Abstract. We compile a worldwide database of about 2,300 shallow intracontinental earthquakes (1900-1994) from original publications and other recent compilations. Information on earlier historical earthquakes from these existing compilations are also included. We consider earthquakes in both the active and stable continental regions. Within active regions we distinguish the earthquakes on the few faults known to have high-long term slip rate >= 10 mm/yr. Besides the basic information about date, origin time, magnitudes, epicentral coordinates, and depth, we include the following source parameters wherever possible: strike, dip, rake of the two nodal planes, seismic moment, source-time function, and dimensions of the rupture zone. We also give a classification of the mechanism type, references and a ranking of the quality for various parameters. The catalog is presented as a flat table in an ASCII file format. The resulting database is suitable for seismotectonic studies and calculations of empirical relationships among the different parameters.
Our goal is to analyze intracontinental shallow earthquakes (depth <= 45 km) on a global basis. In this study we refer to intracontinental earthquakes the ones not related to subduction zones. Thus, we exclude events at subduction zones that are at the plate interface or in close proximity to it in the continental upper plate and those in the down going oceanic lithosphere. Since worldwide catalogs of earthquakes are dominated by subduction zones events, then, on average, relationships among its parameters mainly reflect the characteristics of subduction events. Earthquake catalogs that compile intracontinental seismotectonic information alone on a worldwide basis have been lacking until recently, although detailed seismological studies and compilations existed in several continental areas of the world. In part what has been missing is a simple selection of the boundaries that individualize and separate different intracontinental regions of the world from subduction zones.
On the other hand, for seismotectonic studies there has been a need to incorporate revised locations, magnitudes, seismic moment, and others focal mechanism parameters available in the literature into existing catalogs. The Centroid Moment Tensor Catalog (CMT) of Harvard University  includes the determination of the seismic moment and focal mechanism parameters for worldwide earthquakes of Mw >= 5.0 from 1977 to the present. Therefore, it was desirable to compile similar information for earthquakes in earlier years wherever it was available. The CMT solutions are in general good determinations of various earthquake parameters. In many cases they are the only available source. In some cases, however, other specific studies about particular earthquakes lead to better or at least alternative determinations. The CMT-depth of earthquakes is often a fixed parameter in the inversion, and therefore in those cases it is not entirely reliable. This occurs particularly for shallow earthquakes, which are the most common events in intracontinental areas. Then, even for the period where the CMT solutions are available a revision of its parameters and cross-checking with the ones existing in the literature is needed.
The two recent compilations of intracontinental earthquakes by the Electric Power Research Institute , which we hereafter refer as EPRI  or the EPRI report, and by Wells and Coppersmith  are very important contributions.
The EPRI  report includes nearly 1,000 shallow earthquakes (Mw >= 4.5) from 1900 through 1990 and also information on large earlier historical earthquakes in so- called Stable Continental Regions (SCR). Their compilation includes focal mechanism parameters and dimensions of the rupture wherever possible. The EPRI definition of SCR regions is a restriction to the continental crust of having no significant tectonism, volcanism, or anorogenic intrusions younger than early Cretaceous, and no rifting or major extension or transtension younger than Paleogene. Thus, their definition is based on geology, not on rates of seismicity. Because of uncertainties in the earthquake locations and/or in the SCR boundaries, less than 5% of the earthquakes in the EPRI data set is categorized as "transitional", meaning that the possibility exists that they might have occurred in oceanic or active continental crust.
Wells and Coppersmith  compiled a worldwide database of source parameters for 244 intracontinental earthquakes (depth < 40 km, Mw > 4.5, from 1857 through May, 1993) from which they developed empirical relationships among magnitude, rupture dimensions (length, width, area), and surface displacement. They also gave an additional list of 187 earthquakes with insufficient information or poor- quality data. The earthquake mechanism is indicated only by a classification of the slip type, but strike, dip, and rake of the nodal planes or assumed fault plane are not given. The catalog does not list coordinates, depths, and origin times of the earthquakes but instead, the location by region, locality and/or name of each earthquake is given. The Wells and Coppersmith  compilation is complementary to that of EPRI  in the sense that it provides information on earthquakes in Active Continental Regions (ACR).
Another source of information is the Pacheco and Sykes  compilation of shallow Mw >= 7.0 earthquakes from 1900 through 1989, which includes about 120 intracontinental earthquakes. The catalog is homogeneous in the surface wave magnitude and includes seismic moment determinations and classification of the slip type when available although no criterium was given for this classification.
Since our purpose is to examine the ACR and SCR intracontinental earthquakes, a unique database with earthquake in both regions was needed. We also believed that more information could be added or revisions could be made to the above mentioned sources particularly in the ACR regions. This belief was based on the fact that at the time of the EPRI  and Wells and Coppersmith  publications we already had started a compilation from primary sources on intracontinental earthquakes.
In this study we compile information from seven large continental areas that include most of the continental regions of the world and organize a database that includes about 1,300 and 950 Mw >= 5.0 earthquakes (1900-1994) in active and stable continental regions respectively (Figure 1). Triep and Sykes  use data from the catalog for the period 1978-1994 for Mw >= 5.0 and 1990-1994 for Mw >= 7.0. They describe some additional aspects of data for those two periods, including a map of active and stable continental areas and epicenters in them for those two intervals.
The main data sources are as follows: a) publications on earthquakes in intracontinental regions; b) the EPRI  catalog of earthquakes; c) the Wells and Coppersmith  compilation; d) the CMT catalog [Harvard University, 1995]; e) the Pacheco and Sykes catalog . Information from these sources was combined to produce an unique and improved listing, i.e. we chose what we believe are the best estimates of the source parameters for each event. A large percentage of a) corresponds to the seismological instrumental era which is from 1964 to the present. We did not search for information on earthquakes in previous centuries, but we included the existing information of those earthquakes in the sources b) and c).
Continental Areas Included
We mainly consider the following seven large continental regions: Africa, Asia, Australia, Europe, Middle East, North and South America. The dividing lines between the areas we refer to as Asia, Middle East, and Europe are somewhat arbitrary. We refer to Middle East the Eurasian region with longitudes greater than or equal to 25 E and smaller than 64.5 E (includes Turkey and the Arabian peninsula). The latter longitude is taken as the dividing line from Asia.
We did not compile information from primary sources (i.e., from original information in the literature) for Europe and Antarctica, or small areas of known or questionable continental crust in Indonesia, New Guinea, New Britain, New Ireland, the Philippines, Taiwan, Japan, Sakhalin, New Zealand, and Central America as well as backarc basins (e.g. the Aegean). However, for the earthquakes in Antarctica and SCR of Europe we included the existing information in the EPRI report, and for the mentioned small areas and ACR of Europe the information in the Wells and Coppersmith  catalog.
Stable and Active Continental Regions
We follow the EPRI  report in defining the Stable Continental Regions. Our boundaries for these regions are polygons extracted from the original boundaries on the EPRI's maps (Figure 1). We refer to Active Continental Regions the areas that do not satisfy at least one of the restrictions imposed in the definition of Stable Continental regions (see Introduction). The boundaries of the Active Continental Regions are defined somewhat more arbitrarily (Figure 1). We approximately follow the outer edge of the continental margin or in some cases we follow the lines that separate the seismicity related to subduction from the intracontinental seismicity as is the case in the western boundary of South America.
In a few cases, the boundaries of the active regions are arbitrary lines which exclude the small intracontinental regions mentioned above for which we did not compile information from primary sources. However, these areas do not have stable continental crust. Thus, all of them should be categorized as active, except for part of the New Zeland crust for which the category of "submerged continental crust" may be more appropriate [EPRI, 1994]. For simplicity in our catalog, we refer to these areas by abbreviations of their names without specifying that they are active areas. The areas explicitly catagorized as active in our catalog refer to the areas defined by the corresponding polygons in Figure 1.
Within active areas we do distinguish a few faults that are known to have high long-term slip rates (equal or greater than 10 mm/yr) and call them either Type 1 [Scholz et al., 1986] or interplate faults (and events on them Type 1 earthquakes). These include the San Andreas, San Jacinto, Queen Charlotte, Fairweather, North Anatolian faults, and the most active faults of the Himalayan region. Earthquakes that occur on these faults are categorized separately in our catalog.
Description of the Catalog
The database is organized in an ASCII table with 38 columns or "fields" with only one entry for each earthquake and can be accessed by both this text file and ftp URL: ftp://www.ldeo.columbia.edu/pub/triep/intra.cat The fields are as follows:
2 origin time
3 continental region
5 source of information
6 active, stable or Type 1 ("a", "s", or "i" respectively)
7 latitude (deg. north and south of equator)
8 longitude (deg. east and west of Greenwhich)
9 reference for coordinates
10 mb (body wave magnitude)
11 Ms (surface wave magnitude)
12 Ms reference (or mb reference if there is no Ms)
13 Mw (moment magnitude)
14 Mw reference
15 depth (km)
16 depth reference
17 depth quality (A, B, C, D, or "-")
18 a (factor a of the seismic moment expression: Mo = a x 10**b N m)
19 b (exponent b of the seismic moment expression: Mo = a x 10**b N m)
20 Mo reference
21 Mo quality (A, B, C, D, or "-")
22 strike of nodal plane 1 (deg)
23 dip of nodal plane 1 (deg)
24 rake of nodal plane 1 (deg)
25 strike of nodal plane 2 (deg)
26 dip of nodal plane 2 (deg)
27 rake of nodal plane 2 (deg)
28 mechanism reference
29 mechanism quality (A, B, C, D, or "-")
30 mechanism type
31 source time duration (sec)
32 reference to source time duration
33 source time duration quality (A, B, C, D, or "-")
34 surface rupture length (km)
35 subsurface rupture length (km)
36 rupture width (km)
37 reference to rupture dimentions
38 rupture dimentions quality (see explanation below)
Field 2 is one of the following:
CE Central America
ME Middle East
NZ New Zealand
NG New Guinea, New Britain, and New Ireland
NA North America
SA South America
Field 5 indicates data from the following sources:
trsy Triep and Sykes- a compilation from sources in the literature given in
Table1 plus information from USGS, ISC and ISS catalogs.
epri EPRI 
weco Wells and Coppersmith 
trep Triep-Sykes and EPRI 
trwe Triep-Sykes and Wells-Coppersmith 
epwe EPRI  and Wells-Coppersmith 
tew Triep-Sykes, EPRI , and Wells-Coppersmith 
PS Pacheco and Sykes 
cmt Centroid Moment Tensor catalog (CMT) [Harvard University, 1995]
The classification of mechanism type (field 30) is obtained with the same procedure as EPRI :
Faulting Type B-plunge P-plunge T-plunge
TH 0-25 < 45 (always) > 45 (always)
NO 0-25 > 45 (always) < 45 (always)
SS 65-90 < 45 (always) < 45 (always)
TS 26-44 < 45 (always) T > P (always)
NS 26-44 > 45 (usually) T < 45 (always)
ST 45-64 < 45 (always) P <= T (always)
SN 45-64 < 45 (always) P >= T (always)
where B, P, and T are the null, pressure and tensional axes. The abbreviations are TH, thrust; NO, normal; SS, strike-slip; TS, thrust with a strike-slip component; NS, normal with a strike-slip component; ST, strike-slip with a thrust component; SN, strike-slip with a normal component.
With this classification, it is possible to know if the faulting occurred in a compressional or extensional stress regime if we use the standard convention that the orientation of the maximum compressive stress (inferred from the P-axis) characterizes the stress state. A compressional regime is one in which the P-axis is more horizontal than vertical (P < 45). Hence, TH, SS, TS, and ST are types of faulting that occur in compressional regime, while NO and NS occur in extensional regimes.
When RL or LL is added to the abbreviation of the mechanism type, it means that the fault associated with the earthquake is known to be right-lateral or left-lateral respectively (e.g., STRL implies mechanism with a right-lateral strike-slip motion with a thrust component).
The ranking of the quality of depth determination, seismic moment, mechanism type, and source time function (fields 17, 21, 28 and 33) is a result of the evaluation of the type, quality and quantity of data and method used for the respective determination. For example, we considered parameters obtained from waveform modeling or moment tensor inversion better constrained than those from P-wave first motion alone, and in the cases that SH-waves are included in the analysis we assign a higher quality than those where only the P-waveform is used. The symbol "-" means no qualification. In many cases the assignment of quality A, B, C, D is a highly subjective process, and therefore, the classification has a large uncertainty. This should be kept in mind by the users of the catalog.
The quality of the rupture dimensions (field 38) is given by three letters; the first letter corresponds to the quality of the surface rupture length (field 34), the second to subsurface rupture length (field 35), and the third to the rupture width (field 36). Each letter can be A, U, or N, meaning that the determination is reliable, unreliable, or no data was compiled for the corresponding rupture dimension. For example, UAN means surface rupture length is unreliable, subsurface reliable, and no data were compiled for rupture width. Other possible combinations are AAA, AUA, NAA, etc. The categories of reliable or unreliable depend on the type and quality of data and the method used in the analysis (rupture dimensions based on field observations, aftershock distribution, geodetic modeling, or teleseismic inversion). Here, we also warn the user of the catalog about the subjectivity of the assignment of quality.
Some other useful information for the user of the catalog is as follows: a) When the source of information (field 5) is trsy, epri, weco, or PS, the references for the different parameters if any (fields 9, 12, 14, 16, 20, 28, 32, and 37) are given with the notation of the original source(*). The references for Triep and Sykes (this study) are given below. When the source is from two or three catalogs (e.g., trep and tew respectively) the references for each parameter correspond to the catalog from which it came.
(*)[Note that when the field 5 is weco and the field 12 is PS, these two letters are related to the reference in Wells and Coppersmith  and not to Pacheco and Sykes . In all the other cases PS denotes this last reference].
b) If the field 5 is not epri the reference numbers smaller or equal to 142 correspond to the references given in this study. When the field 5 is epri, those numbers correspond to the references in EPRI . Reference numbers larger than 142 always correspond to EPRI .
c) The moment magnitude, Mw (field 13), does not necessarily correspond to the seismic moment Mo on fields 17 and 18. That is, since Mw can be obtained from Mo through the Eqn. Mw = (2/3)log(Mo) - 6.06 [Hanks and Kanamori, 1979] ( Mo in N m), we do not add a field for the result of this calculation. The field 13 for Mw is reserved for other determinations of Mw, typically the Mw from the EPRI , Wells and Coppersmith , or Pacheco and Sykes  catalog.
d) When the reference of the depth determination (field 16) is "weco", it means that it comes from an unpublished compilation of Donald L. Wells (D.L. Wells, personal communication, 1995). These depths do not have original references and/or quality specifications.
e) The Wells and Coppersmith  and Pacheco and Sykes  catalogs did not include strike, dip, and rake of the nodal planes. We added those types of information wherever possible. We also added the date, origin time, epicentral coordinates, and depth to the former catalog .
f) The catalog is a space separated text. An example of a typical C-shell or perl "printf" sentence is:
printf OUT "%8s %-9s %2s %-25s %4s %1s %7s %8s %4s %3s %3s %4s %6s %4s %4s %4s %1s %6s %2s %4s %1s %5s %4s %6s %5s %4s %6s %5s %1s %4s %4s %4s %1s %6s %6s %6s %5s %3s/n"
An example of a typical line of the catalog is (wrapped because of its length):
19910429 091248.1 ME Caucasus trsy a 42.424 43.664 127 6.2 7.0 109 - - 4.4 127 A 2.25 19 127 B 292.9 24.2 97.7 104.5 66.0 86.6 127 A TH 25 127 B - 70 24 127 NAA
The cut-off magnitude was Mw ~ 5.0-5.3 from 1964 through 1994 (instrumental seismological era), but from 1900 through 1963 most of the compiled earthquakes have larger magnitudes.
Three subsets of the earthquakes in our catalog can be considered to be complete: 1) the Mw >= 7.0 earthquakes from 1900 through 1994, 2) the earthquakes from 1978- 1994 with Mw larger than 5.3 or 5.5 depending on the region under consideration, and 3) the Mw >= 5.0 earthquakes of the SCR regions from 1964 to 1994. Since the SCR earthquakes in our catalog are dominated by the information from EPRI , we accept their statement of completeness for the period 1964-1990 for that range of magnitudes. However, the EPRI report excludes two regions from that statement; a) South America with a magnitude completeness threshold of 5.5 from 1964 through 1968, and 5.0 thereafter, and b) Indochina has a threshold of 5.0 only from 1980 to the present as a consequence of the information gap related to the Vietnam War.
We thank Arch C. Johnston for sending us a preprint of his paper on earthquakes in stable continental regions. We appreciate unpublished compilation that Donald L. Wells provided us for depths of intracontinental earthquakes. This research was supported by National Science Foundation grant EAR-91- 04158 and the Air Force Phillips Lab contract F19628-90-K0059.
Electric Power Research Institute (EPRI), The Earthquakes of Stable Continental Regions, EPRI vol. 1, edited by J.F. Schneider, California, 1-1 to 6-21, 1994.
Hanks, T. C., and H. Kanamori, A moment magnitude scale, J. Geophys. Res., 84, 2348-2350, 1979.
Johnston, A.C., Seismic Moment Assessment of Earthquakes in Stable Continental Regions, Part 1: Instrumental Seismicity, Geophys. J. Int., (in Press 1996).
Pacheco, J. F., and L. R. Sykes, Seismic moment catalog of large shallow earthquakes, 1900 to 1989, Bull. Seismol. Soc. Am., 82, 1306-1349, 1992.
Scholz, C. H., C. A. Avilles, and S. G. Wesnousky, Scaling differences between large interplate and intraplate earthquakes, Bull. Seismol. Soc. Am., 76, 65-70, 1986.
Triep, E.G., and L.R. Sykes, Frequence of Occurrence of Moderate to Great Earthquakes in Intracontinental Regions, submitted to J. Geophys. Res., 1996.
Wells, D. L., and K. J. Coppersmith, New Empirical Relationships among Magnitude, Rupture Length, Rupture Width, Rupture Area, and Surface Displacement, Bull. Seism. Soc. Am., 84, 974-1002, 1994.
References for the Catalog
Figure 1. Intracontinental earthquakes. Light and dark gray areas are Stable (SCR) and Active (ACR) continental regions respectively as defined in text. Circles denote epicenters of earthquakes from 1900 through 1994 with moment magnitudes, Mw, greater than or equal to 5.0. Squares indicate epicenters from 1900 through 1994 with Mw >= 7.0; solid squares the seven of those earthquakes in SCR. Open triangles are Type 1 earthquakes (as defined in text) from 1900 to 1994. Geographical boundaries of the Stable continental regions closely follow those of Electric Power Research Institute . The area referred to as the Middle East extends from 25 E to 64.5 E. The area referred to as Asia extends to the east of 64.5.
Figure 1A. Color version of Figure 1.