Application of Coastal Resilience Metrics at Panama City Beach, Florida

There is extensive documentation on the workflow and steps taken within the CERI toolbox (an ArcGIS Pro-based toolbox) because it not only functions as a resilience toolbox but also provides coastal engineers the ability to generate shorelines and transects, extract morphology data, plot profile data as extracted, and more (Robertson et al. 2018). Throughout this study, the CERI toolbox was under development, and this study provided another opportunity for the developers of the toolset to improve the code and develop patches. Simultaneous to this research, the developers were working on their own CERI-based research that focused on a national scale instead of a local scale such as the Panama City Beach. Regardless of the project scale, the CERI toolset can provide insights and extract critical data for future coastal engineering research.

For this study, the first step was data acquisition for the CERI toolset. The inputs require (at a minimum) a raster surface that can act as a digital elevation model (DEM) and knowledge of the area such as the MHW elevation relative to the DEM vertical datum. This allows the user to generate an MHW shoreline from the DEM and then generate shore-perpendicular transects along that shoreline. These transects are then used to grid data points (extracting elevation data from the DEM and assigning to points along each transect) at a 1-m interval. The gridded points are used as input to the morphology extraction step, and the extracted features are plotted. After verification of the results through user quality assurance and control methods, the extracted features are used for the CERI calculations for each transect. Most edits to the extracted feature points in the present study were very minor and caused a minimal (less than 0.2% difference on average, with a maximum difference of 1.5%) effect on the CERI values. These CERI calculations also use wave and wind lookup tables (provided with the toolbox) to provide a calculation based on the return interval of the user’s choice. For this work, both 10-year and 50-year intervals were calculated for comparison. The CERI is then calculated using the feature points, and values are output to a summary table that can be exported in various formats (.csv, .xlsx, etc.) for further analysis. A more in-depth walkthrough of the methodology for individual CERI steps can be found in Spurgeon (2022).

DEM coverage for Panama City Beach was downloaded from the National Oceanic and Atmospheric Administration's (NOAA) Digital Coast Data Access Viewer and included data from the US Army Corps of Engineers (USACE), Florida State Agencies, the National Aeronautics and Space Administration (NASA), United States Geological Survey (USGS), and the NOAA. To ensure consistency and repeatability of the boundary for this project, the default selection of boundary boxes when typing “Panama City Beach” into the Digital Coast web application was used. In this manner, the identical data sets could be accessed at any time by any party. DEM data were available at various instances in time: 1998, 2001, 2005 (post-Dennis), 2005 (post-Katrina), 2006, 2007, 2010, 2015, 2017, 2018, and 2020. Of these data sets, only the 2010 data set was deemed unusable because the near-shore region had considerable data gaps. These data gaps can occur depending on near-shore conditions at the time of surveying, particularly when high-wave-energy conditions increase turbidity, making it more difficult for the LiDAR systems to accurately measure bathymetry. While this problem is common in DEM data sets (Dunkin et al. 2020) and has known fixes, the 2010 data set was deemed too sparse to justify interpolation efforts. It should be noted that each of the CERI toolbox steps was completed for each DEM data set, holding all inputs constant other than the DEM itself.

As noted in the “Introduction” section, Beach-fx is a planning model that evaluates economic benefits and physical performance of CSRM projects. An extensively used model within the USACE to decide if projects are economically justifiable, Beach-fx is often used to compare nourishment alternatives at a location regarding economic and damage factors. Beach-fx uses the Storm-Induced BEAch CHange (SBEACH) model or the Cross Shore (CSHORE) numerical model to model cross-shore morphological changes caused by storms (Johnson and Sanderson 2020). The present study uses SBEACH, which is a cross-shore numerical model used to calculate dune and beach erosion based on storm impacts (Larson and Kraus 1989). The empirical model is driven by engineering data such as wave height, wave period, and water level time series. Additional inputs such as the median grain size and profile shape are used to further inform the model. Dally’s breaker decay model by Dally et al. (1985) forms the basis of the SBEACH model’s approach to monochromatic waves and was eventually expanded to include a Rayleigh distribution to handle random waves, as presented in Larson and Kraus (1989). The model uses the equilibrium beach profile as presented by Bruun (1954) and modified by Dean (1977) and is shown in the form ofwhere h = water depth; A = shape parameter; and x = horizontal distance from the shoreline. SBEACH calculates the cross-shore transport rate as follows:where q = net cross-shore sediment transport rate in m³/m/s; K = empirical sediment transport rate coefficient; D = wave energy dissipations per unit water volume; D_eq = equilibrium wave-energy dissipations per unit water volume; and $\epsilon$ = slope-related sediment transport rate coefficient. The wave-energy dissipation per unit volume, D, is given by the change in wave energy flux and is calculated as follows:where E = wave-energy density; and C_g = wave group velocity. The equilibrium wave energy dissipation per unit volume, D_eq, is calculated in terms of the equilibrium profile shape parameter as follows:where ρ = density of water; g = gravitational acceleration; and γ = ratio of breaking wave height to the water depth. The SBEACH model outputs were used as inputs for the Beach-fx study at Panama City Beach.

Beach-fx calculates the damages and costs from storm events and simulates emergency and planned nourishment alternatives. The model uses the SBEACH or CSHORE model results to simulate many iterations of storm seasons, employing a relative storm probability table to complete a Monte Carlo simulation. The Monte Carlo method follows a Poisson distribution on the number of storms alongside the relative storm probabilities. No profile response computations are completed by Beach-fx—instead, the responses are kept in a lookup table in the storm response database, which is populated by SBEACH or CSHORE (Johnson and Sanderson 2020). The morphology timeline output of the Beach-fx model can be used as the direct input for the BW metric, and with 300 iterations per alternative, the overall trends emerge.

Beach-fx was run by the USACE mobile district and provided the model output for the present study. Beach-fx uses reaches to split up areas of interest with different morphological and land-use/economic conditions, and the district data contain four reaches. Information from a recent survey showed that Reaches 2–4 closely approximate the morphology at Panama City Beach. This was determined by leveraging a Python-based toolset that was created as part of Spurgeon (2023) named the Representative Beach Profile Generator (RBPG). This toolset imports the CERI toolbox output (extracted elevation points and CERI extracted feature points), aligns the profiles in the cross-shore, and plots them. Once plotted, the user inputs their own values to approximate the beach morphology in a trapezoidal format. This tool allowed the 2017 profile data (Fig. 3) extracted in the CERI methodology to be analyzed and compared with the district-created reaches.

Ultimately, Reach 2 was considered for further analysis because it had a historical erosion rate of 0.283 m/year that was most similar to the 0.3 m/year erosion rate calculated from a shoreline analysis of historical DEM data. There were no other pertinent details that differentiated Reach 2 from Reaches 3 and 4, because this study does not consider the economic impacts generated by Beach-fx. Therefore, Beach-fx results for Reach 2 were used in this study to evaluate the four nourishment strategies shown in Table 1. These nourishment types include “emergency” and “planned” nourishment alternatives. It should be noted that the emergency nourishment options were either the “null” alternative of no nourishment or the “emergency” nourishment, which has a blank template. A blank emergency nourishment template within Beach-fx defaults to the emergency nourishment fill density specification, which is a user-specified density of cubic meters per 0.305 m of shoreline (or cubic yards/linear foot) longshore that will be placed in the case of an emergency nourishment trigger. This fill density was set to 3.06 m³ per 0.305 m. In the planned alternatives, each nourishment extended the berm width to at least 9.1 m to bring the profile to the template width. The “3 dune” and “6 dune” alternatives nourish the dune width an additional 3.048 m (10 ft) and 6.096 m (20 ft), respectively, as seen by the change from the 9.1 berm alternative (13.81-m dune width) to the 9.1 berm 3 dune (16.81-m dune width) and the 9.1 berm 6 dune alternative (19.81-m dune width). The only configuration inputs that changed between alternatives were the nourishment types, both with planned and emergency nourishment.

Beach-fx was run a total of five times (Table 1). Each alternative included three sea-level change rates. The sea-level change rates were those defined by the district: the “low” case introducing no change, the “intermediate” case introducing a positive 0.117 m of sea-level change, and the “high” change case introducing a positive 0.490 m of sea-level change. All alternatives used an already calibrated set of scenario data and also the “calibration run” option to eliminate damage calculations, which decreases computational time. In addition, each time Beach-fx was run, 300 iterations of that model setup were run in order to obtain an average response. Most Beach-fx studies include economic considerations of damage to property protected by the beach, but this information is not needed for the BW metric and has no effect on the overall result of this resilience study. The cost of beach renourishment is the only economic consideration that is considered.

Once the Beach-fx alternatives were completed, the Beach-fx morphology timeline output was used to calculate the BW. This was done using the output dune height, dune width, berm width, and dune slope values. The berm height is not included in the morphology timeline because Beach-fx treats the berm only as a horizontally eroding feature and thus never changes the elevation. For the purposes of the BW calculation, a constant value of 2.01 m was used for the berm elevation. While keeping a berm elevation constant may artificially steepen the dune slope, Beach-fx adjusts for this by eroding the dune slope and redistributing it across the berm width. The BW was calculated for each time step.

The CERI results are focused on the calculated metric for each survey. While each individual transect did have its own CERI value assigned to it for each survey data set, averaging the data across all transects allows for an overview of the study site temporally. The direct export from the CERI toolbox includes this data set averaging as well as the generation of summary tables of the averaged data. Table 2 lists the 10- and 50-year extreme water level and wave return period data sets and survey notes. It should be noted that the 2017 survey started on April 9, 2017 and ended on May 17, 2017. The nourishment that occurred in 2017 was completed on May 19, 2017, hence the “During Nourishment” designation. The survey notes are included in Table 2 for contextual purposes: the CERI tool, when used in a temporal setting, is best suited whenever there is information about storm impacts, nourishments (planned or emergency), and any other events that might influence the resilience of the area. For further visualization of the results in Table 2, Fig. 4 shows both the 10- and the 50-year return interval points, nourishment efforts, and storm events.

The impact of Hurricane Earl in 1998 left Panama City beach with a relatively low CERI value, which was increased significantly after the large sediment placement was completed in 1999. Hurricane Dennis reduced the overall CERI value but did not reduce it below the original 1999 low value. The nourishment in 2006 helped raise the CERI value, and nourishments in 2011 and 2017 align with consistent CERI values. Hurricane Michael in 2018 had a small impact on the CERI value, although previous sections mention that Hurricane Michael made impact within 25 mi of the beach and was the highest category storm to impact the area in half of a century. The 2017 and 2018 surveys were taken about 1 year apart, thereby isolating the impacts of Hurricane Michael for analysis. This seemingly small change is something to be noted, considering the major impact of Hurricane Michael at the location. Houston (2023) found that PCB beaches experienced 1.9 m of surge due to Hurricane Michael and that no coastal damage occurred, except those due to wind. The nourishment prevented more than 70% of the potential damage. While storm statistical analysis is not within the scope of this work, it is a topic that should be considered for future work with the CERI toolbox. For example, the 1999 nourishment of 4.9 × 10⁶ m³ showed a 21.3% increase in CERI resilience, while Hurricane Dennis resulted in a 7.4% decrease in the same metric. After the 2006 nourishment (a 4.9% CERI increase), changes were minimal (less than 2.5%). These large-scale metrics (percent changes across the entire domain) are just one example of the quantitative metrics that are provided through the CERI. Coastal scientists are also able to focus on a more microscale, down to the individual transect CERI value. Defining the most morphologically resilient profile for an area or domain may help coastal managers decide on a design template. Percent changes in the values of individual transects or sections of the beach could provide an insight into the likelihood of change of an area.

Other less direct metrics from the CERI toolbox are also available in the exported values. The extraction of key features, their elevations, and their cross-shore locations is an asset for other coastal engineering calculations. For example, the elevation differences in key features between the oldest and the lowest 10-year CERI data set (1998), the highest 10-year CERI data set (2015), and the most recent data set (2020) can be seen in Fig. 5. The extraction of these key features (berm and dune features alike) provides the basis for numerous opportunities for further analysis that can be performed using coastal engineering calculations: the CERI metric is just one of many. For this work, data analysis focused on the CERI metric and explained the final numerical values that it generated with a more in-depth look into the extracted morphology (see “Discussion” section).

The primary exports generated by Beach-fx alternatives can be used to directly calculate the BW metric. The results, once calculated and summarized, are separated into two sets for the purpose of displaying results: Null alternatives and Planned Nourishment alternatives. The Null alternatives are those with no planned nourishment of any kind, although they can include emergency nourishment. Planned Nourishment alternatives comprise the rest of the test alternatives and always include a planned nourishment and an emergency nourishment. Fig. 6 shows a temporal evolution of the BW metric from 300 calculations of the BW from the Beach-fx iterations for the 9.1 berm 6 dune alternative. In the figure, there are 300 points in each year (each “column” of the data given is 1 year in the model), and each point refers to a single iteration. All 300 iterations are shown to highlight that, even with differences in the BW between iterations, a general trend exists and is shown with a trend line for each SLR scenario. Similar plots have been created for the other planned nourishment alternatives, and these can be seen in Spurgeon (2022).

To simplify these data into a more understandable format, Fig. 7 shows the BW trend lines for each high SLR nourishment case for the 175th iteration only. While this figure is only a single iteration out of 300 for the study, it displays the individual behavior of each nourishment alternative throughout each storm season and nourishment. It was chosen as a sample iteration at random to help clarify and simplify metric behavior. The BW metric considers a higher value to be equivalent to a higher resilience due to the morphology because there is more cross-shore distance between the water and the back of the dune, thus limiting storm impacts to the infrastructure behind the dune system. Each time a storm event happens, there is a sudden decrease in the BW value, followed by a near-immediate partial recovery, the natural recovery built into Beach-fx. Furthermore, the beach loses the BW between storm events as part of the natural erosion processes that occur at Panama City Beach. Nourishment efforts are observed as instantaneous increases in the BW, as labeled for the 9.1 berm nourished alternative in Fig. 7.

The high SLR trend lines from each alternative are shown in Fig. 8 and allow for a direct comparison between every nourishment alternative across all 300 iterations. The 9.1 berm 6 dune alternative has the most nourishment and shows the highest overall BW value across the entire span of the 51-year timeframe. Alternatively, the null alternative without emergency nourishment has a significantly lower starting value (the “current” condition of the beach) and declines considerably over the course of simulation due to a lack of nourishment to maintain the beach. The same behavior is seen in Fig. 7 with the 9.1 berm 6 dune alternative: while there are various storm and recovery events, the overall value of the most heavily nourished alternative is one of a generally high BW value across the model timeframe. Comparatively, the null alternative without emergency nourishment, as shown in Fig. 7, tracks closely with its trend line: its value starts low and simply decreases as daily impacts and storm erosion occurs at the location. At around the 35-year mark, there is some behavior that shows a sudden increase in the BW; however, the null without the emergency alternative (in yellow) should not increase in the BW because there are no nourishment events that can take place for that alternative. While the BW, as it is measured, does increase in a non-nourished alternative for certain iterations, this is not attributed to an actual increase in overall resilience. Instead, an erosion of the dune height causes an extension of the berm as the sediment moves seaward, resulting in an increase in the BW and the “false positive” behavior at the 35-year mark. This brings up an important point: the BW metric does not address the issue of elevation, rather it addresses only the cross-shore distance. This limitation must be understood for the results to be correctly interpreted. As nourishments start to take place (gray, dark blue, and orange lines in Fig. 7), this limitation is no longer seen since storm events do not reach the top of the dune and the dune elevation rarely changes. Therefore, the increases in the BW are due to nourishment events and display expected behavior. However, in a no-nourishment alternative, the lack of nourishment eventually causes the dune height to be affected and thereby causes the “false positive” effect. The small increases of the BW in the null with emergency nourishment alternative (light blue) are due to the emergency nourishment being triggered, and there is a substantial difference in the amount of BW increase between that and the nourished alternatives during their full planned nourishment. Something else of note in Fig. 7 is that the 9.1 berm 3 dune nourishment alternative never sees a positive jump; therefore, this alternative within this iteration never receives nourishment. An inspection of other iterations shows that the 9.1 berm 3 dune alternative does not always go without nourishment, but the nourishments are far more unlikely to happen. The negative trend line data for the 9.1 berm 3 dune alternative in Fig. 8 displays the overall trend that occurs whenever nourishments are unlikely to trigger.

The nourishment alternative with the most consistent nourishments and highest BW midrange (9.1 berm 6 dune, 51 m) is 129.2% and 95.8% higher, respectively, than the midranges of the null alternatives (Null w/emergency, 22 m and Null w/o emergency 26 m). The same high midrange is 24.3% higher than the midrange of the nourished case that was not frequently nourished (9.1 berm 3 dune, 41 m) and 8.5% higher than the midrange of the other frequently nourished case (9.1 berm, 47 m). There is a clear trend within these BW midrange metrics that strengthen the conclusion that frequency of nourishment is just as important as the template that is used. This BW midrange metric allows coastal managers to understand how well their nourishment alternatives are projected to influence how the location might respond to storm impacts at a given time over the temporal model domain. A low midrange suggests that there will be less cross-shore distance between the infrastructure and the shoreline on average. The ranges themselves may also be considered as valuable metrics; a case with a minimum BW value of 1.5 m is significantly weaker than a case with a minimum BW value of 29 m.

The Beach-fx alternatives also calculated the estimated costs of nourishments. Table 3 considers the high SLR scenario for planned and emergency nourishment average costs of each nourishment alternative. Note that the data reference the average costs for the entire Panama City beach (all reaches) and not only the second reach that was shown in the morphological results. This was done to understand the cost required to nourish the entire Panama City beach. The average cost per reach for each alternative is shown in Table 3. The 9.1 berm 6 dune alternative has the highest price with an average of $6.76 million per nourishment over the 51-year timeframe, and the 9.1 berm 6 dune alternative is the lowest cost alternative with an average of $4.62 million per nourishment over the same timeframe.

The CERI results demonstrate how the Panama City Beach CSRM project has increased the resilience of this area since its inception in 1998–1999 in relation to reducing storm impacts. After Hurricane Earl in 1998, the CSRM project was developed to better protect the area from coastal storm impacts. One measure of success for CSRM projects may be a coastal resilience metric, especially if the resilience metric focuses primarily on how well the beach and dune system responds to coastal storm impacts.

Overall, the CERI metric tracks closely with the coastal engineering understanding of the effects of storm events and nourishments on coastal resilience. Relating back to the original four stages of resilience (prepare, resist, recover, and adapt), the CERI data speak to each, with the focus primarily on the preparation and resistance aspects of coastal resilience in relation to the morphology of the system. The preparations that occurred in 1999 (a CSRM project) increased the resilience of the area according to the CERI metric (Fig. 4). Moreover, whenever the cycle started over after a storm event (note the 2006 project), the nourishment projects reprepared the area by raising the CERI value to previous “high” values. During storm events, the increase in resistance was shown twofold: first, in 2005, the CERI score decreased but was still considered to be a higher resilience value than the original project poststorm in 1998. The initial nourishment in 1999 helped the area better resist storm impacts. Secondly, resistance is increased in that the overall decrease in resilience lessens; Fig. 4 shows that even when impacted by a major storm event such as Hurricane Michael, the CERI for the Panama City Beach area decreases by fewer points than when it was impacted by Hurricane Dennis in 2005. As stated previously, storm statistical analysis has not been completed for the CERI metric, so this observation is currently ambiguous, yet it suggests that the nourishments that took place between 2005 and 2018 were effective in their goal to increase or maintain storm resistance. Each subsequent storm event required less recovery (either via nourishment or natural recovery of the system). It should be noted that the typical increase in resilience due to recovery is seen in the resilience triangle as a reduction of time to recover to the desired level of functionality, and the CERI metric speaks specifically to the level to which recovery is needed, and not how quickly recovery efforts will succeed. Lastly, the adaptations, such as the 1.6-km construction segment that was built in 2011, contributed to storm resistance and therefore to the less need for making recovery efforts. The main focus of the CERI metric is on the resistance aspect through storm impact reduction.

Of the four stages of resilience, the CERI metric struggles to show how nourishment affects recovery. Instead, the key information is the constant CERI values after the CSRM project. As previously stated, nourishment reduces the amount that the system needs to recover (thereby being more a part of the adaptation of the system), but there is no clear evidence that the system recovers better naturally after nourishments. In the 10-year return interval CERI values, there is a slight increase in the metric in the few months between the 2005 surveys but a slight decrease in the metric after the 2018 survey. The 50-year return interval CERI value shows the inverse reaction: in 2005, there is a decrease in values, and post-2018, there is a slight increase in values, although it is probably better considered to be almost constant because the changes in the CERI values are negligible. More surveys would allow for better analysis to be completed using the CERI and how nourishments impact recovery of the beach. Due to the timing of surveys, it is difficult to draw conclusions about the recovery aspect of resilience at Panama City Beach. It is obvious that the act of nourishment is an engineered recovery of the beach, which may instead be seen as the cycle starting over and being “prepared” once again, thereby showing a constant/flat line result.

The BW metric results project into the future and allow the user to understand how resilience may change given a certain set of nourishment parameters. The figures in the “BW Results: A Look into the Future” section make indications and together provide a comprehensive view of future resilience at Panama City Beach through the trend line, range of values, and nourishment costs. Fig. 8 provides an insight into the trend lines of the evaluated alternatives. Nourished alternatives have generally flat or mildly sloping trend lines, indicating fairly consistent resilience over time on the whole. However, this issue is not as simple as nourished alternatives versus non-nourished alternatives. For example, both the 9.1 berm alternative and the 9.1 berm 6 dune alternative indicate future resilience with their flat trend line data, while the trend line for 9.1 berm 3 dune alternative shows a decrease in resilience over time with a negative slope. This result is counterintuitive from a nourishment template perspective because the two alternatives with flat future resilience have both more and less volume density than the decreasing resilience alternative of 9.1 berm 3 dune. Therefore, the nourishment volume in the template is not exclusively dictating the future resilience. It should also be noted that the resilience being considered within this case goes back to the foundation of the BW metric; the more the cross-shore distance between the back of the dune and the shoreline, the more the likelihood that the failure of the system is avoided. Therefore, this view of resilience is also primarily focused on the resistance aspect due to storm impact reduction.

Looking closely at the 9.1 berm 3 dune alternative in Fig. 7, nourishment does not happen, although the profile has eroded substantially. Although the dune and berm had eroded, the 153,000 m³ volume loss threshold required to trigger nourishment was not met in many alternatives. This volume threshold is dependent on the template itself as well as the erosion that is possible at a location. For example, a profile with a relatively wide berm is more likely to lose larger amounts of volume and more easily meet nourishment requirement thresholds because there is more sediment at a lower elevation, which allows for increased erosion from wave energy. For the 9.1 berm 3 dune alternative, most of the iterations are not nourished as frequently as other alternatives, and therefore, the negative trend line in Fig. 8 is explained.

The results for the Null without Emergency Nourishment alternative are expected. Without nourishment occurring when needed, the BW resilience metric substantially decreases over that time. Once nourishment does happen (observed in other iterations), a full recovery to the original BW is possible, further identifying that the nourishment template itself likely does not have any inherent issues. Instead, the frequency at which the nourishment template is implemented is highly relevant. The 9.1 berm 3 dune alternative has a nourishment timeline of 5 years with a triggering threshold of 153,000 m³ (Table 1), and the 9.1 berm alternative has an increment of 3 years and a volume threshold of 117,000 cu yd. Changes in how often the profile is being checked and what is required for nourishment to be triggered have a substantial effect on the frequency of the occurrence of nourishments. For example, the 9.1 berm 6 dune alternative has the same timeline and threshold values as the 9.1 berm 3 dune alternative, and this shows that nourishment is happening as expected. This is due to the nourishment needing more sediment: the extra 3.05 m that the dune is being nourished means that the 9.1 berm 6 dune profile will meet the volume threshold sooner than the 9.1 berm 3 dune alternative given the same exact conditions. From a resilience standpoint, a template with poorly designed triggers will still result in a weak and vulnerable system.

From a financial standpoint, these conclusions remain the same: the 9.1 berm 3 dune alternative costs considerably less than the other two nourished alternatives because the requirements to trigger a nourishment are too strict for that profile. Based on the slope of the trend lines alone, it would be almost as cost-effective to pay for emergency nourishments ($4.66 million) for the next 50 years than it would be to nourish routinely using a stronger profile template ($4.62 million), although this is not recommended because the overall values of the BW are lower for the emergency alternative. Instead, this small cost difference makes the case that consistency in frequency of nourishments is just as important as nourishment templates. These costs are estimates; the delayed nourishment alternative that costs $4.62 million cannot necessarily be directly compared with the other alternatives since the nourishment schedule was drastically different (due to the volume thresholds being different). Therefore, the less-nourished alternative will cost less since there are fewer mobilization costs included over the same life span.

Even with the negative slope, the 9.1 berm 3 dune alternative had a higher BW range (12–70 m) compared with the null alternatives (1.5–43 and 9.1–43 m with/without emergency nourishment, respectively). The consistently frequent nourished alternatives also showed that the 9.1 berm alternative had a lower range (27–67 m) than the 9.1 berm 6 dune alternative (29–73 m). However, the cost estimates for each ($6.43 million versus $6.76 million, respectively) might sway the final decision toward the slightly less robust template since there is consistency based on the trend line data in both alternatives.

The greatest purpose of the CERI and BW metrics is to inform stakeholders about the coast’s resilience in terms of its ability to resist storm damage over time at a location with specific attention on nourishment and other protection features. Panama City Beach had relatively low resilience in the late 1990s after being impacted by Hurricane Earl. Once the CSRM project was introduced, the area became a more resilient system (as shown by the CERI metric) through constant management. The history of this location suggests that the investment in beach nourishment has resulted in the area better resisting storm damage and providing greater protection and recovery for Panama City. The healthy, wide dune systems increase the CERI metric as seen in Fig. 4. The general shape of the beach does not change, but the width of the dune and berm system proportionally increases.

For an already-resilient beach, general management is necessary to keep the beach to the level of resilience established. If nourishment is kept consistent in its frequency, the beach will not only be able to maintain its consistent resilience metric values but actually increase them even when facing SLR. In alternatives where frequent nourishment is abandoned altogether, a decrease in the BW, and therefore resilience, is observed in the area, unless emergency nourishment occurs. The best options are those that have frequent nourishments because they have a larger range of BW values as well as little to no overall degradation in resilience over the course of the next half century. As discussed previously, a nourishment with a strong profile but an infrequent nourishment schedule will still result in lower resilience. This is seen in conjunction with the CERI data: the CERI value stays relatively high after the initial CSRM project construction because the longest period without nourishment was from 1999 to 2006, a period of only 7 years. Consistency in nourishments has helped Panama City Beach historically just as much as the templates used in those nourishments.

The main contributions of this manuscript are applying the combination of the CERI and BW metrics, using a specific location to do so, and demonstrating the benefit that the combined metrics provide, which they would not have been able to provide separately. While the metrics are related and use similar inputs, they provide different temporal scales of the morphologic resilience at a location, and the use of both metrics provides a more complete temporal understanding to coastal scientists. These resilience metrics provide insights and also key information that was not previously accessible by the coastal community. The ability to quantify the strengthening or degradation effects of various events that the area has already experienced can further inform engineers, scientists, and stakeholders about coastal resilience at a location historically. Furthermore, those insights can be applied to future scenarios with greater confidence. In the case of nourishments specifically, alternatives can be compared for an area in a new and unique way. Instead of simply comparing nourishment alternatives based on cost, amount of sediment placed, and damage factors, an understanding of the resilience of the system can be gained using one of the two resilience metrics described herein. These metrics together can help better inform decision-making in the coastal environment and provide a greater understanding of how an area will respond to storm events overall. These metrics, especially when combined to receive the full temporal portrait of a site, “establish a range of metrics and critical measurements needed to monitor performance of and design infrastructure for coastal systems” (Rosati et al. 2015). Fox-Lent and Linkov (2018), as well as many others, have attempted to fulfill this call and have done so partially, but none have completed it in such a simple and direct approach as the CERI and BW metrics when relating to coastal storm impact reduction, sea-level change adaptation, and nourishment efforts. Furthermore, in combination, the CERI and BW metrics provide more information because they are based on different yet complementary sets of coastal engineering equations, theories, and best practices.