Geoethics in Hollywood : How Can We Put More Reality into Fault Hazard Zoning ?

The 15-km-long Hollywood fault extends through some of the most densely developed and expensive areas in southern California. Given that billions of dollars of real estate could be impacted by an earthquake occurring on the fault, and that tens of thousands of people live near it, the seismic hazard posed by the Hollywood fault is concerning. But how much so? The City of West Hollywood has required geological fault investigations and building avoidance of at least 15 meters from the fault’s active trace since the mid-1990s, resulting in numerous site investigations. Geologic studies in support of the Los Angeles Metro subway and subsequent research found as-yet-unconfirmed equivocal evidence of an early Holocene-age (~8 ka) displacement event, and estimated strain rates of 0.3-0.9 mm/yr. In 2014, and following a “not-in-my-backyard” press campaign against the Hollywood Millennium project, the California Geological Survey zoned the eastern Hollywood fault as Holocene-active under their fault zoning program, requiring geological investigations and building setbacks from Holocene-age faults. In the years since the fault was zoned by the City of West Hollywood and the California Geological Survey, millions of dollars have been invested in geological studies of the fault for building projects. All of these dollars have been reluctantly spent by the private sector, and all of this work has been eagerly done by consulting firms. The results? With only four exceptions, two in West Hollywood, and two in the City of Los Angeles, and all having opportunities for alternative interpretations, no Holocene-age fault displacements have been found. Instead, other studies have found definitive evidence that the fault has not ruptured in 10s to 100s of thousands of years. So what is the ethical alternative? Should we ignore the hazard posed by the fault? Are we geo-professionals too enamored of the financial gains such regulatory zoning provides that we are unwilling to suggest changes? Could we move past the “one size fits all” zoning that requires that single-family homes are held to the same standard as 40-story buildings? In the following sections we present several case studies along the Hollywood fault and provide our suggestions for a more progressive program of fault hazard management.


IN TROD UCTION
he H ollyw ood fau lt is m apped as an east-to northeast-trend ing stru ctu re along the sou thern m argin of the Santa Monica Mou ntains (locally referred to as the H ollyw ood H ills), placing late Ju rassic m etam orphic and Cretaceou s to Miocene granitic and sed im entary rocks on the north against the Qu aternary-aged valley fill of the H ollyw ood Basin.Most recent m ap s generally show the H ollyw ood fau lt extend ing from the w est bank of the Los Angeles River w estw ard , throu gh the cities of Los Angeles, West H ollyw ood and Beverly H ills, a d istance of abou t 15 km (Fig. 1 and 2).The fau lt is consid ered part of the approxim ately 230-km-long fau lt system at the sou thern ed ge of the Transverse Ranges physiographic province of sou thern California.N earby fau lts that are also part of this system inclu d e the Raym ond and Sierra Mad re fau lts to the east, and the Santa Monica, and Malibu Coast fau lts to the w est.

MAPPIN G OF THE HOLLYWOOD FAULT THROUGH TIME
As early as 1931, H oots and Kew p rep ared the first com prehensive geologic m ap of the area, and inferred the Santa Monica-H ollyw ood T fau lts as bu ried stru ctu res at or near the sharp break in slope along the sou thern m argin of the Santa Monica Mou ntains.The fau lt, or sections of the fau lt zone, w ere also m ap ped by Lam ar (1970), H ill et al. (1978, 1979), Weber et al. (1980), Crook and Proctor (1992), w ith m ap com pilations by Dibblee (1991a, b) and Yerkes and Cam pbell (2005).The m ost d etailed m app ing of the H ollyw ood -Santa Monica fau lt w as d one by Dolan et al.Dibblee's (1991a, b) geologic maps as a base.N ote that in the east-central portion of the map, most researchers place the fault in almost the same area, whereas to the west and east, the interpreted fault locations vary, sometimes significantly.(1992,1997), w ho u tilized U.S. Geological Su rvey 1920s vintage 5-foot (~1.5 m ) contou r top ograp hic m aps of the area to m ap scarps, fans and other potentially fau lt-related topographic featu res now hid d en by u rban d evelopm ent.Dolan et al. (1992Dolan et al. ( , 1997) id entified several linear scarps and faceted sou th -facing rid ges (see yellow zones in Fig. 2) that they interpreted as the near-su rface expression of the fau lt.Where the scarps are broad featu res 15 to 60+ m w id e, the researchers com bined the geom orphic d ata w ith su bsu rface inform ation obtained from geotechnical and grou nd w ater stu d ies in an attem pt to better constrain the location of the m ajor traces of the H ollyw ood fau lt.H oots (1931) and other early researchers had inferred the fau lt to be prim arily a north -d ip ping reverse stru ctu re responsible for the grow th of the a djacent H ollyw ood H ills. H ow ever, based on the fau lt's linear trace and steep 80to 90-d egree d ip s observed in d ow n -hole logged borings, Dolan et al. (1997Dolan et al. ( , 2000) ) interpreted the fau lt as prim arily left-lateral strike-slip w ith a sm all am ou nt of north -sid e u p, reverse com ponent.N otably, in the West H ollyw ood and Beverly H ills areas, Dolan et al. (1997Dolan et al. ( , 2000) ) interpreted the H ollyw ood fau lt sou th of the m ain break in slope that H oots and Kew (1931) an d m ost other researchers had u sed to infer the location of the fau lt.Several fau lt-location stu d ies in the West H ollyw ood area have been cond u cted since the m id -1990s.Abou t a third of these stu d ies have encou ntered or interpreted fau lts, and the location s of these have rou ghly agreed w ith Dolan et al.'s (1997Dolan et al.'s ( , 2000) ) fau lt scarp s, althou gh the zone of fau lting has been show n to be w id er and m ore com plex, interpreted to consist of both north -d ip ping reversesep aration stru ctu res, and sou th -d ip ping norm al fau lts w ithin the hanging w all.Significan tly, how ever, only one of these fau lts has been interpreted as H olocene active, based on the find ings at the Su nset Millenniu m and H ou se of Blu es sites (on the far left in Fig. 3), w here rad iocarbon-d ated sam ples w ere collected from above and below the fau lt zone in d ow n -hole logged borings.Several of these stu d ies also fou nd that the steep step in the bed rock that w as previou sly interpreted as a fau lt scarp (H oots and Kew , 1931;Dibblee, 1991a, b) is an ancient beach seacliff.Trenches and borings have exposed an old m arine abrasion (w avecu t) platform and associated m arine d ep osits estim ated to be at a m inim u m 400,000 to 900,000 years old (su m m arized in H ernand ez and Treim an, 2014; H ernand ez, 2014).Sim ilar m arine sed im ents have been observed at d epth both to the w est and east, in Beverly H ills and H ollyw ood , respectively.The H ollyw ood fau lt w as interpreted or observed in som e of the transects cond u cted for infrastru ctu re projects su ch as the Los Angeles (LA) Metro Red Line along Cam ino Palm ero-Martel Avenu e, and La Brea Avenu e, the LA MetroRail project along Cahu enga Bou levard , and storm d rain trenches along Vista Street and Fu ller Avenu e (Dolan et al., 1997).Dolan et al. (2000) revisited the Cam ino Palm ero-Martel Avenu e location and u sed a series of closely sp aced d ow n-hole logged borings to better locate and stu d y the fau lt.The find ings from that stu d y su ggested that the m ost recent su rfaceru ptu ring earthqu ake (MRE) on the H ollyw ood fau lt occu rred abou t 7,000 to 9,500 years ago, w ith a possible penu ltim ate event abou t 22,000 years ago, ind icating a long interval betw een earthqu akes, and an u nconfirm ed estim ated strain rate of 0.3-0.9m m / yr.Given its length, the H ollyw ood fau lt is only cap able of generating abou t a Mw ~6.6 earthqu ake.To prod u ce a larger-m agnitu d e earthqu ake requ ires that the fau lt ru ptu re together w ith other fau lts to the w est and east, su ch as the Santa Monica, Malibu Coast and Raym ond fau lts.The existing paleoseism ic d ata, how ever, show that each of these fau lts has a significantly d ifferent earthqu ake history, w ith thou sand s of years b etw een events on d ifferent fau lts.Thu s, the available d ata su ggest that the H ollyw ood fau lt ru ptu res ind epend ently, has a slow rate of slip, and takes thou sand s of years to reach its breaking point.Assu m ing the MRE age and strain rate are correct, the fau lt has accu m u lated b etw een 2.1 and 8.5 m of strain, u nlikely high valu es for su ch a short fau lt.

Hollyw ood Millennium Sites and Hudson Avenue
The p rop osed East and West H ollyw ood Millenniu m sites (far right on Figu re 3) extend across the Yu cca Street section of the H ollyw ood fau lt as m apped by CGS on their Alqu ist-Priolo (A-P) m ap.The geotechnical consu ltants for this p roject (Grou p Delta Consu ltants) pu shed m ore than 100 cone pen etrom eter test (CPT) probes, d rilled and logged m ore than 30 continu ou sly cored borings, and excavated and logged tw o trenches w ith a com bined total length of m ore than 120 m (Fig .4).These stu d ies have show n that the area is u nd erlain by ~10 m of H olocene-aged allu viu m d eposited w ithin a canyon that w as carved into Pleistocene-aged m u d flow d eposits and , in the northern p ortion of the stu d y area, Top anga Form ation (~15Ma) bed rock and old er allu vial fan d eposits that significantly pre-d ate the m u d flow d eposits.These su bsu rface stu d ies su ggest the presence of at least tw o m inor fau lts u nd er the area, w ith last m ovem ent on these fau lts occu rring betw een abou t 100,000 and 150,000 years ago.Thu s, the area is not u nd erlain by active fau lts, contrary to the CGS' m ap that ind icates the active H ollyw ood fau lt extend s throu gh this site.Sim ilarly, CPTs and borings d rilled along a section of H u d son Avenu e encou ntered a d iscontinu ity in the d eeper layers that is either an erosional su rface or a fau lt.The featu re is overlain by several u nbroken layers of old er allu viu m m od ified by soilform ing p rocesses.The d egree of soil d evelopm ent exhibited by these allu vial sed im ents ind icate that if the d iscontinu ity is fau ltcontrolled , the fau lt has not m oved in at least 50,000 years.This estim ate is consid ered a m inim u m valu e given that m ost of the overlying u nbroken soils w ere tru ncated , ind icating significant erosional period s.

Cherokee Avenue Sites
To d ate, three sep arate stu d ies have been cond u cted off Cherokee Avenu e, in H ollyw ood .The first stu d y fou nd fau lted bed rock w ith no overlying allu vial sed im ents from w hich to d eterm ine fau lt activity; the CGS chose to consider the fau lt active (see H ernand ez et al., 2014).In 2016 w e cond u cted a stu d y im m ed iately to the sou th.Ou r easternm ost trenches exposed the H ollyw ood fau lt placing bed rock of the Top anga Form ation (Tt) over Pleistocene age allu viu m (Fig. 5).The fau lt d ipped to the north at abou t 40 d egrees at the bottom of the trenches, bu t rolled over to nearly horizontal, form ing the contact betw een tw o old er allu vial d epo sits.The fau lt tip thru st Qoal2 u nd er Qoal3 d u ring its last su rface-ru ptu ring earthqu ake.The fau lt cou ld not be traced u pw ard into the u pper section of the Qoal2 u nit, as this contact w as intensely biotu rbated .In a parallel trench, how ever, the top of the Qoal1 u nit w as clearly not broken, so w e are certain that the fau lt has not m oved since the soil in the Qoal1 u nit form ed .Ou r age estim ates for that u nit, based on d egree of soil d evelop m ent, range from abou t 160,000 to 230,000 years, so this fau lt has not m oved in at least 160,000 years, a find ing that w as later confirm ed by others on an ad jacent property.

D ISCUSSION
Despite extensive and costly investigations, evid ence that the H ollyw ood fau lt is active r e- California law for active fau lt zoning specifies avoid ance of the fau lt trace as the only perm itted m itigation alternative if the fau lt has exp erienced even a single, m inor H olocene d isplacem ent.This binary, one-size-fits-all choice has led m any ow ners to not u pd ate their existing old er bu ild ings becau se a fau lt stu d y w ou ld be requ ired before they can red evelop.If a H olocene fau lt is fou nd u nd er their site, the ow ner w ou ld likely be prohibited from u nd ertaking stru ctu ral m od ernization for seism ic safety.Thu s, one of the law 's u nintend ed consequ ences is the p reservation and continu ed u se of stru ctu res that are seism ically u nsafe from high-probability regional earthqu ake events becau se of the low -probability of su rface fau lt ru ptu re throu gh their site.We p ropose a m ore p rogressive ap proach to m itigate the hazard of su rface fau lt ru ptu re that takes into accou nt the forty years of earthqu ake geology know led ge gained since the law w as enacted .We su ggest that su rface fau lt ru ptu re m itigation be incorp orated into the bu ild ing cod e, rather than being cod ified into law .In d oing so, geologists w ill be tasked w ith d eterm ining the fau lt's location and ru ptu re kinem atics, tim ing of prior events, and estim ated tim e to the next event.Then, based on the type of stru ctu re proposed , a risk m anagem ent ap proach w ou ld be em ployed scaled to the hazard and the risk, sim ilarly to all other geologic risk factors (seism ic shaking, grou nd failu re, slope failu re, liqu efaction, flood ing, etc.) that are alread y consid ered in the bu ild ing cod e. Thu s, for single-fam ily resid ential hom es, a fau lt w ou ld be su bject to m itigation requ irem ents if the fau lt lies w ithin 500 years of its average recu rrence interval.For larger resid ential and com m ercial stru ctu res, the fau lt w ou ld be su bject to m itigation if it lies w ithin 1000 years of its average recu rrence interval, and for critical infrastru ctu re and facilities, that interval cou ld be 1500 or 2500 years.These hazard w ind ow s are sim ilar to those for seism ic stru ctu ral d esign, liqu efaction and land slid e geotechnical stabilization, and exceed those for flood hazard s.Mitigation alternatives m ight inclu d e stru ctu ral isolation or reinforcem ent, soil strengthening, or avoidance, d epend ing u p on the anticip ated fau lt's d isplacem ent, site cond ition s, and econom ic or social valu e. Continu ing d ow n the forty-year-old p ath of strict avoid ance for any H olocene-age fau lt, d espite its d isplacem ent history or m agnitu d e, can no longer be d efend ed as a cost-effective solu tion to im prove pu blic safety.Intelligent and progressive m itigation of know n and properly qu antified hazard s is the basis of all perform ance-based d esigns, and it is tim e for California's geologists, and the California Geological Su rvey, to m ove on from the A-P Act by either m od ernizing or su nsetting it, and letting the bu ild ing cod e process hand le fau lt ru ptu re hazard as ju st another geologic factor in land planning and d evelopm ent.It is the ethical thing to d o.

ACKNOWLEDGEMEN TS
Thank you to Maria H erzberg of Earth Consu ltants International for preparing Figu res 1 throu gh 3. The report w as im proved follow ing com m ents and ed its to the original m anu script provid ed by Um berto Fracassi, an anonym ou s review er, and Giu sep pe Di Capu a. Thank you for you r valu able and honest feed back.

Figure 1 :
Figure 1: Location of the Hollywood fault at the base of the Hollywood Hills, forming the northern boundary to the Hollywood basin.The fault is considered a short segment of the approximately E-W to N E-W SW system of faults that include the Santa M onica, M alibu Coast, Raymond and Sierra M adre faults shown on this figure.A dditional faults that are thought to be part of this system occur farther west and east, respectively, of the mapped image.Y ellow square shows the area covered in Figures 2 and 3.

Figure 2 :
Figure 2: V arious interpretations of the location of the Hollywood fault usingDibblee's (1991a, b)  geologic maps as a base.N ote that in the east-central portion of the map, most researchers place the fault in almost the same area, whereas to the west and east, the interpreted fault locations vary, sometimes significantly.

Figure 3 :
Figure 3: Part of the zone of required investigation (yellow area) established by the California Geological Survey (CGS, 2014) for the Hollywood fault.A ny property within the yellow zone is required to be investigated for possible active faulting prior to it being developed or re-developed.Blue stars denote sites discussed in the text where faults found to be not active were exposed in trenches or were interpreted from transects of CPTs and borings; red stars show the only sites where radiocarbon-dated sediments have suggested the occurrence of Holocene-active faults.The black lines show the location and extent of transects across the fault zone that were conducted between the 1980s and approximately 2014.Since the fault was zoned, several additional studies have been conducted; only a few of these (such as Cherokee, Hudson, and Hollywood M illennium) are included here.
Follow ing the City of West H ollyw ood example from the m id -1990s, the City of Los Angeles began requ iring fau lt stu d ies in the H ollyw ood area in the early 2000s.Then in 2014, that portion of the H ollyw ood fau lt in the H ollyw ood qu ad rangle w as zoned by the California Geological Su rvey u nd er the provisions of the 1972 Alqu ist-Priolo Earthqu ake Fau lt Zoning Act(Figu re 3;H ernand ez and Treim an, 2014;H ernand ez, 2014; CGS, 2014).As a resu lt of these requ irem ents, several stu d ies have been cond u cted in the H ollyw ood area (su m m arized in H ernand ez and Treim an, 2014; H ernand ez,

Figure 4a :
Figure 4a: M ap showing the extensive data set used to evaluate the Hollywood M illennium sites, including trenches (in light turquoise), CPTs (in red), and continuously sampled borings (in blue).The inferred location of the Y ucca Street section of the Hollywood fault shown by the black dots.Two faults (orange lines) were interpreted from the subsurface data; these faults are minor and have not moved in at least 100,000 years.

Figure 4b :
Figure 4b: Photograph showing a portion of the 9+ m deep trench excavated in the parking lot of the Tower Records building, which is also the East Millennium site.The footprint of the trenches is shown in light turquoise on the figure to the left.A demonstrably not active fault with about 50 cm of total displacement was exposed near the bottom of the trench.Photo courtesy of Group Delta Consultants.

Figure 5 :
Figure 5: Log of southern end of Trench ECI-1 at the Cherokee A venue site showing the fault in red, and the geologic units described in the text.