Probabilistic tsunami hazard assessment and flood modeling: A case study of Manzanillo, Colima, Mexico

Probabilistic tsunami hazard assessment and flood modeling: A case study of Manzanillo, Colima, Mexico | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Probabilistic tsunami hazard assessment and flood modeling: A case study of Manzanillo, Colima, Mexico Lizzeth Vázquez-Camaal, Christian M. Appendini, Ericka Alinne Solano-Hernández, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7257032/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This study presents a probabilistic tsunami hazard assessment (PTHA) for Manzanillo, Colima, a coastal city in Mexico, which is vulnerable to local tsunamis due to its proximity to the subduction zone along the Mesoamerican Trench. Historical records show that the region experienced major tsunamis, notably in 1932 and 1995. Using stochastic earthquake source models and numerical simulations, we calculated the tsunami inundation scenarios for different return periods (50, 100, and 500 years). The study employed the GeoClaw open-source software package, incorporating high-resolution topography and bathymetry, to simulate tsunami wave propagation and coastal inundation. Model validation using the 1995 Colima-Jalisco event demonstrates that the simulation correctly captures the observed tsunami characteristics. The results revealed that lowland areas, particularly near Manzanillo and Santiago Bays, and the Cuyutlán Lagoon, could experience inundation of several kilometers inland in a worst-case scenario. Probability-of-exceedance curves indicate a high likelihood of moderate to significant tsunami wave heights within a 50-year time frame, underscoring the substantial risk to the city. These findings provide crucial information for local authorities to develop effective tsunami risk management strategies including hazard mapping, improved building codes, and emergency preparedness plans. PTHA GeoClaw Mesoamerican Trench Coastal risk Hazard mapping Return period Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1 Introduction Tsunamis are long-wavelength ocean waves generated by sudden disturbances in the water column over a large area, typically caused by seismic events, submarine landslides, or volcanic eruptions. These vertical displacements in the water column release potential energy that is converted into kinetic energy, resulting in radial propagation of tsunami waves (Geist and Oglesby, 2014 ). Among the various geological phenomena that can trigger tsunamis, subduction zone earthquakes are the most common, especially where tectonic plates converge, leading to significant vertical displacement of the seafloor. These types of earthquakes are typically thrust or reverse fault events that are highly effective in generating tsunamis (Geist and Oglesby, 2014 ). The Pacific Ocean's "Ring of Fire" is a well-known source of seismic activity owing to its numerous subduction zones, accounting for approximately 78% of all tsunamis globally (NOAA, 2023). Within this context, Mexico's Pacific coast, particularly the central and southern regions, is highly susceptible to local tsunamis due to its proximity to the Mesoamerican Trench (Fig. 1 a). This trench is formed by the convergence of the Rivera, Cocos, and North American plates, creating a tectonically active zone prone to large earthquakes (Molnar and Sykes, 1969 ; Bandy et al., 2000 ). The resulting seismic activity has historically generated tsunamis that have impacted coastal communities in Mexico, including Manzanillo, a port city in the State of Colima (Castillo-Aja and Ramírez-Herrera, 2017). Manzanillo is particularly vulnerable to tsunamis because of its geographical location, which is adjacent to the Mesoamerican Trench. This region has experienced multiple tsunamis throughout history, with significant events recorded as far back as 1563. Notably, the tsunamis of 1932 and 1995 stand out due to their devastating impacts on Manzanillo, causing severe damage and loss of life (Sánchez and Farreras, 1993 ; Corona and Ramírez-Herrera, 2012 ). The June 22, 1932, earthquake (M ~ 6.9) reportedly produced run-up heights of up to 12 m on nearby coasts (Corona and Ramírez-Herrera, 2012 ), and the 1995 earthquake (M8.0) caused notable damage in Manzanillo Bay due to the tsunami currents (Borerro et al., 1997 ). These historical events underscore the importance of conducting detailed tsunami hazard assessments in this area. Traditionally, tsunami hazard assessments have been performed using deterministic or probabilistic approaches. The deterministic method, often referred to as the "worst-case scenario" approach, relies on the largest recorded events within a region to predict potential future impacts. However, this method may not account for the full range of possible seismic events, particularly in areas with limited historical data on tsunamis (Mori et al., 2018; Jelínek et al., 2012 ). On the other hand, the Probabilistic Tsunami Hazard Assessment (PTHA) approach builds on methodologies developed for seismic hazard analysis (PSHA) and provides a more comprehensive framework for understanding tsunami risk by incorporating a variety of possible earthquake scenarios, including those that might not have been historically observed (Cornell and Vanmarcke, 1969 ; Geist and Parsons, 2006 ; Mori et al., 2018; Miyashita et al., 2020 ). This study aims to perform a PTHA for Manzanillo, Colima, using stochastic earthquake source models and numerical simulations. The assessment was particularly focused on evaluating tsunami hazards for different return periods (e.g., 50, 100, and 500 years). The use of advanced numerical models, including GeoClaw (LeVeque et al., 2011 ), allows for a detailed simulation of tsunami wave propagation and coastal inundation, providing valuable insights into their potential impacts on Manzanillo. The results of this study are intended to aid local authorities in developing effective tsunami risk management and mitigation strategies. Given the region’s tectonic setting and historical tsunami events, it is crucial to understand the probabilistic nature of tsunami hazards. By considering a broad spectrum of possible seismic events (beyond those in the historical record), this study used PTHA to provide a more robust picture of the tsunami hazard. These findings can enhance broader disaster preparedness and coastal management initiatives, ensuring that Manzanillo and its surrounding areas are better prepared to confront future tsunamis. 2 Study Area The study area encompasses the coastal municipality of Manzanillo, which is located along the central Pacific coast of Mexico. Geographically, Manzanillo is positioned between 19.31°N–19.95°N latitude and 104.68°W–104.03°W longitude, covering a territorial area of approximately 1,578 km². The area is located within the tectonically active region influenced by the convergence of the Rivera, Cocos, and North American plates. This geographical location places Manzanillo at significant risk from seismic and tsunami events owing to its proximity to the Mesoamerican Trench, where these tectonic plates interact. The city lies within the Sierra Madre del Sur physiographic province and features diverse topography, including mountainous terrain, intermontane valleys, coastal plains, and lagoons. According to national geographical data (INEGI, 2010), the most prominent physiographic components are high complex sierras (approximately 60.5% of the area), intermontane valleys with hills (12.6%), coastal plains with lagoons (12.1%), and coastal plains with floodable deltas and salt flats (5.6%). Most of the remaining area consists of beaches, coastal bars, and minor valleys. The coastline of Manzanillo is characterized by a series of bays and lagoons that significantly influence local geomorphology and potential tsunami inundation patterns. The main bays from west to east include Carrizales Bay , Higueras (Cenicero) Bay , Santiago Bay , and Manzanillo Bay (Fig. 1 b). The latter two, Santiago and Manzanillo, are particularly important in terms of urban development and population exposure, making them critical areas for tsunami risk. The shores around these bays range from rocky headlands and cliffs to low-lying sandy beaches with elevations that vary and can influence the propagation and run-up of tsunami waves on land. For instance, steep rocky sections may reflect or refract waves, whereas flat coastal plains allow waves to penetrate further inland (González et al., 1995 ; Aránguiz et al., 2019). Manzanillo’s hazard is further exacerbated by the presence of several rivers and lagoons, which add complexity to tsunami flooding scenarios. The Juluapan , Peñitas , and Cuyutlán lagoons are notable hydrological features connected to the ocean through narrow inlets or channels. These water bodies can act as conduits that channel tsunami waves inland, or amplify flooding in adjacent low-lying areas (Zamora et al. 2021 ). In particular, the Cuyutlán Lagoon, a large lagoon on the eastern side of Manzanillo’s urban area (near Campos Beach ), has expansive low-elevation wetlands. During a tsunami, this lagoon and its surroundings could experience amplified inundation due to the funneling effect and low terrain, as water may pour into and spread through the lagoon system, an effect observed in similar coastal environments (Pari et al., 2008 ). Figure 1 illustrates the location and main geographical features of the Manzanillo study area. The key features highlighted include the aforementioned bays (Carrizales, Cenicero, Santiago, and Manzanillo) as well as the lagoons and river outlets. These geographic features are central to understanding tsunami hazards as they can either mitigate or worsen the impact of incoming tsunami waves. For example, Santiago Bay is partially sheltered by headlands, whereas Manzanillo Bay opens directly toward the trench area, and the coastline at Campos (near Cuyutlán Lagoon) faces the open ocean, which is reflected in the hazard results. 3 Data and Methods The methodology for this study involves a combination of data collection, seismic source modeling, and numerical simulation to assess tsunami hazards in the Manzanillo region. In summary, we first gathered the necessary geological and geophysical data (seismic source parameters, bathymetry, and topography) and then generated a suite of plausible earthquake scenarios using stochastic modeling approaches. These scenarios were the inputs for tsunami simulations using a physics-based model (GeoClaw) to propagate the resulting tsunami to the coast and compute inundation on land. Finally, we performed a probabilistic analysis of the results to estimate the likelihood of exceeding certain wave heights (or run-up levels) over given time periods. The methodology includes a validation step, where we compare simulation outputs with observations from a past tsunami event (the 1995 Colima-Jalisco tsunami) to ensure the reliability of the model. 3.1 Data Accurate assessment of tsunami hazards requires various datasets, most importantly, seismic source data to characterize potential earthquakes and a seamless bathymetry-topography dataset to properly model the wave propagation across the ocean and to model onshore inundation. The datasets are briefly described below. 3.1.1 Seismic Source Models To represent the seismic setting of the region, we utilized a detailed three-dimensional model of the subduction interface along the Mesoamerican Trench and a seismic wave velocity model. The subduction zone geometry was based on the Central American Slab2.0 model (Hayes et al., 2018 ), which provides the depth and dip of the subducting plate. The geometry was discretized into a finite mesh down to a depth of approximately 30 km, as large thrust earthquakes in this region typically occur within this depth range. This mesh effectively outlines the locations where fault plane ruptures can occur. The Slab2.0 model integrates data from various sources (e.g., earthquake catalogs and seismic surveys) to ensure an accurate geometric representation of the plate interface. In addition, the seismic wave velocity structure from Mendoza and Hartzell ( 1999 ), which was used to analyze the 1995 Jalisco-Colima earthquake, was implemented to generate the rupture dynamics scenarios presented herein. Our focus is on tsunami generation, which primarily depends on static seafloor deformation, and having a reasonable seismic velocity model aids in creating more physically plausible earthquake slip distributions for the scenarios. 3.1.2 Bathymetry and Topography Bathymetric and topographic data are critical for accurately modeling tsunami wave propagation in the ocean and inundation over land, respectively. For deep-ocean and regional propagation, we used the SRTM30 + global bathymetry dataset (Becker et al., 2009 ), which has a resolution of 30 arc-seconds (~ 900 m). Although relatively coarse, this resolution is sufficient for far-field wave propagation and the initial wavefront development. For nearshore and onshore areas, especially within the bays of Cenicero, Santiago, and Manzanillo and around the Cuyutlán Lagoon, we incorporated high-resolution bathymetric and topographic datasets. High-resolution bathymetric surveys (from nautical charts or local studies) were merged with topographic elevation data (LiDAR or photogrammetry, if available, or 10-m resolution DEMs) to create a detailed Digital Elevation Model (DEM) of the coastal zone. This high-resolution DEM (with grid spacing on the order of tens of meters) captures important small-scale features, such as beach berms and channel connections between the ocean and lagoons. These features strongly influence how a tsunami wave transitions from deep water to the shoreline, and how far inland water can penetrate. All spatial datasets were projected onto a common coordinate system and vertical datum, and we took care to stitch the global and local bathymetry seamlessly so that no artificial gaps or overlaps existed. 3.2 Tsunami Modeling and Hazard Analysis The tsunami hazard assessment was carried out in two main phases: (1) generating tsunami source scenarios using stochastic seismic modeling and simulating each scenario’s tsunami propagation and inundation and (2) performing a probabilistic analysis of the ensemble of simulations to derive hazard curves and maps (i.e., computing the probability that certain tsunami intensity measures will be exceeded in given time frames). In addition, we validated our modeling approach against a known historical event to ensure the robustness of the proposed method. 3.2.1 Seismic Fault Models and Scenario Generation We employed the “FakeQuakes” module of the open-source MudPy code (Melgar et al., 2016 ) to generate stochastic earthquake rupture scenarios on the Mesoamerican trench model. FakeQuakes uses the statistical relations of past earthquakes and geophysical constraints to create realistic slip distributions for hypothetical events. Each synthetic rupture is defined by parameters such as the magnitude, spatial slip distribution on the fault plane, rupture dimensions, depth, and hypocenter location. This approach allowed us to sample a wide range of possible earthquakes, including events larger than those in the historical catalog, with their associated occurrence frequencies, which are crucial for PTHA. Two sets of scenarios were produced. 1. Model Validation Scenarios: First, we focused on the 1995 Colima-Jalisco earthquake (Mw 8.0) as a reference event for model validation, considering the availability of water level observations. We generated 600 synthetic rupture models constrained to the known characteristics of the 1995 event (based on the studies by Mendoza and Hartzell, 1999 ). These scenarios varied in detail, including the exact slip pattern, hypocenter, and rupture extent. From all the results, only 28 scenarios met the criteria of having magnitudes in the range of ~ 7.99–8.01, matching the 1995 event magnitude, which were selected for validation. 2. PTHA Scenarios: Next, probabilistic analysis was used to generate a vast range of plausible large earthquakes in the region. We defined magnitude bins from Mw 7.0 to Mw 8.5 with increments of ΔM = 0.1. For each magnitude bin, multiple random rupture realizations were generated, resulting in a total suite of 4,800 synthetic earthquakes (300 scenarios for each of the 16 magnitude bins between 7.0 and 8.5). These scenarios were broadly centered offshore near Manzanillo (within the segment of the trench facing Colima), but with random variations in the exact hypocenter and rupture area along the trench segment. We did not restrict the hypocenter locations within that segment, allowing ruptures to be initiated anywhere along the portion of the fault adjacent to Manzanillo. We set an upper bound on the slip per patch (60 m maximum) to reflect the physical limits generally observed in subduction earthquakes. Each scenario yielded a different pattern of seafloor deformation, producing a distribution of tsunamis. 3.2.2 Tsunami Simulations (GeoClaw Model) For each generated earthquake scenario, we simulated the resulting tsunami using the GeoClaw numerical model (LeVeque et al., 2011 ). GeoClaw is a well-validated solver for nonlinear shallow water equations and is specifically designed to handle tsunami propagation and inundation with high efficiency through adaptive mesh refinement (AMR). The static seafloor deformation from each synthetic earthquake (essentially, the initial condition for tsunami generation) was computed using Okada’s elastic dislocation model (Okada, 1985, 1992). This model considers the slip on the fault plane and computes the corresponding vertical displacement of the ocean bottom. We applied Okada’s formulas to each subfault in the rupture model to construct the total deformation field. The resulting uplift and subsidence pattern (which can have peak displacements of the order of several meters for Mw ~ 8 events) is then applied instantaneously to the ocean surface to initiate the tsunami in the model. Once the initial wave is generated, GeoClaw simulates its propagation across the ocean, interaction with bathymetric features, and eventually shoaling producing the inundation on land. The key parameters and configurations for the GeoClaw simulations were as follows: Adaptive mesh refinement: We utilized five nested grid levels. The coarsest level (Level 1) covered the entire computational domain (encompassing a large portion of the Pacific, including the wave’s path to Manzanillo) with a grid resolution of 5 arcmin. Intermediate levels (Level 2 at 1 arcminute, Level 3 at 30 arcseconds) gradually increased the resolution as the waves neared the coast. The finest grids (Level 4 at 5 arc-seconds to approximately 150 m, and Level 5 at 1 arc-second to ~ 30 m) were centered on the Manzanillo coastal region. This approach allowed us to capture broad-scale propagation efficiently while still resolving the complex coastline and bathymetry of Manzanillo in high detail for inundation calculations. Notably, Levels 4 and 5 cover the bays and lagoons of interest with ~ 30 m resolution, which is already adequate to model water flow through narrow channels and around small islands or peninsulas. Time step and stability: A Courant–Friedrichs–Lewy (CFL) condition was used to ensure numerical stability. We set the CFL number target to 0.75 (with an allowable maximum of 1.0). This means that the time step automatically adjusts such that the tsunami wave does not travel more than 75% of the grid cells per step on the finest grids. This adaptive time-stepping is important given varying grid sizes. Friction and drag: To account for energy losses due to seafloor roughness and onshore terrain friction (vegetation, buildings, etc.), we applied a Manning’s roughness coefficient of 0.025. This value is commonly used for tsunami modeling over a mix of coastal environments (LeVeque et al., 2011 ) and provides a reasonable approximation of the drag on flowing water. For output and analysis, we recorded simulated water surface elevations at five virtual gauge stations (denoted P1 through P5) located along the Manzanillo coast, each positioned at a 10 m water depth just offshore (Fig. 1 ). These synthetic tide gauge points (including one near the location of the actual Manzanillo tide gauge in the harbor) captured the incoming wave time-series for each scenario. We also saved full-time historical inundation results on the fine grids to determine the run-up distance, that is, how far inland water traveled, and the depths of the resulting flooding. For the validation scenarios (the 28 cases modeling the 1995 event), the time series at a virtual gauge corresponding to Manzanillo’s harbor (MTG in Fig. 1 ) was compared to the 1995 tsunami records (Fig. 2 ). This allowed us to assess the accuracy of the model in terms of the timing, wave height, and overall wave train shape. 3.2.3 Probabilistic Tsunami Hazard Assessment (PTHA) Probability exceedance curves were calculated by accounting for the probability that a tsunami wave surpasses a threshold \(\:{\eta\:}_{c}\) in an exposition window \(\:{T}_{exp}\) for all evaluated magnitudes, as shown in Eq. (1). These curves express the probability that a given tsunami wave height (or run-up height) will be equal to or exceeded at a site within a specified period (e.g., 10-, 50-, 100-year, etc.). $$\:P(\eta\:>{\eta\:}_{c})=1-{\prod\:}_{i=1}^{n}{[1-(1-e}^{-{\nu\:}_{i}{T}_{exp}}\left)P\right(\eta\:>{\eta\:}_{c}\left|{M}_{i}\right)]\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(1\right)$$ Here, \(\:{T}_{exp}\) , is the exposition window and was established in 1 year; \(\:P(\eta\:>{\eta\:}_{c}){M}_{i})\:\) is the probability that a seismic event with magnitude \(\:{M}_{i}\) generates tsunami amplitudes exceeding a certain level \(\:{\eta\:}_{c}\) in the next year; \(\:{\nu\:}_{i}\) is the truncated form of the annual seismic occurrence rate from the Gutenberg-Richter relationship (Cornell and Vanmarcke, 1969 ; Salazar-Monroy et al., 2021 ), as derived from Eq. (2). The hazard assessment in this work was performed for magnitudes ranging from Mw = 7.0 to Mw = 8.5 \(\:(\varDelta\:M=0.1)\) , resulting in the evaluation of 16 magnitude bins. $$\:{\nu\:}_{i}={\nu\:}_{0}\frac{{e}^{-\beta\:({M}_{i}-{M}_{min})}-{e}^{-\beta\:({M}_{max}-{M}_{min})}}{{1-e}^{-\beta\:({M}_{max}-{M}_{min})}},{M}_{min}<{M}_{i}<{M}_{max}\:\:\:\:\:\:\:\:\left(2\right)\:$$ where \(\:{\nu\:}_{0}={e}^{\alpha\:-\beta\:{M}_{min}}\) , \(\:\alpha\:=a{\prime\:}\:ln\left({T}_{r}\right)\) y \(\:\beta\:=b\:ln\left({T}_{r}\right)\) , in this case the return period, \(\:{T}_{r}\) , was fixed to 10 years. In our calculations, we used the regional parameters \(\:a\) and \(\:b\) from the Gutenberg-Richter law as proposed by Zuñiga et al. (2017), which is appropriate for the Colima region. These parameters were originally determined for a larger coastal segment than that of our study area. Thus, to better represent the events attributable to the Manzanillo segment and the size and recurrence of the evaluated rupture area, we applied a correction to the \(\:a\) value. This 30% reduction in total seismicity was implemented by modifying the parameter as \(\:{a}^{{\prime\:}}=a+ln\left(0.70\right)\) . Furthermore, the magnitude range was truncated at Mw 8.5, which is consistent with the historical and tectonic characteristics of this segment. This provides a conservative but reasonable upper bound for hazard analysis without underestimating the tail risk. For practical purposes, we focused on three possible return periods: 50, 100, and 500 years, which correspond to relatively frequent, moderately infrequent, and rare tsunamis, respectively. By evaluating exceedance probabilities for these horizons, we can infer, for example, what wave height has a 10% chance of being exceeded in 10 years (equivalent to a ~ 100-year return period), and so on. The output of this step is a set of hazard curves for different coastal points (P1–P5), indicating the probability that the tsunami height will exceed various levels within a 1-year exposure window for each return period of interest. These curves can be compared with historical data-derived curves (based on past events’ run-up and tide gauge records) to check for consistency. To generate historical curves and compare them with the probability of exceedance curves from simulations, a Tapered Pareto distribution function was fitted to the historical data, and the hazard was then evaluated using the Poisson model. 3.2.4 Inundation Maps To translate the probabilistic results into actionable information, we developed tsunami inundation maps for the 50-year, 100-year, and 500-year return period scenarios. These maps depict inundation levels associated with fixed return periods (or fixed exceedance probabilities) and highlight the extent of flooding in the study area. It is worth noting that, strictly speaking, the inundation levels shown in, for example, the 50-year return period map does not take place at the same time. Despite this, these maps are useful for assessing the relative risk to different parts of the city and surrounding regions, helping identify areas of high vulnerability. We employed two complementary approaches to generate the inundation extents, as described below. Elevation-based mapping (GIS method): Using the hazard curves and recorded wave heights from the PTHA, we inferred the expected run-up or inundation elevation for each return period using the hazard curve data from synthetic tide gauge P3, which is located in the most populated area of Manzanillo. From the synthetic tide-gauges set up in GeoClaw at a 10m isobath to measure the wave amplitude, we applied Green’s law scaling (Eq. 3) to approximate how the wave height would amplify as it shoals to a shallower reference depth (e.g., 1 m depth near the shoreline). $$\:{\eta\:}_{2}={\eta\:}_{1}\sqrt[4]{\frac{{h}_{1}}{{h}_{2}}},\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(3\right)$$ where \(\:{\eta\:}_{1}\) and \(\:{\eta\:}_{2}\) are the wave amplitudes at two points perpendicular to the coastline, and \(\:{h}_{1}\) and \(\:{h}_{2}\) , are the corresponding water depths of those two points. Green’s law provides a first-order approximation to estimate the run-up or nearshore wave heights from the offshore values. We used this method to estimate the run-up heights corresponding to wave heights with a given exceedance probability. Although Green’s law does not capture complexities such as wave breaking, reflection, or bore formation, it is widely used in regional tsunami studies to obtain first-order estimates of run-up when detailed modeling is not available (Davies et al., 2018 ; Mulia et al., 2020 ; Ha et al., 2022 ). We took the estimated inundation heights and delineated areas on a map that lie below these heights and are connected to the coast (to avoid counting inland depressions that would not actually flood because they are isolated by higher ground). This was performed in QGIS using the ~ 30 m resolution DEM. Essentially, we “flooded” the DEM up to the critical elevation and identified all contiguous areas that made contact with the shoreline. We refined these areas using high-resolution coastline data and overlaid OpenStreetMap layers for context (roads and buildings) in the visualization. This produced a set of polygons indicating the potential flood zones for each scenario (50-, 100-, 500-year). Dynamic simulation mapping: To complement the method described above, we directly implemented GeoClaw to simulate specific scenario events that roughly correspond in size to the target return periods. For example, we identified one of the synthetic earthquake scenarios of approximately Mw 7.54, which produced an offshore wave height similar to that of the 50-year hazard level, for which we performed a full 3-hour tsunami inundation simulation for that event on a very fine grid extending up to ~ 25 m elevation inland. Similarly, we simulated a Mw 7.74 event for the 100-year scenario and an Mw ~ 8.49 event for the 500-year scenario. These simulations yielded high-resolution water depth and flow extent data, which were then mapped as maximum inundation footprints. Comparing the GIS-based inundation extents with the simulation-based extents allowed us to gauge the accuracy of the simpler method. In cases where the simulation showed water reaching further inland than in the static method, it indicates that dynamic effects (such as momentum carrying water over small ridges or through narrow passes) are important. By combining these methods, we ensured that the inundation maps were both data-driven (from hazard curves) and physically validated (through hydrodynamic simulations). All maps were referenced to the same vertical datum (mean sea level), and the influence of tides was not explicitly included (effectively assuming that a tsunami occurs at the mean sea level). In practice, if a tsunami arrives at high tide, which in this region reaches a maximum of 1.14 m (Aguilar-Chávez et al., 2009 ), inundation could reach slightly further away. Table 1 summarizes the tsunami amplitudes at each virtual sensor derived from the hazard curves, along with the representative earthquake magnitudes and the resulting offshore peak wave elevations used in the inundation map scenarios for both evaluated methodologies. Table 1. Maximum tsunami amplitudes for the cases used to evaluate inundation in GeoClaw, compared with the values obtained from the PTHA analysis at the virtual tide gauge locations. Return period (years) Hazard curve Database comparable scenario η (m) Mw η (m) P1 P2 P3 P4 P5 P1 P2 P3 P4 P5 50 0.81 1.36 1.42 0.85 1.13 7.54 0.77 1.35 1.38 0.86 1.14 100 1.37 2.25 2.31 1.40 1.84 7.74 1.32 2.27 2.25 1.32 1.91 500 3.16 4.61 4.53 3.08 3.81 8.49 3.41 4.57 4.73 3.09 3.98 4 Results The results of the tsunami hazard analysis for Manzanillo provide crucial insights into the potential impacts of tsunamis on the region. In the following subsections we present the validation of our model against the 1995 tsunami data, the probabilistic exceedance curves derived from the suite of scenario simulations, and the inundation maps for different return periods, highlighting the spatial extent of flooding in each scenario and the implications for the region. 4.1 Validation To verify the accuracy of the tsunami simulations, we compared the model output for the 1995 Colima-Jalisco tsunami with observational data. Figure 2 shows a comparison of the simulated and observed water-level time series at the Manzanillo tide gauge location (MTG). The model results, based on the 28 validation scenarios, are summarized by the mean water surface elevation and a shaded band indicating one standard deviation across those scenarios. The first wave arriving in both the observation and the model is a trough (a lowering of the water level) followed by a peak, consistent with regional subduction movement. The mean peak amplitude of the model is within the observed range, and subsequent oscillations (the trailing wave train) show a similar pattern of decay over time. This suggests that the stochastic slip distributions that we used captured the essential characteristics of the 1995 earthquake tsunami generation. One consistent discrepancy was noted, which is that the arrival time of the tsunami in the simulations tended to be slightly delayed by approximately 3–10 min compared to the actual tide gauge record. Considering that the aim of this study was to determine the maximum amplitudes and inundation extents, the timing offset was not critical, as in the case of an early warning system. All 28 scenario simulations exhibited this delay to a certain extent. This timing offset could be due to slight inaccuracies in the assumed rupture initiation location or the velocity model (which affects the wave speed), or perhaps the bathymetric grid resolution introducing small phase errors. Despite this, the wave period and the overall duration of the event were well reproduced. Importantly, the range between the maximum and minimum simulated wave heights encompasses the observed waveform, indicating the observed event’s characteristics are not outliers relative to our model ensemble. These results provide a confident framework showing that our approach is reliable for predicting tsunami impacts in the region, at least for events of similar scale and mechanism to 1995. As such, the validation exercise confirmed that the chosen modeling framework (FakeQuakes + GeoClaw) is suitable for simulating local tsunamis in Manzanillo. The close match in waveform evolution lends credibility to inundation and hazard predictions from hypothetical events. Minor timing errors in the simulation are acknowledged, but these are not significant enough to influence hazard assessment outcomes, which concern maximum amplitudes and inundation extents. Nonetheless, this indicates a limitation in using this method for emergency response, where timing is essential for effective action. 4.2 Probability of Exceedance Curves Using the generated database of 4,800 synthetic tsunami simulations, exceedance probability curves were constructed for each measurement point (P1–P5) along the coast. Figure 3 presents these curves, showing the estimated relationship between the tsunami amplitude and return period at each point. It is worth mentioning that the probability of exceedance curves was calculated without considering epistemic uncertainties in the input data or model used, which may lead to hazard underestimation. For clarity, we also compared these with analogous curves derived from historical tsunami data for the Manzanillo region, following an approach similar to that of Geist and Parsons ( 2006 ). The following patterns were observed in the PTHA curves: For a 100-year return period (which corresponds to a 40% exceedance probability in 50 years or ~ 99% in 500 years), the highest tsunami amplitudes among the five sites were found at P3, which is situated near the city center (Manzanillo Bay). This suggests that, for relatively moderate events, the central bay experiences the largest waves. However, when we consider longer return periods (more extreme events), the hazard appears to shift spatially. For return periods of over 800 years, site P2 (representing the Santiago Bay area) began to show a higher hazard than P3. The same shift was observed at points P1 and P4. This indicates that extremely large-magnitude events, although rare, are more likely to severely impact Santiago Bay. In addition, P5 shows a slight increase in hazard by a 1,000-year return period, consistent with the idea that open-coast facing segments (such as Campos) are prone to the impact of large tsunamis. When comparing our PTHA curves to those derived from historical data, we observed that our runup-estimated curves have a good fit for small return periods, but this approximation tends to overestimate tsunami heights for long return periods, especially at higher exceedance levels. For instance, for an amplitude of up to 1 m, our analysis shows good agreement with the historical record; however, at higher exceedance amplitudes, our analysis might assign a shorter return period than the historical record. Part of this discrepancy arises from limited historical data, as the historical catalog for Manzanillo is sparse and might underestimate the tail risk, whereas our synthetic approach, by design, explores more extremes. Another interesting observation is the difference between the use of offshore wave data and run-up data for constructing hazard curves. Using the full historical dataset, the hazard curve generated from the run-up estimations (via Green’s law) aligns well with the historical data for smaller tsunami amplitudes. However, at large tsunami amplitudes, we noticed that runup estimations over-predicted the hazard, whereas offshore data curves showed a better correspondence with the historical data. This aligns with known issues where Green’s law can overestimate inland amplification because it does not consider energy losses and complex ground interactions, which is especially important for very large events where inland penetration of water may be of greater importance. From a practical perspective, these exceedance curves provide a complete distribution of tsunami hazards as a function of the wave height estimation. Based on the results, the probability that a tsunami exceeds a given height within a specific exposure window can be estimated using a generalized Poisson model: \(\:P(\eta\:>{\eta\:}_{c}\:in\:{T}_{exp}\:years)={(1-e}^{-{T}_{exp}/{T}_{r}\left({\eta\:}_{c}\right)}),\) where \(\:{T}_{r}\left({\eta\:}_{c}\right)\) is the return period associated with a tsunami height \(\:{\eta\:}_{c}.\) This formulation enables multiple applications crucial for engineering design and planning. Nonetheless, uncertainties should be addressed, as the curves have inherent variability due to assumptions in the seismic rates (Gutenberg–Richter parameters), chosen maximum magnitude, restriction of the rupture area, and number of scenarios that were sampled. It is noteworthy that the results are consistent with evidence from historical run-ups beyond 10 m reported for the 1932 event (Corona and Ramírez-Herrera, 2012 ). 4.3 Inundation Maps Based on the methods described in Section 3.2.4 , we generated inundation maps for the 50-year, 100-year, and 500-year return-period scenarios. For each scenario, we present two maps, one based on the simplified Green’s law-based inundation elevation method using QGIS processing tools (labeled as “GIS method”) and the other set based on the explicit GeoClaw simulation of a representative event (labeled as “simulation results”). These are shown in Figs. 4 – 6 , where the simulation result maps also show the amplitudes recorded at the measurement points and the runup values approximated using Green’s law as a reference. The maps consistently show that low-lying areas near the coast are the most vulnerable to tsunami flooding in Manzanillo. In particular, the flat regions fringing Santiago Bay, the downtown and port area around Manzanillo Bay, and the extensive wetlands around Cuyutlán Lagoon emerge as hotspots of inundation. As the return period (and associated tsunami size) increases from 50 to 500 years, the extent of flooding increases, covering larger areas and reaching further inland. 4.3.1 50-Year Return Period For a relatively frequent tsunami scenario (on the order of a 50-year return period, which might correspond to a local offshore earthquake of roughly low Mw ~ 7.3 based on the Gutenberg-Richter law), inundation is limited but still noteworthy in certain spots (Fig. 4 . a ). Based on offshore hazard values and run-up estimates, flood extents indicate that water primarily affects zones that are immediately adjacent to the shoreline. In Santiago Bay, water intrudes near Juluapan Lagoon, reaching ~ 1.6 km inland along the lagoon’s low-lying channel. In Manzanillo Bay, the most extensive inundation occurs around the Las Garzas/Valle de las Garzas Lagoon, where the inundation is more associated with water intrusion from the port. In this area, water extensions reach roughly 1.8 km inland along the lagoon systems. This area includes urban neighborhoods and the backwater zones of the port. The San Pedrito and Las Garzas lagoons (in central Manzanillo) act as pathways for water, meaning that the surrounding neighborhoods could experience flooding because of the concentration of water in those basins, even if the open coast shows smaller waves. Notably, even in this modest scenario, the Cuyutlán Lagoon area on the eastern side experiences the farthest inland flooding, reaching ~ 5.6 km from the shoreline This is largely because the terrain there is very flat and low; once water overtops the beach or enters the lagoon inlet, it can propagate far across the marshy land with little resistance. Comparison with a simulation of a specific Mw 7.54 earthquake scenario (Fig. 4 . b ) meant to represent a 50-year event, shows a broadly similar pattern at the coastline, with influence restricted to the first ~ 100 m of the shoreline but with markedly less inundation through the lagoon channels. In this simulation, the maximum inundated area occurs at Manzanillo port, with flooding extending ~ 400 m inland and maximum inundation depths reaching ~ 4 m inside the port. In Santiago Bay, floodwaters reach ~ 300 m inland, although with maximum depths remaining below 1 m at Cuyutlán Lagoon; simulation results show ~ 200 m of water intrusion through the lagoon inlet. While the affected zone reaches up to ~ 4 km from the coast, the water depths are shallow (a few centimeters), likely due to frictional and topographic effects. Overall, the simulated inundation footprints were less extensive than the GIS-based footprints, implying that the straightforward elevation method may overestimate the inundation for smaller events. Despite these differences, both methods consistently flag the same zones (Juluapan, downtown/port, and Cuyutlán) as areas of concern, even for relatively frequent tsunamis. These results suggest that even a tsunami expected in a 50-year frame could cause flooding a few kilometers inland in the worst-hit parts of Manzanillo, which is significant for local planning. 4.3.2 100-Year Return Period A less frequent event resulting in inundation maps is shown in Fig. 5 . This set corresponds to a larger event with a 100-year return period, roughly similar to a Mw 7.74 earthquake impacting the area based on the Gutenberg-Richter law. The inundation areas are larger than those in the 50-year scenario. Consistently, the same neighborhoods are affected as in the 50-year scenario, but flooding reaches further inland and covers wider zones. For example, in Manzanillo Bay, using run-up estimation from the GIS evaluation, flooding extends approximately 1.8 km inland. This would likely inundate much of the downtown waterfront district and the tourist areas near the shore. The low-lying part of the city (including port facilities and main roads along the coast) could be under a few meters of water. The higher spots or cliffs (such as the Peninsula de Santiago, dividing Santiago and Manzanillo Bays) remain safe regions, underscoring the importance of small elevation differences. In Santiago Bay, the water can reach up to ~ 1.7 km inland near the Juluapan Lagoon, a slight increase from the 50-year case, but significantly inundating that lagoon’s surroundings and possibly affecting communities along its edges. The Cuyutlán Lagoon shows an increase to ~ 5.8 km of inland flooding in the run-up-based map, indicating significant impacts on its low-lying areas. One area that stands out for the 100-year event is the downtown Manzanillo port area. Given its economic importance, the map indicates that significant portions of port infrastructure are within the 100-year inundation zone. Inundation depths in the port/downtown area could be significant in worst-case scenarios, where tsunami waves are funneled between structures or along channels. The dynamic simulation for a 100-year scenario, particularly the Mw 7.74 scenario, is shown in Fig. 5 . b. This approach highlights remarkable hotspots with elevated flood depths, particularly over 6 m, within the port. In Santiago Bay, the maximum inundation depth reaches ~ 2.9 m in the eastern sector. The greatest inland reach in this bay occurs at Juluapan Lagoon, with tsunami effects extending up to ~ 1.5 km, although with water depths of just a few centimeters. In Manzanillo Bay, sea intrusion reaches ~ 1.5 km, slightly less than the 1.8 km shown by the static method. For Cuyutlán, the simulation showed a ~ 4.3 km inland reach versus ~ 5.8 km in static. Again, the simulation suggests slightly less penetration, possibly owing to energy dissipation and the finite duration of the wave. Despite differences in exact numbers, both approaches clearly indicate that a 100-year tsunami would inundate large parts of the coastal plains and devastate unprepared infrastructure. The results emphasize that the critical infrastructure in Manzanillo, including the port, power facilities, and tourism centers, largely lies within the 100-year tsunami flood zone, underlining an urgent need for mitigation. 4.3.3 500-Year Return Period For the low-probability 500-year scenario, which could correspond to an event approaching the upper magnitude considered (Mw ~ 8.3 as estimated from the Gutenberg-Richter law), the GIS-based inundation map (Fig. 6 . a ) shows extensive flooding at Manzanillo Bay, covering nearly all urban and port areas. The inundation reaches up to ~ 2.6 km inland, sparing only the steepest zones near the bay edges. Around Juluapan Lagoon, flooding may extend ~ 2.5 km inland, affecting the surrounding roads and residential areas. Eastern Santiago Bay experiences significant inundation through the Peñitas Lagoon and Santiago River channel. In Cuyutlán the inundation reached up to ~ 6.2 km, suggesting the entire lagoon would become part of the ocean temporarily, and water would penetrate beyond its normal boundaries into adjacent communities. Campos, the community near Cuyutlán Lagoon’s mouth, and the coastline immediately south of it, facing the open ocean, would likely be ground zero under these extreme events. The simulation for a Mw ~ 8.49 event (Fig. 6 . b ) indicates extreme water depths in certain bays, such as ~ 6.2 m in Cenicero Bay. While this zone did not flood far inland owing to steep terrain, the high values likely reflect local topographic amplification. Similarly, the eastern side of Santiago Bay shows maximum depths of ~ 6.4 m. Inland reach in the simulation is ~ 1.6 km at Juluapan Lagoon and ~ 4.4 km at Cuyutlán, comparable to the 100-year simulation, which suggest that after a certain threshold, excess water tends to rise vertically rather than advance further inland. In Manzanillo Bay, a larger difference is observed, with sea intrusion reaching ~ 1.9 km from the shoreline. The port area records extreme values of ~ 7.8 m overtopping the breakwaters and ~ 6.9 m of inundation inside the port, which could cause severe disruption to operations and infrastructure. One notable detail in this extreme scenario is the extent of the water entering Cuyutlán Lagoon, indicating severe impacts on any settlement along its shores and some industrial facilities settled in the area. The maps indicate potential inundation there, but it is worth highlighting qualitatively that a 500-year tsunami could plausibly send a ~ 10 m high bore over Campos and into the lagoon behind it. Considering the above results, the 500-year inundation scenario delineates the outer envelope of the tsunami flooding, and one might reasonably prepare for the given current knowledge. Key infrastructure that remains unaffected even in the 500-year maps tends to be on high ground, such as some parts of the port, which might be built up or protected by seawalls to higher levels, although this was not explicitly modeled here. 5 Discussion The results provide a detailed picture of tsunami hazards in Manzanillo, highlighting both the spatial variability in risk and the importance of a probabilistic approach. The analysis identified hotspots of high-risk areas directly linked to local geomorphology, as described below. Manzanillo Bay (urban center) has emerged as a hotspot for tsunamis generated by moderately large earthquakes (Mw < 8.0), with return periods under a few hundred years. The bay’s shape and orientation likely contribute to this, as a semi-enclosed bay resonance effect can amplify tsunami wave heights at particular frequencies (Bellotti et al., 2012 ). The infrastructure and population density here mean that even moderate flooding can have outsized consequences, especially in the context of an increasing population and development. Santiago Bay, especially near the Juluapan Lagoon, becomes comparably or more hazardous as we consider rarer, larger tsunamis (~ 800–1000 year events). This might be because extremely large waves from the southwest can wrap around into this bay effectively or even enter through the gap toward the lagoon. The low-land areas at the Cuyutlán Lagoon and the coastline near the entrance of the lagoon are notably exposed. For the largest events we considered (~ 1,000-year return periods), this open-coast site experiences an increase of the run-ups when compared to the amplitudes recorded at P1 and P4. This is consistent with historical accounts, such as the catastrophic 1932 tsunami that produced a ~ 12 m run-up along the Cuyutlán shores (Corona and Ramírez-Herrera, 2012 ). The lack of significant headlands to dissipate energy and the long fetch over the lagoon allow waves to maintain their height. Areas with natural barriers, such as cliffs and rocky outcrops (e.g., parts of the coast near points P1 and P4 in our setup), show reduced inundation in all scenarios. These features reflect or channel water, preventing it from flowing directly inland. However, it’s important to note that while such features protect against smaller tsunamis, they do not eliminate risk in extreme events (De Risi and Goda, 2017 ). It is possible that a sufficiently large tsunami may overtop cliffs or wrap around them. The incorporation of a reduced seismicity model allows for a more representative hazard estimation of the regional tectonic setting and is better suited for realistic scenarios. It is worth mentioning that our scenario suite was based on ruptures directly offshore Manzanillo. Earthquakes further up or down the trench could present a certain rupture directivity, affecting the spatial distribution of ground motion amplitudes and the resulting coseismic seafloor deformation. This variation generates different tsunami wave patterns and hence leads to variations in coastal inundation and damage distribution (for example, Singh et al. 2023 ; Iglesias et al. 2022 ). Nevertheless, the spatial pattern we found aligns with what was intuitively expected and what previous regional models have suggested: the central and eastern parts of Manzanillo’s coast (harbor and Cuyutlán side) are more vulnerable than the western part (around Carrizales/Cenicero bays) (Salazar-Monroy et al., 2021 ; Evangelista et al., 2018 ). From a planning perspective, tsunami mitigation efforts should prioritize high-risk zones, such as reinforcing sea walls and natural buffers in port areas and maintaining mangrove belts around lagoons to dampen waves. Local amplification phenomena, such as harbor oscillations (seiche), could prolong hazards for many hours after the initial tsunami, and our 3-hour simulations in GeoClaw indicate that oscillations persist, which is in line with other port cities’ experiences. The results from synthetic tsunami events were used to generate hazard curves, which were compared with historical run-up data and tide gauge measurements. Hazard estimations derived from Green’s Law tend to overestimate the runup at high return periods, but when compared to only historical run-up data, the results align better for large amplitudes. However, the hazard curve based on run-up data may be unreliable owing to the limited availability of measurements. Generally, there is a scarcity of homogeneous historical data, leading to inconsistencies in PTHA evaluations, and relying only on historical data may lead to an underestimation of the hazard level of a region. An example is the Tohoku-Oki tsunami event in 2011 (Mw 9.0) (Geist and Parsons, 2006 ; Mori et al., 2017 ). In this study, tsunami inundation levels near the coast combined with amplification factors were used, and inundation simulations were performed for comparison. While widely used, Green’s Law oversimplifies inland attenuation; therefore, Glimsdal et al. ( 2019 ) suggested replacing it with more accurate amplification models. This limitation may lead to an overestimation of inundation when using elevation-based assessments without accounting for terrain friction, which differs from flooding extensions when compared to simulations, aligning with findings from similar studies (e.g., Tonini et al., 2021 ). Nevertheless, inundation simulations may exhibit larger values in some regions compared with flood level-based assessments, suggesting that local geomorphological interactions amplify tsunami waves. These findings underscore the importance of using detailed simulation models for regional tsunami risk assessment because simplified methods cannot adequately capture these amplification effects. Exceedance probability curves were constructed using a 1-year time window to facilitate the interpretation of hazard levels over different planning horizons (e.g., 50-, 100-, and 500-year return periods). Discrepancies with previous studies (e.g., Salazar-Monroy et al., 2021 ) arise from differing data resolutions, observation windows, and sampling methods. Higher maximum amplitudes for 1,000-year return periods were found, likely because of finer bathymetry resolution and a fixed number of simulated events (300 scenarios for each magnitude), especially for the largest magnitudes. However, expanding large-magnitude event scenarios should be considered, as historical data confirm run-up values exceeding 10 m along this region. The probabilistic approach in this assessment provides a more comprehensive understanding of tsunami hazards by considering a range of possible earthquake scenarios rather than relying solely on historical events, which are critical for regions such as Manzanillo, where historical records may not fully capture the complete potential of hazards. Flood maps generated using inundation levels in QGIS identified high-risk areas that aligned with previous studies by Evangelista et al. ( 2018 ) and the Atlas Nacional de Riesgos (ANR, 2024) for Manzanillo, Colima, although a smaller range of risk areas was found. In the ANR evaluation, regions up to 20–30 m elevation were classified as medium- and low-risk, while this study found a maximum inundation height of approximately 10 m for a 500-year return period. In contrast, differences with Evangelista et al. ( 2018 ) stem from the use of higher-magnitude seismic sources (up to 9.5 Mw) and different topographic data. The results indicate that some regions of Manzanillo are at higher risk of inundation, particularly during large, infrequent tsunami events. These areas, such as the central bay and the vicinity of Cuyutlán Lagoon, require targeted risk-management strategies, including infrastructure reinforcement, evacuation planning, and public education programs. The information provided by inundation maps can be used by local authorities to prioritize these efforts and ensure that the most vulnerable populations are protected. One of the key strengths of this study is the use of high-resolution topographic and bathymetric data, which allow for detailed simulations of tsunami impacts. This level of detail is essential for accurately predicting the extent of inundation and identifying the specific areas of concern. However, it is important to note that the accuracy of the predictions depends on the quality of the input data and assumptions made in the modeling process. 6 Conclusions The probabilistic tsunami hazard assessment (PTHA) conducted for Manzanillo, Colima, demonstrated that the city is at significant risk from tsunamis, especially those originating near the Mesoamerican Trench. This study incorporated a probabilistic approach, considering a wide range of possible earthquake scenarios rather than relying solely on historical events, leading to a more comprehensive assessment of tsunami hazards. By combining stochastic seismic modeling, hazard exceedance curves, and high-resolution tsunami simulations using GeoClaw, the present study provides essential insights for disaster preparedness, coastal management, and infrastructure planning. The hazard exceedance curves reveal that tsunamis with wave heights of up to 1 m have a more than 80% probability of occurring within the next 50 years. This quantification underscores the need for proactive risk reduction measures, particularly in low-lying coastal areas such as Santiago Bay, Manzanillo Bay (downtown and port zones), and Cuyutlán Lagoon, which are the most vulnerable during inundation. Even a moderate tsunami (with a 50-year return period) could cause flooding of up to a few kilometers inland, whereas a 500-year return period event could inundate areas up to 5–6 km inland through Manzanillo’s water bodies. These findings highlight specific neighborhoods and zones that require prioritized attention for tsunami risk reduction, underscoring the need for targeted disaster preparedness efforts, including infrastructure reinforcement, public education, and the development of comprehensive evacuation plans. Model validation against the 1995 tsunami event confirms that the approach used is robust, with simulated wave heights and inundation extents closely matching historical observations. However, minor timing discrepancies suggest that future studies could refine bathymetric resolution and rupture parameterization to improve model accuracy, particularly for emergency responses. The hazard curves further indicate that certain regions, such as Campos and the Cuyutlán Lagoon area, exhibit greater tsunami risk for longer return periods due to their open-coast exposure and lack of natural barriers. The information provided by this study can directly inform local authorities and stakeholders in developing and refining risk mitigation strategies. For instance, inundation maps can be used to update tsunami evacuation routes and signage, ensuring that they encompass areas at risk for even larger events. Urban development plans can incorporate these hazard zones by avoiding critical new infrastructure in areas shown to flood during 100- or 500-year events. Existing critical facilities within hazard zones should have contingency plans or protective measures such as seawalls and elevated structures. Probability curves and worst-case scenarios can also be used to design tsunami drills and public education campaigns to enhance community preparedness. Future work should continue to refine these models by incorporating new data and improving the accuracy of the predictions, as well as considering the epistemic uncertainties involved. Additionally, integrating tsunami hazard assessments with other coastal risk factors, such as sea-level rise and storm surges, will be crucial for creating a holistic approach to coastal management in Manzanillo. Enhancing early warning systems, particularly by integrating seismic and sea-level monitoring into automated alert systems, will be essential for ensuring rapid response capabilities. Developing evacuation infrastructure, such as vertical evacuation options in the most at-risk zones and conducting regular tsunami drills, will further improve community resilience. Moreover, land use planning must consider tsunami hazard zones when approving new developments, and protective engineering measures such as tsunami barriers and reinforced structures should be implemented for critical infrastructure in vulnerable areas. By implementing these measures, Manzanillo can significantly reduce the potential loss of life and property damage from future tsunamis. The findings of this study serve as a scientific foundation for action, and stakeholders should use this knowledge proactively. The methodology and lessons from this study can also be extended to other communities in tsunami-prone areas that face similar risks. Declarations Acknowledgments This research was supported by the Secretaría de Ciencia, Humanidades, Tecnología e Innovación (formerly CONAHCYT), through a graduate studies scholarship to LVC. CMA was supported by UNAM-DGAPA-PASPA and Aarhus University Research Foundation fellowships. The authors acknowledge IT support from Gonzalo Martín Ruiz and the Gerencia de Ingeniería from Comisión Federal de Electricidad (CFE) for providing the detailed bathymetry data used in this work. Funding This work was supported by the Secretaría de Ciencia, Humanidades, Tecnología e Innovación (formerly CONAHCYT), through a scholarship for LVC. CMA was supported by UNAM-DGAPA-PASPA and Aarhus University Research Foundation fellowships. Competing Interests The authors have no relevant financial or non-financial interests to disclose. Author Contributions Study conception and design were performed by Lizzeth Vázquez Camaal, Christian M. Appendini and Ericka Alinne Solano Hernández. Material preparation and data collection were performed by Lizzeth Vázquez Camaal. Analysis was performed by all authors. The first draft of the manuscript was written by Lizzeth Vázquez Camaal and Christian M. Appendini and all authors commented on previous versions of the manuscript. All authors have read and approved the final manuscript. Data availability The data used in this study will be made available at Zenodo upon acceptance. References Aguilar-Chávez, A., Salas-de-León, D. A., Monreal-Gómez, M. A., Aldeco-Ramírez, J., & Signoret, M. (2009). Circulation and numerical modeling of the Manzanillo harbor, Colima, Mexico. WEH International Workshop on Environmental Hydraulics: Theoretical, Experimental and Computational Solutions, Valencia, Spain. Aranguiz, R., Catalán, P. A., Cecioni, C., Bellotti, G., Henriquez, P., & González, J. (2019). Tsunami resonance and spatial pattern of natural oscillation modes with multiple resonators. Journal of Geophysical Research: Oceans , 124 (11), 7797-7816. https://doi.org/10.1029/2019JC015206 Atlas Nacional de Riesgos (ANR). (2024). Atlas de riesgos de Manzanillo 2024. Retrieved from http://www.atlasnacionalderiesgos.gob.mx/ Bandy, W. L., Mortera-Gutiérrez, C. A., Urrutia-Fucugauchi, J., & Hilde, T. W. C. (2000). The subducted Rivera-Cocos plate boundary: Where is it, what is it, and what is its relationship to the Colima Rift? Geophysical Research Letters , 27(17), 3397-3400. https://doi.org/10.1029/1999GL011247 Becker, J. J., Sandwell, D. T., Smith, W. H. F., Braud, J., Binder, B., Depner, J., Fabre, D., Factor, J., Ingalls, S., Kim, S.-H., Ladner, R., Marks, K., Nelson, S., Pharaoh, A., Trimmer, R., Von Rosenberg, J., Wallace, G., & Weatherall, P. (2009). Global bathymetry and elevation data at 30 arc seconds resolution: SRTM30_PLUS. Marine Geodesy , 32(4), 355-371. https://doi.org/10.1080/01490410903297766 Bellotti, G., Briganti, R., & Beltrami, G. M. (2012). The combined role of bay and shelf modes in tsunami amplification along the coast. Journal of Geophysical Research: Oceans , 117 (C8). https://doi.org/10.1029/2012JC008061 Borerro, J., Ortiz, M., Titov, V., & Synolakis, C. (1997). Field survey of Mexican tsunami produces new data, unusual photos. Eos, Transactions American Geophysical Union , 78 (8), 85-88. https://doi.org/10.1029/97EO00054 Castillo‐Aja, R., & Ramírez‐Herrera, M. T. (2017). Updated tsunami catalog for the Jalisco‐Colima coast, Mexico, using data from historical archives. Seismological Research Letters , 88 (1), 144-158. https://doi.org/10.1785/0220160133 Cornell, C. A., & Vanmarcke, E. H. (1969). The major influences on seismic risk. In Proceedings of the fourth world conference on earthquake engineering , vol. 1, pp. 69-83. Corona, N., & Ramírez-Herrera, M. T. (2012). Tsunami hazard and vulnerability in the Mexican Pacific coast: A review. Natural Hazards , 61(2), 1213-1224. https://doi.org/10.1007/s11069-011-0076-0 Davies, G., Griffin, J., Løvholt, F., Glimsdal, S., Harbitz, C., Thio, H. K., & Baptista, M. A. (2018). A global probabilistic tsunami hazard assessment from earthquake sources. Geological Society, London, Special Publications, 456 , 219–244. https://doi.org/10.1144/SP456.5 De Risi, R., & Goda, K. (2017). Simulation-based probabilistic tsunami hazard analysis: empirical and robust hazard predictions. Pure and Applied Geophysics , 174 , 3083-3106. https://doi.org/10.1007/s00024-017-1588-9 Evangelista, Y., Castrejón, C., Villa, D., Trujillo, K., Monsivais-Huertero, A., & Mendozc, A. (2018, July). Generation of a tsunami hazard map for the coast of Manzanillo, Colima in Mexico using numerical simulation and wave modeling. In IGARSS 2018-2018 IEEE International Geoscience and Remote Sensing Symposium (pp. 7843-7846). IEEE. https://doi.org/10.1109/IGARSS.2018.8517856 Geist, E. L., & Oglesby, D. D. (2014). Tsunamis: Stochastic models of occurrence and generation mechanisms. Journal of Geophysical Research: Solid Earth , 119(9), 7223-7242. https://doi.org/10.1002/2014JB011418 Geist, E. L., & Parsons, T. (2006). Probabilistic analysis of tsunami hazards. Natural Hazards , 37(3), 277-314. https://doi.org/10.1007/s11069-005-4646-z Glimsdal, S., Løvholt, F., Harbitz, C. B., Romano, F., Lorito, S., Orefice, S., Brizuela, B., Selva, J., Hoechner, A., Volpe, M., Babeyko, A., Tonini, R., Wronna, M. & Omira, R. (2019). A new approximate method for quantifying tsunami maximum inundation height probability. Pure and Applied Geophysics , 176 , 3227-3246. https://doi.org/10.1007/s00024-019-02091-w González, F. I., Satake, K., Boss, E. F., & Mofjeld, H. O. (1995). Edge wave and non-trapped modes of the 25 April 1992 Cape Mendocino tsunami. In: Imamura, F., Satake, K. (eds) Tsunamis: 1992–1994 . Pageoph Topical Volumes. Birkhäuser Basel. https://doi.org/10.1007/978-3-0348-7279-9_3 Ha, T., Yoon, J. S., Kyung, J. B., Yoon, S. H., Lee, D. W., & Kim, K. H. (2022). Occurrence of local tsunamis along the eastern coast of the Korean peninsula based on numerical modeling of historical earthquakes. Frontiers in Earth Science , 10 , 836168. https://doi.org/10.3389/feart.2022.836168 Hayes, G. P., Moore, G. L., Portner, D. E., Hearne, M., Flamme, H., Furtney, M., & Smoczyk, G. M. (2018). Slab2, a comprehensive subduction zone geometry model. Science , 362 (6410), 58-61. https://doi.org/10.1126/science.aat4723 Iglesias, A., Singh, S. K., Castro‐Artola, O., Pérez‐Campos, X., Corona‐Fernández, R. D., Santoyo, M. A., Espíndola V. H., Arroyo D., and Franco, S. I. (2022). A source study of the Mw 7.0 Acapulco, Mexico, earthquake of 8 September 2021. Seismological Research Letters , 93 (6), 3205-3218. https://doi.org/10.1785/0220220124 Jelínek, R., Čížková, H., & Spikings, R. (2012). Thermal structure and evolution of the lithosphere in the central Mexican subduction zone: New constraints on the thermal budget of subduction. Earth and Planetary Science Letters , 330-331, 97-108. https://doi.org/10.1016/j.epsl.2012.03.020 LeVeque, R. J., George, D. L., & Berger, M. J. (2011). Tsunami modelling with adaptively refined finite volume methods. Acta Numerica , 20 , 211-289. https://doi.org/10.1017/S0962492911000043 Melgar, D., LeVeque, R. J., Dreger, D. S., & Allen, R. M. (2016). Kinematic rupture scenarios and synthetic displacement data: An example application to the Cascadia subduction zone. Journal of Geophysical Research: Solid Earth , 121(9), 6658-6674. https://doi.org/10.1002/2016JB013314 Mendoza, C., & Hartzell, S. H. (1999). Fault-slip distribution of the 1995 Colima-Jalisco, Mexico, earthquake. Bulletin of the Seismological Society of America , 89(5), 1338-1344. https://doi.org/10.1785/BSSA0890051338 Miyashita, K., Shuto, N., & Imamura, F. (2020). Comprehensive study of tsunami effects on coastal areas. Journal of Coastal Research , 36(1), 44-52. https://doi.org/10.2112/JCOASTRES-D-18-00123.1 Molnar, P., & Sykes, L. R. (1969). Tectonics, seismicity, and morphology of the Caribbean and adjacent regions. Geological Society of America Bulletin , 80(9), 1639-1684. https://doi.org/10.1130/0016-7606(1969)80[1639:TSAMOT]2.0.CO;2 Mori, J., & McCaffrey, R. (2018). Tsunami hazard in the Pacific Ocean. Annual Review of Earth and Planetary Sciences , 46, 77-103. https://doi.org/10.1146/annurev-earth-082517-010031 Mori, N., Muhammad, A., Goda, K., Yasuda, T., & Ruiz-Angulo, A. (2017). Probabilistic tsunami hazard analysis of the Pacific coast of Mexico: case study based on the 1995 Colima earthquake tsunami. Frontiers in built environment , 3 , 34. https://doi.org/10.3389/fbuil.2017.00034 Mulia, I. E., Ishibe, T., Satake, K., Gusman, A. R., & Murotani, S. (2020). Regional probabilistic tsunami hazard assessment associated with active faults along the eastern margin of the Sea of Japan. Earth, Planets and Space , 72 , 1-15. https://doi.org/10.1186/s40623-020-01256-5 National Oceanic and Atmospheric Administration (NOAA). (2023). Historical tsunami database. Retrieved from https://www.ngdc.noaa.gov/hazard/tsu_db.shtml. Pari, Y., Murthy, M. R., Subramanian, B. R., & Ramachandran, S. (2008). Morphological changes at Vellar estuary, India—Impact of the December 2004 tsunami. Journal of environmental management , 89 (1), 45-57. https://doi.org/10.1016/j.jenvman.2007.01.055 Salazar‐Monroy, E. F., Melgar, D., Jaimes, M. A., & Ramirez‐Guzman, L. (2021). Regional probabilistic tsunami hazard analysis for the Mexican subduction zone from stochastic slip models. Journal of Geophysical Research: Solid Earth , 126(6), e2020JB020781. https://doi.org/10.1029/2020JB020781 Sánchez, J., & Farreras, S. F. (1993). Catalog of tsunamis on the western coast of Mexico. World Data Center A for Solid Earth Geophysics Publication, SE-50, National Oceanic and Atmospheric Administration, Geophysical Data Center, Boulder. Singh, S. K., Iglesias, A., Arroyo, D., Pérez-Campos, X., Ordaz, M., Mendoza, C., Corona-Fernández, R. D., Rivera, L., Espíndola, V. H., González-Ávila, D., Martínez-López, R., Castro-Artola, O., Santoyo, M. A., & Franco, S. I. (2023). A seismological study of the Michoacán-Colima, Mexico, earthquake of 19 September 2022 (Mw7.6). Geofísica internacional , 62(2), 445-465. https://doi.org/10.22201/igeof.2954436xe.2023.62.2.1453 Smith, W. H. F., & Sandwell, D. T. (1997). Global sea floor topography from satellite altimetry and ship depth soundings. Science , 277(5334), 1956-1962. https://doi.org/10.1126/science.277.5334.1956 Tonini, R., Di Manna, P., Lorito, S., Selva, J., Volpe, M., Romano, F., & Vittori, E. (2021). Testing tsunami inundation maps for evacuation planning in Italy. Frontiers in Earth Science , 9 , 628061. https://doi.org/10.3389/feart.2021.628061 Zamora, N., Gubler, A., Catalán, P. A., & Carvajal, M. (2021). Tsunami inundation forecast in central Chile using stochastic earthquake scenarios. In Proceedings of the 8th BSC Doctoral Symposium (pp. 72–73). Barcelona Supercomputing Center. https://hdl.handle.net/2117/346622 Zúñiga, F. R., Suárez, G., Figueroa-Soto, Á., & Mendoza, A. (2017). A first-order seismotectonic regionalization of Mexico for seismic hazard and risk estimation. Journal of Seismology , 21 , 1295-1322. https://doi.org/10.1007/s10950-017-9666-0 Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7257032","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":494301459,"identity":"58e844cc-a377-4b17-ba4c-3613be152957","order_by":0,"name":"Lizzeth Vázquez-Camaal","email":"","orcid":"","institution":"Universidad Nacional Autonoma de Mexico Instituto de Ingenieria","correspondingAuthor":false,"prefix":"","firstName":"Lizzeth","middleName":"","lastName":"Vázquez-Camaal","suffix":""},{"id":494301460,"identity":"bfd89470-4e28-47ee-b1ac-5da17fe923a9","order_by":1,"name":"Christian M. Appendini","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEUlEQVRIiWNgGAWjYBAC/mYwxczDwMzY+CDBgAgtEscbYFqYDxsQpcXA5wBYCxCzpUkQ5TADidxjDz7usJYxZ+cxq3hQcMeuX7qBTbqgZhuD7owEHFry0g1nnknnsWzmMbuRYPAseeacA2zSM47dZjA7cwCHlhwzad62wzwGh8FaDicb3Ehgk+ZhA2qBeBNTi/wbM+m/YC383wpAWuzBWv4BtRzG4ZcIoC2MUFsYgFrsDCSAWnjbcNsicQPol962dJAWYwmglgSJG4nN1rx9t3lw+YV/BjDEfrZZ2xucP2P48cefw/b8M5IP3ub5dlsO6DXsLmPgYUPhJjYwMIIdxINDPaYWe9wqR8EoGAWjYKQCAN40XeyOnLyZAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-6044-3351","institution":"Universidad Nacional Autonoma de Mexico Instituto de Ingenieria","correspondingAuthor":true,"prefix":"","firstName":"Christian","middleName":"M.","lastName":"Appendini","suffix":""},{"id":494301461,"identity":"de14ee8b-a469-476e-9076-844afee50392","order_by":2,"name":"Ericka Alinne Solano-Hernández","email":"","orcid":"","institution":"Universidad Nacional Autónoma de México Escuela Nacional de Estudios Superiores Unidad Morelia: Universidad Nacional Autonoma de Mexico Escuela Nacional de Estudios Superiores Unidad Morelia","correspondingAuthor":false,"prefix":"","firstName":"Ericka","middleName":"Alinne","lastName":"Solano-Hernández","suffix":""},{"id":494301462,"identity":"25e33246-a283-4b4f-a9fd-c8cca8b67d1e","order_by":3,"name":"Angel Ruiz-Angulo","email":"","orcid":"","institution":"University of Iceland Institute of Earth Sciences: Haskoli Islands Jarovisindastofnun Haskolans","correspondingAuthor":false,"prefix":"","firstName":"Angel","middleName":"","lastName":"Ruiz-Angulo","suffix":""},{"id":494301463,"identity":"f87da63e-72e9-4258-9881-da630ec033b2","order_by":4,"name":"Mario Ordaz","email":"","orcid":"","institution":"Universidad Nacional Autonoma de Mexico Instituto de Ingenieria","correspondingAuthor":false,"prefix":"","firstName":"Mario","middleName":"","lastName":"Ordaz","suffix":""}],"badges":[],"createdAt":"2025-07-31 01:40:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7257032/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7257032/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89259426,"identity":"a757e2b6-1ed2-41ca-ad48-17fd0f908e62","added_by":"auto","created_at":"2025-08-18 06:27:12","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1035160,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ea)\u003c/em\u003e Tectonic features of the Colima region, including approximate rupture areas for the 1932 and 1995 earthquakes. The June 22, 1932 (M~6.9) earthquake is only indicated by the epicenter (yellow star). \u003cem\u003eb) \u003c/em\u003eMap of the study area in Manzanillo, Colima showing key geographical locations, including coastal bays, lagoons, and river outlets. MTG indicates Manzanillo harbor.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/3e36ecb014aa26c96222e02d.png"},{"id":89259422,"identity":"50f7a743-38e3-4e15-af61-7b13e858bfbb","added_by":"auto","created_at":"2025-08-18 06:27:12","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1604605,"visible":true,"origin":"","legend":"\u003cp\u003eSea level time series at the MTG comparing the observed tide-gauge record of the 1995 Manzanillo tsunami (MTG) and GeoClaw simulation results at the same location. The black curve represents the observed water level, the blue line represents the mean value of the 28 simulated scenarios for that event, and the gray band represents the corresponding mean ±1 std.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/8252dfc1f8c19876f522712e.png"},{"id":89258055,"identity":"bdaa1c1f-35c9-4a8a-91ce-3f85b12fc2d0","added_by":"auto","created_at":"2025-08-18 06:19:12","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":191991,"visible":true,"origin":"","legend":"\u003cp\u003eTsunami probability-of-exceedance curves as a function of maximum tsunami amplitude for coastal points P1–P5 around Manzanillo. Solid lines indicate the PTHA results from this study (with offshore wave heights at 10 m depth converted to run-up estimates), while dashed lines show curves derived from historical data: \u003cem\u003ea)\u003c/em\u003e using combined tide gauge and run-up observations, \u003cem\u003eb)\u003c/em\u003e using only tide gauge data, and \u003cem\u003ec)\u003c/em\u003e using only run-up data.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/09e3ed9fd65503d348bc0d6e.png"},{"id":89260349,"identity":"b75c25b2-b3b8-43d4-9a33-32f84fa7634d","added_by":"auto","created_at":"2025-08-18 06:43:12","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":394470,"visible":true,"origin":"","legend":"\u003cp\u003eInundation maps for the 50-year return period. \u003cem\u003ea)\u003c/em\u003e Inundation map using the GIS method. \u003cem\u003eb)\u003c/em\u003eInundation map using simulation results for a scenario with a 50-year return period. Flooding area from the GIS method (\u003cem\u003ea\u003c/em\u003e) is shown for comparison.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/2238ef597770a3d672fcffc9.png"},{"id":89258061,"identity":"5621f136-b7b3-442e-b3ab-5562c778da1d","added_by":"auto","created_at":"2025-08-18 06:19:12","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":397665,"visible":true,"origin":"","legend":"\u003cp\u003eInundation map for a 100-year return period. \u003cem\u003ea)\u003c/em\u003e Inundation map using the GIS method. \u003cem\u003eb)\u003c/em\u003e Inundation map using simulation results for a scenario with a 100-year return period. Flooding area from the GIS method (a) is shown for comparison.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/9fe96dbd87a3eb1e6b0c7aa1.png"},{"id":89259807,"identity":"42795f2a-c2e8-4623-b95e-6f57fd9f668f","added_by":"auto","created_at":"2025-08-18 06:35:12","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":395157,"visible":true,"origin":"","legend":"\u003cp\u003eInundation map for a 500-year return period. \u003cem\u003ea)\u003c/em\u003e Inundation map using the GIS method. \u003cem\u003eb)\u003c/em\u003e Inundation map using simulation results for a scenario with a 500-year return period. Flooding area from the GIS method (a) is shown for comparison.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/209f293b77782776309ba7b9.png"},{"id":92981718,"identity":"0ba0fe76-5198-42f7-852d-bda1ac354eb7","added_by":"auto","created_at":"2025-10-07 19:49:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4406407,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7257032/v1/3ff33007-d8fa-49a9-a576-361c398a40fe.pdf"}],"financialInterests":"","formattedTitle":"Probabilistic tsunami hazard assessment and flood modeling: A case study of Manzanillo, Colima, Mexico","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eTsunamis are long-wavelength ocean waves generated by sudden disturbances in the water column over a large area, typically caused by seismic events, submarine landslides, or volcanic eruptions. These vertical displacements in the water column release potential energy that is converted into kinetic energy, resulting in radial propagation of tsunami waves (Geist and Oglesby, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Among the various geological phenomena that can trigger tsunamis, subduction zone earthquakes are the most common, especially where tectonic plates converge, leading to significant vertical displacement of the seafloor. These types of earthquakes are typically thrust or reverse fault events that are highly effective in generating tsunamis (Geist and Oglesby, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). The Pacific Ocean's \"Ring of Fire\" is a well-known source of seismic activity owing to its numerous subduction zones, accounting for approximately 78% of all tsunamis globally (NOAA, 2023). Within this context, Mexico's Pacific coast, particularly the central and southern regions, is highly susceptible to local tsunamis due to its proximity to the Mesoamerican Trench (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). This trench is formed by the convergence of the Rivera, Cocos, and North American plates, creating a tectonically active zone prone to large earthquakes (Molnar and Sykes, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e1969\u003c/span\u003e; Bandy et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). The resulting seismic activity has historically generated tsunamis that have impacted coastal communities in Mexico, including Manzanillo, a port city in the State of Colima (Castillo-Aja and Ram\u0026iacute;rez-Herrera, 2017). Manzanillo is particularly vulnerable to tsunamis because of its geographical location, which is adjacent to the Mesoamerican Trench. This region has experienced multiple tsunamis throughout history, with significant events recorded as far back as 1563. Notably, the tsunamis of 1932 and 1995 stand out due to their devastating impacts on Manzanillo, causing severe damage and loss of life (S\u0026aacute;nchez and Farreras, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e1993\u003c/span\u003e; Corona and Ram\u0026iacute;rez-Herrera, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The June 22, 1932, earthquake (M\u0026thinsp;~\u0026thinsp;6.9) reportedly produced run-up heights of up to 12 m on nearby coasts (Corona and Ram\u0026iacute;rez-Herrera, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), and the 1995 earthquake (M8.0) caused notable damage in Manzanillo Bay due to the tsunami currents (Borerro et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e1997\u003c/span\u003e). These historical events underscore the importance of conducting detailed tsunami hazard assessments in this area.\u003c/p\u003e\u003cp\u003eTraditionally, tsunami hazard assessments have been performed using deterministic or probabilistic approaches. The deterministic method, often referred to as the \"worst-case scenario\" approach, relies on the largest recorded events within a region to predict potential future impacts. However, this method may not account for the full range of possible seismic events, particularly in areas with limited historical data on tsunamis (Mori et al., 2018; Jel\u0026iacute;nek et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). On the other hand, the Probabilistic Tsunami Hazard Assessment (PTHA) approach builds on methodologies developed for seismic hazard analysis (PSHA) and provides a more comprehensive framework for understanding tsunami risk by incorporating a variety of possible earthquake scenarios, including those that might not have been historically observed (Cornell and Vanmarcke, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e1969\u003c/span\u003e; Geist and Parsons, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Mori et al., 2018; Miyashita et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThis study aims to perform a PTHA for Manzanillo, Colima, using stochastic earthquake source models and numerical simulations. The assessment was particularly focused on evaluating tsunami hazards for different return periods (e.g., 50, 100, and 500 years). The use of advanced numerical models, including GeoClaw (LeVeque et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), allows for a detailed simulation of tsunami wave propagation and coastal inundation, providing valuable insights into their potential impacts on Manzanillo. The results of this study are intended to aid local authorities in developing effective tsunami risk management and mitigation strategies. Given the region\u0026rsquo;s tectonic setting and historical tsunami events, it is crucial to understand the probabilistic nature of tsunami hazards. By considering a broad spectrum of possible seismic events (beyond those in the historical record), this study used PTHA to provide a more robust picture of the tsunami hazard. These findings can enhance broader disaster preparedness and coastal management initiatives, ensuring that Manzanillo and its surrounding areas are better prepared to confront future tsunamis.\u003c/p\u003e"},{"header":"2 Study Area","content":"\u003cp\u003eThe study area encompasses the coastal municipality of Manzanillo, which is located along the central Pacific coast of Mexico. Geographically, Manzanillo is positioned between 19.31\u0026deg;N\u0026ndash;19.95\u0026deg;N latitude and 104.68\u0026deg;W\u0026ndash;104.03\u0026deg;W longitude, covering a territorial area of approximately 1,578 km\u0026sup2;. The area is located within the tectonically active region influenced by the convergence of the Rivera, Cocos, and North American plates. This geographical location places Manzanillo at significant risk from seismic and tsunami events owing to its proximity to the Mesoamerican Trench, where these tectonic plates interact. The city lies within the Sierra Madre del Sur physiographic province and features diverse topography, including mountainous terrain, intermontane valleys, coastal plains, and lagoons. According to national geographical data (INEGI, 2010), the most prominent physiographic components are high complex sierras (approximately 60.5% of the area), intermontane valleys with hills (12.6%), coastal plains with lagoons (12.1%), and coastal plains with floodable deltas and salt flats (5.6%). Most of the remaining area consists of beaches, coastal bars, and minor valleys.\u003c/p\u003e\u003cp\u003eThe coastline of Manzanillo is characterized by a series of bays and lagoons that significantly influence local geomorphology and potential tsunami inundation patterns. The main bays from west to east include \u003cem\u003eCarrizales Bay\u003c/em\u003e, \u003cem\u003eHigueras (Cenicero) Bay\u003c/em\u003e, \u003cem\u003eSantiago Bay\u003c/em\u003e, and \u003cem\u003eManzanillo Bay\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). The latter two, Santiago and Manzanillo, are particularly important in terms of urban development and population exposure, making them critical areas for tsunami risk. The shores around these bays range from rocky headlands and cliffs to low-lying sandy beaches with elevations that vary and can influence the propagation and run-up of tsunami waves on land. For instance, steep rocky sections may reflect or refract waves, whereas flat coastal plains allow waves to penetrate further inland (Gonz\u0026aacute;lez et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Ar\u0026aacute;nguiz et al., 2019).\u003c/p\u003e\u003cp\u003eManzanillo\u0026rsquo;s hazard is further exacerbated by the presence of several rivers and lagoons, which add complexity to tsunami flooding scenarios. The \u003cem\u003eJuluapan\u003c/em\u003e, \u003cem\u003ePe\u0026ntilde;itas\u003c/em\u003e, and \u003cem\u003eCuyutl\u0026aacute;n\u003c/em\u003e lagoons are notable hydrological features connected to the ocean through narrow inlets or channels. These water bodies can act as conduits that channel tsunami waves inland, or amplify flooding in adjacent low-lying areas (Zamora et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In particular, the Cuyutl\u0026aacute;n Lagoon, a large lagoon on the eastern side of Manzanillo\u0026rsquo;s urban area (near \u003cem\u003eCampos Beach\u003c/em\u003e), has expansive low-elevation wetlands. During a tsunami, this lagoon and its surroundings could experience amplified inundation due to the funneling effect and low terrain, as water may pour into and spread through the lagoon system, an effect observed in similar coastal environments (Pari et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2008\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the location and main geographical features of the Manzanillo study area. The key features highlighted include the aforementioned bays (Carrizales, Cenicero, Santiago, and Manzanillo) as well as the lagoons and river outlets. These geographic features are central to understanding tsunami hazards as they can either mitigate or worsen the impact of incoming tsunami waves. For example, Santiago Bay is partially sheltered by headlands, whereas Manzanillo Bay opens directly toward the trench area, and the coastline at Campos (near Cuyutl\u0026aacute;n Lagoon) faces the open ocean, which is reflected in the hazard results.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"3 Data and Methods","content":"\u003cp\u003eThe methodology for this study involves a combination of data collection, seismic source modeling, and numerical simulation to assess tsunami hazards in the Manzanillo region. In summary, we first gathered the necessary geological and geophysical data (seismic source parameters, bathymetry, and topography) and then generated a suite of plausible earthquake scenarios using stochastic modeling approaches. These scenarios were the inputs for tsunami simulations using a physics-based model (GeoClaw) to propagate the resulting tsunami to the coast and compute inundation on land. Finally, we performed a probabilistic analysis of the results to estimate the likelihood of exceeding certain wave heights (or run-up levels) over given time periods. The methodology includes a validation step, where we compare simulation outputs with observations from a past tsunami event (the 1995 Colima-Jalisco tsunami) to ensure the reliability of the model.\u003c/p\u003e\n\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\n \u003ch2\u003e3.1 Data\u003c/h2\u003e\n \u003cp\u003eAccurate assessment of tsunami hazards requires various datasets, most importantly, seismic source data to characterize potential earthquakes and a seamless bathymetry-topography dataset to properly model the wave propagation across the ocean and to model onshore inundation. The datasets are briefly described below.\u003c/p\u003e\n \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e\n \u003ch2\u003e3.1.1 Seismic Source Models\u003c/h2\u003e\n \u003cp\u003eTo represent the seismic setting of the region, we utilized a detailed three-dimensional model of the subduction interface along the Mesoamerican Trench and a seismic wave velocity model. The subduction zone geometry was based on the Central American Slab2.0 model (Hayes et al., \u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e), which provides the depth and dip of the subducting plate. The geometry was discretized into a finite mesh down to a depth of approximately 30 km, as large thrust earthquakes in this region typically occur within this depth range. This mesh effectively outlines the locations where fault plane ruptures can occur. The Slab2.0 model integrates data from various sources (e.g., earthquake catalogs and seismic surveys) to ensure an accurate geometric representation of the plate interface. In addition, the seismic wave velocity structure from Mendoza and Hartzell (\u003cspan class=\"CitationRef\"\u003e1999\u003c/span\u003e), which was used to analyze the 1995 Jalisco-Colima earthquake, was implemented to generate the rupture dynamics scenarios presented herein. Our focus is on tsunami generation, which primarily depends on static seafloor deformation, and having a reasonable seismic velocity model aids in creating more physically plausible earthquake slip distributions for the scenarios.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e\n \u003ch2\u003e3.1.2 Bathymetry and Topography\u003c/h2\u003e\n \u003cp\u003eBathymetric and topographic data are critical for accurately modeling tsunami wave propagation in the ocean and inundation over land, respectively. For deep-ocean and regional propagation, we used the SRTM30\u0026thinsp;+\u0026thinsp;global bathymetry dataset (Becker et al., \u003cspan class=\"CitationRef\"\u003e2009\u003c/span\u003e), which has a resolution of 30 arc-seconds (~\u0026thinsp;900 m). Although relatively coarse, this resolution is sufficient for far-field wave propagation and the initial wavefront development. For nearshore and onshore areas, especially within the bays of Cenicero, Santiago, and Manzanillo and around the Cuyutl\u0026aacute;n Lagoon, we incorporated high-resolution bathymetric and topographic datasets. High-resolution bathymetric surveys (from nautical charts or local studies) were merged with topographic elevation data (LiDAR or photogrammetry, if available, or 10-m resolution DEMs) to create a detailed Digital Elevation Model (DEM) of the coastal zone. This high-resolution DEM (with grid spacing on the order of tens of meters) captures important small-scale features, such as beach berms and channel connections between the ocean and lagoons. These features strongly influence how a tsunami wave transitions from deep water to the shoreline, and how far inland water can penetrate. All spatial datasets were projected onto a common coordinate system and vertical datum, and we took care to stitch the global and local bathymetry seamlessly so that no artificial gaps or overlaps existed.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\n \u003ch2\u003e3.2 Tsunami Modeling and Hazard Analysis\u003c/h2\u003e\n \u003cp\u003eThe tsunami hazard assessment was carried out in two main phases: (1) generating tsunami source scenarios using stochastic seismic modeling and simulating each scenario\u0026rsquo;s tsunami propagation and inundation and (2) performing a probabilistic analysis of the ensemble of simulations to derive hazard curves and maps (i.e., computing the probability that certain tsunami intensity measures will be exceeded in given time frames). In addition, we validated our modeling approach against a known historical event to ensure the robustness of the proposed method.\u003c/p\u003e\n \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e\n \u003ch2\u003e3.2.1 Seismic Fault Models and Scenario Generation\u003c/h2\u003e\n \u003cp\u003eWe employed the \u0026ldquo;FakeQuakes\u0026rdquo; module of the open-source MudPy code (Melgar et al., \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e) to generate stochastic earthquake rupture scenarios on the Mesoamerican trench model. FakeQuakes uses the statistical relations of past earthquakes and geophysical constraints to create realistic slip distributions for hypothetical events. Each synthetic rupture is defined by parameters such as the magnitude, spatial slip distribution on the fault plane, rupture dimensions, depth, and hypocenter location. This approach allowed us to sample a wide range of possible earthquakes, including events larger than those in the historical catalog, with their associated occurrence frequencies, which are crucial for PTHA. Two sets of scenarios were produced.\u003c/p\u003e\u003cspan\u003e1. Model Validation Scenarios: First, we focused on the 1995 Colima-Jalisco earthquake (Mw 8.0) as a reference event for model validation, considering the availability of water level observations. We generated 600 synthetic rupture models constrained to the known characteristics of the 1995 event (based on the studies by Mendoza and Hartzell,\u0026nbsp;\u003cspan class=\"CitationRef\"\u003e1999\u003c/span\u003e). These scenarios varied in detail, including the exact slip pattern, hypocenter, and rupture extent. From all the results, only 28 scenarios met the criteria of having magnitudes in the range of ~\u0026thinsp;7.99\u0026ndash;8.01, matching the 1995 event magnitude, which were selected for validation.\u003cbr\u003e\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003e2. PTHA Scenarios: Next, probabilistic analysis was used to generate a vast range of plausible large earthquakes in the region. We defined magnitude bins from Mw 7.0 to Mw 8.5 with increments of \u0026Delta;M\u0026thinsp;=\u0026thinsp;0.1. For each magnitude bin, multiple random rupture realizations were generated, resulting in a total suite of 4,800 synthetic earthquakes (300 scenarios for each of the 16 magnitude bins between 7.0 and 8.5). These scenarios were broadly centered offshore near Manzanillo (within the segment of the trench facing Colima), but with random variations in the exact hypocenter and rupture area along the trench segment. We did not restrict the hypocenter locations within that segment, allowing ruptures to be initiated anywhere along the portion of the fault adjacent to Manzanillo. We set an upper bound on the slip per patch (60 m maximum) to reflect the physical limits generally observed in subduction earthquakes. Each scenario yielded a different pattern of seafloor deformation, producing a distribution of tsunamis.\u003c/p\u003e\n \u003c/span\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e\n \u003ch2\u003e3.2.2 Tsunami Simulations (GeoClaw Model)\u003c/h2\u003e\n \u003cp\u003eFor each generated earthquake scenario, we simulated the resulting tsunami using the GeoClaw numerical model (LeVeque et al., \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e). GeoClaw is a well-validated solver for nonlinear shallow water equations and is specifically designed to handle tsunami propagation and inundation with high efficiency through adaptive mesh refinement (AMR). The static seafloor deformation from each synthetic earthquake (essentially, the initial condition for tsunami generation) was computed using Okada\u0026rsquo;s elastic dislocation model (Okada, 1985, 1992). This model considers the slip on the fault plane and computes the corresponding vertical displacement of the ocean bottom. We applied Okada\u0026rsquo;s formulas to each subfault in the rupture model to construct the total deformation field. The resulting uplift and subsidence pattern (which can have peak displacements of the order of several meters for Mw\u0026thinsp;~\u0026thinsp;8 events) is then applied instantaneously to the ocean surface to initiate the tsunami in the model. Once the initial wave is generated, GeoClaw simulates its propagation across the ocean, interaction with bathymetric features, and eventually shoaling producing the inundation on land. The key parameters and configurations for the GeoClaw simulations were as follows:\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eAdaptive mesh refinement: We utilized five nested grid levels. The coarsest level (Level 1) covered the entire computational domain (encompassing a large portion of the Pacific, including the wave\u0026rsquo;s path to Manzanillo) with a grid resolution of 5 arcmin. Intermediate levels (Level 2 at 1 arcminute, Level 3 at 30 arcseconds) gradually increased the resolution as the waves neared the coast. The finest grids (Level 4 at 5 arc-seconds to approximately 150 m, and Level 5 at 1 arc-second to ~\u0026thinsp;30 m) were centered on the Manzanillo coastal region. This approach allowed us to capture broad-scale propagation efficiently while still resolving the complex coastline and bathymetry of Manzanillo in high detail for inundation calculations. Notably, Levels 4 and 5 cover the bays and lagoons of interest with ~\u0026thinsp;30 m resolution, which is already adequate to model water flow through narrow channels and around small islands or peninsulas.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eTime step and stability: A Courant\u0026ndash;Friedrichs\u0026ndash;Lewy (CFL) condition was used to ensure numerical stability. We set the CFL number target to 0.75 (with an allowable maximum of 1.0). This means that the time step automatically adjusts such that the tsunami wave does not travel more than 75% of the grid cells per step on the finest grids. This adaptive time-stepping is important given varying grid sizes.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eFriction and drag: To account for energy losses due to seafloor roughness and onshore terrain friction (vegetation, buildings, etc.), we applied a Manning\u0026rsquo;s roughness coefficient of 0.025. This value is commonly used for tsunami modeling over a mix of coastal environments (LeVeque et al., \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e) and provides a reasonable approximation of the drag on flowing water.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cp\u003eFor output and analysis, we recorded simulated water surface elevations at five virtual gauge stations (denoted P1 through P5) located along the Manzanillo coast, each positioned at a 10 m water depth just offshore (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). These synthetic tide gauge points (including one near the location of the actual Manzanillo tide gauge in the harbor) captured the incoming wave time-series for each scenario. We also saved full-time historical inundation results on the fine grids to determine the run-up distance, that is, how far inland water traveled, and the depths of the resulting flooding. For the validation scenarios (the 28 cases modeling the 1995 event), the time series at a virtual gauge corresponding to Manzanillo\u0026rsquo;s harbor (MTG in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e) was compared to the 1995 tsunami records (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). This allowed us to assess the accuracy of the model in terms of the timing, wave height, and overall wave train shape.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e\n \u003ch2\u003e3.2.3 Probabilistic Tsunami Hazard Assessment (PTHA)\u003c/h2\u003e\n \u003cp\u003eProbability exceedance curves were calculated by accounting for the probability that a tsunami wave surpasses a threshold \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\eta\\:}_{c}\\)\u003c/span\u003e\u003c/span\u003e in an exposition window \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{T}_{exp}\\)\u003c/span\u003e\u003c/span\u003e for all evaluated magnitudes, as shown in Eq. (1). These curves express the probability that a given tsunami wave height (or run-up height) will be equal to or exceeded at a site within a specified period (e.g., 10-, 50-, 100-year, etc.).\u003c/p\u003e\n \u003cdiv id=\"Equa\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e$$\\:P(\\eta\\:\u0026gt;{\\eta\\:}_{c})=1-{\\prod\\:}_{i=1}^{n}{[1-(1-e}^{-{\\nu\\:}_{i}{T}_{exp}}\\left)P\\right(\\eta\\:\u0026gt;{\\eta\\:}_{c}\\left|{M}_{i}\\right)]\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\left(1\\right)$$\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003eHere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{T}_{exp}\\)\u003c/span\u003e\u003c/span\u003e, is the exposition window and was established in 1 year; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:P(\\eta\\:\u0026gt;{\\eta\\:}_{c}){M}_{i})\\:\\)\u003c/span\u003e\u003c/span\u003eis the probability that a seismic event with magnitude \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{M}_{i}\\)\u003c/span\u003e\u003c/span\u003e generates tsunami amplitudes exceeding a certain level \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\eta\\:}_{c}\\)\u003c/span\u003e\u003c/span\u003e in the next year; \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\nu\\:}_{i}\\)\u003c/span\u003e\u003c/span\u003e is the truncated form of the annual seismic occurrence rate from the Gutenberg-Richter relationship (Cornell and Vanmarcke, \u003cspan class=\"CitationRef\"\u003e1969\u003c/span\u003e; Salazar-Monroy et al., \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e), as derived from Eq. (2). The hazard assessment in this work was performed for magnitudes ranging from Mw\u0026thinsp;=\u0026thinsp;7.0 to Mw\u0026thinsp;=\u0026thinsp;8.5 \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:(\\varDelta\\:M=0.1)\\)\u003c/span\u003e\u003c/span\u003e, resulting in the evaluation of 16 magnitude bins.\u003c/p\u003e\n \u003cdiv id=\"Equb\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e$$\\:{\\nu\\:}_{i}={\\nu\\:}_{0}\\frac{{e}^{-\\beta\\:({M}_{i}-{M}_{min})}-{e}^{-\\beta\\:({M}_{max}-{M}_{min})}}{{1-e}^{-\\beta\\:({M}_{max}-{M}_{min})}},{M}_{min}\u0026lt;{M}_{i}\u0026lt;{M}_{max}\\:\\:\\:\\:\\:\\:\\:\\:\\left(2\\right)\\:$$\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\nu\\:}_{0}={e}^{\\alpha\\:-\\beta\\:{M}_{min}}\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\alpha\\:=a{\\prime\\:}\\:ln\\left({T}_{r}\\right)\\)\u003c/span\u003e\u003c/span\u003e y \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:=b\\:ln\\left({T}_{r}\\right)\\)\u003c/span\u003e\u003c/span\u003e, in this case the return period, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{T}_{r}\\)\u003c/span\u003e\u003c/span\u003e, was fixed to 10 years. In our calculations, we used the regional parameters \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:a\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:b\\)\u003c/span\u003e\u003c/span\u003e from the Gutenberg-Richter law as proposed by Zu\u0026ntilde;iga et al. (2017), which is appropriate for the Colima region. These parameters were originally determined for a larger coastal segment than that of our study area. Thus, to better represent the events attributable to the Manzanillo segment and the size and recurrence of the evaluated rupture area, we applied a correction to the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:a\\)\u003c/span\u003e\u003c/span\u003e value. This 30% reduction in total seismicity was implemented by modifying the parameter as \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{a}^{{\\prime\\:}}=a+ln\\left(0.70\\right)\\)\u003c/span\u003e\u003c/span\u003e. Furthermore, the magnitude range was truncated at Mw 8.5, which is consistent with the historical and tectonic characteristics of this segment. This provides a conservative but reasonable upper bound for hazard analysis without underestimating the tail risk.\u003c/p\u003e\n \u003cp\u003eFor practical purposes, we focused on three possible return periods: 50, 100, and 500 years, which correspond to relatively frequent, moderately infrequent, and rare tsunamis, respectively. By evaluating exceedance probabilities for these horizons, we can infer, for example, what wave height has a 10% chance of being exceeded in 10 years (equivalent to a\u0026thinsp;~\u0026thinsp;100-year return period), and so on. The output of this step is a set of hazard curves for different coastal points (P1\u0026ndash;P5), indicating the probability that the tsunami height will exceed various levels within a 1-year exposure window for each return period of interest. These curves can be compared with historical data-derived curves (based on past events\u0026rsquo; run-up and tide gauge records) to check for consistency. To generate historical curves and compare them with the probability of exceedance curves from simulations, a Tapered Pareto distribution function was fitted to the historical data, and the hazard was then evaluated using the Poisson model.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e\n \u003ch2\u003e3.2.4 Inundation Maps\u003c/h2\u003e\n \u003cp\u003eTo translate the probabilistic results into actionable information, we developed tsunami inundation maps for the 50-year, 100-year, and 500-year return period scenarios. These maps depict inundation levels associated with fixed return periods (or fixed exceedance probabilities) and highlight the extent of flooding in the study area. It is worth noting that, strictly speaking, the inundation levels shown in, for example, the 50-year return period map does not take place at the same time. Despite this, these maps are useful for assessing the relative risk to different parts of the city and surrounding regions, helping identify areas of high vulnerability. We employed two complementary approaches to generate the inundation extents, as described below.\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eElevation-based mapping (GIS method): Using the hazard curves and recorded wave heights from the PTHA, we inferred the expected run-up or inundation elevation for each return period using the hazard curve data from synthetic tide gauge P3, which is located in the most populated area of Manzanillo. From the synthetic tide-gauges set up in GeoClaw at a 10m isobath to measure the wave amplitude, we applied Green\u0026rsquo;s law scaling (Eq.\u0026nbsp;3) to approximate how the wave height would amplify as it shoals to a shallower reference depth (e.g., 1 m depth near the shoreline).\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cdiv id=\"Equc\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e$$\\:{\\eta\\:}_{2}={\\eta\\:}_{1}\\sqrt[4]{\\frac{{h}_{1}}{{h}_{2}}},\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\left(3\\right)$$\u003c/div\u003e\n \u003c/div\u003e\n \u003cdiv class=\"BlockQuote\"\u003e\n \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\eta\\:}_{1}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\eta\\:}_{2}\\)\u003c/span\u003e\u003c/span\u003e are the wave amplitudes at two points perpendicular to the coastline, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{h}_{1}\\)\u003c/span\u003e\u003c/span\u003eand \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{h}_{2}\\)\u003c/span\u003e\u003c/span\u003e, are the corresponding water depths of those two points. Green\u0026rsquo;s law provides a first-order approximation to estimate the run-up or nearshore wave heights from the offshore values. We used this method to estimate the run-up heights corresponding to wave heights with a given exceedance probability. Although Green\u0026rsquo;s law does not capture complexities such as wave breaking, reflection, or bore formation, it is widely used in regional tsunami studies to obtain first-order estimates of run-up when detailed modeling is not available (Davies et al., \u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e; Mulia et al., \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e; Ha et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). We took the estimated inundation heights and delineated areas on a map that lie below these heights and are connected to the coast (to avoid counting inland depressions that would not actually flood because they are isolated by higher ground). This was performed in QGIS using the ~\u0026thinsp;30 m resolution DEM. Essentially, we \u0026ldquo;flooded\u0026rdquo; the DEM up to the critical elevation and identified all contiguous areas that made contact with the shoreline. We refined these areas using high-resolution coastline data and overlaid OpenStreetMap layers for context (roads and buildings) in the visualization. This produced a set of polygons indicating the potential flood zones for each scenario (50-, 100-, 500-year).\u003c/p\u003e\n \u003c/div\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eDynamic simulation mapping: To complement the method described above, we directly implemented GeoClaw to simulate specific scenario events that roughly correspond in size to the target return periods. For example, we identified one of the synthetic earthquake scenarios of approximately Mw 7.54, which produced an offshore wave height similar to that of the 50-year hazard level, for which we performed a full 3-hour tsunami inundation simulation for that event on a very fine grid extending up to ~\u0026thinsp;25 m elevation inland. Similarly, we simulated a Mw 7.74 event for the 100-year scenario and an Mw\u0026thinsp;~\u0026thinsp;8.49 event for the 500-year scenario. These simulations yielded high-resolution water depth and flow extent data, which were then mapped as maximum inundation footprints.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n \u003cp\u003eComparing the GIS-based inundation extents with the simulation-based extents allowed us to gauge the accuracy of the simpler method. In cases where the simulation showed water reaching further inland than in the static method, it indicates that dynamic effects (such as momentum carrying water over small ridges or through narrow passes) are important. By combining these methods, we ensured that the inundation maps were both data-driven (from hazard curves) and physically validated (through hydrodynamic simulations). All maps were referenced to the same vertical datum (mean sea level), and the influence of tides was not explicitly included (effectively assuming that a tsunami occurs at the mean sea level). In practice, if a tsunami arrives at high tide, which in this region reaches a maximum of 1.14 m (Aguilar-Ch\u0026aacute;vez et al., \u003cspan class=\"CitationRef\"\u003e2009\u003c/span\u003e), inundation could reach slightly further away. Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e summarizes the tsunami amplitudes at each virtual sensor derived from the hazard curves, along with the representative earthquake magnitudes and the resulting offshore peak wave elevations used in the inundation map scenarios for both evaluated methodologies.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eTable 1.\u0026nbsp;\u003c/strong\u003eMaximum tsunami amplitudes for the cases used to evaluate inundation in GeoClaw, compared with the values obtained from the PTHA analysis at the virtual tide gauge locations.\u003c/p\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"587\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003eReturn period (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"5\" valign=\"top\" style=\"width: 240px;\"\u003e\n \u003cp\u003eHazard curve\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 282px;\"\u003e\n \u003cp\u003eDatabase comparable scenario\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" valign=\"top\" style=\"width: 240px;\"\u003e\n \u003cp\u003e\u0026eta; (m)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 46px;\"\u003e\n \u003cp\u003eMw\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"5\" valign=\"top\" style=\"width: 236px;\"\u003e\n \u003cp\u003e\u0026eta; (m)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003eP1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003eP2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003eP3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003eP4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003eP5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003eP1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003eP2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003eP3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003eP4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003eP5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e1.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e1.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e1.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e7.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e0.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e1.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e1.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e0.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e1.14\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e1.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e2.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e2.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e1.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e1.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e7.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e1.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e2.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e2.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e1.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e1.91\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 65px;\"\u003e\n \u003cp\u003e500\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e3.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e4.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e4.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e3.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 49px;\"\u003e\n \u003cp\u003e3.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e8.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e3.41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e4.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e4.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 46px;\"\u003e\n \u003cp\u003e3.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e3.98\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"4 Results","content":"\u003cp\u003eThe results of the tsunami hazard analysis for Manzanillo provide crucial insights into the potential impacts of tsunamis on the region. In the following subsections we present the validation of our model against the 1995 tsunami data, the probabilistic exceedance curves derived from the suite of scenario simulations, and the inundation maps for different return periods, highlighting the spatial extent of flooding in each scenario and the implications for the region.\u003c/p\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e4.1 Validation\u003c/h2\u003e\u003cp\u003eTo verify the accuracy of the tsunami simulations, we compared the model output for the 1995 Colima-Jalisco tsunami with observational data. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows a comparison of the simulated and observed water-level time series at the Manzanillo tide gauge location (MTG). The model results, based on the 28 validation scenarios, are summarized by the mean water surface elevation and a shaded band indicating one standard deviation across those scenarios. The first wave arriving in both the observation and the model is a trough (a lowering of the water level) followed by a peak, consistent with regional subduction movement. The mean peak amplitude of the model is within the observed range, and subsequent oscillations (the trailing wave train) show a similar pattern of decay over time. This suggests that the stochastic slip distributions that we used captured the essential characteristics of the 1995 earthquake tsunami generation. One consistent discrepancy was noted, which is that the arrival time of the tsunami in the simulations tended to be slightly delayed by approximately 3\u0026ndash;10 min compared to the actual tide gauge record. Considering that the aim of this study was to determine the maximum amplitudes and inundation extents, the timing offset was not critical, as in the case of an early warning system. All 28 scenario simulations exhibited this delay to a certain extent. This timing offset could be due to slight inaccuracies in the assumed rupture initiation location or the velocity model (which affects the wave speed), or perhaps the bathymetric grid resolution introducing small phase errors. Despite this, the wave period and the overall duration of the event were well reproduced. Importantly, the range between the maximum and minimum simulated wave heights encompasses the observed waveform, indicating the observed event\u0026rsquo;s characteristics are not outliers relative to our model ensemble. These results provide a confident framework showing that our approach is reliable for predicting tsunami impacts in the region, at least for events of similar scale and mechanism to 1995. As such, the validation exercise confirmed that the chosen modeling framework (FakeQuakes\u0026thinsp;+\u0026thinsp;GeoClaw) is suitable for simulating local tsunamis in Manzanillo. The close match in waveform evolution lends credibility to inundation and hazard predictions from hypothetical events. Minor timing errors in the simulation are acknowledged, but these are not significant enough to influence hazard assessment outcomes, which concern maximum amplitudes and inundation extents. Nonetheless, this indicates a limitation in using this method for emergency response, where timing is essential for effective action.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e4.2 Probability of Exceedance Curves\u003c/h2\u003e\u003cp\u003eUsing the generated database of 4,800 synthetic tsunami simulations, exceedance probability curves were constructed for each measurement point (P1\u0026ndash;P5) along the coast. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents these curves, showing the estimated relationship between the tsunami amplitude and return period at each point. It is worth mentioning that the probability of exceedance curves was calculated without considering epistemic uncertainties in the input data or model used, which may lead to hazard underestimation. For clarity, we also compared these with analogous curves derived from historical tsunami data for the Manzanillo region, following an approach similar to that of Geist and Parsons (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). The following patterns were observed in the PTHA curves:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eFor a 100-year return period (which corresponds to a 40% exceedance probability in 50 years or ~\u0026thinsp;99% in 500 years), the highest tsunami amplitudes among the five sites were found at P3, which is situated near the city center (Manzanillo Bay). This suggests that, for relatively moderate events, the central bay experiences the largest waves. However, when we consider longer return periods (more extreme events), the hazard appears to shift spatially.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eFor return periods of over 800 years, site P2 (representing the Santiago Bay area) began to show a higher hazard than P3. The same shift was observed at points P1 and P4. This indicates that extremely large-magnitude events, although rare, are more likely to severely impact Santiago Bay. In addition, P5 shows a slight increase in hazard by a 1,000-year return period, consistent with the idea that open-coast facing segments (such as Campos) are prone to the impact of large tsunamis.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eWhen comparing our PTHA curves to those derived from historical data, we observed that our runup-estimated curves have a good fit for small return periods, but this approximation tends to overestimate tsunami heights for long return periods, especially at higher exceedance levels. For instance, for an amplitude of up to 1 m, our analysis shows good agreement with the historical record; however, at higher exceedance amplitudes, our analysis might assign a shorter return period than the historical record. Part of this discrepancy arises from limited historical data, as the historical catalog for Manzanillo is sparse and might underestimate the tail risk, whereas our synthetic approach, by design, explores more extremes.\u003c/p\u003e\u003cp\u003eAnother interesting observation is the difference between the use of offshore wave data and run-up data for constructing hazard curves. Using the full historical dataset, the hazard curve generated from the run-up estimations (via Green\u0026rsquo;s law) aligns well with the historical data for smaller tsunami amplitudes. However, at large tsunami amplitudes, we noticed that runup estimations over-predicted the hazard, whereas offshore data curves showed a better correspondence with the historical data. This aligns with known issues where Green\u0026rsquo;s law can overestimate inland amplification because it does not consider energy losses and complex ground interactions, which is especially important for very large events where inland penetration of water may be of greater importance.\u003c/p\u003e\u003cp\u003eFrom a practical perspective, these exceedance curves provide a complete distribution of tsunami hazards as a function of the wave height estimation. Based on the results, the probability that a tsunami exceeds a given height within a specific exposure window can be estimated using a generalized Poisson model: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:P(\\eta\\:\u0026gt;{\\eta\\:}_{c}\\:in\\:{T}_{exp}\\:years)={(1-e}^{-{T}_{exp}/{T}_{r}\\left({\\eta\\:}_{c}\\right)}),\\)\u003c/span\u003e\u003c/span\u003e where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{T}_{r}\\left({\\eta\\:}_{c}\\right)\\)\u003c/span\u003e\u003c/span\u003e is the return period associated with a tsunami height \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\eta\\:}_{c}.\\)\u003c/span\u003e\u003c/span\u003e This formulation enables multiple applications crucial for engineering design and planning. Nonetheless, uncertainties should be addressed, as the curves have inherent variability due to assumptions in the seismic rates (Gutenberg\u0026ndash;Richter parameters), chosen maximum magnitude, restriction of the rupture area, and number of scenarios that were sampled. It is noteworthy that the results are consistent with evidence from historical run-ups beyond 10 m reported for the 1932 event (Corona and Ram\u0026iacute;rez-Herrera, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e4.3 Inundation Maps\u003c/h2\u003e\u003cp\u003eBased on the methods described in Section \u003cspan refid=\"Sec11\" class=\"InternalRef\"\u003e3.2.4\u003c/span\u003e, we generated inundation maps for the 50-year, 100-year, and 500-year return-period scenarios. For each scenario, we present two maps, one based on the simplified Green\u0026rsquo;s law-based inundation elevation method using QGIS processing tools (labeled as \u0026ldquo;GIS method\u0026rdquo;) and the other set based on the explicit GeoClaw simulation of a representative event (labeled as \u0026ldquo;simulation results\u0026rdquo;). These are shown in Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, where the simulation result maps also show the amplitudes recorded at the measurement points and the runup values approximated using Green\u0026rsquo;s law as a reference. The maps consistently show that low-lying areas near the coast are the most vulnerable to tsunami flooding in Manzanillo. In particular, the flat regions fringing Santiago Bay, the downtown and port area around Manzanillo Bay, and the extensive wetlands around Cuyutl\u0026aacute;n Lagoon emerge as hotspots of inundation. As the return period (and associated tsunami size) increases from 50 to 500 years, the extent of flooding increases, covering larger areas and reaching further inland.\u003c/p\u003e\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\u003ch2\u003e4.3.1 50-Year Return Period\u003c/h2\u003e\u003cp\u003eFor a relatively frequent tsunami scenario (on the order of a 50-year return period, which might correspond to a local offshore earthquake of roughly low Mw\u0026thinsp;~\u0026thinsp;7.3 based on the Gutenberg-Richter law), inundation is limited but still noteworthy in certain spots (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003cem\u003ea\u003c/em\u003e). Based on offshore hazard values and run-up estimates, flood extents indicate that water primarily affects zones that are immediately adjacent to the shoreline. In Santiago Bay, water intrudes near Juluapan Lagoon, reaching\u0026thinsp;~\u0026thinsp;1.6 km inland along the lagoon\u0026rsquo;s low-lying channel. In Manzanillo Bay, the most extensive inundation occurs around the Las Garzas/Valle de las Garzas Lagoon, where the inundation is more associated with water intrusion from the port. In this area, water extensions reach roughly 1.8 km inland along the lagoon systems. This area includes urban neighborhoods and the backwater zones of the port. The San Pedrito and Las Garzas lagoons (in central Manzanillo) act as pathways for water, meaning that the surrounding neighborhoods could experience flooding because of the concentration of water in those basins, even if the open coast shows smaller waves. Notably, even in this modest scenario, the Cuyutl\u0026aacute;n Lagoon area on the eastern side experiences the farthest inland flooding, reaching\u0026thinsp;~\u0026thinsp;5.6 km from the shoreline This is largely because the terrain there is very flat and low; once water overtops the beach or enters the lagoon inlet, it can propagate far across the marshy land with little resistance.\u003c/p\u003e\u003cp\u003eComparison with a simulation of a specific Mw 7.54 earthquake scenario (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e.\u003cem\u003eb\u003c/em\u003e) meant to represent a 50-year event, shows a broadly similar pattern at the coastline, with influence restricted to the first\u0026thinsp;~\u0026thinsp;100 m of the shoreline but with markedly less inundation through the lagoon channels. In this simulation, the maximum inundated area occurs at Manzanillo port, with flooding extending\u0026thinsp;~\u0026thinsp;400 m inland and maximum inundation depths reaching\u0026thinsp;~\u0026thinsp;4 m inside the port. In Santiago Bay, floodwaters reach\u0026thinsp;~\u0026thinsp;300 m inland, although with maximum depths remaining below 1 m at Cuyutl\u0026aacute;n Lagoon; simulation results show\u0026thinsp;~\u0026thinsp;200 m of water intrusion through the lagoon inlet. While the affected zone reaches up to ~\u0026thinsp;4 km from the coast, the water depths are shallow (a few centimeters), likely due to frictional and topographic effects. Overall, the simulated inundation footprints were less extensive than the GIS-based footprints, implying that the straightforward elevation method may overestimate the inundation for smaller events. Despite these differences, both methods consistently flag the same zones (Juluapan, downtown/port, and Cuyutl\u0026aacute;n) as areas of concern, even for relatively frequent tsunamis. These results suggest that even a tsunami expected in a 50-year frame could cause flooding a few kilometers inland in the worst-hit parts of Manzanillo, which is significant for local planning.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section3\"\u003e\u003ch2\u003e4.3.2 100-Year Return Period\u003c/h2\u003e\u003cp\u003eA less frequent event resulting in inundation maps is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. This set corresponds to a larger event with a 100-year return period, roughly similar to a Mw 7.74 earthquake impacting the area based on the Gutenberg-Richter law. The inundation areas are larger than those in the 50-year scenario. Consistently, the same neighborhoods are affected as in the 50-year scenario, but flooding reaches further inland and covers wider zones. For example, in Manzanillo Bay, using run-up estimation from the GIS evaluation, flooding extends approximately 1.8 km inland. This would likely inundate much of the downtown waterfront district and the tourist areas near the shore. The low-lying part of the city (including port facilities and main roads along the coast) could be under a few meters of water. The higher spots or cliffs (such as the Peninsula de Santiago, dividing Santiago and Manzanillo Bays) remain safe regions, underscoring the importance of small elevation differences.\u003c/p\u003e\u003cp\u003eIn Santiago Bay, the water can reach up to ~\u0026thinsp;1.7 km inland near the Juluapan Lagoon, a slight increase from the 50-year case, but significantly inundating that lagoon\u0026rsquo;s surroundings and possibly affecting communities along its edges. The Cuyutl\u0026aacute;n Lagoon shows an increase to ~\u0026thinsp;5.8 km of inland flooding in the run-up-based map, indicating significant impacts on its low-lying areas. One area that stands out for the 100-year event is the downtown Manzanillo port area. Given its economic importance, the map indicates that significant portions of port infrastructure are within the 100-year inundation zone. Inundation depths in the port/downtown area could be significant in worst-case scenarios, where tsunami waves are funneled between structures or along channels.\u003c/p\u003e\u003cp\u003eThe dynamic simulation for a 100-year scenario, particularly the Mw 7.74 scenario, is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003cem\u003eb.\u003c/em\u003e This approach highlights remarkable hotspots with elevated flood depths, particularly over 6 m, within the port. In Santiago Bay, the maximum inundation depth reaches\u0026thinsp;~\u0026thinsp;2.9 m in the eastern sector. The greatest inland reach in this bay occurs at Juluapan Lagoon, with tsunami effects extending up to ~\u0026thinsp;1.5 km, although with water depths of just a few centimeters. In Manzanillo Bay, sea intrusion reaches\u0026thinsp;~\u0026thinsp;1.5 km, slightly less than the 1.8 km shown by the static method. For Cuyutl\u0026aacute;n, the simulation showed a\u0026thinsp;~\u0026thinsp;4.3 km inland reach versus ~\u0026thinsp;5.8 km in static. Again, the simulation suggests slightly less penetration, possibly owing to energy dissipation and the finite duration of the wave. Despite differences in exact numbers, both approaches clearly indicate that a 100-year tsunami would inundate large parts of the coastal plains and devastate unprepared infrastructure. The results emphasize that the critical infrastructure in Manzanillo, including the port, power facilities, and tourism centers, largely lies within the 100-year tsunami flood zone, underlining an urgent need for mitigation.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section3\"\u003e\u003ch2\u003e4.3.3 500-Year Return Period\u003c/h2\u003e\u003cp\u003eFor the low-probability 500-year scenario, which could correspond to an event approaching the upper magnitude considered (Mw\u0026thinsp;~\u0026thinsp;8.3 as estimated from the Gutenberg-Richter law), the GIS-based inundation map (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. \u003cem\u003ea\u003c/em\u003e) shows extensive flooding at Manzanillo Bay, covering nearly all urban and port areas. The inundation reaches up to ~\u0026thinsp;2.6 km inland, sparing only the steepest zones near the bay edges. Around Juluapan Lagoon, flooding may extend\u0026thinsp;~\u0026thinsp;2.5 km inland, affecting the surrounding roads and residential areas. Eastern Santiago Bay experiences significant inundation through the Pe\u0026ntilde;itas Lagoon and Santiago River channel. In Cuyutl\u0026aacute;n the inundation reached up to ~\u0026thinsp;6.2 km, suggesting the entire lagoon would become part of the ocean temporarily, and water would penetrate beyond its normal boundaries into adjacent communities. Campos, the community near Cuyutl\u0026aacute;n Lagoon\u0026rsquo;s mouth, and the coastline immediately south of it, facing the open ocean, would likely be ground zero under these extreme events.\u003c/p\u003e\u003cp\u003eThe simulation for a Mw\u0026thinsp;~\u0026thinsp;8.49 event (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e.\u003cem\u003eb\u003c/em\u003e) indicates extreme water depths in certain bays, such as ~\u0026thinsp;6.2 m in Cenicero Bay. While this zone did not flood far inland owing to steep terrain, the high values likely reflect local topographic amplification. Similarly, the eastern side of Santiago Bay shows maximum depths of ~\u0026thinsp;6.4 m. Inland reach in the simulation is ~\u0026thinsp;1.6 km at Juluapan Lagoon and ~\u0026thinsp;4.4 km at Cuyutl\u0026aacute;n, comparable to the 100-year simulation, which suggest that after a certain threshold, excess water tends to rise vertically rather than advance further inland. In Manzanillo Bay, a larger difference is observed, with sea intrusion reaching\u0026thinsp;~\u0026thinsp;1.9 km from the shoreline. The port area records extreme values of ~\u0026thinsp;7.8 m overtopping the breakwaters and ~\u0026thinsp;6.9 m of inundation inside the port, which could cause severe disruption to operations and infrastructure. One notable detail in this extreme scenario is the extent of the water entering Cuyutl\u0026aacute;n Lagoon, indicating severe impacts on any settlement along its shores and some industrial facilities settled in the area. The maps indicate potential inundation there, but it is worth highlighting qualitatively that a 500-year tsunami could plausibly send a\u0026thinsp;~\u0026thinsp;10 m high bore over Campos and into the lagoon behind it.\u003c/p\u003e\u003cp\u003eConsidering the above results, the 500-year inundation scenario delineates the outer envelope of the tsunami flooding, and one might reasonably prepare for the given current knowledge. Key infrastructure that remains unaffected even in the 500-year maps tends to be on high ground, such as some parts of the port, which might be built up or protected by seawalls to higher levels, although this was not explicitly modeled here.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"5 Discussion","content":"\u003cp\u003eThe results provide a detailed picture of tsunami hazards in Manzanillo, highlighting both the spatial variability in risk and the importance of a probabilistic approach. The analysis identified hotspots of high-risk areas directly linked to local geomorphology, as described below.\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003eManzanillo Bay (urban center) has emerged as a hotspot for tsunamis generated by moderately large earthquakes (Mw\u0026thinsp;\u0026lt;\u0026thinsp;8.0), with return periods under a few hundred years. The bay\u0026rsquo;s shape and orientation likely contribute to this, as a semi-enclosed bay resonance effect can amplify tsunami wave heights at particular frequencies (Bellotti et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The infrastructure and population density here mean that even moderate flooding can have outsized consequences, especially in the context of an increasing population and development.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eSantiago Bay, especially near the Juluapan Lagoon, becomes comparably or more hazardous as we consider rarer, larger tsunamis (~\u0026thinsp;800\u0026ndash;1000 year events). This might be because extremely large waves from the southwest can wrap around into this bay effectively or even enter through the gap toward the lagoon.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eThe low-land areas at the Cuyutl\u0026aacute;n Lagoon and the coastline near the entrance of the lagoon are notably exposed. For the largest events we considered (~\u0026thinsp;1,000-year return periods), this open-coast site experiences an increase of the run-ups when compared to the amplitudes recorded at P1 and P4. This is consistent with historical accounts, such as the catastrophic 1932 tsunami that produced a\u0026thinsp;~\u0026thinsp;12 m run-up along the Cuyutl\u0026aacute;n shores (Corona and Ram\u0026iacute;rez-Herrera, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The lack of significant headlands to dissipate energy and the long fetch over the lagoon allow waves to maintain their height.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003eAreas with natural barriers, such as cliffs and rocky outcrops (e.g., parts of the coast near points P1 and P4 in our setup), show reduced inundation in all scenarios. These features reflect or channel water, preventing it from flowing directly inland. However, it\u0026rsquo;s important to note that while such features protect against smaller tsunamis, they do not eliminate risk in extreme events (De Risi and Goda, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). It is possible that a sufficiently large tsunami may overtop cliffs or wrap around them.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eThe incorporation of a reduced seismicity model allows for a more representative hazard estimation of the regional tectonic setting and is better suited for realistic scenarios. It is worth mentioning that our scenario suite was based on ruptures directly offshore Manzanillo. Earthquakes further up or down the trench could present a certain rupture directivity, affecting the spatial distribution of ground motion amplitudes and the resulting coseismic seafloor deformation. This variation generates different tsunami wave patterns and hence leads to variations in coastal inundation and damage distribution (for example, Singh et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Iglesias et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Nevertheless, the spatial pattern we found aligns with what was intuitively expected and what previous regional models have suggested: the central and eastern parts of Manzanillo\u0026rsquo;s coast (harbor and Cuyutl\u0026aacute;n side) are more vulnerable than the western part (around Carrizales/Cenicero bays) (Salazar-Monroy et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Evangelista et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). From a planning perspective, tsunami mitigation efforts should prioritize high-risk zones, such as reinforcing sea walls and natural buffers in port areas and maintaining mangrove belts around lagoons to dampen waves. Local amplification phenomena, such as harbor oscillations (seiche), could prolong hazards for many hours after the initial tsunami, and our 3-hour simulations in GeoClaw indicate that oscillations persist, which is in line with other port cities\u0026rsquo; experiences.\u003c/p\u003e\u003cp\u003eThe results from synthetic tsunami events were used to generate hazard curves, which were compared with historical run-up data and tide gauge measurements. Hazard estimations derived from Green\u0026rsquo;s Law tend to overestimate the runup at high return periods, but when compared to only historical run-up data, the results align better for large amplitudes. However, the hazard curve based on run-up data may be unreliable owing to the limited availability of measurements. Generally, there is a scarcity of homogeneous historical data, leading to inconsistencies in PTHA evaluations, and relying only on historical data may lead to an underestimation of the hazard level of a region. An example is the Tohoku-Oki tsunami event in 2011 (Mw 9.0) (Geist and Parsons, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Mori et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). In this study, tsunami inundation levels near the coast combined with amplification factors were used, and inundation simulations were performed for comparison. While widely used, Green\u0026rsquo;s Law oversimplifies inland attenuation; therefore, Glimsdal et al. (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) suggested replacing it with more accurate amplification models. This limitation may lead to an overestimation of inundation when using elevation-based assessments without accounting for terrain friction, which differs from flooding extensions when compared to simulations, aligning with findings from similar studies (e.g., Tonini et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Nevertheless, inundation simulations may exhibit larger values in some regions compared with flood level-based assessments, suggesting that local geomorphological interactions amplify tsunami waves. These findings underscore the importance of using detailed simulation models for regional tsunami risk assessment because simplified methods cannot adequately capture these amplification effects.\u003c/p\u003e\u003cp\u003eExceedance probability curves were constructed using a 1-year time window to facilitate the interpretation of hazard levels over different planning horizons (e.g., 50-, 100-, and 500-year return periods). Discrepancies with previous studies (e.g., Salazar-Monroy et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) arise from differing data resolutions, observation windows, and sampling methods. Higher maximum amplitudes for 1,000-year return periods were found, likely because of finer bathymetry resolution and a fixed number of simulated events (300 scenarios for each magnitude), especially for the largest magnitudes. However, expanding large-magnitude event scenarios should be considered, as historical data confirm run-up values exceeding 10 m along this region. The probabilistic approach in this assessment provides a more comprehensive understanding of tsunami hazards by considering a range of possible earthquake scenarios rather than relying solely on historical events, which are critical for regions such as Manzanillo, where historical records may not fully capture the complete potential of hazards.\u003c/p\u003e\u003cp\u003eFlood maps generated using inundation levels in QGIS identified high-risk areas that aligned with previous studies by Evangelista et al. (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) and the Atlas Nacional de Riesgos (ANR, 2024) for Manzanillo, Colima, although a smaller range of risk areas was found. In the ANR evaluation, regions up to 20\u0026ndash;30 m elevation were classified as medium- and low-risk, while this study found a maximum inundation height of approximately 10 m for a 500-year return period. In contrast, differences with Evangelista et al. (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) stem from the use of higher-magnitude seismic sources (up to 9.5 Mw) and different topographic data.\u003c/p\u003e\u003cp\u003eThe results indicate that some regions of Manzanillo are at higher risk of inundation, particularly during large, infrequent tsunami events. These areas, such as the central bay and the vicinity of Cuyutl\u0026aacute;n Lagoon, require targeted risk-management strategies, including infrastructure reinforcement, evacuation planning, and public education programs. The information provided by inundation maps can be used by local authorities to prioritize these efforts and ensure that the most vulnerable populations are protected. One of the key strengths of this study is the use of high-resolution topographic and bathymetric data, which allow for detailed simulations of tsunami impacts. This level of detail is essential for accurately predicting the extent of inundation and identifying the specific areas of concern. However, it is important to note that the accuracy of the predictions depends on the quality of the input data and assumptions made in the modeling process.\u003c/p\u003e"},{"header":"6 Conclusions","content":"\u003cp\u003eThe probabilistic tsunami hazard assessment (PTHA) conducted for Manzanillo, Colima, demonstrated that the city is at significant risk from tsunamis, especially those originating near the Mesoamerican Trench. This study incorporated a probabilistic approach, considering a wide range of possible earthquake scenarios rather than relying solely on historical events, leading to a more comprehensive assessment of tsunami hazards. By combining stochastic seismic modeling, hazard exceedance curves, and high-resolution tsunami simulations using GeoClaw, the present study provides essential insights for disaster preparedness, coastal management, and infrastructure planning.\u003c/p\u003e\u003cp\u003eThe hazard exceedance curves reveal that tsunamis with wave heights of up to 1 m have a more than 80% probability of occurring within the next 50 years. This quantification underscores the need for proactive risk reduction measures, particularly in low-lying coastal areas such as Santiago Bay, Manzanillo Bay (downtown and port zones), and Cuyutl\u0026aacute;n Lagoon, which are the most vulnerable during inundation. Even a moderate tsunami (with a 50-year return period) could cause flooding of up to a few kilometers inland, whereas a 500-year return period event could inundate areas up to 5\u0026ndash;6 km inland through Manzanillo\u0026rsquo;s water bodies. These findings highlight specific neighborhoods and zones that require prioritized attention for tsunami risk reduction, underscoring the need for targeted disaster preparedness efforts, including infrastructure reinforcement, public education, and the development of comprehensive evacuation plans.\u003c/p\u003e\u003cp\u003eModel validation against the 1995 tsunami event confirms that the approach used is robust, with simulated wave heights and inundation extents closely matching historical observations. However, minor timing discrepancies suggest that future studies could refine bathymetric resolution and rupture parameterization to improve model accuracy, particularly for emergency responses. The hazard curves further indicate that certain regions, such as Campos and the Cuyutl\u0026aacute;n Lagoon area, exhibit greater tsunami risk for longer return periods due to their open-coast exposure and lack of natural barriers.\u003c/p\u003e\u003cp\u003eThe information provided by this study can directly inform local authorities and stakeholders in developing and refining risk mitigation strategies. For instance, inundation maps can be used to update tsunami evacuation routes and signage, ensuring that they encompass areas at risk for even larger events. Urban development plans can incorporate these hazard zones by avoiding critical new infrastructure in areas shown to flood during 100- or 500-year events. Existing critical facilities within hazard zones should have contingency plans or protective measures such as seawalls and elevated structures. Probability curves and worst-case scenarios can also be used to design tsunami drills and public education campaigns to enhance community preparedness.\u003c/p\u003e\u003cp\u003eFuture work should continue to refine these models by incorporating new data and improving the accuracy of the predictions, as well as considering the epistemic uncertainties involved. Additionally, integrating tsunami hazard assessments with other coastal risk factors, such as sea-level rise and storm surges, will be crucial for creating a holistic approach to coastal management in Manzanillo. Enhancing early warning systems, particularly by integrating seismic and sea-level monitoring into automated alert systems, will be essential for ensuring rapid response capabilities. Developing evacuation infrastructure, such as vertical evacuation options in the most at-risk zones and conducting regular tsunami drills, will further improve community resilience. Moreover, land use planning must consider tsunami hazard zones when approving new developments, and protective engineering measures such as tsunami barriers and reinforced structures should be implemented for critical infrastructure in vulnerable areas. By implementing these measures, Manzanillo can significantly reduce the potential loss of life and property damage from future tsunamis. The findings of this study serve as a scientific foundation for action, and stakeholders should use this knowledge proactively. The methodology and lessons from this study can also be extended to other communities in tsunami-prone areas that face similar risks.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported by the Secretar\u0026iacute;a de Ciencia, Humanidades, Tecnolog\u0026iacute;a e Innovaci\u0026oacute;n (formerly CONAHCYT), through a graduate studies scholarship to LVC. CMA was supported by UNAM-DGAPA-PASPA and Aarhus University Research Foundation fellowships. The authors acknowledge IT support from Gonzalo Mart\u0026iacute;n Ruiz and the Gerencia de Ingenier\u0026iacute;a from Comisi\u0026oacute;n Federal de Electricidad (CFE) for providing the detailed bathymetry data used in this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Secretar\u0026iacute;a de Ciencia, Humanidades, Tecnolog\u0026iacute;a e Innovaci\u0026oacute;n (formerly CONAHCYT), through a scholarship for LVC. CMA was supported by UNAM-DGAPA-PASPA and Aarhus University Research Foundation fellowships.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStudy conception and design were performed by Lizzeth V\u0026aacute;zquez Camaal, Christian M. Appendini and Ericka Alinne Solano Hern\u0026aacute;ndez. Material preparation and data collection were performed by Lizzeth V\u0026aacute;zquez Camaal. Analysis was performed by all authors. The first draft of the manuscript was written by Lizzeth V\u0026aacute;zquez Camaal and Christian M. Appendini and all authors commented on previous versions of the manuscript. All authors have read and approved the final manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used in this study will be made available at Zenodo upon acceptance.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAguilar-Ch\u0026aacute;vez, A., Salas-de-Le\u0026oacute;n, D. A., Monreal-G\u0026oacute;mez, M. A., Aldeco-Ram\u0026iacute;rez, J., \u0026amp; Signoret, M. (2009). Circulation and numerical modeling of the Manzanillo harbor, Colima, Mexico. WEH International Workshop on Environmental Hydraulics: Theoretical, Experimental and Computational Solutions, Valencia, Spain.\u003c/li\u003e\n\u003cli\u003eAranguiz, R., Catal\u0026aacute;n, P. A., Cecioni, C., Bellotti, G., Henriquez, P., \u0026amp; Gonz\u0026aacute;lez, J. (2019). Tsunami resonance and spatial pattern of natural oscillation modes with multiple resonators. \u003cem\u003eJournal of Geophysical Research: Oceans\u003c/em\u003e, \u003cem\u003e124\u003c/em\u003e(11), 7797-7816. https://doi.org/10.1029/2019JC015206\u003c/li\u003e\n\u003cli\u003eAtlas Nacional de Riesgos (ANR). (2024). Atlas de riesgos de Manzanillo 2024. Retrieved from http://www.atlasnacionalderiesgos.gob.mx/\u003c/li\u003e\n\u003cli\u003eBandy, W. L., Mortera-Guti\u0026eacute;rrez, C. A., Urrutia-Fucugauchi, J., \u0026amp; Hilde, T. W. C. (2000). The subducted Rivera-Cocos plate boundary: Where is it, what is it, and what is its relationship to the Colima Rift? \u003cem\u003eGeophysical Research Letters\u003c/em\u003e, 27(17), 3397-3400. https://doi.org/10.1029/1999GL011247\u003c/li\u003e\n\u003cli\u003eBecker, J. J., Sandwell, D. T., Smith, W. H. F., Braud, J., Binder, B., Depner, J., Fabre, D., Factor, J., Ingalls, S., Kim, S.-H., Ladner, R., Marks, K., Nelson, S., Pharaoh, A., Trimmer, R., Von Rosenberg, J., Wallace, G., \u0026amp; Weatherall, P. (2009). Global bathymetry and elevation data at 30 arc seconds resolution: SRTM30_PLUS. \u003cem\u003eMarine Geodesy\u003c/em\u003e, 32(4), 355-371. https://doi.org/10.1080/01490410903297766\u003c/li\u003e\n\u003cli\u003eBellotti, G., Briganti, R., \u0026amp; Beltrami, G. M. (2012). The combined role of bay and shelf modes in tsunami amplification along the coast. \u003cem\u003eJournal of Geophysical Research: Oceans\u003c/em\u003e, \u003cem\u003e117\u003c/em\u003e(C8). https://doi.org/10.1029/2012JC008061\u003c/li\u003e\n\u003cli\u003eBorerro, J., Ortiz, M., Titov, V., \u0026amp; Synolakis, C. (1997). Field survey of Mexican tsunami produces new data, unusual photos. \u003cem\u003eEos, Transactions American Geophysical Union\u003c/em\u003e, \u003cem\u003e78\u003c/em\u003e(8), 85-88. https://doi.org/10.1029/97EO00054\u003c/li\u003e\n\u003cli\u003eCastillo‐Aja, R., \u0026amp; Ram\u0026iacute;rez‐Herrera, M. T. (2017). Updated tsunami catalog for the Jalisco‐Colima coast, Mexico, using data from historical archives. \u003cem\u003eSeismological Research Letters\u003c/em\u003e, \u003cem\u003e88\u003c/em\u003e(1), 144-158. https://doi.org/10.1785/0220160133\u003c/li\u003e\n\u003cli\u003eCornell, C. A., \u0026amp; Vanmarcke, E. H. (1969). The major influences on seismic risk. In \u003cem\u003eProceedings of the fourth world conference on earthquake engineering\u003c/em\u003e, vol. 1, pp. 69-83.\u003c/li\u003e\n\u003cli\u003eCorona, N., \u0026amp; Ram\u0026iacute;rez-Herrera, M. T. (2012). Tsunami hazard and vulnerability in the Mexican Pacific coast: A review. \u003cem\u003eNatural Hazards\u003c/em\u003e, 61(2), 1213-1224. https://doi.org/10.1007/s11069-011-0076-0\u003c/li\u003e\n\u003cli\u003eDavies, G., Griffin, J., L\u0026oslash;vholt, F., Glimsdal, S., Harbitz, C., Thio, H. K., \u0026amp; Baptista, M. A. (2018). A global probabilistic tsunami hazard assessment from earthquake sources. \u003cem\u003eGeological Society, London, Special Publications, 456\u003c/em\u003e, 219\u0026ndash;244. https://doi.org/10.1144/SP456.5\u003c/li\u003e\n\u003cli\u003eDe Risi, R., \u0026amp; Goda, K. (2017). Simulation-based probabilistic tsunami hazard analysis: empirical and robust hazard predictions. \u003cem\u003ePure and Applied Geophysics\u003c/em\u003e, \u003cem\u003e174\u003c/em\u003e, 3083-3106. https://doi.org/10.1007/s00024-017-1588-9\u003c/li\u003e\n\u003cli\u003eEvangelista, Y., Castrej\u0026oacute;n, C., Villa, D., Trujillo, K., Monsivais-Huertero, A., \u0026amp; Mendozc, A. (2018, July). Generation of a tsunami hazard map for the coast of Manzanillo, Colima in Mexico using numerical simulation and wave modeling. In IGARSS 2018-2018 IEEE International Geoscience and Remote Sensing Symposium (pp. 7843-7846). IEEE. https://doi.org/10.1109/IGARSS.2018.8517856\u003c/li\u003e\n\u003cli\u003eGeist, E. L., \u0026amp; Oglesby, D. D. (2014). Tsunamis: Stochastic models of occurrence and generation mechanisms. \u003cem\u003eJournal of Geophysical Research: Solid Earth\u003c/em\u003e, 119(9), 7223-7242. https://doi.org/10.1002/2014JB011418\u003c/li\u003e\n\u003cli\u003eGeist, E. L., \u0026amp; Parsons, T. (2006). Probabilistic analysis of tsunami hazards. \u003cem\u003eNatural Hazards\u003c/em\u003e, 37(3), 277-314. https://doi.org/10.1007/s11069-005-4646-z\u003c/li\u003e\n\u003cli\u003eGlimsdal, S., L\u0026oslash;vholt, F., Harbitz, C. B., Romano, F., Lorito, S., Orefice, S., Brizuela, B., Selva, J., Hoechner, A., Volpe, M., Babeyko, A., Tonini, R., Wronna, M. \u0026amp; Omira, R. (2019). A new approximate method for quantifying tsunami maximum inundation height probability. \u003cem\u003ePure and Applied Geophysics\u003c/em\u003e, \u003cem\u003e176\u003c/em\u003e, 3227-3246. https://doi.org/10.1007/s00024-019-02091-w\u003c/li\u003e\n\u003cli\u003eGonz\u0026aacute;lez, F. I., Satake, K., Boss, E. F., \u0026amp; Mofjeld, H. O. (1995). Edge wave and non-trapped modes of the 25 April 1992 Cape Mendocino tsunami. In: Imamura, F., Satake, K. (eds) \u003cem\u003eTsunamis: 1992\u0026ndash;1994\u003c/em\u003e. Pageoph Topical Volumes. Birkh\u0026auml;user Basel. https://doi.org/10.1007/978-3-0348-7279-9_3 \u003c/li\u003e\n\u003cli\u003eHa, T., Yoon, J. S., Kyung, J. B., Yoon, S. H., Lee, D. W., \u0026amp; Kim, K. H. (2022). Occurrence of local tsunamis along the eastern coast of the Korean peninsula based on numerical modeling of historical earthquakes. \u003cem\u003eFrontiers in Earth Science\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e, 836168. https://doi.org/10.3389/feart.2022.836168\u003c/li\u003e\n\u003cli\u003eHayes, G. P., Moore, G. L., Portner, D. E., Hearne, M., Flamme, H., Furtney, M., \u0026amp; Smoczyk, G. M. (2018). Slab2, a comprehensive subduction zone geometry model. \u003cem\u003eScience\u003c/em\u003e, \u003cem\u003e362\u003c/em\u003e(6410), 58-61. https://doi.org/10.1126/science.aat4723\u003c/li\u003e\n\u003cli\u003eIglesias, A., Singh, S. K., Castro‐Artola, O., P\u0026eacute;rez‐Campos, X., Corona‐Fern\u0026aacute;ndez, R. D., Santoyo, M. A., Esp\u0026iacute;ndola V. H., Arroyo D., and Franco, S. I. (2022). A source study of the Mw 7.0 Acapulco, Mexico, earthquake of 8 September 2021. \u003cem\u003eSeismological Research Letters\u003c/em\u003e, \u003cem\u003e93\u003c/em\u003e(6), 3205-3218. https://doi.org/10.1785/0220220124 \u003c/li\u003e\n\u003cli\u003eJel\u0026iacute;nek, R., Č\u0026iacute;žkov\u0026aacute;, H., \u0026amp; Spikings, R. (2012). Thermal structure and evolution of the lithosphere in the central Mexican subduction zone: New constraints on the thermal budget of subduction. \u003cem\u003eEarth and Planetary Science Letters\u003c/em\u003e, 330-331, 97-108. https://doi.org/10.1016/j.epsl.2012.03.020\u003c/li\u003e\n\u003cli\u003eLeVeque, R. J., George, D. L., \u0026amp; Berger, M. J. (2011). Tsunami modelling with adaptively refined finite volume methods. \u003cem\u003eActa Numerica\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e, 211-289. https://doi.org/10.1017/S0962492911000043\u003c/li\u003e\n\u003cli\u003eMelgar, D., LeVeque, R. J., Dreger, D. S., \u0026amp; Allen, R. M. (2016). Kinematic rupture scenarios and synthetic displacement data: An example application to the Cascadia subduction zone. \u003cem\u003eJournal of Geophysical Research: Solid Earth\u003c/em\u003e, 121(9), 6658-6674. https://doi.org/10.1002/2016JB013314\u003c/li\u003e\n\u003cli\u003eMendoza, C., \u0026amp; Hartzell, S. H. (1999). Fault-slip distribution of the 1995 Colima-Jalisco, Mexico, earthquake. \u003cem\u003eBulletin of the Seismological Society of America\u003c/em\u003e, 89(5), 1338-1344. https://doi.org/10.1785/BSSA0890051338\u003c/li\u003e\n\u003cli\u003eMiyashita, K., Shuto, N., \u0026amp; Imamura, F. (2020). Comprehensive study of tsunami effects on coastal areas. \u003cem\u003eJournal of Coastal Research\u003c/em\u003e, 36(1), 44-52. https://doi.org/10.2112/JCOASTRES-D-18-00123.1\u003c/li\u003e\n\u003cli\u003eMolnar, P., \u0026amp; Sykes, L. R. (1969). Tectonics, seismicity, and morphology of the Caribbean and adjacent regions. \u003cem\u003eGeological Society of America Bulletin\u003c/em\u003e, 80(9), 1639-1684. https://doi.org/10.1130/0016-7606(1969)80[1639:TSAMOT]2.0.CO;2\u003c/li\u003e\n\u003cli\u003eMori, J., \u0026amp; McCaffrey, R. (2018). Tsunami hazard in the Pacific Ocean. \u003cem\u003eAnnual Review of Earth and Planetary Sciences\u003c/em\u003e, 46, 77-103. https://doi.org/10.1146/annurev-earth-082517-010031\u003c/li\u003e\n\u003cli\u003eMori, N., Muhammad, A., Goda, K., Yasuda, T., \u0026amp; Ruiz-Angulo, A. (2017). Probabilistic tsunami hazard analysis of the Pacific coast of Mexico: case study based on the 1995 Colima earthquake tsunami. \u003cem\u003eFrontiers in built environment\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e, 34. https://doi.org/10.3389/fbuil.2017.00034\u003c/li\u003e\n\u003cli\u003eMulia, I. E., Ishibe, T., Satake, K., Gusman, A. R., \u0026amp; Murotani, S. (2020). Regional probabilistic tsunami hazard assessment associated with active faults along the eastern margin of the Sea of Japan. \u003cem\u003eEarth, Planets and Space\u003c/em\u003e, \u003cem\u003e72\u003c/em\u003e, 1-15. https://doi.org/10.1186/s40623-020-01256-5\u003c/li\u003e\n\u003cli\u003eNational Oceanic and Atmospheric Administration (NOAA). (2023). Historical tsunami database. Retrieved from https://www.ngdc.noaa.gov/hazard/tsu_db.shtml. \u003c/li\u003e\n\u003cli\u003ePari, Y., Murthy, M. R., Subramanian, B. R., \u0026amp; Ramachandran, S. (2008). Morphological changes at Vellar estuary, India\u0026mdash;Impact of the December 2004 tsunami. \u003cem\u003eJournal of environmental management\u003c/em\u003e, \u003cem\u003e89\u003c/em\u003e(1), 45-57. https://doi.org/10.1016/j.jenvman.2007.01.055\u003c/li\u003e\n\u003cli\u003eSalazar‐Monroy, E. F., Melgar, D., Jaimes, M. A., \u0026amp; Ramirez‐Guzman, L. (2021). Regional probabilistic tsunami hazard analysis for the Mexican subduction zone from stochastic slip models. \u003cem\u003eJournal of Geophysical Research: Solid Earth\u003c/em\u003e, 126(6), e2020JB020781. https://doi.org/10.1029/2020JB020781\u003c/li\u003e\n\u003cli\u003eS\u0026aacute;nchez, J., \u0026amp; Farreras, S. F. (1993). Catalog of tsunamis on the western coast of Mexico. \u003cem\u003eWorld Data Center A for Solid Earth Geophysics Publication, SE-50, National Oceanic and Atmospheric Administration, Geophysical Data Center, Boulder.\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eSingh, S. K., Iglesias, A., Arroyo, D., P\u0026eacute;rez-Campos, X., Ordaz, M., Mendoza, C., Corona-Fern\u0026aacute;ndez, R. D., Rivera, L., Esp\u0026iacute;ndola, V. H., Gonz\u0026aacute;lez-\u0026Aacute;vila, D., Mart\u0026iacute;nez-L\u0026oacute;pez, R., Castro-Artola, O., Santoyo, M. A., \u0026amp; Franco, S. I. (2023). A seismological study of the Michoac\u0026aacute;n-Colima, Mexico, earthquake of 19 September 2022 (Mw7.6). \u003cem\u003eGeof\u0026iacute;sica internacional\u003c/em\u003e, 62(2), 445-465. https://doi.org/10.22201/igeof.2954436xe.2023.62.2.1453\u003c/li\u003e\n\u003cli\u003eSmith, W. H. F., \u0026amp; Sandwell, D. T. (1997). Global sea floor topography from satellite altimetry and ship depth soundings. \u003cem\u003eScience\u003c/em\u003e, 277(5334), 1956-1962. https://doi.org/10.1126/science.277.5334.1956\u003c/li\u003e\n\u003cli\u003eTonini, R., Di Manna, P., Lorito, S., Selva, J., Volpe, M., Romano, F., \u0026amp; Vittori, E. (2021). Testing tsunami inundation maps for evacuation planning in Italy. \u003cem\u003eFrontiers in Earth Science\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e, 628061. https://doi.org/10.3389/feart.2021.628061\u003c/li\u003e\n\u003cli\u003eZamora, N., Gubler, A., Catal\u0026aacute;n, P. A., \u0026amp; Carvajal, M. (2021). Tsunami inundation forecast in central Chile using stochastic earthquake scenarios. In \u003cem\u003eProceedings of the 8th BSC Doctoral Symposium\u003c/em\u003e (pp. 72\u0026ndash;73). Barcelona Supercomputing Center. https://hdl.handle.net/2117/346622\u003c/li\u003e\n\u003cli\u003eZ\u0026uacute;\u0026ntilde;iga, F. R., Su\u0026aacute;rez, G., Figueroa-Soto, \u0026Aacute;., \u0026amp; Mendoza, A. (2017). A first-order seismotectonic regionalization of Mexico for seismic hazard and risk estimation. \u003cem\u003eJournal of Seismology\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e, 1295-1322. https://doi.org/10.1007/s10950-017-9666-0\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":true,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"PTHA, GeoClaw, Mesoamerican Trench, Coastal risk, Hazard mapping, Return period","lastPublishedDoi":"10.21203/rs.3.rs-7257032/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7257032/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study presents a probabilistic tsunami hazard assessment (PTHA) for Manzanillo, Colima, a coastal city in Mexico, which is vulnerable to local tsunamis due to its proximity to the subduction zone along the Mesoamerican Trench. Historical records show that the region experienced major tsunamis, notably in 1932 and 1995. Using stochastic earthquake source models and numerical simulations, we calculated the tsunami inundation scenarios for different return periods (50, 100, and 500 years). The study employed the GeoClaw open-source software package, incorporating high-resolution topography and bathymetry, to simulate tsunami wave propagation and coastal inundation. Model validation using the 1995 Colima-Jalisco event demonstrates that the simulation correctly captures the observed tsunami characteristics. The results revealed that lowland areas, particularly near Manzanillo and Santiago Bays, and the Cuyutl\u0026aacute;n Lagoon, could experience inundation of several kilometers inland in a worst-case scenario. Probability-of-exceedance curves indicate a high likelihood of moderate to significant tsunami wave heights within a 50-year time frame, underscoring the substantial risk to the city. These findings provide crucial information for local authorities to develop effective tsunami risk management strategies including hazard mapping, improved building codes, and emergency preparedness plans.\u003c/p\u003e","manuscriptTitle":"Probabilistic tsunami hazard assessment and flood modeling: A case study of Manzanillo, Colima, Mexico","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-18 06:19:07","doi":"10.21203/rs.3.rs-7257032/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"03e5a638-a829-48df-a213-c5d4d44405f4","owner":[],"postedDate":"August 18th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-18T14:25:18+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-18 06:19:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7257032","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7257032","identity":"rs-7257032","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}