Updated Seismic Hazard Curves, Maps, and Spectra for the Northern Dominican Republic using a Probabilistic Seismic Hazard Analysis

doi:10.21203/rs.3.rs-2082413/v1

Download PDF

Research Article

Updated Seismic Hazard Curves, Maps, and Spectra for the Northern Dominican Republic using a Probabilistic Seismic Hazard Analysis

https://doi.org/10.21203/rs.3.rs-2082413/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this older preprint version

Read the latest preprint version →

This article presents updated seismic hazard curves, maps, and spectra of ground motion intensity parameters for the northern region of the Dominican Republic (DR) obtained using a probabilistic seismic hazard analysis (PSHA). The analysis uses as input data an earthquake recurrence model based on fault slip rates derived from GPS measurements published in the aftermath of the 2010 Haiti Earthquake. The seismicity rates data is used to calibrate a composite characteristic earthquake model, which is combined with a Poisson process to provide a temporal characterization of earthquakes occurrence. The PSHA accounts for uncertainties in the analysis by modeling magnitude and site-to-rupture distance as random variables, considering the aleatory uncertainty associated to the ground motion prediction equations (GMPE) used for spatial characterization of earthquake-induced ground motions. The seismic hazard curves and maps include (horizontal) peak ground acceleration and pseudo-spectral response accelerations at 0.2s and 1.0s periods for 5% damping and firm rock sites. The results show that the ground motion parameters with a 2% probability of exceedance (PE) in 50 years determined in this study are up to 46% larger than the corresponding parameters specified in the current DR building code for the northern DR. Moreover, the design response spectra for a site in the city of Santiago specified in the code is significantly lower than the 2% PE in 50 years uniform hazard spectra determined in this study for periods smaller than 0.5s, a range that includes the majority of the structures that define the built environment of the DR.

Probabilistic seismic hazard analysis

PSHA

earthquakes

Dominican Republic

seismic hazard

The Hispaniola Island is located in a seismically active region bounded by two thrust systems defined by the convergence of the North American and Caribbean plates. Two strike-slip faults, the Septentrional fault and the Enriquillo-Plantain Garden fault, bound the internal Gonâve microplate as shown in Fig. 1. The damage potential of the Hispaniola’s seismic faults was demonstrated after the 2010 and 2021 Haiti earthquakes, the former considered one of the most devastating disasters caused by a natural event in recent history. The 2010 Haiti earthquake caused a coastal uplift with no significant surface rupture along the Enriquillo-Plantain Garden fault. The earthquake resulted in major life and economic losses with a death toll estimated at 300,000 casualties and an economic loss estimated at nearly US$14 billion exceeding the gross domestic product of the country (Prentice et al., 2010; DesRoches et al., 2011).

Despite recent seismic activity along the Enriquillo-Plantain Garden fault in the southwestern region of the Hispaniola, the northern region of the island has long been recognized as a high seismic hazard zone due to the Septentrional fault (Calais et al., 1998; DeMets et al., 2000; Calais et al., 2002; Mann et al., 2002). The Septentrional fault is a left-lateral strike-slip fault that crosses the northern region of the Hispaniola and accommodates part of the motion between the North America and Caribbean plates; a segment of the fault runs near several densely populated cities of the Dominican Republic (DR). Trenches excavated in the area and radiocarbon analyses have shown that the fault has accumulated at least 5m of deformation, and that the most recent severe surface rupture in the central portion of the fault likely occurred before 1260 (Mann et al., 1998). The recurrence interval of the fault is estimated to be between 800 and 1200 years (Prentice et al., 2003), and the slip rate is estimated to be 12 mm/year with a characteristic magnitude or maximum moment magnitude of rupture of 7.8 (Calais et al., 2002; Frankel et al., 2011). Based on the conclusions of the aforementioned the Septentrional fault may be in the late phase of its rupture cycle (Calais et al., 2002).

The Septentrional fault is the main contributor to the seismic hazard and seismic risk in the northern region of the DR due to its closeness to densely populated cities such as Santiago and Puerto Plata (Fig. 2). For this reason it is of interest to study and quantify the characteristics of ground motions induced by a potential rupture of this and other nearby faults, and its effects on the built environment (Erazo, 2019; Erazo, 2020; Rojas-Mercedes et al., 2020; Erazo and Taveras, 2021).

The quantitative characterization of the potential earthquake-induced ground motions at a site or a region is known as a Seismic Hazard Analysis (SHA). The first SHA proposed in the literature were based on a deterministic analysis, performed using an estimate of the maximum expected earthquake that a source can generate and the minimum distance from a site to the source. A deterministic SHA based on a “worst case scenario” has several limitations, including the difficulty to rigorously combine the effects from multiple sources, a lack of uniformity and consistency when comparing the results for different sites, and an analysis that might result in over-conservative parameters when used in the design of civil structures.

The use of a Probabilistic Seismic Hazard Analysis (PSHA) was proposed by Cornell (1968) to address some of the limitations of deterministic seismic hazard analyses. A PSHA provides a rigorous and consistent framework to estimate earthquake-induced ground motion parameters at different spatial locations, considering the hazard contribution from multiple sources and accounting for the likelihood of each individual source to generate earthquakes of different magnitudes (McGuire, 1995; Rathje and Saygili, 2008; Han and Davidson, 2012; Rodriguez-Marek et al., 2014; Abrahamson et al., 2019; Gerstenberger et al., 2020; Baker et al., 2021).

The objective of PSHA is to determine the probability (or rate) of exceeding various ground motion intensity parameters, accounting for the main sources of uncertainty identified by the analyst. PSHA allows to account for uncertainties related to the likelihood of earthquake magnitudes that can be generated by a source, as well as uncertainties related to rupture length, distance from a site to the potential rupture locations, ground motion prediction equations, among others. Despite some of the challenges involved in defining the inputs to probabilistic seismic hazard analyses, such as limited strong earthquakes data for some regions, the outputs of PSHA have shown to be robust and useful in engineering practice when applied to the design of civil structures, helping to protect public safety against earthquakes. PSHA is the most widely accepted tool to quantify earthquake-induced ground motion intensity parameters and its predictions have been tested with observations (Kammerer and Ake, 2012).

This article presents the results of a PSHA performed for the northern Dominican Republic. The PSHA performed is a convolution-based analysis where the hazard integral is solved directly rather than using a numerical approach. The analysis accounts for the contribution to the seismic hazard of the main sources of known significant seismicity within a 200km range from the sites of interest. The considered sources are (refer to Fig. 2): the Septentrional fault − (3)SFZ, the Northern Hispaniola Subduction Zone – (1)NHFZ, the Matheux-Neiba fault – (10)NFZ, the Enriquillo-Plantain Garden fault – (11)EPGEZ, and the Muertos Trough Subduction Zone – (14)MTFZ. No background seismicity is considered in this study, and although excluding background seismicity underestimates the overall seismic hazard, the impact is typically negligible at the hazard level of a 2% probability of exceedance in 50 years that is the focus of this study.

The PSHA performed uses as input data fault slip rates derived from GPS measurements for the region published in the aftermath of the 2010 Haiti Earthquake (Frankel et al., 2011). The rates are used to calibrate a composite characteristic earthquake model of the type proposed by Youngs and Coppersmith (1985) in which the neighborhood of the characteristic earthquake magnitude is modeled as uniform and the low-moderate magnitude range as an exponential tail. The resulting model is combined with a Poisson time-independent model for earthquake recurrence. No time correlation or hazard from aftershocks are considered.

The ground motion intensity parameters of interest include horizontal peak ground acceleration (PGA), and pseudo-spectral (absolute) response accelerations (PSA) at 0.2s and 1.0s periods for 5% damping in firm rock (class B) sites. The rest of the paper is organized as follows. The next sections discuss the recurrence model and ground motion prediction models employed in this study for spatio-temporal characterization of earthquakes and the resulting ground motions. Then, seismic hazard curves are presented for two sites located in Santiago and Puerto Plata, including the uniform hazard spectra with a 2% PE in 50 years for both sites. Finally, uniform hazard maps with a 2% PE in 50 years for the northern region of the DR are presented for PGA and PSA parameters used in engineering practice in the structural design of civil infrastructure. The results are compared to the current design parameters specified in the DR building code.

The recurrence of strong earthquakes in a region can be quantified using the mean cumulative annual rate of exceedance (also referred to as the cumulative rate of earthquakes or frequency of earthquakes) ${\lambda }_{m}$ defined as the ratio of the number of earthquakes of (moment) magnitude M_W greater than m, divided by the time interval under consideration. Traditionally the Gutenberg-Richter (GR) model has been used for the magnitude-recurrence relationship of earthquakes in a region, which results on an exponential probability distribution for earthquake magnitude at a site or location of interest.

The Gutenberg-Richter model is unbounded with no maximum earthquake magnitude that can be generated at a fault. In order to address this drawback the model can be modified to include lower and upper thresholds to limit the earthquake magnitude to an interval, resulting on a truncated exponential probability model for earthquake magnitude (Youngs and Coppersmith, 1985). A limitation of the bounded GR model is that it tends to underestimate the observed rate of occurrence of large earthquakes and geological data (Youngs and Coppersmith, 1985). Paleoseismicity studies suggest that individual faults periodically generate earthquakes of approximately the same magnitude (characteristic earthquake), and that this magnitude is the maximum that can be generated at the fault (Youngs and Coppersmith, 1985). In the composite characteristic earthquake model the low-moderate magnitude range is modeled by an exponential tail, while the magnitude range close to the characteristic earthquake is modeled as a uniform (constant) distribution. The transition between the exponential and uniform regions takes place at a magnitude ${{m}_{c}=m}_{char}-{\Delta }{m}_{c}$ with the exceedance rate of the uniform region defined as the rate at ${m{\prime }=m}_{c}-{\Delta }m{\prime }$ where ${\Delta }{m}_{c}$ and ${\Delta }{m}^{{\prime }}$ are model parameters and ${m}_{char}$ is the fault characteristic magnitude (Youngs and Coppersmith, 1985; Gupta, 2007).

Given that earthquake recurrence varies among seismic sources, a recurrence model needs to be calibrated for each source. The seismic sources considered for the PSHA presented herein are: the Septentrional fault, the Northern Hispaniola Subduction Zone, the Matheux-Neiba fault, the Enriquillo-Plantain Garden fault, and the Muertos Trough Subduction Zone. The characteristic earthquake magnitude, slip rates, and predicted earthquake annual exceedance rates for the sources obtained by Frankel et al. (2011) based on GPS data constrained by historic large earthquakes are shown in Table 1; the information is used as input to calibrate a recurrence model for the sources.

Table 1

Seismic sources exceedance rates; reproduced from Frankel et al. (2011).
Crustal fault or subduction zone	Slip Rate (mm/yr)	m_char	λ6.5	λ7.0
Septentrional fault	12	7.8	0.021	0.0078
Northern Hispaniola SZ*	11	8.0	0.005	0.005
Matheux-Neiba fault	1	7.7	0.0034	0.0012
Enriquillo-Plantain Garden fault	7	7.7	0.016	0.0053
Muertos Trough SZ**	7	8.0	0.005	0.005
*rate of the eastern segment used for the complete SZ
**rate of the central segment used for the complete SZ

As will be shown in a following section, the seismic hazard in the northern region of the DR is dominated by the Septentrional fault due to a combination of higher recurrence rates and its close distance to sites in the region; for this reason the recurrence model of the Septentrional fault will be discussed with more detail. The exceedance annual rates for the Septentrional fault for magnitudes of 6.5 and 7 are, respectively, ${\lambda }_{6.5}=0.021$ and ${\lambda }_{7.0}=0.0078$ (Frankel et al., 2011). The rates are based on a model where 67% of the seismic moment is released by characteristic earthquakes, and the remaining 33% of the seismic moment follows an exponential tail. The exceedance rates and the fault slip rate (12mm/yr) are used as constraints to calibrate a characteristic earthquake model of the composite type proposed by Youngs and Coppersmith (1985) for the fault. The model calibration was performed using a constrained optimization with a maximum error of 5% in the recurrence rates and slip rates predicted by the model, with tuning parameters ${\Delta }{m}_{c}$ and ${\Delta }{m}^{{\prime }}$; the calibration best fit parameters were ${\Delta }{m}_{c}=0.35$ and ${\Delta }{m}^{{\prime }}=0.35$ resulting on a slip rate of 12mm/yr and a seismic moment release distributed as 70% by the characteristic plateau and 30% by the exponential tail. The resulting model is shown in Fig. 3, which shows the model recurrence rates and the corresponding return period. The characteristic earthquake model is compared to the Guttenberg-Richter model adopted in the current DR building code in Figure A1.

Using the same procedure described before a composite characteristic model was calibrated for the Matheux-Neiba fault and the Enriquillo-Plantain Garden fault. Given the lack of additional information, the model adopted for the Northern Hispaniola and Muertos Trough Subduction Zones consisted of a characteristic-only model with the rates defined in Table 1. The subduction zones are treated as dipping fault planes, consistent with the approach adopted in Frankel et al. (2011).

Seismic hazard characterization in a time interval

The magnitude-recurrence model discussed in the previous section allows to estimate the recurrence of earthquakes and the earthquake magnitude probability distribution at a site without considering a time interval in the analysis. For engineering applications it is also of interest to characterize the potential occurrence of earthquakes in a given time interval that involves the lifetime of a civil structure or a time period also known as the exposure time (e.g., 50 years, 100 years, 1,000 years, etc.). After the exposure time is defined, the objective is to determine the probability of exceeding an earthquake magnitude or earthquake-induced ground motion intensity parameter (such as peak ground or response spectral accelerations) in order to design civil structures for a parameter value with a “small” probability of being exceeded in the defined time interval.

Defining an exposure time implies that a model for earthquake occurrence in time is needed. The Poisson process is typically used to model events related to earthquake occurrence in time. In this model the number of occurrences ${X}_{T}$ of an event in the time interval [0, T] is modeled as a Poisson distribution. The variable ${X}_{T}$ can be, for example, the number of earthquakes that exceed a given magnitude, or the number of times that a ground motion intensity parameter exceeds a given value. The probability of exceedance (PE) is defined as the probability that ${X}_{T}$ occurs at least once in the time interval considered, and it is given by $PE\equiv p({X}_{T}\ge 1)=1-{e}^{-\lambda T}$where $\lambda$ is the mean annual rate of occurrence of ${X}_{T}$.

It is worth to point out that modeling earthquake occurrence as a Poisson process is not in agreement with some of the physical processes that generate earthquakes and the elastic rebound theory, since it assumes that earthquakes generated at a source are independent of each other. Given the strain energy accumulation and release mechanisms that take place at a fault, there is a time correlation (clustering) among earthquakes generated at a source, which implies that the mean occurrence rate is time-dependent (Ogata 1999; Cramer et al., 2000; Sykes and Menke, 2006; Headquarters for Earthquake Research Promotion HERP, 2010). Despite this limitation of the Poisson process model, the results of PSHA have shown to be robust to the recurrence model (including time-dependent models), and provide a systematic framework to define engineering design parameters which justifies its use in earthquake engineering applications (Chang et al., 2017).

Ground motion intensity parameters

A fundamental objective of PSHA is the estimation of ground motion intensity parameters at a site and their associated recurrence. The characteristics of earthquake-induced ground motions vary significantly for different earthquakes and sites. Moreover, the ground motion characteristics that have the most significant impact in the response and behavior of civil structures are duration, amplitude (intensity) and frequency content. Some of the factors affecting these parameters are

Type of seismic fault and fault orientation, rupture depth, and rupture direction with respect to the site.
Earthquake magnitude, which in turn depends on the area of the ruptured zone, fault displacement, and shear modulus of the rock adjacent to the fault, among others.
Characteristics of the propagation medium, local soil properties and topography.

Given the difficulty to analytically model the ground motions induced by earthquakes using mechanics-based models, empirical models known as attenuation models or ground motion prediction equations (GMPE) are used to perform this task. GMPE are developed using regression analyses and past seismicity data to relate predictor variables, such as earthquake magnitude and distance, to ground motion intensity parameters (Bozorgnia et al., 2014). When sufficient seismicity data is available GMPE can be calibrated for a specific site of interest; however, this is usually not the case given the small occurrence rate of large earthquakes, and thus data from different regions with similar fault mechanisms and local characteristics needs to be combined.

In this study two sets of GMPE are used for crustal faults and subduction zones. For crustal earthquakes the NGA-West2 GMPE are adopted (Bozorgnia et al., 2014). The models were developed using measured data from a significant expansion of the PEER (Pacific Earthquake Engineering Research Center) strong motion database recorded from shallow crustal earthquakes in active tectonic regions. The predictor variables include magnitude scaling, geometric attenuation, type of fault, hanging-wall effects, shallow linear/nonlinear site response, basin response, hypocentral depth, rupture dip, among others. The models used (with equal weights) are the models by Abrahamson et al. (2014) (ASK14), Boore et al. (2014) (BSSA14), Campbell and Bozorgnia (2014) (CB14) and Chiou and Youngs (2014) (CY14).

For earthquakes on the interface of subduction zones the GMPE employed are the models by Zhao et al. (2006), Atkinson and Boore (2003), and Youngs et al. (1997) weighted by factors of 0.5, 0.25 and 0.25, respectively; a higher weight is typically assigned to the Zhao et al. model given the larger database used in its calibration, particularly for distances smaller than 100km. The weights are based on the 2008 USGS weights used for the U.S. national seismic hazard maps.

The GMPE are parameterized by a site-to-rupture zone distance. For nearly vertical faults (such as the Septentrional fault) the Joyner-Boore distance (closest distance from the surface projection of the rupture zone to the site) is given by ${r}_{JB}=\sqrt{{r}_{e}^{2}+{r}_{min}^{2}}$ where ${r}_{min}$ is the smallest distance from the site to the surface projection of the fault, and ${r}_{e}$ is the epicenter location measured with respect to the point in the surface projection of the fault closest to the site. The epicenter location ${r}_{e}$ is modeled as a uniform random variable along the length of the fault, implying that the likelihood of a rupture initiation at different locations along the fault is constant. This is equivalent to floating or moving the rupture zone assuming that the likelihood of the epicenter location is constant along the longitudinal direction of the fault. For GMPE that use the closest distance to the rupture plane, ${r}_{rup}$, the distance metric is given by (for vertical faults) ${r}_{rup}=\sqrt{{r}_{jb}^{2}+{Z}_{TOR}^{2}}$ where ${Z}_{TOR}$ is the depth to the top of rupture; a probability distribution for ${Z}_{TOR}$is defined as a function of magnitude using the results in Wells and Coppersmith (1994) and Kaklamanos et al. (2011). For dipping fault planes in subduction zones the closest distance of the rupture to the site is employed.

The GMPE are used for global regions and uniform site conditions with an average shear-wave velocity v_s30 of 760 m/s, which corresponds to Class B firm rock sites; the rest of the GMPE parameters are selected as the recommended values defined for each model (Gregor et al., 2014). The analysis requires the characteristics of the fault rupture, such as depth to the top of the rupture area, hypocenter depth and rupture length; parameters related to fault rupture are estimated using the results by Wells and Coppersmith (1994) and Kaklamanos et al. (2011).

Seismic hazard curves for ground motion intensity parameters

The main outputs of a PSHA are the probabilities (or rates) of exceeding different ground motion intensity parameters at a site or region. The parameters of interest in engineering applications typically include (horizontal) peak ground acceleration and pseudo-spectral response accelerations for different vibration periods and 5% damping; In particular the spectral accelerations at 0.2s and 1s are used in building codes to define the design response spectra and seismic loads for the design of civil structures.

For a fault modeled as a line source the annual rate of exceedance of the event$Y>y$ (which is related to exceeding various ground motion intensity parameters) is given by

$${\lambda }_{Y>y}={\lambda }_{min}*P(Y>y)={\lambda }_{min}\iint P\left(Y>y | m,r\right)P\left(r \right| m)P\left(m\right)drdm$$

where $P\left(Y>y | m,r\right)$ is defined by the GMPE, $P\left(m\right)$ is the PDF of earthquake magnitude (see Appendix), and $P\left(r \right| m)$ is the PDF of the seismogenic site-to-rupture location distance conditional on earthquake magnitude; as mentioned before the seismogenic distance $r$ and the earthquake magnitude $m$ are related and thus stochastically correlated random variables.

The rate ${\lambda }_{min}$ is the rate of earthquakes magnitudes greater than the minimum magnitude considered in the analysis. For multiple sources the total rate is computed by adding the rates estimated for the individual sources. The annual rate of exceedance ${\lambda }_{Y>y}$ is used to determine the probability of occurrence of the event $Y>y$ in a time interval or exposure time.

The seismic hazard curves for different ground motion intensity parameters resulting from the procedure discussed in the previous sections are presented next. The hazard curves are developed for sites located near the center of the two main cities in the northern region of the DR: Santiago and Puerto Plata.

Results for a site in Santiago

Figure 4 shows the seismic hazard curves for (horizontal) peak ground acceleration (PGA) for a site close to the center of Santiago (Fig. 2); the site is located at a distance ${r}_{min}=8 km$ from the Septentrional fault, measured to the surface projection of the fault. The hazard curves show the contribution of all the sources to the total hazard, and as can be seen, the total hazard curve is dominated by the Septentrional fault. This is attributed to the close distance between the Septentrional fault and Santiago, with the second closest source being the Northern Hispaniola Subduction Zone located at a distance ${r}_{min}=62 km$ and with a significantly lower recurrence rate (see Table 1). Figure 4 also shows the sensitivity of the hazard curve based on the Septentrional fault to the four GMPE employed in the PSHA (with equal weights) for crustal faults.

The PGA values with a 2% and 10% PE in 50 years are, respectively, 0.83g and 0.50g. According to the DR building code the corresponding PGA values with 2% and 10% exceedance probabilities for this site are 0.60g and 0.43g (Reglamento para el análisis y diseño sísmico de estructuras R-001, 2011). The PGA corresponding to a 2% PE in 50 years obtained in this study is 38% higher than the value specified in the DR building code. The 2% PE in 50 years PGA value estimated in this study is in agreement with the results in Frankel et al. (2011), which estimate the corresponding PGA for the site to be in the range 0.80g − 1.00g.

In addition to the PGA, other ground motion intensity parameters of importance in engineering applications include the pseudo-spectral (absolute) acceleration (PSA) for different vibration periods and 5% damping. In particular the PSA values at periods of 0.2s (PSA_0.2) and 1.0s (PSA₁) are used in building codes to define the design response spectrum needed to specify the design earthquake loads for civil structures (Villaverde, 2009). The design spectrum is typically defined using the PSA_0.2 and PSA₁ values for class B firm rock sites, and then it is further modified in building codes by a set of factors to account for local site effects and soil properties different from rock.

Figure 5 shows the seismic hazard curves for PSA_0.2 for the same site in the center of Santiago. Based on the analysis a PSA_0.2 value of 2.05g has a 2% PE in 50 years, while the PSA_0.2 value with a 10% PE in 50 years is 1.15g. According to the DR building code the PSA_0.2 values with 2% and 10% probability of exceedance for this site are, respectively, 1.40g and 1.00g (Reglamento para el análisis y diseño sísmico de estructuras R-001, 2011). The PSA_0.2 corresponding to a 2% PE in 50 years obtained in this study is 46% higher than the value specified in the DR building code. The 2% PE in 50 years PSA_0.2 value estimated in this study is in agreement with the results in Frankel et al. (2011), which estimate the corresponding PSA_0.2 for this site to be in the range 1.5g – 2.40g.

Figure 6 shows the seismic hazard curves for PSA₁. Based on the analysis the PSA₁ value with a 2% PE in 50 years is 0.69g; similarly, the PSA₁ value with a 10% PE in 50 years is 0.36g. According to the DR building code the PSA₁ values with 2% and 10% exceedance probabilities for the site are, respectively, 0.72g and 0.48g (Reglamento para el análisis y diseño sísmico de estructuras R-001, 2011).

Using the same procedure the seismic hazard curves for PSA and other vibration periods can be readily obtained. A uniform hazard spectra (UHS) is formulated by combining the PSA values corresponding to different vibration periods and the same probability of exceedance (Loh et al., 1994; Wen and Wu, 2001; Lee et al., 2013). A uniform hazard spectra with a 2% PE in 50 years is used to define the seismic loads employed for the seismic design of most conventional civil structures (Gavin and Dickinson, 2011). This design spectrum is referred to as the “Maximum Considered Earthquake (MCE)” in building codes, although the terminology is misleading given that the output of the PSHA is an envelope that considers the likelihood of different earthquake magnitudes and the potential ground motion characteristics induced by these earthquakes as estimated by the GMPE.

Figure 7 shows the uniform hazard spectrum for a 2% PE in 50 years determined for the site under consideration in the city of Santiago. The figure also shows the design spectra for the site based on the specifications in the current seismic building code of the DR for rock sites; the design spectra is also based on a 2% PE in 50 years and both spectra correspond to class B firm rock sites with no additional factors applied in order to compare the spectra directly. As can be seen, the UHS obtained in this study significantly exceeds the design spectra specified in the code for vibration periods smaller than 0.5s, which is the range of the vibration periods of most low-rise structures and buildings under five stories that constitute the majority of the built environment of the city and the rest of the country. The results imply that these structures are currently being designed for significantly smaller forces than the design forces resulting from a response spectrum with a 2% PE in 50 years, as stipulated in the code. As can be observed the low frequency (large period) range of the spectra are consistent. This is attributed to the fact that the Campbell 1997 model predicts significantly larger spectral accelerations than the NGA West2 models for vibration periods larger than 0.5s, compensating for the overall reduction in seismic hazard that results from the recurrence model as discussed in Appendix A. For vibration periods smaller than 0.5s the median parameters estimated by the Campbell 1997 (C97) and the NGA West2 models are consistent, but the larger standard deviation associated to the NGA West2 models, combined with the adopted characteristic earthquake recurrence model, result in an increase of the seismic hazard in this region of the spectra.

The difference in the results presented herein and the parameters specified in the DR building code is attributed to two main factors:

The parameters in the DR building code are based on an analysis that employs an unbounded GR model for earthquake recurrence, which significantly underestimates the probability of occurrence of large magnitude earthquakes associated to large PSA values (see Appendix). Moreover, the recurrence model was calibrated using only past seismicity data which, in addition to be scarce, included mostly data of earthquake magnitudes in the M_W 3-5 range (SODOSISMICA, 2009). The resulting probability density function (shown in Appendix Figure A1) shows that earthquakes of magnitude M_W > 6 (and the ground motions induced by them) receive a small weight when integrated in Equation 1, since they have a very small probability associated with them. The characteristic earthquake model employed in this study is based on earthquake rates that include information related to the fault slip rate, as well as historic records used to constrain the occurrence rates of strong earthquakes which results on a significantly higher probability associated to large earthquakes.
The parameters in the DR building code are based on the ground motion prediction equation proposed by Campbell in 1997 (Campbell, 1997), which is outdated and has been significantly refined over the last 20 years using strong motion data of a significantly larger earthquakes database (Campbell and Bozorgnia, 2014). The standard deviation of the ground motion intensity parameters estimated by recent models are considerably larger than the estimates provided by the Campbell 1997 model (see Appendix). The smaller uncertainty associated to parameters predicted by the Campbell 1997 model implies that ground motion parameter values larger than the median (which are associated to strong shaking) receive a small weight when integrated in Equation 1, resulting on a small probability and thus a small rate for such parameter values.

These factors contribute to decrease the estimates of ground motion parameters associated to large earthquakes. The UHS in Fig. 7 stresses the importance to update the current seismic hazard specifications in the DR building code, especially for the northern region of the country. It is worth to point out that an analysis with uniform hazard as a target does not result on a uniform probability of collapse (uniform risk) at different locations as shown in Li et al. (2010) for light-frame wood residential structures due to uncertainty in the capacity (strength) of the structure and variability in the shape of the hazard curves. To achieve a uniform risk based design the parameters are typically adjusted by a factor in the 0.85–1.15 range (Luco et al., 2007).

Results for a site in Puerto Plata

Figure 8 shows the seismic hazard curves for (horizontal) peak ground acceleration (PGA) for a site close to the center of Puerto Plata. The site is located at a distance ${r}_{min}=26 km$ to the Septentrional fault, and a distance ${r}_{min}=26 km$ to the Northern Hispaniola Subduction Zone. Based on the analysis the PGA value with a 2% PE in 50 years is 0.50g. Similarly, the PGA with a 10% PE in 50 years is 0.30g. The curves show the contribution of the various sources to the total hazard, and as can be seen the Septentrional fault and the Northern Hispaniola Subduction Zone have a comparable contribution to the total hazard in Puerto Plata. The PGA for various hazard levels in Puerto Plata are not specified in the DR building code.

The seismic hazard curves for PSA_0.2 are shown in Fig. 9. Based on the analysis the PSA_0.2 value with a 2% PE in 50 years is 1.20g. Similarly, the PSA_0.2 with a 10% PE in 50 years is 0.70g. The PSA hazard values are also not specified in the DR building code for Puerto Plata. Instead, the code suggests using the values for Santiago, assuming that these parameters are higher (and thus provide an upper bound) than the corresponding Puerto Plata parameters.

The seismic hazard curves for PSA₁ are shown in Fig. 10. Based on the analysis the PSA₁ value with a 2% PE in 50 years is 0.43g, while the PSA₁ value with a 10% PE in 50 years is 0.22g. The PSA₁ hazard values are also not specified in the code.

Figure 11 shows the uniform hazard spectra with a 2% PE in 50 years determined in this study for a site in Puerto Plata. A mentioned before, a 2% PE in 50 years design spectrum is not defined on the DR building code for this site and it is suggested to use the design spectrum for Santiago. It is worth to point out that in 2003 Puerto Plata was struck by a M_W 6.4 earthquake that caused significant structural damage and the collapse of multiple houses and low-rise buildings (Dolan and Bowman, 2004). In particular several public schools showed partial or total collapse; fortunately the earthquake occurred afterhours and the school buildings were empty.

The uniform hazard spectra for a site in the cities of Santiago and Puerto Plata were presented in the previous section. Given that a site of interest can be in any spatial location within a region, uniform hazard maps provide a useful tool for implementation in building codes. The maps are defined based on lines of constant parameter values, which makes them suitable for engineering practice.

Figure 12a shows the uniform hazard map for PGA with a 2% PE in 50 years and firm rock sites based on the PSHA performed in this study, while Fig. 12b shows the corresponding PGA hazard map specified in the current DR building code. As can be seen the hazard map in the building code shows significant lower hazard levels than this study, especially in the near-fault regions. It is worth to point out that based on the results in Frankel et al. (2011) the PGA value with a 2% PE in 50 years along the Septentrional fault is in the 1.0-1.8g range; the PGA value of 1.15g determined in this study is consistent with this range, while the 0.64g value specified in the code is significantly smaller than range lower bound.

Figure 13 and Fig. 14 show, respectively, the uniform hazard maps for PSA_0.2 and PSA₁ with a 2% PE in 50 years and firm rock sites. As mentioned previously the PSA_0.2 and PSA₁ values for a site are used to define the design response spectrum and seismic loads used in the structural design of civil structures to resist the strong ground motions induced by large earthquakes. The maps show that the seismic hazard specified in the current building code are underestimated throughout the northern region of the DR, which is attributed to the outdated models used to develop these maps. According to the results in Frankel et al. (2011) the value of the PSA_0.2 parameter along the Septentrional fault is in the 2.4g-3.2g range, which is in agreement with the 2.85g value determined in this study. The corresponding value specified in the DR building code is 1.51g, which again is significantly smaller than range lower bound. Note that the seismic hazard maps presented do not have a deterministic upper limit near the faults, since this is typically defined when developing a building code. For example, the ASCE 7–22 maps adopt a 1.8 times the estimated median response to the characteristic earthquake of the faults to define an upper limit.

The results of this study could potentially be used by regulating authorities of the Dominican Republic as a baseline to update the current seismic hazard maps specified in the DR building code, Moreover, the results provide a rigorous basis for comparison of future seismic hazard studies in the Dominican Republic, results which are lacking in the literature.

This paper presented the results of a probabilistic seismic hazard analysis (PSHA) performed for the northern region of the Dominican Republic (DR). The PSHA used as input recurrence models calibrated using fault slip rates derived from GPS measurements published in the aftermath of the 2010 Haiti Earthquake. The recurrence models were combined with a Poisson process model and a set of recently proposed ground motion prediction equations (GMPE) for spatio-temporal characterization of ground motion intensity parameters at a site. The PSHA performed accounted for uncertainties related to earthquake magnitude, distance from a site to the potential rupture locations along the fault, and the aleatory uncertainty associated to the regression analyses performed to obtain the GMPE.

The estimated ground motion parameters included horizontal peak ground acceleration (PGA) and pseudo-spectral (absolute) response accelerations at 0.2s (PSA_0.20) and 1.0s (PSA₁) periods for 5% damping in firm rock (class B) sites. A set of seismic hazard curves were developed for the PGA, PSA_0.20 and PSA₁ parameters for two sites in the center of the cities of Santiago and Puerto Plata. The results included the uniform hazard spectra with a 2% probability of exceedance (PE) in 50 years for both sites. The results showed that ground motion parameters (such as PGA and PSA for vibration periods smaller than 0.5s) are underestimated in the current DR building code, and inconsistent with the corresponding parameters with a 2% PE in 50 years. This implies that low-rise civil structures with vibrations periods smaller than 0.5s are currently being designed for seismic forces significantly smaller than the target forces expected from a design response spectrum with a 2% PE in 50 years. Structures in this class include one to four stories houses and buildings, which represent the majority of the built environment of the DR.

In addition to hazard curves and spectra, uniform hazard maps with a 2% PE in 50 years were developed for the PGA, PSA_0.20 and PSA₁ parameters for the northern region of the Dominican Republic. The maps were compared to the current hazard maps specified in the DR building code. It was shown that the response parameters specified in the code are underestimated and inconsistent with a 2% PE in 50 years hazard level throughout the northern region of the DR. The results obtained in this study showed to be consistent with the results by Frankel et al. (2011). The results highlight the urgency to update the current design seismic hazard maps for the DR given the high strain accumulation levels in the Septentrional fault reported in the literature.

Future work is aimed at including spatially smoothed seismicity to account for earthquakes not attributed to crustal faults and subduction zones, and to extending the analysis to the rest of the country.

DATA AND RESOURCES

All data used in this paper came from published sources listed in the references.

DECLARATION OF COMPETING INTERESTS

The author declares no competing interests.

ACKNOWLEDGMENTS

This research was partially supported by the Instituto Tecnológico de Santo Domingo (INTEC) and the Ministerio de Educación Superior Ciencia y Tecnología (MESCYT) of the Dominican Republic through the FONDOCYT program.

Abrahamson, N. A., Silva, W. J., and Kamai, R. (2014). Summary of the ASK14 ground motion relation for active crustal regions, Earthquake Spectra 30, 1025–1055.

Abrahamson, N. A., Kuehn, N. M., Walling, M., and Landwehr, N. (2019). Probabilistic seismic hazard analysis in California using nonergodic ground‐motion models. Bulletin of the Seismological Society of America, 109(4), 1235-1249.

Atkinson, G. M., and Boore, D. M. (2003). Empirical ground-motion relations for subduction zone earthquakes and their application to Cascadia and other regions, Bulletin of the Seismological Society of America 93, 1703–1729.

Baker, J.B., Bradley, B., and Stafford, P. (2021). Seismic Hazard and Risk Analysis. Cambridge University Press.

Boore, D. M., Stewart, J. P., Seyhan, E., and Atkinson, G. M. (2014). NGA-West2 equations for predicting PGA, PGV, and 5% damped PSA for shallow crustal earthquakes, Earthquake Spectra 30, 1057–1085.

Bozorgnia, Y., Abrahamson, N. A., Atik, L. A., Ancheta, T. D., Atkinson, G. M., Baker, J. W., ... and Youngs, R. (2014). NGA-West2 research project. Earthquake Spectra, 30(3), 973-987.

Calais, E., Perrot, J., and De Lepinay, B. M. (1998). Strike-slip tectonics and seismicity along the northern Caribbean plate boundary from Cuba to Hispaniola. SPECIAL PAPERS-GEOLOGICAL SOCIETY OF AMERICA, 125-142.

Calais, E., Mazabraud, B., Mercier de Le´pinay, Mann, P., Mattioli, G., and Jansma, P. (2002). Strain partitioning and fault slip rates in the northeastern Caribbean from GPS measurements. Geophysical Research Letters, 29(18), 1856, doi: 10.1029/2002GL015397

Campbell (1997). Empirical Near-Source Attenuation Relationships for Horizontal and Vertical Components of Peak Ground Acceleration, Peak Ground Velocity, and Pseudo-Absolute Acceleration Response Spectra. Seismiological Research Letters, 68(1), 154-179.

Campbell, K. W., and Bozorgnia, Y. (2014). NGA-West2 ground motion model for the average horizontal components of PGA, PGV, and 5% damped linear acceleration response spectra. Earthquake Spectra, 30(3), 1087-1115.

Chang, Y. W., Loh, C. H., and Jean, W. Y. (2017). Time-predictable model application in probabilistic seismic hazard analysis of faults in Taiwan. Terrestrial, Atmospheric and Oceanic Sciences, 28(6).

Chiou, B. S.-J., and Youngs, R. R. (2014). Update of the Chiou and Youngs NGA model for the average horizontal component of peak ground motion and response spectra, Earthquake Spectra 30, 1117–1153.

Cornell, C.A. (1968). Engineering seismic risk analysis. Bulletin of the Seismological Society of America, 58, 1583-1606.

Cramer, C. H., Petersen, M. D., Cao, T., Toppozada, T. R., and Reichle, M. (2000). A time-dependent probabilistic seismic-hazard model for California. Bulletin of the Seismological Society of America, 90(1), 1-21.

DeMets, C., Jansma, P. E., Mattioli, G. S., Dixon, T. H., Farina, F., Bilham, R., ... and Mann, P. (2000). GPS geodetic constraints on Caribbean‐North America plate motion. Geophysical Research Letters, 27(3), 437-440.

DesRoches, R., Comerio, M., Eberhard, M., Mooney, W., and Rix, G. (2011), “Overview of the 2010 Haiti earthquake”, Earthquake Spectra, 27(S1), 1–21. https://doi.org/10.1193/1.3630129.

Dolan, J. F. and Bowman, D. D. (2004). Tectonic and seismologic setting of the 22 September 2003, Puerto Plata, Dominican Republic earthquake: implications for earthquake hazard in northern Hispaniola. Seismological Research Letters, 75(5), 587-597.

Erazo, K. (2019). Probabilistic seismic hazard analysis and design earthquake for Santiago, Dominican Republic. Ciencia, Ingenierías y Aplicaciones, 2(1), 67-84.

Erazo, K. (2020). Análisis probabilístico de peligro sísmico y terremoto de diseño para Santiago-República Dominicana. Ciencia, Ingenierías y Aplicaciones, 3(1), 7-30.

Erazo, K. and Taveras, A. (2021). Demandas estructurales inducidas por vientos huracanados y terremotos en un edificio flexible en la República Dominicana. Ciencia, Ingenierías y Aplicaciones, 4(1), 57-78.

File:Gonâve microplate.png. (2020, November 5). Wikimedia Commons, the free media repository. Retrieved 18:12, December 28, 2021 from https://commons.wikimedia.org/w/index.php?title=File:Gon%C3%A2ve_microplate.png&oldid=510319836.

Frankel, A., Harmsen, S., Mueller, C., Calais, E., and Haase, J. (2011). Seismic hazard maps for Haiti. Earthquake Spectra, 27(1_suppl1), 23-41.

Gavin, H. P., and Dickinson, B. W. (2011). Generation of uniform-hazard earthquake ground motions. Journal of Structural Engineering, 137(3), 423-432.

Gerstenberger, M. C., Marzocchi, W., Allen, T., Pagani, M., Adams, J., Danciu, L., ... and Petersen, M. D. (2020). Probabilistic seismic hazard analysis at regional and national scales: State of the art and future challenges. Reviews of Geophysics, 58(2), e2019RG000653.

Gregor, N., Abrahamson, N. A., Atkinson, G. M., Boore, D. M., Bozorgnia, Y., Campbell, K. W., ... & Youngs, R. (2014). Comparison of NGA-West2 GMPEs. Earthquake Spectra, 30(3), 1179-1197.

Gupta, I. D. (2007). Probabilistic seismic hazard analysis method for mapping of spectral amplitudes and other design-specific quantities to estimate the earthquake effects on man-made structures. ISET Journal of Earthquake Technology, 44(1), 127-167.

Headquarters for Earthquake Research Promotion HERP (2010). National seismic hazard maps for Japan 2010, Earthquake Research Committee (K. Abe, chair), Headquarters for Earthquake Research Promotion, available from www.jishin.go.jp/main/chousa/10yosokuchizu/index.htm (in Japanese).

Kaklamanos, J., Baise, L. G., and Boore, D. M. (2011). Estimating unknown input parameters when implementing the NGA ground-motion prediction equations in engineering practice. Earthquake Spectra, 27(4), 1219-1235.

Kammerer, A. and Ake, J. (2012). Practical Implementation Guidelines for SSHAC Level 3 and 4 Hazard Studies, NUREG-2117, U.S. Nuclear Regulatory Commission, Washington, D.C.

Lee, V., Herak, M., Herak, D., and Trifunac, M. (2013). Uniform hazard spectra in western Balkan Peninsula. Soil Dynamics and Earthquake Engineering, 55, 120.

Li, Y., Yin, Y., Ellingwood, B. R., and Bulleit, W. M. (2010). Uniform hazard versus uniform risk bases for performance‐based earthquake engineering of light‐frame wood construction. Earthquake Engineering and Structural Dynamics, 39(11), 1199-1217.

Loh, C. H., Jean, W. Y., and Penzien, J. (1994). Uniform‐hazard response spectra—An alternative approach. Earthquake engineering and structural dynamics, 23(4), 433-445.

Luco, N., Ellingwood, B. R., Hamburger, R. O., Hooper, J. D., Kimball, J. K., and Kircher, C. A. (2007). Risk-targeted versus current seismic design maps for the conterminous United States. SEAOC 2007 Convention Proceedings.

Mann, P., Prentice, C., Burr, G., Peña, L., and Taylor, F.W. (1998). Tectonic geomorphology and paleoseismology of the Septentrional fault system, Dominican Republic. The Geological Society of America. doi: 10.1130/0-8137-2326-4.63

Mann, P., Calais, E., Ruegg, J. C., DeMets, C., Jansma, P. E., and Mattioli, G. S. (2002). Oblique collision in the northeastern Caribbean from GPS measurements and geological observations. Tectonics, 21(6), 7-1.

McGuire, R. K., 1995: Probabilistic seismic hazard analysis and design earthquakes: Closing the loop. Bull. Seismol. Soc. Am., 85, 1275-1284.

Ogata, Y., 1999: Estimating the hazard of rupture using uncertain occurrence times of paleoearthquakes. J. Geophys. Res., 104, 17995-18014, doi: 10.1029/1999JB900115

Prentice, C.S., Mann, P., Peña, L.R. and Burr, G. (2003). Slip rate and earthquake recurrence along the central Septentrional fault, North American–Caribbean plate boundary, Dominican Republic. Journal of Geophysical Research 108: doi: 10.1029/2001JB000442.

Prentice, C. S., Mann, P., Crone, A. J., Gold, R. D., Hudnut, K. W., Briggs, R. W., ... & Jean, P. (2010). Seismic hazard of the Enriquillo–Plantain Garden fault in Haiti inferred from palaeoseismology. Nature Geoscience, 3(11), 789-793.

Rathje, E. M., and Saygili, G. (2008). Probabilistic seismic hazard analysis for the sliding displacement of slopes: scalar and vector approaches. Journal of Geotechnical and Geoenvironmental Engineering, 134(6), 804-814.

Reglamento para el análisis y diseño sísmico de estructuras R-001. (2011). Ministerio de Obras Públicas y Comunicaciones (MOPC). Santo Domingo, República Dominicana.

Rodriguez‐Marek, A., Rathje, E. M., Bommer, J. J., Scherbaum, F., and Stafford, P. J. (2014). Application of single‐station sigma and site‐response characterization in a probabilistic seismic‐hazard analysis for a new nuclear site. Bulletin of the Seismological Society of America, 104(4), 1601-1619.

Rojas-Mercedes, N. J., Di Sarno, L., Simonelli, A. L., and Penna, A. (2020). Seismic risk of critical facilities in the Dominican Republic: case study of school buildings. Soft Computing, 24(18), 13579-13595.

SODOSISMICA (2009). Estudio de amenaza sísmica de la República para las normas sísmica.

Sykes, L. R., and Menke, W. (2006). Repeat times of large earthquakes: Implications for earthquake mechanics and long-term prediction. Bulletin of the Seismological Society of America, 96(5), 1569-1596.

Villaverde. (2009). Fundamental Concepts of Earthquake Engineering. Boca Raton, FL. USA. CRC Press, Taylor and Francis Group.

Wen, Y. K., and Wu, C. L. (2001). Uniform hazard ground motions for mid-America cities. Earthquake spectra, 17(2), 359-384.

Wells, D. L., & Coppersmith, K. J. (1994). New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement. Bulletin of the seismological Society of America, 84(4), 974-1002.

Youngs, R. R., Chiou, S. -J., Silva, W. J., and Humphrey, J. R. (1997). Strong ground motion attenuation relationships for subduction zone earthquakes, Seismological Research Letters 68, 58–73

Youngs, R.R. and Coppersmith, K.J. (1985). Implications of fault slip rates and earthquake recurrence models to probabilistic seismic hazard estimates. Bulletin of the Seismological Society of America, 75, 939-964.

Zhao, J. X., Zhang, J., Asano, A., Ohno, Y., Oouchi, T., Takahashi, T., Ogawa, H., Irikura, K., Thio, H. K., Somerville, P. G., Fukushima, Y., and Fukushima, Y. (2006). Attenuation relations of strong ground motion in Japan using site classification based on predominant period, Bull. Seism. Soc. of Am. 96, 898–913

APPENDIX.docx

Download PDF

Version 1

posted

You are reading this older preprint version

Read the latest preprint version →

Updated Seismic Hazard Curves, Maps, and Spectra for the Northern Dominican Republic using a Probabilistic Seismic Hazard Analysis

Archived Versions:

Version 1

Abstract

Figures

Introduction

Magnitude-recurrence Model

Seismic hazard characterization in a time interval

Ground motion intensity parameters

Seismic hazard curves for ground motion intensity parameters

Psha Results For Santago And Puerto Plata

Results for a site in Santiago

Results for a site in Puerto Plata

Uniform Seismic Hazard Maps

Conclusions

Declarations

References

Supplementary Files

Archived Versions:

Version 1