Paleoseismic Assessment of the Late Holocene Rupture History of the Rose Canyon Fault in San Diego
December, 2012
Prepared for
Southern California Edison San Onofre Nuclear Generating Station Seismic Source Characterization Research Project
Rev 0
Acknowledgements 1.0
Introduction ................................................................................................................................1
2.0
Rose Canyon fault Slip Rate ...................................................................................................5
3.0
Aerial Image Analysis ..............................................................................................................5
4.0
Permitting Process ...................................................................................................................6
5.0
Cone Penetrometer (CPT) Surveys .....................................................................................7
6.0
Paleoseismic Trenching in Old Town.................................................................................7 6.1
Stratigraphy ................................................................................................................................ 9
7.0
Details on the Radiocarbon Results................................................................................. 15
8.0
Trenching Results – Interpretation of Rupture History ........................................... 16
9.0
Discussion, Implications and Needs for Additional Work ....................................... 17
10.0 References ................................................................................................................................ 19 11.0 Glossary of Terms and Acronyms ..................................................................................... 21 List of Figures 1
Regional map showing the relationship of the Rose Canyon and Newport Inglewood faults in the plate boundary system of fault.
2
Principal elements of the coastal zone of faults, with the Newport-Inglewood-Rose Canyon-Descanso-Agua Blanca fault zone.
3
Map of the Rose Canyon fault zone in San Diego and the new trench site in Old Town.
4
Interpreted earthquakes from the Lindvall and Rockwell (1995) trench site in Rose Creek.
5
(upper) Log of trench T4 from Lindvall and Rockwell (1995); (lower) Reconstructed log of trench T4.
6
Two potential trench sites (northern and southern) on Caltrans property where Holocene alluvial fan deposition may record the late Holocene rupture history of the Rose Canyon fault.
7
Fault-related geomorphology interpreted in 1928 aerial photography, with location of the golf course, CPT lines, and trench T1.
8
CPT locations in the golf course in Old Town.
9
CPT cross-section and interpreted correlation of stratigraphy and faults for CPTs 16.
10
CPT cross-section of CPT line 7-12 on the south side of the golf course in Old Town.
December 2012 Rev 0
Page i
11
Detailed log of a portion of the south face of trench T1, photographed under natural light.
12
Detailed log of a portion of the south face of trench T1, photographed under halogen light.
13
Detailed log of a portion of the north face of trench T1, photographed under halogen light.
14
Composite stratigraphic section from trench T1.
15
OxCal model of the 13 samples listed in Table 1.
16
Probability distributions for events 1 and 2, as determined in the OxCal model.
List of Plates Plate 1
Log of Trench 1 South Face
List of Tables 1 Radiocarbon ages on detrital charcoal and marine shell specimens recovered from the strata exposed in trench T1. Glossary of Terms and Acronyms SONGS
San Onofre Nuclear Generating Station
ka
Thousand years
MIS
Marine isotope stage
MRE
Most recent earthquake
OxCal
Software for use in radiocarbon calibration and analysis by Professor Christopher Bronk Ramsey at University of Oxford
PDF
Probability distribution function
December 2012 Rev 0
Page ii of 20
ACKNOWLEDGEMENTS The research and work in preparing this document was completed by a team of geologists, geomorphologists, paleoseismologists, paleontologists, and pedologists who have focused their careers on the fault and seismic hazards in southern California, particularly along the coast. The team was led by Dr. Thomas K. Rockwell from San Diego State University and Mr. Monte Murbach from Murbach Geotech, and included Diane Murbach with the City of San Diego, Kim Rockwell, Eric Haaker with Earth Consultants International, Alexandra Sarmiento of GeoPentech, and a team of archeologists with Brian F. Smith and Associates. The team was assembled and coordinated by Mr. Tom Freeman of GeoPentech, who also served as the report’s editor. This team was supported by Justin Zumbro from GeoPentech.
December 2012 Rev 0
Page iii
Paleoseismic Assessment of the Late Holocene Rupture History of the Rose Canyon Fault in San Diego 1.0
Introduction
The Rose Canyon Fault in San Diego County, interpreted as the southern continuation of the historically active Newport-Inglewood Fault in Orange and Los Angeles Counties (Figure 1), is a major component of the coastal system of strike-slip faults that together transfer 5-7 mm/yr of the crustal plate boundary deformation (DeMets et al., 1990; Bennett et al., 1996). Historical and paleoseismic activity on this zone suggests that much or all of the primary fault elements of the Newport-Inglewood/Rose Canyon Fault Zone sustained rupture in a sequence of earthquakes over the past few hundred years (Grant and Rockwell, 2002) (Figure 2), and that this sequence included faults farther south in Mexico, including the onshore Agua Blanca fault (Rockwell et al., 1993).
Figure 1. Regional map showing the relationship of the Rose Canyon and Newport Inglewood faults in the plate boundary system of fault. SAFZ = San Andreas fault zone, IF = Imperial fault, SJFZ = San Jacinto fault zone, EFZ = Elsinore fault zone, CPF = Cerro Prieto fault, LSF = Laguna Salada fault zone, SMFZ = San Miguel fault zone, ABFZ = Agua Blanca fault zone, RCF = Rose Canyon fault zone, NIFZ = Newport Inglewood fault zone, CBFZ = Coronado Bank fault zone, SDTF = San Diego Trough fault, SCFZ = San Clemente fault zone, WF = Whittier fault zone, ORF = Oak Ridge fault, SCF = San Cayetano fault, SYF = Santa Ynez fault, SCIF = Santa Cruz Island fault, SRIF = Santa Rosa Island fault, GF = Garlock fault.
The frequency and size of earthquakes on the Newport-Inglewood/Rose Canyon Fault Zone are key parameters in the seismic ground motion hazard analysis of SONGS, and there is a paucity of data on the late Holocene rupture history of large earthquakes on the NewportInglewood/Rose Canyon Fault Zone. The Rose Canyon Fault has sustained at least one late Holocene rupture, with the date of the earthquake estimated to be after AD 1450 (Murbach,
December 2012 Rev 0
Page 1 of 20
2000; URS, 1994; Grant and Rockwell, 2002) and most likely prior to construction of the San Diego Mission in 1769, as a large historical Rose Canyon earthquake would likely have destroyed that mission.
Figure 2. Principal elements of the coastal zone of faults, with the Newport-Inglewood-Rose Canyon-DescansoAgua Blanca fault zone delineated with a bold line (from Grant and Rockwell, 2002). Also shown is the Palos Verdes (PVF) – Coronado Bank (CBF) fault zone.
Previous paleoseismic investigations of the Rose Canyon fault at Rose Creek (Figure 3) documented that ruptures had recurrently produced slip in the early Holocene, but sedimentation at the Rose Creek site ended between 7,000 and 8,000 years ago, precluding development of a complete Holocene fault rupture history (Rockwell et al., 1991; Lindvall and Rockwell, 1995).
Figure 3. Map of the Rose Canyon fault zone in San Diego (modified from Lindvall and Rockwell, 1995). Note the location of their trench in Rose Creek, and the new site in Old Town (this study).
December 2012 Rev 0
Nevertheless, Rockwell (2010) interpreted the occurrence of several early Holocene earthquakes between about 9 ka and 5-7 ka, and also interpreted that there had been a hiatus in activity until the late Holocene event a few hundred years ago (Figure 4). This interpretation was based primarily on the observation that the most recent surface rupture displaced a relatively strong soil at the Rose Creek site that exhibited both an argillic (B) and albic (E) horizon, and that this soil capped and was not displaced by the earlier Holocene fault strands (Figure 5). Based on the strength of soil development, Rockwell (2010) estimated that it probably required on the order of 5000 years to attain the degree of observed development, which in turn required a hiatus in fault activity. However, a Page 2 of 20
weakness in this model is that the exposed fault strands capped by the soil exhibited primarily or solely strike-slip, whereas the strand attributed with the late Holocene activity expressed a dip-slip component. It was therefore plausible that later Holocene events may have not vertically displaced the soil and simply been poorly expressed in the trench exposures. Figure 4. Interpreted earthquakes from the Lindvall and Rockwell (1995) trench site in Rose Creek, as presented in Rockwell (2010).
Figure 5. (upper) Log of trench T4 from Lindvall and Rockwell (1995), with event evidence interpreted by Rockwell (2010). (lower) Reconstruction of the soil developed across the fault zone, which suggests that only the eastern part of fault strand 4 was active in the MRE.
December 2012 Rev 0
Thus, there was a need to develop a trench site in late Holocene sediments to expose observations that might be suitable to resolve whether the Rose Canyon Fault has experienced multiple late Holocene ruptures, which would imply smaller magnitude but more frequent surface rupturing earthquakes. New data may also indicate whether the Rose Canyon Fault behaves in a clustered fashion, as suggested by Rockwell (2010), or whether it has behaved in a more quasi-periodic fashion. A major implication is whether the displacement per event inferred by Rockwell (2010) is typical of Rose Canyon Fault events, or whether many more ruptures have produced the cumulative displacement observed on the Rose Canyon Fault at Rose Creek (Lindvall and Rockwell, 1995), implying a shorter average recurrence interval of smaller magnitude earthquakes. Towards this end, we conducted a number of studies to determine the feasibility of acquiring additional new data on the timing of past Rose Canyon Fault surface ruptures. These included: 1) analysis of early aerial imagery of the onshore portion of the Rose Canyon Fault, with much of the imagery pre-dating urban development; 2) selection of sites that Page 3 of 20
likely have had late Holocene sedimentation, and therefore had the potential to record a late Holocene earthquake record; 3) investigation into permitting requirements at the three sites that we identified with a potential Holocene record, two of which are on Caltrans property where permitting was found to be unrealistic in the timeframe of this project; the third site is on property owned by the City of San Diego at the Presidio Hills golf course in Old Town; 4) permitting with the City of San Diego and the City’s current leasee of the Presidio Hills golf course; 5) a CPT survey at the Presidio Hills Golf Course site to more precisely locate the fault in order to lessen the impact and cost of trenching; and 6) the excavating, logging and backfilling of an investigation trench in the Presidio Hills golf course site across the main strand of the Rose Canyon fault. In this report, we first present an overview of the Rose Canyon Fault, including information on its slip rate. We also present and discuss the three potential trench sites, followed by presentation of the CPT data from the Presidio Hills Golf Course. We then present the results of the new paleoseismic work and discuss its implications to our understanding of the rupture history of the Rose Canyon Fault. 2.0
Rose Canyon Fault Slip Rate
The slip rate on the Rose Canyon fault is not well constrained. Lindvall and Rockwell (1995) determined a minimum early Holocene to present slip rate of 1.1 mm/yr, with a best estimate of 1.5+0.5/-0.4 mm/yr based on 3D trenching in Rose Creek and interpretation of geomorphology. Their rate, however, included activity of only one strand of the fault system, the Mt. Soledad strand, and both the Rose Canyon and Country Club strands have some geomorphic expression. The precise relationship between the Mt. Soledad and Rose Canyon strands is complex, so it is possible that most or all of these two fault strands combine to pass slip through Lindvall and Rockwell’s trench site. However, the Country Club strand is expressed as a lineament across the same terrace deposits trenched by Lindvall and Rockwell (1995), suggesting additional Holocene displacement and implying that 1.5+0.5/-0.4 mm/yr is a minimum rate. More recently, Rockwell (2010) analyzed early aerial imagery in the Old Town area and interpreted two deflected streams as being offset about 250 m, with both incised into a terrace underlain by Bay Point Formation deposits. The age of the Bay Point Formation varies within San Diego, but most is attributed to deposition during the last major interglacial, including during the MIS 5a and 5e (Bird Rock and Nestor terrace equivalents at ca 80 and 120 ka, respectively). If correct, and if the deflected streams reflect actual displacement that post-dates incision into the MIS 5e terrace, then this implies a long-term slip rate of about 2 mm/yr for the Rose Canyon fault. Additional work to collect and summarize existing information on the subsurface geology of the Old Town area would help to further test and quantify this inferred rate. 3.0
Aerial Imagery Analysis
We acquired and analyzed several vintages of early aerial photographs to assess the location of the active strands of the fault through San Diego. The earliest photography available for the San Diego area are the 1928 Fairchild photographs, which provide imagery that predates much of the development of metropolitan San Diego. We also analyzed a set of 1941 stereo pairs that cover
December 2012 Rev 0
Page 4 of 20
much of the onshore fault from Old Town to La Jolla, a set of 1953 USDA stereo pairs, and modern Google Earth imagery. Together, these images provide excellent coverage of the City, prior to the rapid development after World War II, as well as coverage of the City as it exists today. We specifically searched for areas where Holocene sedimentation may preserve a Holocene record of surface faulting. The early Holocene terrace in Rose Creek Canyon is known to have stopped aggradation about 7-8,000 years ago, based on the trenching by Lindvall and Rockwell (1995). However, there are several alluvial fans that emanate from side drainages feeding into Rose Canyon that bury the early Holocene terrace, and these alluvial fans may record fault activity into the late Holocene where the fans cross the fault. We found two potential fault investigation trench sites on the late Holocene fans, both on Caltrans property along the I-5 corridor. One largely undisturbed site is beneath the bridges in the I-5/SR52 interchange (Northern Caltrans site on Figure 6). The second site on the late Holocene alluvial fan is adjacent to the Karl Strauss Brewing Company on Santa Fe Street between I-5 and Santa Fe Street (Southern Caltrans site on Figure 6). At these two Caltrans sites, the fault is expressed as a zone of linear scarps and vegetation lineaments. The Southern Caltrans site appears to have several deflected drainages that may indicate that deposition on this fan surface ceased prior to the past one to several events. The precise location of the fault strands at the Southern Caltrans site was to have been determined by pushing an alignment of CPTs at a high angle to the fault zone, but due to permitting issues, this site was not explored further.
Figure 6. Two potential trench sites (northern and southern) on Caltrans property where Holocene alluvial fan deposition may record the late Holocene rupture history of the Rose Canyon fault.
December 2012 Rev 0
We have been aware of the third site in Old Town at the City’s Presidio Hills Golf Course for many years, but due to past leasees of the golf course, trenching was not an option until recently. We began an extensive permitting process, as briefly described below, ultimately resulting in access to the site.
Page 5 of 20
The location of the fault was interpolated by projecting its geomorphic expression (Figure 7) from a few blocks to the south where a scarp crosses Juan Street in the 1928 imagery, and where recent geotechnical studies have shown the location of active fault strands (Leighton Consulting, 2007, Artim, 2000), northward to the Morena area, where the fault is well-expressed as a scarp, sags, deflected drainages and other geomorphic features. In the area of the proposed Old Town trench site, the fault had no expression, indicating that the active strand is buried. This is ideal for resolving the timing of past earthquakes, as it indicates an area of late Holocene sedimentation. It was this area that we chose to focus our efforts. 4.0
Permitting Process
Permitting the Old Town Presidio Hills Golf Course trench site turned out to be problematic from the start, as archeological and local resident concerns were considered of paramount importance. Specifically, the site lies near the base of a hill that rises to the Old Mission and Presidio, and the flood plain in the area of interest has been occupied or utilized from the earliest days of California’s European historical period. The earliest photography of this area shows Figure 7. Fault related geomorphology interpreted in 1928 building structures, and the modern aerial photography, with location of the Old Town golf course, road layouts that appear on the earliest CPT lines, and trench T1. maps available. Further, scattered shell fragments around the golf course indicated the possible presence of pre-European cultural resources. Consequently, there was substantial reason to believe that cultural resources would be impacted by this fault trenching investigation, which resulted in extensive review of the project plan by the City and the Presidio Hills Golf Course’s current leasee. The fault investigation on the golf course was conducted in multiple phases, with CPTs first pushed to better locate the fault. The first permit was granted to allow the CPTs because this method of investigation had minimal impact on the site or archeology. After completing the CPT investigations, as discussed below, we were required to reenter the permitting process in order to obtain permission to achieve the necessary detail from the trenching phase of the investigation. This second phase of the permitting process required an archeological study to be completed and carried through the trench excavations. The permit for the archeology and paleoseismic work was granted in late October, after which the trench alignment was cleared by an archeological team from Brian F. Smith and Associates. The actual fault investigation trench December 2012 Rev 0
Page 6 of 20
excavation was finally initiated in middle November and the logging and back filling of the trench was completed by early December, 2012. 5.0 Cone (CPT) Surveys
Penetrometer
Two fault investigation alignments were chosen for the CPTs at the Presidio Hills Golf Course. These alignments were selected where there was the best potential to achieve the goal of the paleoseismic trench investigations (Figure 7 and 8) with minimal disturbance to the City’s golf course and the surrounding park. Specifically, the two alignments were along the northwestern and southeastern property boundaries of the golf course so as to minimize impact to the golf course operations. Both of these areas are in “the rough,” although the northern alignment potentially affected a tee-off, depending on the precise location of the fault. The CPTs were pushed to a depth of about 90 feet, or to refusal, to identify where the sediments were displaced, or to where the bedrock stepped (Figures 9 and 10). From these two CPT alignments, the fault was inferred to be located between CPTs 3 and 4 along the northwestern alignment, and between CPTs 8 and 11 along the southeastern alignment (Figure 10). Figure 8. CPT locations in the golf course in Old Town.
6.0
Paleoseismic Trenching in Old Town
A trench was opened to a depth of about 4 meters across the area investigated with the CPTs (Figure 9) where there was an apparent step on the depth to refusal and in the apparent stratigraphic layers, which also approximately coincided with the projection of the fault in the early aerial imagery. The trench was shored with temporary shoring by the excavation contractor, D.P. Reynolds Corporation. Initially, the section of trench that we expected to encounter the fault appeared unbroken, and the trench was extended to the west until an early (Spanish or Mexican period) foundation wall and tile floor were encountered. The foundation had been excavated into the natural fluvial overbank
December 2012 Rev 0
Page 7 of 20
deposits, and more recent flood deposits buried the upper portions of the foundation wall. This observation required that the entire section that we had determined to be unfaulted was, in fact, very young. We then examined a clay layer exposed intermittently at the bottom of the trench approximately between trench station 10 and 15 on Plate 1, and found the fault. The fault was buried by about 1.5 m of unfaulted late Holocene stratigraphy, and another 2 m of mechanically emplaced fill. The fill was apparently emplaced in the 1930’s when the golf course was originally graded and built.
Figure 9. CPT cross-section (sleeve friction) and interpreted correlation of stratigraphy and faults for CPTs 1-6. The fault was interpreted to be between CPTs 3 and 4, and that is where it was found in trench T1.
About four meters of the trench were handexcavated an additional 60-70 centimeters (cm) in depth to near the top of groundwater to expose nearly a meter of the fault zone. The walls were cleaned with scrappers and ceramists’ clayscrolling tools, and the contacts were etched so as to be visible when photographed. In many cases, etching was not necessary if there was a significant color contrasts between different alluvial units. For the stratigraphy at Old Town, some primary units were easily distinguished and these were not etched, but the lower meter or so of strata are weakly bedded sandy silt and clayey silt strata that are similar in color. Consequently, although contacts could be readily observed when cleaned, they would not be visible in ordinary photography so they were etched. The trench face was gridded with string at halfmeter intervals for the four meters of fault zone that were studied in more detail. Each half-meter panel was photographed under different light conditions; once with natural light and again twice (with two different cameras and an I-Pad) under artificial light. The photographs of each panel were imported into Adobe Photoshop and rectified into squares, and then pieced together to make a mosaic of the entire gridded area. This mosaic became the logging surface upon which the final interpretations were made (Figures 11, 12, and 13).
Figure 10. CPT cross-section of CPT line 712 on the south side of the golf course in Old Town. The fault is interpreted where correlated strata make apparent steps. We interpret the main strand to be between CPTs 9 and 10.
December 2012 Rev 0
In addition to the detailed logging in the fault zone, the entire south face of the trench was logged at a scale of 1:20 to document the relationships between the natural strata, the artificial fill layers, and the foundation and tile floor of the Spanish/Mexican period structure (Plate 1). For this log, only the Page 8 of 20
primary unit contacts were described, whereas we chose one panel between shores to make a photomosaic of the entire exposed section (Figure 14). 6.1
Stratigraphy
In general, the natural stratigraphy exposed in the trench consisted of stratified sand, sandy silt and clayey silt deposits that are interpreted as the result of overbank sedimentation (Figure 14). This interpretation is based, in part, on the complete lack of axial channel gravel deposits, as are present in the San Diego River. Above the natural overbank deposits, three distinct fill layers are present that represent mechanically emplaced material from two or three separate sources. The upper fill, unit 1, is a 10-30 cm-thick layer of organically-enriched, dark sandy silt with abundant shell fragments and scattered brick fragments. The lower contact is abrupt and all of the shells are broken, both consistent with the layer being mechanically deposited. We interpret this layer to be redeposited midden material of unknown origin, but it is probably locally derived. Below the redeposited midden topsoil material, unit 2a is a meter-thick layer of reddish brown cobbles derived from the Linda Vista or San Diego Formation. This layer is dense, which we infer to represent its compaction during grading. There are abundant remnants of clay films, indicating that this deposit was derived from the subsoil (B horizon) of a Pleistocene (or older) stratum. The cobbles also have scattered brick fragments throughout, confirming an artificial fill origin. The lower part of unit 2 (2b) is composed of dark, fine-grained sandy silt and clay and is at least partly derived from stripping of local topsoil. It is organic rich and dense, indicating mechanical compaction. The partial ear of an earthenware urn handle (or some other pot) was found in this stratum, confirming its fill origin, or at least an historical component. There were also angular chunks of B horizon material (oxidized, clay-enriched) embedded locally in this layer, and the upper part buries the intact house floor and foundation wall of the Spanish/Mexican structure encountered at the west end of trench T1. We interpret this stratum as partly or wholly fill that was locally pushed around or derived topsoil material that was graded when the golf course was constructed, whereas the overlying oxidized gravel was probably imported to bring the site up to a certain grade level. The lower portion of this unit is possibly locally intact A horizon, but more likely, the living surface at the time of historical occupation is the buried soil or organic silty clay stratum of unit 3b. If correct, this implies that unit 3a is an historical overbank deposit and that the entirety of unit 2 is mechanical fill. The natural strata are divided into two principal units, units 3 and 4, although each represents multiple flooding events (Figure 14). Unit 3 is further subdivided into three thinner units, although further subdivision may be warranted. Unit 3a is a ~30 cm thick light brown, clean (well-sorted) alluvial sand interpreted as overbank sediments. The absence of an A horizon capping this unit suggests that the overlying fill of unit 2b is possibly the redeposited A horizon on the floodplain, or that it is very young and there was insufficient time to develop an A horizon. However, unit 3a appears to pond against or bury the Spanish/Mexican foundation wall, and along with a complete lack of pedogenesis, is interpreted as historical in age. A single sample of detrital charcoal recovered from unit 3a yielded a radiocarbon age of AD 1660-1950
December 2012 Rev 0
Page 9 of 20
December 2012 Rev 0
Page 10 of 20
Figure 11. Detailed log of a portion of the south face of trench T1, photographed under natural light. Unit descriptions are in the text, and summarized in figure 14.
December 2012 Rev 0
Page 11 of 20
Figure 12. Detailed log of a portion of the south face of trench T1, photographed under halogen light. Unit descriptions are in the text, and summarized in Figure 14.
December 2012 Rev 0
Page 12 of 20
Figure 13. Detailed log of a portion of the north face of trench T1, photographed under halogen light. Unit descriptions are in the text, and summarized in Figure 14.
Figure 14. Composite stratigraphic section from trench T1. The upper portion is from about station 17 (Plate 1), whereas the lower portion is from the deeper detailed logged part of the trench (Figure 11).
December 2012 Rev 0
Page 13 of 20
Table 1. Radiocarbon ages on detrital charcoal and marine shell specimens recovered from the strata exposed in trench T1. The upper 13 dating results are used in the OxCal model (Figure 15), whereas the lower four dates are not as they yielded ages that are several hundred years older than those on samples recovered from deeper strata.
Sample #
Unit
Material
C14 Age
19 3a historical flood 27 3c (lower) 29 3c (lower buried A) 45 3c?
charcoal
195+15
AD 1661-1951
charcoal charcoal charcoal
1005+15 1165+15 1165+15
AD 993-1034 AD 780-944 AD 780-944
3c? shell 3c (base, lower clean sand) shell 3c/4a contact shell
1430+15 1515+15 1510+15
AD 602-650 AD 469-602 AD 538-601
4a 4a
charcoal shell
1400+70 1735+15
AD 442-776 AD 246-379
4b 4c 4c 4d
charcoal charcoal charcoal charcoal
1510+15 1725+30 1815+15 2265+15
AD 538-601 AD 243-393 BC 162-AD 547 BC 392-234
3c (lower buried A) 3c (lower clean sand) 3c (lower clean sand 3c (lower clean sand)
charcoal charcoal charcoal charcoal
1725+15 1730+15 1805+15 1800+15
AD 255-381 AD 251-381 AD 135-249 AD 136-313
46 1 2 Event 1 8 3 Event 2 12 35 13 5 Didn’t use 38 4 24 43
Calib 2Age
(Table 1), consistent with deposition during the historical era. However, based on the construction of the Derby Dike in 1853, constructed to protect Old Town from flooding (implying that historical flooding had, in fact, occurred), we infer the age of unit 3a to be between AD 1769 and 1853. Unit 3b is stiff, dark brown sandy silt with minor clay of clear alluvial origin. This unit may be a buried A horizon associated with the living surface at the time of early historical occupation. Unit 3c is a 1.3 m-thick sheet of sand that probably represents at least two major flooding events, as there is hint of a buried soil about midway within this section (Figure 11). This unit can be subdivided into two sub-units. The lowest part is stratified with local cross-bedding, and has many krotovina (filled burrows) punctuating the otherwise natural-appearing cross-bedded sand. The upper meter is massive, has many krotovina, is enriched in organics and a darker brown color than the underlying stratified sand, and is interpreted as one and possibly two buried A December 2012 Rev 0
Page 14 of 20
horizons. Together, we interpret these to represent one or two large flooding events followed by a hiatus in deposition during which weak A horizon soils developed. The Spanish foundation was excavated into these deposits, so we infer the age of this stratum to be pre-1769. Unit 3c contained abundant detrital charcoal and we dated several; all samples yielded ages that range between AD 200 to 1000, indicating that the top of unit 3c is nearly a thousand years in age, assuming no detrital age inheritance. Unit 4 is a sequence of stratified silty clay, clayey silt sandy silty, and muddy sand that we interpret as overbank sedimentation on the San Diego River floodplain. We divide this sequence into at least four subunits, based on both stratigraphic and structural reasons, in that this is the part of the section that is faulted and records the earthquake history of the Rose Canyon fault in Old Town, as discussed below. In general, the stratigraphy is relatively poorly-expressed, as there are quite a few krotovina that are readily visible, and probably many more in the more massive parts of this unit. Locally, individual strata can be traced laterally for meters, whereas in other areas, the detailed stratigraphy is absent, mostly likely due to bioturbation. Fortunately, the stratigraphy is reasonably well preserved on the south face of the trench in the vicinity of and to the east of the fault, whereas on the north face, the stratigraphy is relatively massive at the east end of the deepened exposure, making interpretation difficult. Radiocarbon ages from unit 4 indicate that it ranges between about AD 700 and 500 BC, with the top of unit 4 constrained to about AD 700, consistent with our inferred age of unit 3. 7.0 Details on the Radiocarbon Results
Figure 15. OxCal model of the 13 samples listed in Table 1, along with the placement of events 1 and 2, as interpreted from exposures in trench T1. Shell dates are in green, whereas those on charcoal are dark gray.
December 2012 Rev 0
The results from the 14C dating program appear to be generally coherent, although they conflict to some degree with previous results from other studies in terms of the timing of the most recent rupture. There is clearly a component of detrital age inheritance for some samples, as dates for unit 3c range from about 1000 to 1800 years B.P. (Table 1). Nevertheless, the age dates Page 15 of 20
are generally older for stratigraphically lower strata, as expected (Figure 15). Further, the dates on shells (green symbols in Figure 15) agree quite well with those on detrital charcoal, and the shell is likely of cultural origin and not likely to have had a long residence period prior to burial. Age inheritance is due to the fact that a radiocarbon date records the timing of wood growth, rather than it’s burning or burial. If a sample is derived from the heart of a tree, or if the wood sat around for an appreciable amount of time prior to its burning and incorporation into the deposit, the age will be older by some amount that is equal to the difference between the age of the growth ring and the age of deposition. It is therefore expected to have some inheritance for most or all radiocarbon ages on detrital charcoal, although the amount of inherited age may be minor. Furthermore, if the charcoal is related to or derived from burning of wood by Native Americans, as is commonly the case, the resulting radiocarbon ages are almost always older than the strata that contain them (archeologists refer to this as the “old wood problem”), as most cooking fires will use older or dead wood rather than green wood. The radiocarbon dates on the shells provide some cross-correlation, as the shells are almost certainly of cultural origin and likely did not have a long residence in the system prior to burial. The four shell dates are all in stratigraphic order, providing some confidence as to whether they have been reworked or sat around in the system prior to burial. Further, they are in excellent agreement with the majority of the detrital charcoal ages. The youngest natural stratum, unit 3a, is interpreted to be historical in age based on the observation that it is stratigraphically above the level of the foundation stones and likely buries the living surface associated with the Spanish/Mexican structure, as discussed above. The single radiocarbon date from this stratum is consistent with this interpretation, as the calibrated age is AD 1680 to 1874 (we use the absence of the adobe structure in the 1874 photograph as evidence that it had already been obliterated by this time.) Detrital charcoal from Unit 3c yielded a range of ages between about AD 200 to 1000. Considering that a date on charcoal from the underlying unit 4a yielded a calibrated age of AD 600, we interpret that some of the charcoal from unit 3c may have as much as 400-500 years of inheritance. All things considered, however, most of the dates appear to be in stratigraphic order and record little or no evidence of significant age inheritance. This has important implications to the rupture history of the Rose Canyon fault, as discussed below. 8.0
Trenching Results – Interpretation of Rupture History
The expression of the fault in the trench exposures is different on opposite walls. This may be due, in part, to our interpretation that zones of massive bedding are most likely caused by bioturbation, as the same stratigraphic level is locally well-bedded elsewhere in the exposure. The south wall exposed a zone of folded and faulted strata, with evidence for two separate deformation levels. Units 4b through 4d are folded and faulted in a 2 m-wide zone that appears to be transpressive. This section was then capped by horizontally bedded strata of unit 4a, which was faulted at a later time. The overall expression of faulting is much more severe for the lower deformation event, in that is involved a broader zone with multiple fault splays, included folding of strata, and involved several secondary fault strands over a 2-3 m wide zone. In contrast, the
December 2012 Rev 0
Page 16 of 20
most recent faulting event occurred along a narrow, localized fault that abruptly offsets the top of unit 4 and is capped by finely stratified sand of unit 3. The north face of trench T1 exposed a narrower primary fault zone with no evidence of significant folding. The strata east of the main fault are massive and it is possible that folding may have been missed, but it appears that a lower contact east of the fault is flat-lying and not folded. This implies, then, that the folding observed on the south face is a very localized meterscale feature and may not be a good indicator of the relative size of each interpreted event. The most recent event is again seen to rupture to the top of unit 4 and is capped by fine-grained stratified sand of unit 3. The penultimate deformation event, interpreted as an earlier surface rupture, is not very evident on the north face, although there are hints of its existence. Several small faults terminate at the same level as the top of the folded section exposed on the south face, and there is a possible fissuring of strata up through unit 4b. Together, these observations are supportive evidence for the penultimate event, but they would be too weak to interpret a surface rupture based on their own merits. Last event. In conclusion, based on the two exposures, we interpret that at least two surface ruptures are preserved in T1 at the Old Town trench site. The best evidence is from the south wall, where there are clearly two levels of deformation, with the penultimate event faulting and folding up through unit 4b and capped by unit 4a. The ages of these two identified events are constrained by the radiocarbon results presented above. We calibrated the radiocarbon ages in OxCal, including the inferred event horizons within the model. The modeled PDFs of the event ages are shown in Figure 16 and suggest that both events occurred with a several hundred year window. Further, if the dates are taken at face value, the most recent surface rupture occurred over a thousand years ago, which conflicts with the inferred age of the MRE at both the downtown San Diego site (URS, 1994) and the La Jolla site (Rockwell and Murbach, 1996; Murbach, 2000, both of which indicate an age that is after AD 1450. We discuss the possible implications of this in the discussion section below.
Figure 16. Probability distributions for events 1 and 2, as determined in the OxCal model presented in Figure 15.
December 2012 Rev 0
9.0 Discussion, Implications and Needs for Additional Work The results from the Old Town golf course trenching program indicate the occurrence of two late Holocene surface ruptures on the
Page 17 of 20
Rose Canyon fault through Old Town. These observations can be interpreted in multiple ways, as outlined here, and it will take further work to resolve which model is correct. The first obvious issue is that the inferred age of the most recent surface rupture in trench T1 appears to be at least 500 years older than the results from previous studies. In La Jolla, Rockwell and Murbach (1996) (also cited in Grant and Rockwell, 2002) determined the age of young faulting by dating marine mollusk shells in a faulted Indian midden; the shells filled a fissure along the fault and samples were collected from the fissure for dating. Two individual shells were dated (conventional 14C dates of 920+50 and 880+50), yielding calibrated ages (using the published R correction of 225+35 years; Stuiver et al., 1986) of AD 1644+70/-160 and 1667+146/-144. The consistency of the two ages, both collected from deep in the fault fissure, lends credence to a post-AD 1500 age for the most recent surface rupture in La Jolla. Similarly, a trench investigation in downtown San Diego for a proposed sports arena site exposed a strand of the Rose Canyon fault displacing soil material (URS, 1994). Fissured soil contained detrital charcoal, which yielded a 13C-corrected conventional 14C age of 390+60 years B.P., which in turn leads to a calibrated age of AD 1420 to 1650. As the Old Town site lies between these two other trench sites, and as the entire onshore section of the Rose Canyon fault is less than 20 km in length, it was generally assumed that the similar dates from downtown San Diego and La Jolla represent the same surface rupture. In contrast, the most recent surface rupture at Old Town appears to be at least 500 years older, which leads to several possibilities. First, it is possible that we did not expose the entire width of the Rose Canyon fault in the golf course. The CPT data support the presence of a primary fault strand in the same area that we encountered the fault in T1, but the CPT data are complicated to the east because of the presence of abundant coarse material. It is therefor plausible that another, more recent fault strand is present east of the end of trench T1. Alternatively, it is possible that all of the radiocarbon samples recovered from unit 3c have between 500 and 1000 years of inherited age and that the most recent rupture we observed at the top of unit 4 actually occurred in past several hundred years. This would require, however, a remarkable coincidence to have most of the dates in stratigraphic order, and we consider this to be a low probability. It is also possible that the radiocarbon dates from La Jolla and downtown San Diego are affected by some contamination source, and that the ages of the most recent surface ruptures in those studies are actually older than reported. However, this also seems unlikely as there is no other evidence to support this conclusion. Furthermore, a study conducted for the U.S. Navy Pier 700 project (URS, 1998) determined faulted stratigraphy in San Diego Bay using seismic reflection profiling techniques. Cores from the fault strata yielded samples for radiocarbon dating which, in turn, yielded a radiocarbon age that is younger than 800 years B.P. All of these other observations support the occurrence of a young surface rupture on the Rose Canyon fault in the past 500 years, making the Old Town results somewhat problematic. The simplest explanation seems to be that we did not expose the most recently active strand of the fault at the golf course and that a longer trench is warranted.
December 2012 Rev 0
Page 18 of 20
The other major conclusion from this study is that there have been at least two surface ruptures on the Rose Canyon fault during the late Holocene. This apparently conflicts with the model of Rockwell (2010), where only one late Holocene event was recognized at Rose Creek. There are at least three plausible explanations to account for these collective observations. First, it is possible that the two events identified at Old Town are the result of a main shock and subsequent after-slip, which would require deposition of unit 4a almost immediately after the primary rupture. Another explanation may be that the upper (MRE) event is the result of triggered slip (perhaps by rupture of an adjacent segment) whereas the lower (event 2) is a primary rupture. Both of these models are supported by the observation on the south wall that event 2 seems to have broken a much broader fault zone and that event 1 is confined to a narrow slip zone. Alternatively, it is possible that the stratigraphy at the Rose Creek site of Lindvall and Rockwell (1995) was insufficient to resolve the occurrence of two closely timed events. If the dates on event 1 and 2 from this study are taken at face value, these two events are separated by only a few centuries and it is plausible that two such closely-timed events would appear as a single event, as there was no deposition on the Holocene terrace at their trench site on Rose Creek after about 7 ka. If this is the case, then this has significant implications as to the average recurrence interval on the Rose Canyon fault, as well as its recurrence behavior. Two events in the past 1500 years indicates either that the model proposed by Rockwell (2010) is incorrect and that the average recurrence interval is significantly shorter than currently believed, or that the fault is behaving in a clustered mode and that there have been two closely-timed events in the late Holocene. A third possibility is a combination of the above possibilities; that the two events identified at the Old Town golf course site are in addition to the most recent event identified at La Jolla and downtown San Diego, in which case there have been three events in the past 1500 years. In this scenario, the recurrence interval during the current cluster of large surface-rupturing earthquakes may be in the range of 400-500 years, suggesting a shorter average recurrence interval of smaller magnitude earthquakes with the MRE nearly 500 years ago. In summary, the new results have provided much food for thought in terms of both the recurrence pattern of surface rupturing earthquakes on the Rose Canyon fault, as well as the near-term likelihood of future earthquakes. 10.0
References Cited
Artim, E.R., 2000, Soil/Geological Investigation, Proposed 28 Unit Addition to Hacienda Hotel, Existing County Parking Lot and Adjacent Slope Areas, SW Corner of Intersection of Juan and Harney Streets, San Diego, California, October 22. Bennett, R. A., Rodi, W., and Reilinger, R. E., 1996, Global Positioning System constraints on fault slip rates in southern California and northern Baja, Mexico: Journal of Geophysical Research, v. 101, no. B10, p. 21,943-21,960. DeMets, C., Gordon, R.G., Argus, D.F. and Stein, S., 1990, Current Plate Motions: Geophysical Motions: Geophysical Journal International, v. 101, p. 425-478.
December 2012 Rev 0
Page 19 of 20
Grant, L.B. and Rockwell, T.K, 2002, A northward Propagating Earthquake Sequence in Coastal Southern California?: Seismological Research Letters, v. 73, no. 4, p. 461-469. Grant, L.B., Ballenger, L.J., and Runnerstrom, E.E., 2002, Coastal uplift of the San Joaquin Hills, southern Los Angeles basin, California, by a large earthquake since A.D. 1635: Bulletin of the Seismological Society of America, v. 92, no. 2, p. 590-599. Leighton Consulting, Inc., 2007, Geologic Investigation Report – Earthquake Fault Hazard Study, 2510 Juan Street, San Diego, California, June 8. Murbach, M.L., 2000, The Rose Canyon Fault Zone: New Evidence for Holocene Earthquake Activity in La Jolla, California: Master’s Thesis, San Diego State University. Lindvall, S. and Rockwell, T.K., 1995, Holocene Activity of the Rose Canyon fault, San Diego, California: Journal of Geophysical Research, v. 100, no. B12, p. 241221-24132 Rockwell, T.K., 2010, The Rose Canyon Fault Zone in San Diego: Fifth International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, May 24-29, Paper No. 7.06c, 9 pp. Rockwell et al., 1993, Late Quaternary Slip Rates Along the Agua Blanca Fault, Baja California, Mexico: in (P.L. Abbott, ed.) Geological Investigations of Baja California: South Coast Geological Society, Annual Field Trip Guidebook No. 21, p. 53-92 Rockwell, T.K., and Murbach, M.L.,1996, Holocene Earthquake History of the Rose Canyon Fault Zone: U.S. Geological Survey Final Technical Report for Grant No. 1434-95-2613. Rockwell, T.K., Lindvall, S.C., Haraden, C.C., Hirabayashi, C.K., and Baker, E., 1991, Minimum Holocene slip rate for the Rose Canyon fault in San Diego, California in Abbott, P.L., and Elliott, W.J., eds., Environmental Perils San Diego Region: San Diego, San Diego Association of Geologists, p. 37-46. Stuiver, M., Pearson, G.W., and Brazunias, T., 1986, Radiocarbon age calibration of marine samples back to 900 cal yr BP: Radiocarbon, v. 28, no. 2B, p. 980-1021. URS Greiner Woodward Clyde, 1994, Report of Fault Hazard for the Entertainment and Sports Center, San Diego, California. URS Greiner Woodward Clyde, 1998, Additional Fault Hazard Investigation CVN Berthing Wharf-Phase II (P-700A) Naval Air Station North Island Coronado, California.
December 2012 Rev 0
Page 20 of 20
