TPWT

双平面波

data

1. 按照地震目录切分

data 文件夹下是z 分量 1Hz 的 sac 文件。

切分事件的脚本：

ls data/*.sac > data_z.lst
mktraceiodb -L done_z.lst -O data_z.db -LIST data_z.lst -V
cutevent -V -ctlg event.cat -tbl data_z.db -b +0 -e +10800 -out SAC
rm done_z.lst data_z.db data_z.lst

需要准备 event.cat，且相邻时间不应小于3小时：

2023/01/01,10:10:10
2023/01/02,20:20:20

10800 为截取地震时间开始3小时内的数据，输出文件夹在 SAC 下，按照事件进行了分类。

2. 准备台站和事件信息

event.lst 包含 事件(12位)、经度、纬度、震级 四列 station.lst 包含 台站、经度、纬度 三列

3. 修改格式

• 上述步骤产生的地震事件是14位的，要改成12位，即年月日时分，不要秒。
• sac 的头文件信息可能被改了，确保有 dist，这需要台站位置和事件位置。
• sac 的 channel 要修改为 LHZ
• 最终单个 sac 文件路径为：SAC/202301011010/202301011010.staname.LHZ.sac

台站名要和头文件的 kcpnm 以及 station.lst 的台站信息对应。

run

01

01.bash

Generate reference phase velocities for each pair of event and station, then extract dispersion curves for each sac file.

Prepare files:

• event.lst: event lo la ma
• station.lst: station lo la

1. use

   mk_pathfile match event and station

   TPWT/utils/RAYL_320_80_32000_8000.disp raylei global model

   TPWT/utils/LOVE_400_100.disp love global model

   aftani_c_pgl extract snr curves.

   spectral_snr contral by snr, need use csh, I have modified the .csh to .bash.

2. set

3. generate

flag_GDM52 无意义，标志 01 跑完了

path/ event_station.PH_PRED and GDM52_dispersion.out

SAC/event/ 1.bash, event.station.sac_1_DISP.0, event.station.sac_1_DISP.1, event.station.sac_2_DISP.0, event.station.sac_2_DISP.1. Only event.station.sac_1_DISP.0 will be used

event.station.sac_snr.yyj.txt filelist, flag_SNR, flag_aftan, param.dat, step.txt, temp.dat bash 01.bash

02

02.bash

1. use

   find_phvel_amp_earthquake format the pinsan according to event and zhouqi. generate .ph.txt correct_tt_select_data input mid points coordinate to remove bad points. Keep points which can satisfied less than the tmisfit.

2. set

   cutoff_per=0.7  # 有效台站记录在70%以上，则事件有效
   snr=10
   dist=3500  # the longest distance between two stations.
   tmisfit=10  # 走时误差超过 10s 则舍弃 

   correct_tt_select_data 设定参考台站坐标 plot_phase_time 设定反演范围

3. generate

   sta_dist.lst

   all_events_20150/{t}sec{snr}snr_3500dis/ {event}_am.txt, {event}_am.txt.HD{event}_am.txt_v1, {event}_am.txt_v1.HD, {event}.ph.txt, {event}.ph.txt.HD, {event}.ph.txt_v1, {event}.ph.txt_v1.HD, tomo.grd

   {event}.ph.txt will be used in 02.bash self.

03

1. use

   • wc
   • sdiff

2. set

   • nsta_cutoff = 15
   • Region: la and lo of region, the range should < 10 degree

3. generate

   • SAC/ tempeq and eqlistper: total number of stations to one event and the order number of the event.

04

1. use

   • creatgridnode_TPWT : create fan yan wang ge in invrsnodes.
   • avgvel : average v
   • echo 5 {smooth_len} 0.25 0.25 >>{new_eqlistper} :0.25 can be changed to 0.2 or not.
   • rd : select event, should notice that the event number in sacfile name must be 12.

2. set

   • snrcutoff = 10
   • tcutoff = 10
   • dist_cutoff = 3500 need > 3500
   • nsta_cutoff = 15
   • smooth_len = 80
   • damping = 0.2

3. generate

   • sens_dir/
   • TPW_{snrcutoff}snr_{tcutoff}t_{dist_cutoff}_dist
   • invrsnodes fan yan wang ge jie dian.
   • eqlistper (new): old eqlistper + some new information. detail; summar; invrnodes; pham - resort of .ph.txt generated before.

05

repeat 04, chose pham.V3

Results are in new_2d directory.

06

check board.

final result

The final result is 2 grd.* files in new_2d, .ave is perturbation.

And there are two std* files representing standard deviation in the new_2d.