s relation Mp f p 0 81 2 22 log fp 0 0011 D Stacor G Component correction 0 005 D 40 if D lt 40 km 0 0006 D 350 if D gt 350 km 0 014 Z 10 ifZ gt 10km Supply the coefficient of the D DBRKM1 term Supply DBRKM1 the distance less than which to apply the D DBRKM1 term Supply the coefficient of the D DBRKM2 term Supply DBRKM2 the distance greater than which to apply the D DBRKM2 term Supply the coefficient of the Z ZBRKM term Supply ZBRKM the depth greater than which to apply the Z ZBRKM term Example DU2 6 0 apply no corrections the default or DU2 0 005 40 0 0006 350 0 014 10 Eaton s Md formula Choose the compo
210. was defined to see if the epicenter lies in its inner circle If it does lie in an inner circle it is assigned 100 to that model even if it is also in the inner circle for a later node 2 It is best to have the inner regions surrounding nodes be the high seismicity areas such as faults and the transition regions between models be areas of low seismicity One then eliminates the complex mixing of different models from influencing the patterns of seismicity 3 Allow wide transition zones between the inner circles of different models In other words avoid having 10 km radius inner circles with 1 km wide transition zones This will cause rapid or even sharp transitions between models that may cause epicenters to locate on arcs or form other strange patterns Figure 2 shows a sample of regions defined for Northern California The transition zones between models are probably too narrow in many places but they are generally locations of low seismicity Each crustal model is read separately with a CRH or CRT command This associates each model with a model number and a 3 letter code the beginning of the model name This 3 letter code labels the model in all of the output files Each model has its own set of station delays The DEL command reads a file containing station codes and delays one station per line The DEL command also associates the delays in the file with a model number You may have a different delay file for each model recommended or put
211. y of the station code and one attenuation value The attenuation is in db and is a multiple of 6 0 6 12 60 The old format does not have the space for the century but the Y2000 format does The choice of which format is used depends on the selection made with the 200 command Specifying a calibration factor of 0 or leaving the field blank means that no magnitudes will be computed for this station even if an amplitude is supplied The relation between the CAL factor and attenuation setting is log CAL 0 05 atten 1 35 atten 0 6 12 18 24 30 36 42 48 56 60 cal 25 2 11 65 5 576 2 795 1 4 0 702 0 352 0 176 0 0885 0 044 0 0222 21 atten 66 T2 78 cal 0 0111 0 00557 0 00279 Attenuation history file format pre Y2000 Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 2X Station net code 11 3 A3 1X Station component code 15 2 12 1X First attenuation setting Must be a multiple of 6 db 18 8 412 3X Y M D H expiration date and hour of first attenuation Minutes if supplied will not be used Leave the expiration of the last attenuation blank to indicate no expiration 29 2 12 1X Second attenuation 32 8 412 3X Second expiration date Format repeats 7 12 1X 412 3X Attenuation history file format Y2000 century compatible Start Fortran Col Len Format Data 1 5 A5 1X Station site code 7 2 A2 Station net code 9 2 A2 Station location code 11 3 A3 1X Station co

Download Pdf Manuals

Related Search

Related Contents

Configuração  

Copyright © All rights reserved. Failed to retrieve file