This directory contains event locations in southern California as computed by the source-specific station term, waveform cross-correlation and cluster analysis methods. The catalog is described in: Shearer, P., E. Hauksson and G. Lin, Southern California hypocenter relocation with waveform cross-correlation, Part 2: Results using source-specific station terms and cluster analysis, Bull. Seismol. Soc. Am., 95, 904-915, doi:10.1785/0120040168, 2005. The following 6 files should be in this directory: SHLK_1.0 -- File with locations (340330 events, 42 Mbytes) see below for format description SHLK_1.01 -- Update with some minor corrections (340291 events) (see below for details) SHLK_1.02 -- Update with corrected magnitude information (see below for details) SHLK_IMP_1.0 -- File with Imperial Valley locations (22989 events) computing using a special velocity model. These locations are preferred for this region. locmap.ps -- Postscript map of locations (16 Mbytes) This was generated using GMT. SSST locations are colored by year; the waveform cross-correlation locations are plotted in black. This plot is scaled for regular size printers; for large plotters, add, for example, the line 4 4 scale near the beginning of the file in order to increase the size by a factor of 4. polymap.jpg -- Map showing polygon boundaries (see below) in jpeg format FORMAT DESCRIPTION The locations are in the following custom format with 125 character lines (127 in SHLK_1.02): 1 2 3 4 5 6 7 8 9 10 11 12 1234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567 year mo dy hr mn sec cuspid lat lon dep mag np ns rms rmed clnum nclst nlnk err_h err_z 1987 12 4 21 32 52.400 737950 33.01205 -115.84647 6.009 2.4 5 1 0.07 0.01 5 0 1 220 100 23 0.026 0.068 l where 1987 = year 12 = month 4 = day 21 = hour 32 = minute 52.400 = second 737950 = SCSN cuspid (up to 9 digits) 33.01205 = latitude -115.84647 = longitude 6.009 = depth (km) 2.4 = SCSN assigned magnitude (0.0 if unassigned) (field shifted to left in SHLK_1.02 to allow 0.01 precision) 5 = number of P picks used for SSST location 1 = number of S picks used for SSST location 0.07 = rms residual (s) for SSST location 0.01 = median absolute value residual (s) for SSST location 5 = polygon region number for location (see below) 0 = local day/night flag (=0 for day, =1 for night in Calif.) 1 = location method flag (=0 for SSST, =1 for waveform cross-correlation) 220 = similar event cluster identification number (unique for each polygon) 100 = number of events in similar event cluster 23 = number of links to other events used to location this event 0.026 = est. std. error (km) in horz. position relative to other events in cluster 0.068 = est. std. error (km) in depth relative to other events in cluster l = SCSN flag for event type (l=local, r=regional, q=quarry) (this field is present only in the SHLK_1.02 file) Notes: (1) The last 6 fields are set to zero in the case of SSST locations. (2) 207628 (61%) of the locations are in similar event clusters and use the waveform cross-correlation times. (3) SSST locations are only computed to 0.0001 degree in lat/lon. (4) No standard errors or overall quality ratings are provided for SSST locations, but location quality can be estimated from np, ns, rms and rmed. (5) The median absolute value residual is often zero for small numbers of picks. (6) Estimated standard errors can become very large for small numbers of links -99.00 is used when the error is undefined or exceeds the output format. **** 140.007 in horz. position error should also be -99.00 (undefined) This is a bug in SHLK_1.0, fixed in SHLK_1.01 **** (7) The day/night flag is based on UT=3 to 15 being "night" in Calif. (8) Absolute location accuracy is poor (particularly in depth) in regions not well represented by the 1-D velocity model, such as the Salton Trough and the offshore Borderland region. (9) In SHLK_1.0, there are 39 events for which the rms residual = NaN These locations are not reliable and have been deleted in SHLK_1.01 CHANGES FOR the SHLK_1.02 CATALOG The magnitudes in the 1.0 and 1.01 catalogs are from the original STP phase files. Later we noticed that these magnitudes did not agree with those in the final SCSN catalogs, indicating that they were preliminary and probably not as accurate as they could be. To fix this problem, we obtained magnitudes from the catalog files at the SCEDC website at www.data.scec.org/research.html, using the SCEDC format, which gives the magnitudes to a precision of 0.01. At the same time, we decided to add a flag indicating the event type because it is useful to know which events are probably quarry blasts. Here are some numbers describing updates in the SHLK_1.02 catalog: 50,906 events had zero magnitudes in the SHLK_1.01 catalog and most of these now have non-zero magnitudes in SHLK_1.02. 229 events were changed to M = 0 from non-zero magnitude in SHLK_1.01. 1080 events did not have a cuspid in the SCEDC catalog that matched the original cuspid, but there were matching events within 10 s and 0.08 degrees. The cuspid is updated in SHLK_1.02 for these events but is given as a negative number. 3568 events (about 1% of the total) in SHLK_1.01 could not be found in the SCEDC catalogs. This is rather mysterious! We retained these events and their original magnitudes in SHLK_1.02, flagging them with M (for missing) in the final column. VELOCITY MODELS Locations for SHLK are based on the following 1-D velocity model: Depth (km) Vp (km/s) Vs (km/s) 0.0 4.0 2.31 6.0 6.2 3.58 32.0 6.8 3.93 32.0 7.8 4.50 60.0 7.9 4.55 Please note that this is NOT a layer-cake model; there are linear velocity gradients between model points. For example, Vp is 5.1 km/s at 3.0 km depth. Locations for SHLK_IMP are based on the following model: Depth (km) Vp (km/s) Vs (km/s) 0.00 1.80 0.7595 1.80 3.00 1.402 4.55 5.10 2.851 5.00 5.3475 3.0874 5.55 5.65 3.262 11.30 5.95 3.435 11.30 6.60 3.812 12.30 7.2 4.157 20.00 7.76 4.480 30.00 7.86 4.538 POLYGON PARAMETERIZATION SHLK locations for were first performed in five different polygon shaped regions. The file polymap.jpg shows these regions. Note that we do not include results from polygon 1 in our final catalog because it is on the edge of the southern California network and we did not obtain good results for these events. The boundaries of the polygons are as follows: 1 8 15081 -122.327 38.9108 -115.330 38.9108 -115.330 37.7977 -116.833 36.7096 -118.891 36.7096 -119.820 35.0829 -122.327 35.0829 -122.327 38.9108 2 10 96578 -118.966 36.8096 -116.883 36.8096 -115.380 37.8977 -114.300 37.8977 -114.300 35.3635 -117.266 35.3635 -117.266 34.7019 -119.891 34.7019 -119.891 35.1684 -118.966 36.8096 3 9 70949 -122.113 35.1840 -119.792 35.1840 -119.792 34.7948 -116.786 34.7948 -116.404 33.7880 -117.175 33.7049 -117.834 33.2430 -122.113 33.2430 -122.113 35.1840 4 7 84561 -117.342 35.4230 -114.000 35.4230 -114.000 33.6492 -116.395 33.6492 -116.813 34.7191 -117.346 34.7191 -117.342 35.4230 5 9 75885 -117.868 33.2964 -117.229 33.7441 -116.420 33.8312 -116.367 33.7074 -113.600 33.7074 -113.600 30.8569 -119.671 30.8569 -119.671 33.2964 -117.868 33.2964 The polygon number is given, followed by the number of vertices (plus one, as the first and last vertices are the same) and the number of events (1984-2002) contained inside. The polygon boundaries overlap slightly with each other. Because of this overlap, some events were located in more than one polygon. For the final catalog, one of the possible locations was randomly selected.