[cig-commits] r21092 - in seismo/3D/ADJOINT_TOMO/measure_adj: . USER_MANUAL

[liuqy at geodynamics.org]

Author: liuqy
Date: 2012-11-30 14:02:43 -0800 (Fri, 30 Nov 2012)
New Revision: 21092

Modified:
   seismo/3D/ADJOINT_TOMO/measure_adj/USER_MANUAL/measure_adj_manual.pdf
   seismo/3D/ADJOINT_TOMO/measure_adj/USER_MANUAL/measure_adj_manual.tex
   seismo/3D/ADJOINT_TOMO/measure_adj/measure_adj.f90
Log:

Take sac headers from synthetic files.



Modified: seismo/3D/ADJOINT_TOMO/measure_adj/USER_MANUAL/measure_adj_manual.pdf
===================================================================
(Binary files differ)

Modified: seismo/3D/ADJOINT_TOMO/measure_adj/USER_MANUAL/measure_adj_manual.tex
===================================================================
--- seismo/3D/ADJOINT_TOMO/measure_adj/USER_MANUAL/measure_adj_manual.tex	2012-11-30 12:11:31 UTC (rev 21091)
+++ seismo/3D/ADJOINT_TOMO/measure_adj/USER_MANUAL/measure_adj_manual.tex	2012-11-30 22:02:43 UTC (rev 21092)
@@ -272,7 +272,7 @@
 MTMs are rejected (\verb+ user_trace = .false.+) based on:
 \begin{itemize}
 \item number of cycles in the window
-\item number of frequency points are given within \verb+[fstart,fend]+. Even when \verb+RUN_BANDPASS+ is turned off, the \verb+TLONG+ and \verb+TSHORT+ are still converted to \verb+fstart,fend+. They are best set to the values used in \verb+FLEXWIN+ 
+\item number of frequency points given within \verb+[fstart,fend]+. Even when \verb+RUN_BANDPASS+ is turned off, the \verb+TLONG+ and \verb+TSHORT+ are still converted to \verb+fstart,fend+. They are best set to the values used in \verb+FLEXWIN+ 
 \item \verb+tshift_cc <= dt+: too small a time shift
 \item \verb+ abs(dtau_w(j)) > Tvec(j)/DT_FAC+: $\tau$ at a specific frequency is too high
 \item \verb+ err_dt(j) > Tvec(j)/ERR_FAC+: error for a specific frequency is too high 
@@ -639,11 +639,15 @@
 \subsection{DO\_RAY\_DENSITY\_SOURCE}
 Both \verb+sigma_tau+ and \verb+sigma_dlnA+, as well as measurements \verb+Delta_dtau/DlnA+ are set to be 1, 
 and carried all the way to the adjoint source \verb+tr/amp_adj_src(:)+. \verb+DO_RAY_DENSITY_SOURCE+ automatically 
-sets \verb+Error_type=0+.
+sets \verb+Error_type=0+. It provides a good indication of the volumetric coverage of the model domain by the
+given measurement sets.
 
 \subsection{MTM}
 \verb+dtau_mtm(:)+ starts from $i > 1$, i.e., ignore the time shift at very long periods, and equivalently,
 the first point of \verb+err_dt(1) = LARGE+ so that it does not contribute to adjoint sources and chi values.
+
+\subsection{Questions}
+Should we weigh data by their noise level (before the first arrival)?
 %-------------------------------------------------------------------
 
 \pagebreak

Modified: seismo/3D/ADJOINT_TOMO/measure_adj/measure_adj.f90
===================================================================
--- seismo/3D/ADJOINT_TOMO/measure_adj/measure_adj.f90	2012-11-30 12:11:31 UTC (rev 21091)
+++ seismo/3D/ADJOINT_TOMO/measure_adj/measure_adj.f90	2012-11-30 22:02:43 UTC (rev 21092)
@@ -69,7 +69,6 @@
 
   read(11,*,iostat=ios) npairs
   if (ios /= 0) stop 'Error reading number of pairs of data/syn'
-
   print *, 'reading in the data and synthetics...'
 
   ! output files
@@ -93,8 +92,9 @@
 
     ! read data and syn (in double precision)
     ! LQY: data is read last to be stored in memory for header retrieval later
+
+    call drsac1(datafile,data,npt1,t01,dt1)
     call drsac1(synfile,syn,npt2,t02,dt2)
-    call drsac1(datafile,data,npt1,t01,dt1)
 
     if (DISPLAY_DETAILS) then
        print *
@@ -130,8 +130,8 @@
        call bandpass(syn,npts,dt,fstart0,fend0)
     endif
 
-    ! find out station/network/comp names,etc
-    call get_sacfile_header(trim(datafile),yr,jda,ho,mi,sec,net,sta, &
+    ! find out station/network/comp names,etc from synthetics
+    call get_sacfile_header(trim(synfile),yr,jda,ho,mi,sec,net,sta, &
                           chan_dat,dist,az,baz,slat,slon)
 
     ! theoretical surface wave arrival time