Worldwide trends in research on the San Andreas Fault System

Abstract

This paper deals with the bibliometric study of the research concerning the San Andreas Fault System (SAFS), an extraordinary example of a complex boundary between the North of America and the Pacific plates that can be seen, followed and studied on land for hundreds of kilometers through California, in the USA. The bibliometric analysis discussed here considers the time span of 23 years, from 1991 to 2013. The bibliographic databases taken as a reference are the Science Citation Index Expanded (SCI-EXPANDED) and the Social Sciences Citation Index (SSCI) accessed via Web of Science Core Collection. After having selected the proper search terms to fix the query and extracting the useful information from the two Web of Science (WoS) databases, we performed the analysis of the outputs paying attention to the document typology and languages, journals, subject categories, authors, articles, countries, institutions, and keywords. The results of this study can (1) contribute to see how the research on SAFS has changed over the time and suggest clues about the future investigation trends, (2) help scientists or institutions to build a research network and find strategic partners, (3) suggest scientists and institutions the dissemination strategies, and (4) provide helpful information to researchers wishing to embark on work in this area.

Full text

For Author's Approval
ORIGINAL PAPER
Worldwide trends in research on the San Andreas Fault System
Fabrizio Terenzio Gizzi
1
Received: 22 September 2014 /Accepted: 12 March 2015
#
Saudi Society for Geosciences 2015
Abstract This paper deals with the bibliometric study of the
research concerning the San Andreas Fault System (SAFS),
an extraordinary example of a complex boundary between the
North of America and the Pacific plates that can be seen,
followed and studied on land for hundreds of kilometers
through California, in the USA. The bibliometric analysis
discussed here considers the time spa n of 23 years, from
1991 to 2013. The bibliographic databases taken as a refer-
ence are the Science Citation Index Expanded (SCI-
EXPANDED) and the Social Sciences Citation Index (SSCI)
accessed via Web of Science Core Collection. After having
selected the proper search terms to fix the query and extracting
the useful information from the two Web of Science (WoS)
databases, we performed the analysis of the outputs paying
attention to the document typology and languages, journals,
subject categories, authors, articles, countries, institutions, and
keywords. The results of this study can (1) contribute to see
how the research on SAFS has changed over the time and
suggest clues about the future investigation trends, (2) help
scientists or institutions to build a research network and find
strategic partners, (3) suggest scientists and institutions the
dissemination strategies, and (4) provide helpful information
to researchers wishing to embark on work in this area.
Keywords San Andreas Fault System
.
California
.
?Tectonic study
.
Earthquake
.
Bibliometry
Introduction
The San Andreas Fault System (SAFS) is a complex of faults
having mainly right-lateral strike-slip movements that accom-
modate most of the relative N-S motion between the North
America and the Pacific plates. It runs approximately for
1300 km, from the connection with the Mendocino triple junc-
tion fault zone on the north to the head of the Gulf of Califor-
nia to the southeast (Fig. 1)(Ellsworth1990; Wallace 1990).
The dextral strike-slip San Andreas Fault zone is the main
element of the San Andreas Fault System. The name San
Andreas fault was adopted for the first time by Lawson
(1895 ) even if the fault gained importance only after the 1906
great San Francisco earthquake (M=7.7) (Wallace 1990).
San Andreas Fault zone is divided into ten sections that are,
from the north to the south, the Shelter Cove, North Coast, Pen-
insula, Santa Cruz Mountains, Creeping, Parkfield, Cholame-
Carrizo, Mojave, San Bernardino Mountains, and Coachella sec-
tions (Bryant and Matthew 2002). Each of them consists of
smaller elements such as branches or strand faults. For example,
the San Bernardino Mountain section hosts Quaternary and Ho-
locene faults such as Mission Creek, Mill Creek, and Banning
faults; Mojave section includes Nadeau, Punchbowl, and Little
Rock fault strands (Bryant and Matthew 2002). Other significant
elements of the San Andreas Fault System are, from the north to
the south, the Bartlett Springs, Maacama, Rodgers Creek, Green
Valley, Calaveras, Hayward, San Gregorio, San Jacinto, Elsinore,
and Imperial fault zones, in their turn including several tectonic
lineaments (Bryant and Matthew 2002).
The motion of the North of America and the Pacific plates
induces an elastic strain in SAFS that causes strong earth-
quakes once it is released. As a matter of fact, numerous seis-
mic events happened, historically speaking, such as the Fort
Tejon January 9, 1857 (M=7.8) and the San Francisco April
18, 1906 one (M=7.7), the two most significant earthquakes
* Fabrizio Terenzio Gizzi
f.gizzi@ibam.cnr.it
1
Institute of Archaeological and Monumental Heritage (IBAM,
National Research Council), 85050 Tito Scalo, PZ, Italy
Arab J Geosci
DOI 10.1007/s12517-015-1878-4

For Author's Approval
happened along the San Andreas fault proper, and other strong
Hayward fault-generated earthquakes were those of June 10,
1836 (M=6.7) and October 21, 1868 (M=7.0) that both hit the
city of San Francisco seriously (Ellsworth 1990).
Due to the unique features of the SAFS plate margin that
can be studied and analyzed on land and the need to improve
the knowledge of the California seismic hazard, SAFS has
received the attention of numerous researchers and, as a result,
has become the strike-slip fault system most in-depth studied
all over the world. Despite the importance of such tectonic
boundary, there are no previous studies aimed at analyzing
systematically from the statistical view point, the worldwide
scientific production concerning SAFS. Therefore, this work
aims to fill this gap partially, dealing with the bibliometric
analysis of the scientific production relating to the San
Andreas Fault System considering the last 23 years (1991–
2013).
The bibliometric studies of the literature based on the anal-
ysis of citations, countries, institutions, journals, keywords
chosen by the authors or Keyword Plus, language of the pa-
pers is a method that is being considered more and more in
many disciplines such as medicine, economy, chemistry, biol-
ogy, engineering, geography, and earth sciences in order both
to see how the research has changed over the time and get
some clues about the future trends (e.g., Garcia-Ramon and
Caballé 1998;Grantetal.2000;ChiuandHo2007;
Franceschini and Maisano 2010; Liu et al. 2012;Marxand
Bornmann 2013; Niu et al. 2014).
With this in mind, this work analyzes the research on SAFS
considering several aspects such as document typology and
languages, publication trend and citations, subject categories
and journals, productivity of the authors and their influence in
the scientific community, the most cited articles, countries and
institutions, author keywords and hot issues. The Thomson
Reuters’ Science Citation Index Expanded (SCI-
EXPANDED) and the Social Sciences Citation Index (SSCI)
accessed via the Web of Science Core Collection have been
the bibliographic databases used to perform this study. It is
necessary to emphasize that all the findings and discussions
included below have to be strictly considered as the reflection
of the content analysis of the two ISI databases taken as a
reference here. For that reason, for example, relevant papers
(with their Citations) and leading topics may be omitted if they
are not incorporated in the WoS indexes.
Data and methods
In order to perform the bibliometric study, we have used the
SCI-EXPANDED and SSCI citation indexes. According to
the Journal Citation Report (JCR), the SCI-EXPANDED da-
tabase indexes 8478 journals, while SSCI indexes 3045 in
2013. These values were respectively 5550 and 1699 in
1999 with an associated mean annual increase of journals
equal to 195.2 and 89.7 in the 1999–2013 period.
With the intention of pulling out the suitable records from
the databases, we have considered some strings to be used to
search for titles, abstracts, or keywords of the indexed papers.
To select the proper search keywords, we have taken ad-
vantage of the Quaternary fault and fold database of the
USA, compiled by the US Geological Survey (http://
earthquakes.usgs.gov/hazards/qfaults, accessed 24 April
2014). That database lists 11 major fault zones making part
of SAFS (Bryant et al. 2002). For each of the fault zone, the
BComplete report^ supplied by the US Geological Survey
database gives a literature overview including the names of
the principal faults of the fault zone itself (Hart 1998; Treiman
1998; Treiman and Treiman 1998; Black and Hecker 1999;
Treiman 1999; Treiman et al. 1999; Bryant and Cluett 1999a,
b; Bryant 2000; Bryant and Cluett 2000 ; Hart and Bryant
2001; Bryant and Cluett 2002; Bryant and Matthew 2002).
For example, the San Andreas Fault zone—San Bernardino
section—includes, among the others, Mill Cr eek, Mission
Creek, and the Banning faults; the Hayward fault
Fig. 1 Overview of the San Andreas Fault System in California. The red colour puts into evidence the San Andreas fault proper (after Wallace 1990,
modified)
Arab J Geosci

For Author's Approval
zone—Northern Hayward section—incorporates Mission,
Evergreen, Quimby, the Crosley, and Clayton faults. The
reports also allowed to consider the different names given by
scientists to a fault over the time, as for the Maacama fault-
north section-designed as the Talmadge fault in the past (Hart
et al. 2001).
The first search string used has been Andreas fault*,
followed by the names of the major faults as derived from
the analysis of the US Geological Survey database (Table 1).
On the whole, the search counts about one hundred terms.
However, as much as≈ 86 % of the total documents or articles
retrieved from the two databases came from the mother search
that has made use only of the Andreas fault* as string. There-
fore, the other terms have contributed to the final results with
about four results each, on average. In order to avoid possible
fake results, the papers resulted in the search output have been
analyzed manually each time that a new search string has been
added to the previous.
Once the records have been selected from the two data-
bases, they have been downloaded as plain text including
the names of the authors, the address, the title, the publication
year, the author keywords and keyword plus, the abstract, the
journal name, the Web of Science categories of the paper, the
citations, and the references.
The articles from England, Northern Ireland, Scotland, and
Wales have been reclassified as derived from the UK. The
analysis of the collaboration patterns has been determined
through the authors’ address, and therefore, the phrase Bsingle
country article^ has been assigned if the scientists’ addresses
were from the same country, while the term Binternationally
collaborated article^ has been associated to the articles
coauthored by researchers from multiple countries.
The traditional analysis of the data such as document ty-
pologies, languages, publication trends, subject categories,
journals, authors, countries, institutions, and keywords has
been made through the MS Excel software, the bibliographic
coupling through VOSviewer software that builds distance-
based maps (www.vosviewer.com; Waltman et al. 2010).
Lastly, the network analysis of the international
collaborations has been performed with the IntColl.EXE tool
authored by Loet Leydesdorff (http://www.leydesdorff.net/
software.htm,accessed20June2014).
Results and discussion
Document typology and languages
The output of the search gives 3402 total publications during
the time span of 23 years. Article, including articles published
as proceeding papers or book chapters, was the most frequent
document typology with 3126 records, amounting to about
92 % of the entire collection. Reviews reflected about 6 %
(204) of the collected works. The other documents concerned
Editorial materials (37; 1 %), news items (9; 0.3 %), correc-
tions (7; 0.2 %), book reviews (5; 0.1 %), notes (5; 0.1 %),
letters (4; 0.1 %), meeting abstracts (3; 0.1 %), and discussions
(2; 0.1 %). Considering that the articles were predominant in
the entire dataset, only this type of document has been used to
perform the analysis discussed in the following sections.
However, considering that most of the articles before 1991
do not include abstracts, we have restricted our analysis to
the 1991–2013 period.
Tabl e 1 Search terms used to
extract the records from the SCI-
EXPANDED and SSCI citation
indexes
Search strings
Andreas fault*; San Andreas system; BMill creek fault^; BBanning fault*^; BPunchbowl fault*^; BNadeau
fault*^; BLittle Rock^ fault*; BMission Creek fault*^; BGarnet Hill^ fault*; BGandy Ranch^ Fault*; BSan
Gorgonio Pass Fault*^; BSan Gorgonio Fault*^; BEtsel Ridge fault*^; BUpdegraff Ridge Fault*^; BBartlett
Springs thrust fault*^; B Bartlett Springs^ fault*; B Hunting Creek-Berryessa fault*^; BLake Mountain^ fault*;
BTalmadge fault*^; BMaacama fault*^; BHealdsburg fault*^; BRodgers Creek fault*^; BDiablo Thrust fault*^;
BDiablo fault*^; BSuisun fault*^;
BGreen Valley fault*^; BPaicines fault*^; BCalaveras fault*^; BSan Benito^
fault*; BMission fault*^ California; BEvergreen fault*^; BQuimby fault*^; BCrosley fault*^; BClayton fault*^
California Hayward fault*; B Coyote Creek fault*^; BLytle Creek fault*^; BSan Jacinto^ fault*; BGlen Helen
fault*^; BRialto-Colton fault*^; BClaremont fault*^; B Casa Loma^ fault*; BClark fault*^; BBuck Ridge^
fault*; BCoyote Mountain^ fault*; BSuperstition Hill^ fault*; BSuperstition Mountain^ fault*; BLoma Linda^
fault*; BCucamonga fault*^; BPark Hill^ fault*; BAnza Fault*^;
BHorse Canyon^ Fault*; BSanta Rosa
Fault*^; BArroyo Salada Fault*^; BSan Gregorio Fault*^; BSeal Cove Fault*^; BFrijoles Fault*^; B Coastways
Fault*^; BGreyhound Rock Fault*^; BCarmel Canyon Fault*^; BDenniston Creek Fault*^; BAño Nuevo
Fault*^; BSur Fault*^; BPalo Colorado Fault*^; BPalo Colorado-San Gregorio Fault*^; BImperial Fault*^;
BElsinore Fault*^; BMain Street Fault*^ California; BChino Fault*^ California; BWhittier Fault*^; BFresno
Fault*^; BTin Mine^ Fault*; BMain Street^ Fault*; BWildomar Fault*^; BGlen Ivy North Fault*^; B
Glen Ivy
South Fault*^; BGlen Ivy Fault*^; BWillard Fault*^; BWolf Valley Fault*^; B Laguna Salada^ Fault*; BFresno-
Eagle Fault*^; BEagle Fault*^; BEarthquake Valley Fault*^; BSan Felipe Fault*^; BSan Felipe Valley Fault*^;
BCanyon Royo Fault*^; BMurrieta Creek Fault*^; BBig Pine fault^; BGarlock fault^; BPinto Mountain fault^;
BBlue Cut fault^; BFernando fault^; BMonica fault^; BNewport fault^; BNewport-Inglewood fault*^;
BInglewood fault^; BCerro Prieto Fault*^; BMiguel fault*^
Arab J Geosci

For Author's Approval
An overview of the publication language shows that En-
glish was widely predominant (3102 records, 99.2 %). This is
consistent both with the fact that English is the predominant
language in the scientific field and the SCI and SSCI index
journals publish the full-text, or at least the abstract, in English
(Gotze 1997). Moreover, the percentage of the articles pub-
lished in English was particularly high because the San
Andreas fault is located in the USA as well as the main insti-
tutions that investigated it, as we will see later in this paper.
Other four languages were represented in a very few doc-
uments, such as Chinese (15; 0.5 %), French (4; 0.1 %), Span-
ish (3; 0.1 %), and Russian (2; 0.1 %), which had a very low
impact on the scientific community as emerges from the total
number of citations of the above mentioned articles that, taken
on the whole, did not reach the number of 50.
Publication trends and citations
In order to have a more detailed sight about the publications,
we have analyzed the number of articles over the time. From
Fig. 2 we can distinguish that the trend was characterized by
two main periods. Indeed, after a linear growth (R
2
=0.9993)
in 13 years, from 1991 to 2003, where the total number of
publications reached about 1200, the trend took an exponen-
tial form (R
2
=0.9981) reaching the total number of outputs
equal to 3126, that is 2.6 times of the total articles counted at
the end of the linear time window. The exponential pattern
roughly started when the number of scientists that contributed
to the study of SAFS increased sharply (2002) (see Authors’
productivity section and Fig. 6). This is probably due to the
increase of the interdisciplinary studies on the subject, the
significant development of new approaches and technologies
to investigate SAFS, the increase of interests in some aspects
of research, and the occurrence of earthquakes that stimulated
new investigations, as we will see better in the next sections.
All the articles reached about 77,000 citations in Web of
Science Core Collection, showing a trend that changed in
quite a complex way over the time (Fig. 3). As a matter of
fact, over the 23 years covered by our analysis, three different
periods should be considered: 1991–2000, 2000–2005, and
2005–2013. For each of these, the number of citations fits a
linear trend, but the rate of growth decreases as the publica-
tions become more recent. In particular, in the last part of the
third period, especially 2011–2013, the number of citations
decreased quickly because the time interval during which
the most recent articles can be cited reduced progressively.
This aspect can also be inferred from the number of the cita-
tions received from the articles published in 2012 and 2013
that, taken on the whole, did not reach 900 against the yearly
average of about 3400 over the 23 years.
Subject categories and journals
The research on SAFS spanned over 59 ISI subject categories
during the 1991–2013 period. Geochemistry and Geophysics
was the most important category with 1971 items, covering
57 % of the entire categories. The classes that exceed 100
items were other three, Multidisciplinary geosciences (767;
22 %; second), Geology (201; 6 %; third), and Multidisciplin-
ary sciences (114; 3 %; fourth).
These four categories covered about 88 % of the total from
1991 to 2013. As can be noticed from Fig. 4, the trend of the
articles classified as belonging to Multidisciplinary
geosciences and Geology categories was quite constant over
time. Conversely, Geochemistry and Geophysics category had
a sharp increase in the second part of the 23-year period under
investigation. That increase seems acted as a guide of the
R² = 0.9993
R² = 0.9981
0
500
1000
1500
2000
2500
3000
3500
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Cumulative number of articles
Years
linear trend
exponential trend
Fig. 2 Publication output over
the 23 years time-window inves-
tigated in this paper
Arab J Geosci

For Author's Approval
temporal trend observed in all the articles, as seen in the pre-
vious paragraph (see Fig. 2).
Seven (12 %) other categories that exceed ten items were
Engineering geology, Engineering-civil, Geography-physical,
Water resources, Oceanography, Physic-multidisciplinary,
and Mineralogy. Finally, as many as 18 (30 %) were the cat-
egories that counted only one item.
Articles were published in 231 journals, but those having a
number of published works that exceeded 100, covering about
66 % of the whole data-set were only eight (3.5 %), thus
showing that the hottest research studies are expected to be
published in a few high-impact journals (Table 2).
The Journal of Geophysical research— Solid
earth—published most of the articles (683; 21.8 %) followed
by the Bulletin of the Seismological Society of America
(443;14.2 %) and the Geophysical Resear ch Letters with 226
items (7.2 %). Figure 5, which reports the a nnual amount of
publications for each of these three journals, shows that JGR
always published the highest number of articles, except for
1991, 2002, 2006, and 2008 when BSSA was the first journal.
GRL was the journal with a more constant publication
trend until 2003, after which two significant peaks were re-
corded (2004 and 2007). From the comparison of the three
publication patterns it can be observed that a maximum (rela-
tive or absolute) publication trend of one journal frequently
occurred when a minimum publication pattern of the other/s
journal/s was recorded. This can also be due to the publication
of special issues, such as that published by BSSA in 2002
focused on the 16 October 1999M =7.1 Hector Mine
earthquake.
Regarding the total citations, Journal of Geophysical re-
search—Solid earth—held the highest number (21600), but
Nature (84.32) and Science (77.36) had the highest average
citations for article.
R² = 0.9969
R² = 0.9946
R² = 0.9291
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Cumulative total citations
Years
Fig. 3 Citation trends of the
articles during 1991–2013
0
20
40
60
80
100
120
140
160
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
number of articles
Years
GEOCHEMISTRY GEOPHYSICS GEOSCIENCES MULTIDISCIPLINARY GEOLOGY
Fig. 4 Trend of publications of
the three main WoS subject
categories over the 1991–2013
time span
Arab J Geosci

For Author's Approval
Based on the analysis of the 2012 Journal Citation Report
(JCR) Science edition, it emerges that among the top 30
journals, the most (28) had the impact factor (IF) equal or
greater than 1; moreover, 44 % of these articles were
Tabl e 2 Features of the top 30
most productive journals during
1991–2013
Journals TA TA(%) TC TC/TA IF
Journal of Geophysical Research-Solid Earth 683 21.8 21,600 31.63 3.174
Bulletin of the Seismological Society of America 443 14.2 11,742 26.51 1.940
Geophysical Research Letters 226 7.2 4659 20.62 3.982
Geophysical Journal International 197 6.3 3482 17.68 2.820
Tectonophysics 138 4.4 2837 20.56 2.684
Geology 134 4.3 4942 36.88 4.087
Journal of Structural Geology 124 4.0 4167 33.60 2.285
Earth and Planetary Science Letters 113 3.6 1860 16.46 4.349
Geological Society of America Bulletin 80 2.6 1941 24.26 4.286
Tectonics 75 2.4 1867 24.89 3.487
Pure and Applied Geophysics 72 2.3 1377 19.13 1.617
Science 53 1.7 4100 77.36 31.027
Geochemistry, Geophysics, Geosystems 38 1.2 283 7.45 2.939
Nature 37 1.2 3120 84.32 38.597
Earth Planets and Space 30 1.0 229 7.63 2.921
International Geology Review 24 0.8 582 24.25 3.359
Physics of the Earth and Planetary Interiors 24 0.8 468 19.50 2.383
Geophysics 22 0.7 354 16.09 1.723
Geosphere 22 0.7 239 10.86 2.023
Earthquake Spectra 19 0.6 96 5.05 1.079
Geomorphology 19 0.6 183 9.63 2.552
Journal of Geodynamics 18 0.6 111 6.17 2.967
Journal of Geophys Research-Solid Earth and Planets 17 0.5 881 51.82 3.174
Seismological Research Letters 16 0.5 226 14.13 3.036
Chinese Journal of Geophysics-Chinese Edition 14 0.4 21 1.50 0.667
Lithosphere 14 0.4 32 2.29 2.169
Nature Geoscience 14 0.4 291 20.79 12.367
Annals of Geophysics 11 0.4 119 10.82 1.138
Terrestrial Atmospheric and Oceanic Sciences 11 0.4 74 6.73 0.705
Izvestiya, Physics of the Solid Earth 10 0.3 3 0.30 0.402
TA total articles, TC total citations, IF impact factor, 2012
0
5
10
15
20
25
30
35
40
45
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Number of articles
JGR BSSA GRL
Fig. 5 Publications of the top
three most productive journals
during 1991–2013
Arab J Geosci

For Author's Approval
published in journals having an IF greater than 3. The two
journals with the highest IF were Nature (38.597) and Science
(31.027), where 90 papers were published (2.9 %).
Among the other journals, Geophysical Journal
International ranked fourth with 197 articles (6.3 %). Fifth
to eighth positions were taken by Tectonophysics (138;
4.4 %), Geology (134; 4.3 %), Journal of Structural Geology
(124; 4.0 %), and Earth and Planetary Science Letters (112;
3.6 %).
Productivity of the authors and their influence
on the scientific community
The number of the single authors involved in the SAFS re-
search was 5135, and the average of the collaboration index,
that is the number of authors for article, was 3.3 in the overall
time span considered.
The amount of the single-author number had an increase
from 214, in 1991, up to 882, in 2013 (Fig. 6). In particular,
the number of authors increased slightly until 2001, after
which at least three sharp raises were recorded in 2002,
2010, and 2013, respectively.
The same analysis can be made for the total number of
authors involved in the articles, whose trend was quite similar
to that of the single authors, especially until 2001. The two
patterns stayed comparable in the next years, but the increase
in value difference from 2002 suggests more collaborations
between the authors, as also indicated by the mean data of the
collaboration index ranging from 2.31 in 1991 to 4.08 in 2013
(Fig. 7). However, this analysis can be influenced by problems
affecting the authorship. As a matter of fact, two or more
authors may have the same name or may use more than one
name to sign their own articles (Tang and Walsh 2010).
The number of authors that contributed greatly to the SAFS
research was a small percentage of the total researchers. The
first 100 most prolific authors published about 54 % of the
total items and the first 30 about 27 %, as can be seen from
Table 3. In order to face up the variation of personal spelling
authorship problem, Table 3 (and Table 4) also shows the
equivalence between different names used by the same author
in his/her articles of the collection. To do this, we analyzed
mainly the Curriculum Vitae (CV) of the authors under inves-
tigation available at the institutional web pages. For example,
our collection of the articles on SAFS contains some papers
published by Rockwell TK and others by Rockwell T. In order
to make sure that the two different names conducted to the
same author, we analyzed Rockwell TK’s CV (available at
http://www.earthconsultants.com/ resumes/Rockwell.pdf,
accessed 11 August 2014) and we found that the articles
authored by Rockwell TK and Rockwell T could be traced
back to the same author. Therefore, the total number of the
articles has been reported in Table 3 (and Table 4) under the
double name attribution.
The most productive authors were Burgmann with 56 arti-
cles, closest followed by Ben-Zion with 52, Thurber with 43,
Nadea u and Rockwell with 40 article s. The other authors
followed with a number of articles that gradually tends to
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Number of authors
Years
Single Authors Total Authors
Fig. 6 Number of authors
involved in the SAFS research
during the 1991–2013 period
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Mean number of authors
Years
Fig. 7 Average of the number of authors for article involved in the SAFS
research during the 1991–2013 period
Arab J Geosci

For Author's Approval
20, the boundary of the 30th position. Most of these re-
searchers collaborated with two or more authors. Very few
were the articles with only one author (57, ~7 % of the list),
and only one researcher stood out of the others for his high
number of single-author papers (Parson, 14).
About 47 % (14) of the top 30 most productive scientists
were also among the most cited authors and as many as
five occupied the first ten positions (Table 4). The number of
the articles signed by only one scientist included in this list
was still lower than those of the top 30 most productive author
list (24, ~4 % against 57, ~7 %), confirming that collaborative
papers are more likely to be cited.
Stein held the highest number of citations in the time span
considered, while Forster held the highest average of citations
of a paper of his (214) with a lower number of publications (4)
on the list. Among the first three most cited authors, we can
notice that the first was not included in the list of the most
prolific authors, but his articles have a citation medium rate
that fell among the first three of the most cited authors.
Analyzing the bibliographic coupling between the authors
that signed or co-signed at least five articles, it can be noticed
that the scientists can be grouped into 12 clusters, each of them
with a different size ranging from 79 to 2 authors. Figure 8
shows this, also putting into evidence that the circle size is
directly proportional to the number of the total articles of an
author, the color represents the membership cluster, and the
distance between each circle shows how strong the relation-
ship with the other scientists is—the shorter is the distance, the
Tabl e 3 Top 30 most productive authors involved in the research on
SAFS during 1991–2013
Author TA SA TC TC/TA
Burgmann R 56 2 1816 32.4
Ben-Zion Y 52 3 1667 32.1
Thurber CH/C 43 2 1173 27.3
Nadeau RM/R 40 0 1660 41.5
Rockwell TK/T 40 0 668 16.7
Hauksson E 31 2 1443 46.5
Shearer PM/P 31 2 852 27.5
Arrowsmith JR/R 29 0 374 12.9
Parsons T 28 14 413 14.8
Zoback MD/M 28 1 1057 37.8
Savage JC 27 4 666 24.7
Segall P 26 1 1405 54.0
Simpson RW/R 26 0 1375 52.9
Ellsworth WL/W 25 0 2175 87.0
Peng ZG/Z 24 0 563 23.5
Zhang HJ/H 24 0 565 23.5
Furlong KP/K 23 1 469 20.4
Hardebeck JL/J 23 4 889 38.7
Fialko Y 22 4 644 29.3
Fuis GS/G 22 1 498 22.6
Li YG 22 1 669 30.4
Lockner DA/D 22 0 1210 55.0
Marone C 22 1 670 30.5
Pollitz FF/F 22 7 511 23.2
Rundle JB 22 1 357 16.2
Sandwell D/DT 22 0 402 18.3
KE/KE 21 1 1000 47.6
Lienkaemper JJ 20 1 466 23.3
Sammis CG/C 20 0 1065 53.3
Wyss M 20 4 811 40.6
Authors in bold are also among top 30 most cited scientists (see Table 4)
TA total articles, SA single-author articles, TC total citations
Tabl e 4 Top 30 most cited authors relating to the research on SAFS
during 1991–2013
Author TA SA TC TC/TA
Stein RS/R 16 0 2345 146.6
Ellsworth WL/W 25 0 2175 87.0
Burgmann R 56 2 1816 32.4
Evans JP 18 0 1748 97.1
Lin J/JA 10 0 1725 172.5
Ben-Zion Y 52 3 1667 32.1
Nadeau RM/R 40 0 1660 41.5
King GCP/G 12 0 1616 134.7
Waldhauser F 14 0 1493 106.6
Hauksson E 31 2 1443 46.5
Segall P 26 1 1405 54.0
Eberhart-Phillips D 14 0 1383 98.8
Simpson RW/R 26 0 1375 52.9
Lockner DA/D 22 0 1210 55.0
Thurber CH/C 43 2 1173 27.3
Chester FM 10 1 1122 112.2
Sammis CG/C 20 0 1065 53.3
Zoback MD/M 28 1 1057 37.8
Reasenberg PA/P 7 0 1046 149.4
McEvilly TV 15 0 1014 67.6
Sieh K/KE 21 1 1000 47.6
Byerlee JD/J 13 2 957 73.6
Beroza GC 16 1 923 57.7
Rice JR 11 0 895 81.4
Hardebeck JL/J 23 4 889 38.7
Michael AJ/A 14 1 869 62.1
Jackson DD/D 15 0 867 57.8
Harris RA/R 10 1 857 85.7
Forster CB 4 0 856 214.0
Shearer PM/P 31 2 852 27.5
Authors in bold are also among top 30 most productive scientists (see
Table 3)
TA total articles, SA single-author articles, TC total citations
Arab J Geosci

For Author's Approval
higher is the connection. The first and the third wider clusters
were those led by the first two most productive authors—
Burgmann (the third cluster with 59 authors) and Ben-Zion
(the widest cluster, 79 authors). The map shows that the au-
thors were grouped into clusters where they are close to each
other, thus underlying that there is a good statistical probabil-
ity that they treated a quite similar subject matter.
Most cited articles
The most cited article was Static stress changes and the trig-
gering of earthquake published by King GCP et al., in 1994 in
the Bulletin of the Seismological Society of America, and was
cited 813 times (from Web of Science Core Collection) in the
1991–2013 time span. Observing the yearly-citations trend of
this paper (Fig. 9), we notice that a sharp increase was record-
ed when the article had 2 years of life, and the number of
citations increased constantly until the seventh year. The
highest yearly value of citations (74) was reached 17 years
after the publication.
The second most cited article (728) was A double-
difference earthquake location algorithm: Method and appli-
cation to the northern Hayward fa ult, Cali fornia by
Waldhauser, F and Ellsworth, WL published in 2000 in the
Bulletin of the Seismological Society of America, too. This
article reached rapidly a high number of yearly citations and
the highest values respectively 3 and 13 years after the
publication.
The third (575 Citations) and the fourth (428 Citations)
articles in the rank were Fault zone architecture and perme-
ability structure by Caine et al., published in the Geology in
1996 and Internal structure and weakening mechanism of the
San Andreas Fault by Chester et al., published in the Journal
of Geophysical Research in 1993. All the four articles retained
a high number of yearly citations testifying that the attention
of the researchers to these issues was constant over the whole
time span taken as a reference here.
Fig. 8 Bibliographic coupling of
the authors having a minimum
number of articles equal to 5
0
20
40
60
80
100
120
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Number of citations
Years
King et al., 1994 Waldhauser and Ellswort, 2000 Caine et al., 1996 Chester et al., 1993
Fig. 9 The four most cited
articles on the SAFS during
1991–2013. The data for each
article starts from the year of
publication
Arab J Geosci

For Author's Approval
Where the publications were born: country and institution
analysis
Among 3126 articles, 319 (10.2 %) have no author address
information, and therefore, these data have been excluded
both from the analysis of the publication country and the in-
stitutional distribution of research outputs.
Seventy-three countries/territories all around the world
were involved in the SAFS research in the 1991–2013 period.
Taking into consideration the most productive top 30 coun-
tries (Table 5), one can see that four continents were active in
the SAFS research. Europe topped the rank with 13 countries,
Asia counted 8 countries, America 5, and Oceania 2, while
Africa was not ranked.
The country that held the leadership was the USA, with
1937 records (69.0 %). Among the European countries,
France occupied the second place but significantly distant
from the first, signing 234 papers (8.3 %). Considering the
countries with a minimum number of 100 records, it emerges
that France was followed by Japan with 209 articles (7.4 %),
and Germany (147; 5.2 %), Italy (122; 4.3 %), China (120;
4.3 %), and the UK with 108 articles (3.8 %).
The USA collaborated with the highest number of coun-
tries (64) even if the percentage of the international collabo-
rated articles was the lowest (24.0 %). Other countries had an
international collaborative percentage that ranged from 33 %
of Belgium to 100 % of Chile and Indonesia, with a high mean
of ~64 %. For example, France and Germany, which signed
internationally collaborated articles with 42 and 36 countries
respectively, had a percentage of multi-country items, ranging
from 70 to 74 %. To have a clearer picture of these aspects,
Fig. 10 shows the complex network of international collabo-
rations between all 73 countries. All these data imply that the
SAFS research calls for teamwork between countries, with a
specific reference to the USA.
The analysis of the institution contribution has been per-
formed considering the affiliation of at least one author of the
articles. As said before, there were 319 articles without insti-
tution information. Therefore, the total number of articles con-
sidered for this analysis was 2807. Furthermore, the examina-
tion has left out of the investigation the subdivision of the
institutions in branches.
The top 30 institutions active in the past 23 years are listed in
Table 6 from which it emerges that the SAFS research studies
were performed mainly by the institutions having the seat in the
USA. As a matter of fact, among the first 30 most prolific world-
wide institutions, as many as 22 (73.3 %) were from the USA.
The most prolific institution was the US Geological Survey
(USGS) with 495 articles (17.6 %) published in the 1991–2013
period. The California Institute of Technology (Caltech) followed
with 163 (5.8 %) and five other universities of California. The
first non-USA institution was the Italian Istituto Nazionale di
Geofisica e Vulcanologia (INGV) that occupied the eighth posi-
tion with 63 papers (2.2 %). Among the other EU-country insti-
tutions, we single out the University of Gr enobl e 1 (France)
ranked 16th with 47 articles (1.7 %) and the
GeoForschungsZentrum (GFZ) of Potsdam (Germany) ran ked
19thwith41papers(1.5%),whilethefirstAsiaticinstitutionwas
the University of Tokyo (Japan) that occupied the 10th position
with 59 publications (2.1 %). The highest mean value of citations
for article was held by Harvar d University (48.1) followed by
Columbia University (40.5), while the mean of the citations for
article, considering all the 30 institutions, was about 25.
Author keywords
According to the previous experiences, the author keywords can
help to get the trend regarding a specific research field (Chiu and
Ho 2007; Wang et al. 2014). Therefore, we performed the
Tabl e 5 Top 30 countries/territories involved in the research on SAFS
during 1991–2013
Countries/territories TA TA(%) IC IC(%) TC TC/TA
USA 1937 69.0 464 24.0 52,265 27.0
France 234 8.3 164 70.1 4231 18.1
Japan 209 7.4 92 44.0 3698 17.7
Germany 147 5.2 109 74.1 1659 11.3
Italy 122 4.3 70 57.4 2331 19.1
People's R China 120 4.3 78 65.0 1787 14.9
UK 108 3.8 65 60.2 2458 22.8
Canada 87 3.1 54 62.1 1354 15.6
Taiwan 74 2.6 52 70.3 827 11.2
New Zealand 60 2.1 56 93.3 1325 22.1
Switzerland 52 1.9 45 86.5 1789 34.4
Mexico 46 1.6 33 71.7 703 15.3
Australia 44 1.6 32 72.7 667 15.2
Turkey 44 1.6 36 81.8 1479 33.6
India 32 1.1 8 25.0 319 10.0
Norway 31 1.1 22 71.0 538 17.4
Israel 29 1.0 22 75.9 438 15.1
Spain 28 1.0 20 71.4 213 7.6
Greece 26 0.9 14 53.8 310 11.9
Netherlands 25 0.9 20 80.0 556 22.2
Russia 24 0.9 12 50.0 204 8.5
Iran 14 0.5 10 71.4 99 7.1
South Korea 12 0.4 7 58.3 243 20.3
Ireland 11 0.4 6 54.5 188 17.1
Chile 9 0.3 9 100.0 98 10.9
Brazil 8 0.3 3 37.5 111 13.9
Indonesia 7 0.2 7 100.0 143 20.4
Austria 6 0.2 3 50.0 94 15.7
Belgium 6 0.2 2 33.3 22 3.7
Denmark 5 0.2 0 0.0 81 16.2
TA total articles, IC internationally collaborated articles, TC total citations
Arab J Geosci

For Author's Approval
analysis of the author keywords to try to show up the most
significant issues dealt with in the articles over the years. There-
fore, we have considered the changes of the keywords in four 5-
year periods, starting from 1994. On 2845 total articles,
1134 (39.9 %) included one or more keywords. In detail,
724 (63.8 %) of these had a number of keywords between five
and twelve; the remainder (410) publications had a number
varying from 1 and 4.
We have identified 3037 unique author keywords, with a
total of 5632 occurrences. Few (68, 2.2 %) were the words that
appeared more than or equal to ten times. Two thousand and
three hundred forty (77.0 %) were the words that occurred
only once showing a possible episodic nature of some re-
search activities or the lacking of standardization of the terms
chosen by the authors (Chuang et al. 2007).
In order to get a closer view, we have selected the 40
(1.3 %) most frequent keywords that, taken on the whole,
occurred 932 times, equal to 16.5 % of the total occurrences.
The 40 most frequent keywords can be divided into two
main groups: those reflecting the tectonic studies and those
that can be referred to as seismological works (Table 7). The
former can be identified, for example, by the words fault(s),
faulting, dynamic and mechanic of faulting, active fault, frac-
tures and faults, fault gouge, the latter by earthquake
dynamics, wave propagation, body waves, paleoseismology,
seismic anisotropy,andseismic hazard. A third category
regards the remote-sensing tools such as InSAR or GPS.
However, the keywords had a different importance over the
years. As a matter of fact, excluding San Andreas fault,
earthquake/s, seismicity,andfriction, that retained high posi-
tions in the rank all over the four temporal windows consid-
ered, most of the top 40 keywords were not classified nor had
little relevance in the 1994–1998 period. This can be due, for
example, to the fact that some studies had pioneering feature
at that time. For example, the InSAR technique that was not
ranked in the 1994–1998 period was used for the first time in
the 1960s to observe the planets of the solar system (Rogers
and Ingalls 1969). The 1980s saw the applications of InSAR
to get DEMs of the Earth’s surface, but the technique was
improved only in the 1990s. Just in these years, there was a
strong rise of the InSAR applications, among which those in
the tectonic field that were inaugurated with the analysis of the
co-seismic displacement brought about by the 1992 Landers
earthquake (Massonnet et al. 1993; Goldstein et al. 1998). A
few years after that earthquake and from the end of the 1990s,
InSAR was considered more extensively for the study of
SAFS and was also used to monitor creep and strain accumu-
lation (e.g., Lyons and Sandwell 2003).
Other words included in the list were
paleoseismology,
GPS, tomography, magnetotellurics, seismic anisotropy, wave
propagation,andSAFOD. The GPS, jointly with the InSAR
technique or seismicity-derived stress orientations were used,
for example, to get data relating to the vertical motions in the
San Francisco Bay Area or the fault slip rates of the tectonic
lineaments in the southern California plate-boundary region
(Becker et al. 2005; Burgmann et al. 2006). The second half of
the last decade also saw the start of the publication of the
studies related to the activity of the San Andreas Fault
Fig. 10 Collaboration network
between the countries during
1991–2013. The elaboration has
been performed for all 73
countries using the IntColl.exe
tool, and the visualization of the
network has been obtained
through the VOSviewer software
(see text). The figure zooms on
the core of the network
Arab J Geosci

For Author's Approval
Observatory at Depth (SAFOD) whose drilling activities be-
gan in 2004 and ended in 2007 (http://www-icdp.icdp-online.
org/front_content.php?idcat=712,accessedon14July2014).
The investigation results regarded the rock or fluid sample age
or physical, geochemical, mineralogical features, or the
velocity model identification around the SAFOD site placed
close to the town of Parkfield (e.g., Bradbury et al. 2007;
Wiersberg and Erzinger 2007; Zhang et al. 2009; Janssen
et al. 2011; Spencer et al. 2012).
Hot issues
As said before, only 39.9 % of the articles had the author
keywords, and still a smaller percentage of the articles had a
number of keywords sufficient to represent their content. This
can lead to terms indexing problems (Garfield 1990). There-
fore, in order to overcome these limitations and capture the hot
issues and main research trends, we have performed a joint
analysis of the author keywords, keywords plus, and title
words thus supply ing the top 40 most frequent term list
(Table 8). As for the author keywords, we have considered
the changes of the terms in four 5-year periods, starting from
1994 so as to be able to perform a comparison between the
two groups of terms.
The terms Slip (5th), Parkfield (12th), Basin (24th), Hay-
ward fault (30th), heat-flow (39th), Punchbowl fault (36th)
were some of the words that got in the list, while
paleoseismology, InSAR, magnetotelluric,andSAFOD were
some of the author keywords that lost significance and, con-
sequently, went out of the list. Slip retained a high importance
in all the four 5-year time windows analyzed as well as
Parkfield that reached the highest position (11th) in the
2004–2008 period. These terms also coincided with the hot
issues dealt with by the three most cited authors. As a matter
Tabl e 6 Top30institutions
involved in the research on SAFS
during 1991–2013
Institutions TA TA(%) TC TC/TA
USGS, USA 495 17.6 14,630 29.6
Caltech, USA 163 5.8 5062 31.1
University of Southern California, USA 146 5.2 3907 26.8
University of California, Berkeley, USA 137 4.9 3804 27.8
Stanford University, USA 136 4.8 3725 27.4
University of California, San Diego, USA 105 3.7 3145 30.0
University of California, Los Angeles, USA 79 2.8 2230 28.2
INGV, Italy 63 2.2 867 13.8
State University of San Diego, USA 63 2.2 1221 19.4
University of Wisconsin, USA 61 2.2 1400 23.0
University of Tokyo, Japan 59 2.1 909 15.4
University of California, Santa Barbara, USA 58 2.1 2141 36.9
University of Nevada, USA 58 2.1 1504 25.9
Columbia University, USA 53 1.9 2147 40.5
China Earthquake Administration, China 51 1.8 545 10.7
MIT, USA 50 1.8 1750 35.0
University of California, Davis, USA 48 1.7 905 18.9
University of Grenoble 1, France 47 1.7 687 14.6
Pennsylvania State University, USA 45 1.6 975 21.7
Lawrence Livermore National Laboratory, USA 44 1.6 1442 32.8
University of Colorado, USA 44 1.6 1158 26.3
GFZ, Potsdam, Germany 41 1.5 838 20.4
Kyoto University, Japan 37 1.3 723 19.5
Oregon State University, USA 36 1.3 691 19.2
Arizona State University, USA 35 1.2 511 14.6
Harvard University, USA 34 1.2 1637 48.1
Insti Phys Globe, France 33 1.2 896 27.2
University of Oregon, USA 33 1.2 788 23.9
University Montpellier 2, France 32 1.1 454 14.2
University of California Santa Cruz, USA 30 1.1 938 31.3
TA total articles, TC total citations
Arab J Geosci

For Author's Approval
of fact, Stein RS cosigned a paper investigating the slip
caused by the 1989 Loma-Prieta ea rthquake compar ing
the leveling surveys performed before and after the quake
itself (Marshall et al. 1991). The same author analyzing
the 1983M=6.5 Coalinga and M =6.0 Nuñez earthquakes
inferred about the shear and Coulomb stress change in-
duced on the Parkfield segment so as to calculate the
probability of an event with a magnitude similar to six
Parkfield earthquakes during the 2001–2011 period (Toda
and Stein 2002). Moreover, Ellsworth WL, taking advance
Tabl e 7 Top40keywordsused
by authors in their own articles
during 1994–2013 and four 5-
year periods
Total 1994–1998 1999–2003 2004–2008 2009–2013
Word Cnts Rank Cnts Rank Cnts Rank Cnts Rank Cnts Rank
San Andreas fault 92 1 8 1 20 1 34 1 30 1
Earthquakes 59 2 5 2 18 2 19 2 17 4
Seismicity 45 3 5 2 11 3 11 4 18 3
Earthquake 29 4 3 4 6 7 9 6 11 9
Faults 2854358123713
Friction 28 5 2 5 6 7 10 5 10 10
Crustal structure 27 6 2 5 3 10 7 8 15 6
Fault 26 7 –– 86 78 119
GPS 26 7 –– 4 9 12 3 10 10
Paleoseismology 26 7 –– 10 4 11 4 5 15
Tectonics 26 7 –– 95 96 812
Faulting 258258678812
Seismic tomography 24 9 –– 49 411165
Tomography 24 9 2 5 3 10 9 6 10 10
Faultzone 2210166796614
Seismicity and tectonics 22 10 –– –– –– 22 2
California 21 11 1 6 11 3 4 11 5 15
Crustal deformation 21 11 –– 5 8 12 3 4 16
Dynamics and mechanics
of faulting
20 12 –– –– 411165
Fractures and faults 19 13 –– –– 114183
INSAR 19 13 –– 49 78 812
Magnetotellurics 19 13 –– 112105 812
Earthquake dynamics 18 14 –– –– 312156
Seismic anisotropy 18 14 1 6 1 12 6 9 10 10
Wave propagation 18 14 –– 2 11 2 13 14 7
Body waves 17 15 –– 2 11 1 14 14 7
Seismic hazard 17 15 1 6 3 10 9 6 4 16
Strike-slip fault 17 15 –– 211105 515
Active fault 16 16 1 6 2 11 7 8 6 14
Creep 16162558411515
Permeability 16 16 –– 49 411812
Earthquake interaction 15 17 –– 2 11 2 13 11 9
Geodesy 15 17 –– –– 87 713
Slip rate 15 17 –– 21196 416
Stress 15 17 –– 67 69 317
Transform faults 15 17 2 5 –– 411911
Fault gouge 14 18 2 5 2 11 3 12 7 13
Fluids 14 18 –– 21178 515
Mechanics 14 18 –– 2 11 2 13 10 10
SAFOD 14 18 –– –– 213128
Cnts,Counts
Arab J Geosci

For Author's Approval
of the high-quality measurements obtained after the 28
September 2004 Parkfield earthquake, dealt with new earth-
quake models in order to better estimate the size and the loca-
tion of the future ground shaking (Bakun et al. 2005). Finally,
Burgman R put into evidence, for example, how powerfully
the fault interaction can influence the slip gradients
(Burgmann et al. 1994).
As concerns the Hayward fault, one can show that it gained
importance during the 1999–2003 period, when it ranked 39th
and in the subsequent two periods when it was classified at
Tabl e 8 Top 40 words as derived
by the elaboration of author
keywords, keywords plus, and
title words during 1994–2013 and
four 5-year periods
Total 1994–1998 1999–2003 2004–2008 2009–2013
Word Cnts Rank Cnts Rank Cnts Rank Cnts Rank Cnts Rank
California 1457 1 297 1 315 1 400 1 445 2
San-Andreas fault 1401 2 235 2 260 2 349 2 557 1
Fault 1183 3 194 3 224 3 329 3 436 3
Earthquake 890 4 143 4 160 4 270 4 317 4
Slip 603 5 100 8 112 6 163 5 228 5
Zone 563 6 84 10 110 8 158 6 211 6
Deformation 54971076111714781847
Southern(−)California 519 8 104 7 127 5 135 9 153 10
Earthquakes 456 9 94 9 87 10 121 10 154 9
Stress 431 10 117 5 87 10 102 13 125 12
Andreas 347 11 46 18 62 12 110 12 129 11
Parkfield 334 12 48 17 50 17 116 11 120 13
System 304 13 42 21 67 11 92 14 103 15
Seismic 293 14 49 16 51 16 85 15 108 14
Seismicity 272155813481879168719
Model 260 16 51 15 54 14 74 18 81 20
Evolution 258 17 52 14 56 13 62 22 88 18
Rupture 256 18 44 19 47 19 68 19 97 16
Faults 236194320392157239716
Tectonics 201 20 51 15 52 15 54 24 44 34
Structure 185 21 25 32 39 21 65 20 56 25
Implications 174 22 38 24 32 23 34 35 70 21
Zones 167 23 40 22 32 23 43 29 52 28
Basin 163 24 36 25 41 20 41 30 45 33
Region 161 25 27 31 29 26 51 26 54 26
Inversion 155 26 39 23 30 25 43 29 43 35
Strain 146274419272836333938
Recurrence 144282433262953254136
Models 143 29 22 35 30 25 34 35 57 24
Hayward fault 141 30 –– 15 39 63 21 63 22
Strike-slip 139 31 17 39 27 28 39 32 56 25
Friction 138322036282740315030
Crustal 135 33 29 29 27 28 32 37 47 32
Velocity 130341343223243295228
Japan 126 35 15 41 16 38 45 28 50 30
Shear 126 35 23 34 34 22 30 39 39 38
Punchbowl fault 124 36 3 53 28 27 49 27 44 34
Valley 118373226282733362551
Motion 117382036322339322650
Heat-flow 116 39 28 30 14 39 34 35 40 37
Cnts,Counts
Arab J Geosci

For Author's Approval
least 22nd. The same analysis can be made for the Punchbowl
fault, against the trend that inversion had, which diminished in
importance over the time.
A closer analysis of the temporal changes of some of
the new terms intr oduced by the joint analysis can be
gained from Fig. 11, from which one can see that the trend
is quite constant for all the words, except for Hayward
fault whose related papers gained weight from 2002, and
slip and Parkfield which had an increase after 2004 and
2003, respectively.
Regarding the terms excluded from the list, paleoseismology
ranked 100th over the entire period. Its rank was 48th in the
1994–1998 period, during which the term was not indexed in
the author keyword top ranked list, increasing slightly in sig-
nificance in the 1999–2003 period and gaining the 43rd posi-
tion. In these years, the paleo-seismological studies were per-
formed, for example, on the San Andreas fault or its branches in
order to determine the earthquake average recurrence interval
(e.g., Dolan et al. 2000;Dawsonetal.2003). The last two
time-windows saw a progressi ve light decrease of this re-
search field down to the 55th position of the 2009–20 13
period, confirming partially the general trend that ap-
peared from the only analysis of the author keywords. In
its turn, InSAR ranked 98th i n the multi-term analysis list
against 13th as derived from the author keyword analysis.
The term gained importance in the last 15 years when it
passed progressively from the 50th position of the 1999–
2003 period to the highest position (27th) recorded in the
2009–2013 period.
Analogous considerations can be made for SAFOD that
ranked 80th on the joint-term analysis list. The word had an
increase up to the 41th position in the 2004–2008 period with
an increase in the ranking (29th) in the 2009–2013 period,
similarly to the outputs of the analysis restricted only to the
author keywords.
Conclusions
The paper has presented an overview of the global research on
the San Andreas Fault System during the 1991–2013 period
based on the Thomson Reuters’ Science Citation Index Ex-
panded (SCI-EXPANDED) and the Social Sciences Citation
Index (SSCI) accessed via the Web of Science Core Collec-
tion. The aspects analyzed have been the typology and lan-
guages, the journals, the subject categories, the authors, the
articles, the countries, the institutions, and the hot issues.
The study has highlighted that 3126 articles were pub-
lished. The publication trend over time indicates that the in-
crease in publication number occurred since 2003, when an
exponential tendency replaced the linear growth. The expo-
nential pattern, which roughly started when the number of
researchers that contributed to the studies on SAFS increased
sharply, if used as future forecast may suggest that about 2000
articles will be published until 2020.
The Geochemistry and Geophysics was the most important
subject category covering 57 % of the entire categories. With
regard to the journals, the Journal of Geophysical research—
Solid earth—published most of the articles followed by the
Bulletin of the Seismological Society of America and the Geo-
physical Research Letters.
The most cited article was Static stress changes and the
triggering of earthquake published by King GCP et al., in
1994 in the Bulletin of the Seismological Society of America.
Seventy-three countries/territories all around the world
contributed to the SAFS research in the 1991–2013 period.
The country that held the leadership was the USA that also
collaborated with the highest number of countries.
Regarding the institutions, there was a clear predominance
of those having the seat in the USA. The most prolific insti-
tution was the US Geological Survey (USGS) followed by the
California Institute of Technology (Caltech).
0
100
200
300
400
500
600
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
Cumulative number of counts
Years
HEAT-FLOW SLIP PARKFIELD
BASIN STRESS HAYWARD FAULT
FRICTION PUNCHBOWL FAULT
Fig. 11 Visualization of some of
the most significant terms ranked
in the top 40 list obtained through
the joint analysis of the author
keywords, keywords plus, and
title words (total data as in
Table 8)
Arab J Geosci