When using the grid search option for start location, an extra ouput is given in the print.out file:
Grid rms smaller than solution rms 1.953 3.391
Hypocentral dist. (km) from epicenter to grid point min. 28.3
Depth change 10.00
Iterate again with new start location: 62.303 5.760 15.0
This information gives how far away the minimum is and what the RMS is.
The most likely local minimum is happening for the depth. The start depth (on control line in STATION0.HYP) is therefore important and should be what is expected for the area. If there is a large variation in depths, like in a subduction zone, and the network is small, it might be an advantage to use several start depths. This can be done in two ways:
1: Set a range of start depths on the control line. In this case a complete location is done with each start depth. Setting many depth will increase the processing time proportionally.
2: Set RESET TEST(96) to 1.0. With this parameter set, the program will, once a location is found, look for the depth with the smallest RMS using the loction found. A new location is now made with the a start depth location at the minimum rms depth. This procedure is fast since the mimumum RMS depth is found with a fixed epicenter and no new location is found, however, since a new location is not found, it will not always work as well as option 1. The procedure for this option is to start with a depth 3 times the current depth (set to at least 40km) and then calculate rms in 40 steps from that depth to the surface. In SEISAN before version 12.0, the procedure was to start with current depth and the go in steps of 10 to the surface but this did not work well with shallow earthquakes.
Peter Voss : Fri Nov 12 10:33:10 UTC 2021