Long-Term Performance of Cyclic Beach Nourishment and Roles of Erosional Hotspots Free

Barrier islands are among the most desirable places to live and visit. Beaches on barrier islands are recreationally and ecologically valuable, providing habitats and nesting grounds for shorebirds and land and marine animals. Wide beaches and healthy dune systems offer the first line of defense against storms by absorbing wave energy and protecting often heavily developed coastal communities.

Populations in coastal areas have grown exponentially. Low-lying coastal communities (less than 10 m elevation) represent only 2% of Earth’s landmass but contain 10% of the world’s population (Joy and Gopinath, 2021). These areas are highly vulnerable to sea-level rise and storm impact. Barrier islands are vital to Florida’s economy, with an average of 19 million tourists traveling to coastal cities annually (Office of Economic & Demographic Research, 2015). Nationwide, beach tourism contributes tremendously to the U.S. economy (Houston, 2024), even during the coronavirus 2019 (COVID-19) pandemic (Houston, 2023).

The acceleration in sea-level rise will intensify storm surge flooding, beach erosion, and coastal structural damage (Sweet et al., 2022; Zhang and Leatherman, 2011). Beach nourishment has become the dominant sandy shoreline mitigation approach in the United States over the past 40 years (Elko et al., 2021), overtaking the previous approach of constructing hard engineering structures such as groins and breakwaters. From a sediment budget point of view, beach erosion is the result of a sand deficit (Wang and Beck, 2022). Beach nourishment is the sole approach that directly addresses the issue of a sand deficit by restoring the sediment into the littoral system and allowing natural forces to continue adjusting (NRC, 1995). Beach nourishment is considered sacrificial at times, in that it must be repeatedly conducted in certain time intervals that are controlled by the site-specific background erosion rates and lateral diffusion of the placed sand and storm impacts (Dean, 2002). Beach segments with higher erosion rates than those in adjacent segments, known as erosional hotspots, can persist due to several conditions (Kraus and Galgano, 2001). Management of such areas depends on the reasons for and types of erosion hotspots. In many cases, more frequent nourishment is needed, or a wider beach needs to be constructed during each renourishment at erosional hotspots (Benedet, Finkl, and Hartog, 2007; Cheng and Wang, 2022).

As of 2004, the Florida Gulf Coast had received over 38 million cubic meters of sediment for beach nourishment since the 1960s (Campbell and Benedet, 2006). This number has increased substantially since then, mostly due to renourishment projects (Elko et al., 2021). The Pinellas County coast (Figure 1) represents one of the earliest and largest nourishment programs in Florida. The three heavily developed barrier islands, Sand Key, Treasure Island, and Long Key, receive regular beach nourishments. Sand Key has received island-wide nourishment since 1999. The nourishment along Treasure Island was conducted roughly every 4 years since 2000, while Sunset Beach at the southern end hosted one of the earliest nourishment projects in 1969. Since 2006, systematic beach nourishment monitoring through bimonthly to quarterly beach-profile surveys have been conducted by the University of South Florida’s Coastal Research Laboratory (USF-CRL) at the three barrier islands. Beach-profile data before 2006 were collected by USF-CRL and various entities but sometimes along different survey azimuths. Although not completely comparable with the post-2006 data, the earlier data are still applicable for assessing long-term beach nourishment performance. In this study, only Sand Key and Treasure Island nourishments were analyzed. The nourishment performance at Long Key (Elko, Holman, and Gelfenbaum, 2005; Elko and Wang, 2007) is influenced by four T-groins at the northern end of the island that were constructed during the 2006 and 2012 nourishment cycles.

The overall goal of this study was to assess the long-term performance of the nourishment cycles since 1990 to identify if beach nourishment practices can eventually amend the sand deficit along the developed barrier islands, or if nourishment acts as a long-term maintenance strategy. In addition to assessing overall nourishment performance, three specific issues were addressed, including benefits to adjacent beaches, nourishment design and performance at erosional hotspots, and processes that cause erosional hotspots. Several sections of the beach along the two barrier islands that are either between or directly adjacent to the nourished sections were also examined, with the goal of determining if and how the cyclic nourishment projects benefited nearby nonnourished sections. Three erosional hotspots were identified based on the time-series beach-profile data. More sand was placed along two of the erosional hotspots as compared to typical beach sections. The performance of the enhanced beach nourishment at these erosional hotspots was evaluated. The processes that are responsible for the erosional hotspots were investigated based on the beach-profile data and numerical wave and flow modeling.

This study examined long-term beach nourishment performance along two heavily developed barrier islands, Sand Key and Treasure Island, in west-central Florida (Figure 1). Sediment along the west-central Florida coast is composed of both siliciclastic and carbonate grains. The siliciclastic sediment is primarily fine quartz sand, whereas the carbonate grains are mostly shell fragments of different sizes (Cheng and Wang, 2022; Wang, Davis, and Kraus, 1998; Wang, Kraus, and Davis, 1998). Dunes are small on this coast, with typical elevations ranging from 1 to 2 m above the dry beach due to low wind speeds and low sediment supply (Davis, 1994; Wang et al., 2024).

West-central Florida has a low-wave-energy microtidal coast, which allows for a simplified approach to beach-profile data collection as opposed to a high-wave-energy macrotidal coast (Cheng, Wang, and Guo, 2016). Spring tides are typically diurnal with ranges up to 1 m. Neap tides are semidiurnal with ranges of 0.4 m. Wave energy is mild along this stretch of the Gulf of Mexico coast, averaging 0.25–0.30 m in nearshore wave height (Beck et al., 2020; Tanner, 1960). Wave properties along this coast vary seasonally. Higher waves are typically associated with the passages of winter cold fronts and occasional passages of tropical storms/hurricanes in the summer (Walton, 1973). Northerly winds during the frequent passages of winter cold fronts, roughly every 10–14 days (Walton, 1973), often produce higher waves and result in more active longshore sand transport from north to south as compared to the south to north transport produced by the predominantly smaller southward incident waves. This results in a generally net annual north to south longshore sediment transport on the order of 30,000 to 60,000 m³/y (Beck and Wang, 2019; Walton, 1973).

Sand Key is the longest barrier island on the Florida peninsular west coast and spans 22 km. It is bounded by two heavily structured inlets: Clearwater Pass to the north and John’s Pass to the south. Both inlets are classified as mixed energy with large ebb shoals (Beck and Wang, 2019; Gibeaut and Davis, 1993). Sand Key extends across a broad headland roughly in the middle with a total shoreline orientation change of 65° controlled by antecedent geology, i.e. outcropping Miocene to Pliocene Tampa Limestone (Cherry, Stewart, and Mann, 1970; Davis and Barnard, 2003; Masetti, Fagherazzi, and Montanari, 2008). The headland and the complex beach-inlet interactions at both ends of the island, coupled with the dense oceanfront residential development, which extends from roughly 25 m to 200 m from the shoreline, have significant influences on beach management and shore protection (Beck and Wang, 2019).

Beach nourishment on Sand Key began in the late 1980s and early 1990s. By 2000, repeated renourishments every 4–7 years had become island-wide. Systematic beach nourishment monitoring by USF-CRL began in 2006, directly following the nourishment that year. To account for spatial variations caused by the two tidal inlets at both ends and the headland in the middle, the analysis of the long Sand Key nourishment is divided into five segments, including two erosional hotspots (Figure 1). The North Sand Key stretch is 2.1 km long and is a persistent erosional hotspot (Cheng and Wang, 2022; Roberts and Wang, 2012). Typically, the dry beach is widened by ∼70 m per nourishment along this stretch. The Belleair Shores stretch is 1.8 km long and has never been nourished due to management issues associated with land ownership. This gap provides an opportunity to study longshore spreading from two nourished sections to the north and south and its potential benefits along this nonnourished section. The Indian Rocks stretch is 2.4 km long and located along the northern flank of the headland. Typically, the dry beach is widened by ∼27 m per nourishment. The Headland stretch is another erosional hotspot. It is 3.9 km long and typically widened by ∼27 m per nourishment. South Sand Key is 3.3 km long and located along the southern flank of the headland, and it is typically widened by ∼27 m per nourishment as well. The Indian Rocks, Headland, and South Sand Key are sections of one continuous nourishment project. They are discussed separately due to their different shoreline orientations.

Treasure Island is 5.4 km long and is bounded by John’s Pass to the north and the Blind Pass to the south (Figure 1). Both inlets are heavily structured with jetties and seawalls. The opening of John’s Pass in 1848 decreased the tidal prism at Blind Pass and induced the southern migration of the inlet, which was artificially stopped by a series of seawall construction projects starting in 1937 (Beck and Wang, 2019; Davis and Barnard, 2003).

Beach nourishment on Treasure Island began in the late 1960s and represents one of the earliest nourishments along the Gulf Coast of Florida. Renourishment has been ongoing since then and has become more regular since the year 2000, occurring every 3–6 years. Nourishment sections were concentrated at the northern and southern ends of the island. The nourishment project at Sunshine Beach at the north end is short, with a project span of 0.5 km and various nourishment widths depending on the prenourishment state of the beach. It was not included in this study. The Sunset Beach nourishment project at the southern end is 1.2 km long and is typically widened ∼30 m per nourishment cycle.

Borrow areas utilized for nourishment events are primarily from offshore deposits or ebb shoals. The 2006 and 2018 nourishments on Sand Key used sand from the eastern Egmont shoal at the mouth of Tampa Bay roughly 20 to 30 km south of the project sites. The 2012 nourishment used sand from offshore borrow areas with high contents of shell fragments, resulting in overall coarser sediment. The Treasure Island nourishment used sand from both east and west Egmont shoals, John’s Pass ebb shoal, and John’s Pass channel. Except for the 2012 nourishment, similar sediments were used in each nourishment cycle.

Bimonthly to quarterly beach profiles spaced ∼300 m alongshore were surveyed by USF-CRL from 2006 to 2022 along the two regularly nourished islands. The beach monitoring used R-monuments and their survey azimuths (roughly shore perpendicular), established by the Florida Department of Environmental Protection (FDEP). In total, 74 beach profiles on Sand Key (R-55 to R-124) and 17 on Treasure Island (R-127 to R-143) were surveyed over 17 years. The beach-profile monitoring covered the entire islands, including both nourished and nonnourished sections. The surveys encompassed the beach from the dune field or seawall to ∼−3 m water depth, or roughly 1 m above the depth of closure in this area (Royer, Wang, and Cheng, 2023; Wang and Davis, 1999). Although the seaward limit of the survey was shallower than the closure depth, the time-series profiles largely converged at the offshore limit. In addition to regular monitoring, surveys were also conducted directly after significant storm impacts. The survey followed a standard level-and-transit procedure using an electronic total survey station and a 4 m survey rod (Cheng, Wang, and Guo, 2016). All surveys were completed using the Florida State Plane West coordinate system in meters. The elevation was referenced to the North American Vertical Datum of 1988 (NAVD88), which is 8.2 cm above mean sea level in this area.

Earlier beach nourishment monitoring by USF-CRL, specifically from 1988 to 2005, followed a different procedure in that only the distance and elevation were measured along the survey azimuth, and the State Plane easting and northing locations were not collected before the availability of the precise real-time kinematic global positioning system (RTK-GPS). Furthermore, the profile survey only extended to roughly 1.5 m water depth, terminating mostly landward of the nearshore bar. The offshore portion of the profile was surveyed annually and separately from the beach-profile survey using a vessel-towed sled. This data set, although not directly comparable to the post-2006 surveys, was used for long-term nourishment performance analysis. Additional data before 1988 were obtained from various sources, including FDEP and the U.S. Army Corps of Engineers (USACE).

The Regional Morphology Analysis Package (RMAP), developed by the U.S. Army Engineer Research and Development Center (ERDC), was used to extract contour locations and calculate profile-volume changes. Profile-volume and contour-location changes were calculated using the postnourishment profile and a winter profile approximately 4 years after beach fill. The fixed time duration of 4 years after beach fill was chosen to maintain consistency among different nourishment cycles with different time intervals. Sand Key is typically nourished every 4 to 6 years, while Treasure Island is nourished mostly every 4 years.

A dry beach zone was defined for shoreline-contour determination with the goal of eliminating influences of tidal fluctuations during profile analysis. A +1.0 m contour was used as the seaward limit of the dry beach. This is the area of the beach that waves would not reach under average conditions. Based on individual characteristics of each profile and field observations, the seaward edge of the dune or seawall was identified and used as the most landward boundary for profile-volume calculation. This landward limit, excluding most of the densely vegetated dune field, was used because the dune is highly three dimensional, often with steep slopes. The level-and-transit profile-survey procedure used here did not provide adequate spatial resolution to accurately capture these three-dimensional features. The distance from the landward edge of the profile to the +1.0 m contour represents the dry beach width. The contour changes between the beginning and end of each nourishment were used to evaluate the performance of each nourishment.

Profile-volume changes were calculated to depict the erosional hotspots along the study area for the 2018 nourishment at Sand Key and Treasure Island. A 0 m elevation and a −3 m elevation (NAVD88) were used to analyze the profile-volume change at the roughly mean sea-level shoreline and to the depth of closure, respectively.

The most up-to-date version (2022) of the Coastal Modeling System (CMS), specifically the CMS-Wave and CMS-Flow models, was used to depict processes that cause erosional hotspots. The CMS model, developed by the Coastal Inlets Research Program (CIRP) at the USACE, is a collection of numerical models for simulating flow, waves, sediment transport, and morphology change in coastal areas (Buttolph et al., 2006; Larson, Camenen, and Nam, 2011; Lin, Demirbilek, and Mase, 2011; Reed et al., 2011; Sánchez and Wu, 2011; Sánchez et al., 2014; Wu, Sánchez, and Zhang, 2011).

The CMS is capable of computing wave field, current field, sediment transport, and morphology change. Since the goal of this study was to depict oceanographic processes that cause erosional hotspots, sediment transport and morphology change were not computed due to large uncertainties associated with calculating sediment transport by breaking waves and complications induced by sediment availability. Wave field, flow field driven by tides only, and flow field driven by combined wave and tides, i.e. longshore current, were computed and analyzed.

The construction of the numerical modeling grid used a combination of bathymetry data in the nearshore area collected by USF-CRL and National Oceanic and Atmospheric Administration (NOAA) bathymetry data in the offshore area (https://www.ncei.noaa.gov/maps/bathymetry/). USF-CRL conducts annual surveys of ebb shoals, with the most recent being in the summer of 2023. USF-CRL also conducted detailed surveys of the back-barrier bays from 2014 to 2018. An offshore survey extending to roughly 1 km seaward of the shoreline was conducted by USF-CRL after Hurricane Ian in 2022. These data, in addition to the August 2023 beach-profile data, were used to construct the CMS model grid.

The present beach condition along most of Sand Key and Treasure Island is largely controlled by repeated beach nourishment. This is clearly illustrated by pre- and postnourishment example photos (Figure 2). Before the island-wide beach nourishment, the seawall and associated riprap were exposed at the waterline and were susceptible to constant wave attack (Figure 2, left panels). The beach nourishment successfully restored a relatively wide beach and a dune field (Figure 2, right panels). Presently, almost all seawalls are buried under the restored beach or dune, with almost no sections exposed at the shoreline.

The total sand volume placed during each nourishment event and volume placed per year at Sand Key and Treasure Island are summarized in Figure 3. The nourishment density varied substantially alongshore. The 2.1-km-long North Sand Key received much more sand per unit length of beach than the rest of the nourished segments. There is a gap in the nourishment between North Sand Key and the rest of the nourished sections. The Sand Key nourishment is divided into five segments based on the nourishment and coastal conditions (Figure 1). The design of each renourishment at each segment remained similar.

Similar total volumes of sand were placed along the entire Sand Key in 1988 and 1998, 1.87 million and 2.00 million m³, respectively. The amounts of sand placed in 2006 and 2012 nourishments were less than the previous nourishments, 1.30 million and 0.96 million m³, respectively. The 2018 total nourishment volume had a slight increase to 0.99 million m³. Since the nourishment time intervals are different, the volume per year was also calculated from the end of a specific nourishment to the beginning of the next nourishment. The 1998 nourishment had the highest volume per year of 250,000 m³/y. The four other nourishments ranged from 159,000 m³/y in 1988 to 217,000 m³/y in 2006.

The Treasure Island nourishment is concentrated along the northern and southern ends of the island. The very short 0.5 km Sunshine Beach project at the north end was not included in this study. The 1.2 km Sunset Beach project has been nourished since 1969. Projects before 1986 were not well documented. The largest volume of sand of 421,000 m³ was placed in 1986. Only a 39,000 m³ volume was placed in 1996, followed by 267,000 m³ in 2000. The total nourishment volume was 172,000 m³ in 2004; 81,000 m³ in 2006; 96,000 m³ in 2010; 178,000 m³ in 2014; and 183,000 m³ in 2018. Compared to the Sand Key projects, the volume of sand placed per year on Sunset Beach varied substantially. Excluding the 1996 nourishment, the volume per year ranged from 24,000 to 140,000 m³/y.

Since 1996, Sunset Beach has been nourished every 4 years, except for the emergency nourishment in 2006, as compared to every 6 years on Sand Key since 1999. The influence of the different placement volumes was analyzed. The temporal variation of Treasure Island’s placement volume was greater than that at Sand Key (Figure 3).

Five nourishment cycles on Sand Key and six cycles on Treasure Island were examined. Because the beach nourishment was systematically monitored by USF-CRL since 2006, the performance of the three nourishments in 2006, 2012, and 2018 on Sand Key and four nourishments on Treasure Island in 2006, 2010, 2014, and 2018 were analyzed in more detail. The nourishments before 2006 are discussed separately due to different monitoring schemes.

North Sand Key is a chronic erosional hotspot. More sand was placed along this section compared to all other sections in the study area. Profile R-61 is used as an example to illustrate time-series beach changes in this section (Figure 1). The most recent 2018 nourishment shifted the shoreline farther seaward compared to the 2006 and 2012 nourishments (Figure 4A). However, at the end of the three nourishments, the beach profile retreated to a similar position. Hurricane Hermine impacted the study area near the end of the 2012 nourishment cycle and eroded the already depleted beach. For this section, only one nourishment was conducted before 2006 (in 1999). The performance of the 1999 beach nourishment is illustrated in Figure 5A. There was no subaerial beach before the first nourishment in 1999, with a seawall and riprap exposed directly at the shoreline. The 1999 nourishment widened the beach to a similar extent compared to the following three nourishments. At the end of 1999 nourishment, the beach retreated further landward as compared to the following nourishments. However, compared to the prenourishment conditions, the first nourishment in 1999 maintained a beach, although moderately narrower beach than the following cycles.

Directly south of the North Sand Key section, Belleair Shores represents a gap in the nourishment. Profile R-69 provides a representative example to illustrate beach changes in this section (Figure 4B). Minimal changes occurred on the dry beach in terms of volume or beach width. This is true for all four nourishment events since 1999 with considerable different grain sizes of fill material (Figure 5B). In contrast, the subaqueous portion of the profile showed an accretionary trend. During the four cycles, the widths of both the dry and intertidal portions of the beach remained similar, while the nearshore area became shallower.

Different from North Sand Key, which extended the dry beach seaward up to 70 m, the rest of the Sand Key nourishment widened the dry beach roughly 27 m. Profile R-74 is used here as an example to illustrate beach changes in the Indian Rocks section (Figure 4C). For the three recent nourishments since 2006, after each nourishment, the beach retreated to a similar position. More sand was placed during the 2018 nourishment; however, the beach still retreated to similar locations as compared to previous nourishments after 4 years. Different from the North Sand Key case, the Indian Rocks beach was nourished twice before 2006. A much wider beach was maintained as compared to the prenourishment 1990 condition (Figure 5C). At the end of the first nourishment in 1998, or 8 years after, the beach retreated roughly 8 m farther landward than the subsequent nourishment. Similar to the North Sand Key case, except the first nourishment, the beach retreated to a similar location after the following nourishments.

The Headland segment was also extended roughly 27 m seaward during the recent three nourishments in 2006, 2012, and 2018. Profile R-84 is used as an example to illustrate beach-profile changes in this segment (Figure 4D). After each nourishment, the beach retreated to a similar position. Hurricane Hermine caused more erosion along the headland stretch as compared to other segments of Sand Key (Figure 4). However, the intertidal zone of the profile was roughly 20 m wider after Hurricane Hermine than the 1998 and 2005 profiles, which represent the end of the first and second nourishments, respectively (Figure 5D). For the previous two nourishments in 1990 and 1999, the beach retreated to a similar location as compared to the following three nourishments. The beach was constructed much wider during the 1990 nourishment than all other nourishments (Figure 5D). The postnourishment beach remained much wider than the prenourishment beach before 1990, regardless of being at the beginning or end of a nourishment cycle.

For the South Sand Key segment, the beach widened roughly 27 m during each nourishment. Using profile R-106 as an example (Figure 4E), the end-of-cycle profile in 2010 retreated further seaward than the two following nourishments. Profile R-106 is located just 300 m north of the end of the entire Sand Key nourishment project. Longshore spreading may contribute significantly to profile changes. Before the regular nourishment started in 1988, the seawall was exposed at the shoreline (Figure 5E). Much less sand was placed during the first nourishment in 1988. After 10 years, the seawall was no longer exposed. More sand was retained at the end of the second nourishment in 2005. The following three nourishments were able to maintain a wider beach.

Sunset Beach, located along the southern section of Treasure Island, is a persistent erosional hotspot. Unlike Sand Key, Treasure Island has been nourished four times since 2006. Using profile R-139 as an example, the beach profile was constructed further seaward during each nourishment (Figure 4F). However, the entire profile retreated further landward at the end of each cycle. This suggests the erosion rate has increased over time. Before 2006, Sunset Beach had been nourished many times dating back to 1969. A narrow beach has been maintained since the late 1990s (Figure 5F).

The example profiles discussed here represent one location along each segment. Considerable alongshore variation occurred at each segment. Figure 6 depicts the dry beach width taken at a +1 m contour level. These charts quantify the dry beach width change between the beginning and end of the most recent three nourishments cycles (i.e., 2006–20, 2012–16, 2018–22). The contour change calculations were not conducted for nourishment cycles before 2006.

The North Sand Key segment represents an erosional hotspot with much greater changes than other segments (Figure 6A). Profile R-61 (Figure 4A), as discussed above, represents roughly the largest changes, although other profiles show a similar trend but with a smaller magnitude. The dry beach width changes range from about 10 m at R-58 to nearly 60 m at R-61 (Figure 6A). The beach change during the second nourishment cycle was greater than the other two cycles due to Hurricane Hermine’s impact at the end. The average beach width change, excluding R-58 and R-65A at both ends, was −26 m, −49 m, and −35 m in 2006, 2012, and 2018, respectively. Overall, although there was a modest alongshore variation, the repeated beach nourishment did not result in a progressively wider dry beach. The beach retreated to generally similar locations at the end of each nourishment.

Along the nonnourished Belleair Shores segment, the dry beach width changes ranged from a roughly 7 m gain to a 10 m loss, both occurring at R-67 (Figure 6B). The dry beach width widened during the 2006 nourishment cycle but showed loss during the 2012 and 2018 cycles. The average beach width change, excluding R-66 and R-71 at both ends, was +4 m, −8 m, and −8 m in 2006, 2012, and 2018, respectively. Overall, the nonnourished section, despite being located between two nourished sections, did not gain a persistent amount of sand on the dry beach.

Along the Indian Rocks segment, the width of the dry beach remained fairly uniform alongshore (Figure 6C). The dry beach width loss ranged from 2.5 m to 57 m, both occurring at R-72, indicating a substantial temporal variable end effect. The average beach width loss, excluding R-72 and R-80 at both ends, was 8 m, 19 m, and 24 m in 2006, 2012, and 2018, respectively. The third nourishment cycle widened the shoreline the furthest but experienced the largest beach width loss. Overall, this section experienced an increasing trend in beach width loss over the three nourishments.

The Headland segment protrudes further seaward than all other segments (Figure 1). At the end of each cycle, the beach profiles retreated to a similar shoreline position. The impact of Hurricane Hermine caused more erosion along this section compared to all other sections (Figure 4D). Unlike the three sections to the north, there was considerable longshore variation in beach width (Figure 6D). Profiles R-86 and R-87 were not nourished during the 2018 cycle due to landownership issues and saw a gain in beach width at the end of the cycle due to the longshore spreading of adjacent nourished profiles. Along this section, the dry beach width losses ranged from 3.5 m to 29 m (Figure 6D). The average beach width loss, excluding R-82 and R-93 at both ends, as well as R-86 and R-87, was 15 m, 22 m, and 20 m in 2006, 2012, and 2018, respectively. Overall, the dry beach width change was greater during the 2012 and 2018 cycles than during the 2006 cycle. The large loss in 2016, or the end of the 2012 nourishment cycle, was related to the distal passage of Hurricane Hermine.

South Sand Key has a SW-facing shoreline (Figure 1). Considerable longshore variation occurred along this section (Figure 6E). The example profile at R-106 (Figures 4 and 5) is located near the south end of the Sand Key nourishment project. The performance of this section is influenced by a shore-parallel breakwater in the vicinity of R-101. The beach was not nourished near the breakwater because the intertidal beach extends almost to the structure since its construction in 1985. The breakwater induced considerable longshore variation. The average beach width loss, excluding R-94 and R-107 at both ends, was 25 m, 22 m, and 17 m in 2006, 2012, and 2018, respectively (Figure 6E).

The Sunset Beach segment on Treasure Island represents an aggressive erosional hotspot. Four beach nourishments were conducted since 2006. Except at the north end of the nourishment, at R-136, beach changes were relatively uniform alongshore (Figure 6F). The average beach width loss, excluding R-136 and R-141 at the two ends, was 24 m, 20 m, 34 m, and 28 m in 2006, 2010, 2014, and 2018, respectively, illustrating a general increasing trend over time.

Three erosional hotspots can be identified based on time-series beach changes (Figures 7 and 8). For the convenience of discussion, an erosional hotspot is defined qualitatively here based on two characteristics: (1) large amounts of change occurring during each nourishment cycle, and (2) greater change at the section as compared to the changes at adjacent sections.

The North Sand Key section is an aggressive erosional hotspot. Figure 7A illustrates changes at profile R-61 during the 2018 nourishment cycle. The entire profile shifted landward during the 4 year period with much more change during the first year. The 2018 nourishment at this section extended substantially further seaward as compared to the 2006 and 2012 nourishments (Figure 4A). The study area was influenced by distal passages of Hurricane Eta in 2020 and Hurricane Ian in 2022. These factors had a significant influence on the profile change. Figure 7B illustrates the profile change during the 2012 cycle. The landward shift of the entire profile is more uniform temporally than the 2018 cycle. Figure 8 (upper panel) illustrates the profile-volume change along the entire Sand Key. For the North Sand Key section, the profile-volume change above the entire profile (−3 m NAVD88) is much greater than the volume change above the shoreline (0 m NAVD88). The landward shift of the entire profile is reflected in the similar alongshore pattern of the above two profile-volume change values. This landward shift of the entire profile was caused by a gradient in longshore sediment transport.

The Headland section represents another erosional hotspot, although not as aggressive as North Sand Key based on the magnitude of profile-volume change (Figure 8, upper panel). Some of the sand that was eroded from the dry beach and intertidal area was deposited on the nearshore sand bar (Figure 7C). This is also reflected in the different spatial pattern of the volume change above the entire profile and shoreline. This observation suggests that cross-shore sand transport as part of the postnourishment profile equilibration played a significant role in addition to longshore-transport gradient.

The Sunset Beach section on Treasure Island is an aggressive erosional hotspot. Figure 7D illustrates changes at profile R-139 during the 2018 nourishment cycle. The entire profile shifted landward progressively during the 4 year period, although the magnitude of change during the first year was much larger. The landward shift of the entire profile is reflected in the similar spatial pattern of the profile-volume change along this section (Figure 8, lower panel), reflecting erosion caused by a persistent gradient of longshore sediment transport.

A coupled wave-and-current numerical model was applied to compute the nearshore wave conditions and longshore current velocities, with the goal of quantifying the processes that are responsible for erosional hotspots. Specifically, the numerical model was applied to identify alongshore gradients in nearshore wave height and longshore current velocity. Although the CMS model has the capability, computation of nearshore sediment transport and morphology change was not included because of the limited understanding and subsequently large uncertainties associated with computing nearshore sediment transport, in addition to complications due to sediment availability.

Accurate computations of flow and wave fields are controlled by accurate representation of bathymetry (Figure 9). The offshore bathymetry is quite complicated, with numerous shallow ridges extending at an oblique angle to the shoreline. These bathymetric characteristics would influence wave propagation patterns. The ebb shoals at Clearwater Pass and John’s Pass are well captured. The complex bathymetry in the back-barrier bay, e.g., the Intercoastal Waterway, which has significant influence on flow patterns (Wang, Beck, and Roberts, 2011), was also well captured.

The model domain extends from roughly 5 km north of Clearwater Pass to just north of the main entrance of Tampa Bay (Figure 9), including not only Sand Key and Treasure Island, but also several adjacent barrier islands (Figure 1). The seaward boundary is about 20 km from the Sand Key headland at a water depth of roughly 13 m. In addition to the three inlets that border the two studied barrier islands, Sand Key and Treasure Island, two more inlets to the south were also included in the model domain. The back-barrier bays that are served by these five inlets were also included. In order to efficiently represent the large model domain, a telescoping model grid was constructed, with smaller 15 × 15 m grid cells at the inlets, beaches, and nearshore, 30 × 30 m grids at ebb shoals and back bays, and larger 60 × 60 m and 120 × 120 m cells in the offshore area.

Since the main goal of the numerical modeling effort was to quantify processes that induce erosional hotspots, schematic model runs based on statistical wave conditions were conducted. Computed hourly wave data from 1980 to 2022 were extracted from the Wave Information Study (WIS; https://wisportal.erdc.dren.mil/#) developed by USACE at station ST3266 with a depth of 13 m, which coincides with the seaward boundary of the model domain. The 43 year wave data were organized into sixteen 22.5 degree angle brackets. Offshore-directed waves were not included because they should have minimal influences on nearshore processes. Fifteen-day predicted tides, representing one neap-spring cycle, at the NOAA station CWBF1 near Clearwater Pass were extracted and used in the flow simulation.

The flow computation was calibrated and verified based on measurements conducted at John’s Pass and Blind Pass, respectively (Figure 10). The measurements were conducted in 2014 for an inlet management study at the two inlets (Horwitz, 2017). The friction coefficient (Manning’s number, n) was the only parameter adjusted during calibration. Three n values, 0.015, 0.025, and 0.035, were used to calibrate the computed flow velocity in comparison with the measured velocity. The default CMS-Flow Manning’s (n) value of 0.025 provided the best fit among the three values tested and was used in the production model runs. The computed flow velocity was verified at Blind Pass.

The input schematic wave conditions developed based on the statistical analysis of WIS wave data are summarized in Table 1. The numerical modeling focused on energetic conditions. In addition to using the statistical waves for wave-only simulation, these schematic wave conditions were also applied for coupled wave-flow runs. For the highly oblique incident waves, i.e., NNW, NW, SW, and SSW, the peak wave period was reduced by 1 to 2 seconds (as listed in Table 1) to reduce the degree of wave refraction and produce a faster longshore current velocity. It is acknowledged here that the computed longshore current velocity using the schematic waves was therefore artificially enhanced. However, the slightly modified wave conditions were not unrealistic. For the shore-perpendicular waves, i.e., WNW, W, and WSW, the wave period was not altered.

Since the longshore sediment transport rate is a strong function of incident wave angle, the net longshore transport along the two studied barrier islands can be significantly influenced by the 65° shoreline orientation change around the headland because it has significant influence on the breaker angle. South of the headland, where the shoreline faces roughly southwest, the S and SSW waves drive northward longshore transport, while N and NNW waves drive southward longshore transport. The SW waves approach the shoreline perpendicularly. The net southward longshore transport is supported by the statistical wave analysis of the 43 year WIS data set (Table 1), where the NW and NNW waves are higher than the S and SSW waves in terms of both average and storm wave heights. Although the S waves have a higher frequency of occurrence, the wave height is significantly smaller than the NW and NNW waves. Since the longshore transport rate is proportional to the wave height to the 2–2.5 power (Equations [3] and [4]), this small difference in wave height can signify a large variance in longshore sediment transport rate.

For the sections of the beach north of the headland, i.e. Indian Rocks and North Sand Key, the shoreline faces roughly west-northwest. S, SW, and SSW incident waves would drive northward longshore sediment transport, while NW and NNW waves would drive a southward sediment transport. Different from the shoreline south of the headland, SW waves, which have higher average and storm wave heights than the S and SSW waves, play a significant role in northward longshore transport. Therefore, the balance between northward and southward longshore sediment transport is not as apparent.

The above general assessment is qualitative based on statistical analysis of offshore wave conditions. The wave modeling results provide quantitative information and an improved spatial resolution. Figure 11 illustrates the wave field associated with NW and SW incident storm waves for the entire region. For the NW incident waves, the longshore transport is to the south along the entire coast. For the SW incident waves, longshore transport north of the headland is toward the north, but the longshore transport south of the headland is toward the south due to the significant wave refraction and the shoreline orientation change. The wave refraction (Figure 11B) results in nearly E-W wave propagation. However, the shoreline is facing southwest, therefore easterly propagating waves would induce southward longshore transport. The wave modeling results indicate that the above general analysis based on the statistical offshore wave angles failed to reveal the southward longshore transport for the SW incident waves.

For the SW incident waves, a divergent zone of longshore sediment transport is identified along the south flank of the broad headland. Due to the shoreline orientation change, longshore transport is to the north along the northern portion of this beach section, based on the angle between nearshore waves and the shoreline, while longshore transport is to the south along the southern portion of this section. A closer examination of the wave field at the southern end of the headland, and just to the south of where the shoreline orientation changes from facing roughly west to southwest, is shown in Figure 12. A significant change in longshore transport occurs based on the computed wave angle relative to the shoreline. At the northern end of Figure 12, the longshore transport appears to be directed to the north associated with the SW incident waves. Along the southern section, the longshore transport is directed to the south due to the significant shoreline orientation change and the refraction of the SW incident waves.

The above divergence of longshore sediment transport caused more erosion along those sections than that along the adjacent sections (Figures 6E and 8, profiles R-96 to R-103). However, this erosional hotspot was not further analyzed due to the presence of a shoreline breakwater at R-101. North of the breakwater, beach-profile volume loss was comparable to the headland erosional hotspot. Near the breakwater, a volume gain occurred, while volume loss resumed south of the breakwater. The breakwater was constructed in 1985, just before the large-scale beach nourishment project along this section in 1988. Before the construction of the breakwater, this section of the beach was rather depleted of sand (Figure 13). This was likely the reason for the construction of the breakwater. The nourishment performance along this section of the beach is significantly influenced by the presence of the breakwater and the fact that it is located along the southern and downdrift end of the nourishment.

The processes causing the three erosional hotspots at the North Sand Key, Headland, and Sunset Beach sections were examined based on both wave and current fields. Figure 14 illustrates an example of the computed flow field from the coupled CMS-Flow and CMS-Wave simulation under the NW incident storm wave condition. Near the tidal inlets, the flow field is controlled by both tides and waves. While along the beach, away from the inlets, the longshore current is mostly driven by breaking waves.

The North Sand Key section represents an erosional hotspot, which receives more than twice the dry beach width of sand placement as compared to the rest of the sections on Sand Key. Figure 15A,C illustrates the wave field associated with the storm waves approaching from the NW and SW, respectively. For the NW incident waves, the sheltering by the Clearwater Pass ebb shoal results in a southward-increasing wave height. Wave refraction around the ebb shoal results in more shore-perpendicular waves closer to the inlet, while the oblique wave angle increases toward the south (Figure 15A). This southward-increasing wave height and wave angle result in significant southward-increasing longshore current (Figure 15B), with the largest gradient between roughly R-60 and R-62, where the most erosion occurs (Figure 8). Based on Equation (2), the increasing longshore current (u) would result in an increasing rate of longshore sand transport, while the increasing wave height would lead to increasing sediment concentration (c). Because the erosional hotspot is over 3 km south of the inlet, the gradient in longshore current velocity is not significantly influenced by tidal flow.

For the SW incident waves, the sheltering by the protruding broad headland and associated offshore bathymetry (Figure 9) results in a northward-decreasing wave height (Figure 15C), although the gradient is smaller than the NW case. However, the incident wave angle increases toward the north. The computed longshore current is directed toward the north. A gradient can be identified at the erosional hotspot caused by increased incident wave angle, although the wave height decreases slightly (Figure 15D). The gradient in longshore current velocity is not as significant as the NW incident waves. Since the beach to the south, i.e. Belleair Shores, has never been nourished and has been depleted of sand since the late 1990s, it is not capable of providing sand to the North Sand Key section under conditions of SW incident waves. Therefore, the small gradient in northward longshore current can have exaggerated influence at the erosional hotspot due to the lack of sand supply from the south.

In summary, the numerical modeling results indicate that along the North Sand Key erosional hotspot, a large gradient of southward sand transport occurs when waves are approaching from the northerly direction. When waves are approaching from the southwesterly direction, a lack of sand supply from the nonnourished beach to the south combined with a modest gradient in longshore current result in sand leaving this section of the beach and being transporting north. Therefore, sand has a tendency of leaving this section under both northerly and southerly incident waves, leading to the aggressive erosional hotspot.

The protruding headland is exposed to higher waves as compared to the rest of the coast. Under energetic NW waves, a strong longshore current is computed (Figure 16). However, the longshore current is rather uniform along this section. Therefore, although the waves are higher and the longshore current is stronger under storm conditions, the gradient in alongshore sediment transport is not as large as that in the North Sand Key section. Furthermore, the beaches directly to the north (Indian Rocks section) and to the south (South Sand Key section) have been nourished regularly since the 1990s, providing sand supply in both directions of longshore transport. Because of these factors, the headland erosional hotspot is not as aggressive as those at North Sand Key and Sunset Beach (Figure 8). However, if the present nourishment cycle were to be discontinued, it is reasonable to deduce that the sand supply from north and south would become depleted, and this erosional hotspot may return to the prenourishment case (Figure 2, upper panel).

The bathymetry offshore Treasure Island is complicated (Figure 17). Before the opening of John’s Pass by a storm in 1848, Blind Pass was the main tidal inlet serving this portion of Boca Ciega Bay (Beck and Wang, 2019; Horwitz, 2017). The opening of John’s Pass and subsequent capture of most of the tidal prism led to southward migration of Blind Pass. The large relic Blind Pass ebb shoal and its evolution associated with the southward migration, in addition to the nearshore dredge pit from the 1969 Sunset Beach nourishment (Figure 17), resulted in the complex bathymetry that has significant influence on the nearshore wave field.

The simulated wave field associated with NW incident storm waves is strongly influenced by the complex bathymetry (Figure 18A). Shoaling of the long-period and high storm waves occurred over the shallower water at the John’s Pass ebb shoal and the relic Blind Pass ebb shoal. The large John’s Pass ebb shoal also provides a rather extensive wave shadow zone in the middle of Treasure Island for the NW-approaching storm waves. The wave shadow zone terminates roughly at the northern boundary of the Sunset Beach erosional hotspot. In addition, the wave angle along the shoreline changes from roughly shore-perpendicular directions in the middle of Treasure Island to a more oblique direction along Sunset Beach due to decreasing wave refraction. These spatial variations in wave height and wave angle result in a significant longshore current gradient along Sunset Beach (Figure 18B). Because Sunset Beach is located over 3 km south of John’s Pass, the modeled longshore current is controlled mostly by incident wave conditions, rather than by tides.

For SW incident storm waves along Treasure Island (Figure 18C), due to the significant wave refraction and the southwest-facing shoreline, the simulated longshore current is directed to the south, driving southward longshore transport (Figure 18D). A mild gradient in longshore current is simulated. The wide beach to the north as well as the shallow John’s Pass and its attachment point provide adequate sand supply for the southward longshore transport. Therefore, this aggressive erosional hotspot is caused by a persistent gradient in southward longshore transport instead of depletion of sand supply. The transport gradient is greater under northerly approaching waves than under southerly approaching waves.

The five sections on Sand Key and one on Treasure Island represent a variety of beach conditions including three erosional hotspots, one narrow beach section that has never been nourished, and two typical erosive beaches. The factors controlling the erosional hotspots as well as the performance of consecutive cycles of renourishment are discussed next.

The North Sand Key erosional hotspot is caused by a persistent gradient in longshore sediment transport. Under northerly approaching waves, the stronger southward longshore current along this section than that along the section to the north results in more sand leaving this section toward the south. This gradient is caused by wave refraction induced by Clearwater Pass’s ebb shoal. Under southerly approaching waves, a modestly stronger northward longshore current combined with minimal sediment supply from the narrow beach to the south result in more sand leaving the hotspot toward the north. Therefore, the aggressive erosion at this hotspot is caused by the longshore sediment transport gradient that occurs under both northerly and southerly approaching waves.

The erosion caused by the longshore transport gradient explains the measured landward shift of the entire profile while the profile maintains a relatively consistent shape (Figure 7A,B). This happens for all beach nourishment cycles. Since the sand that was placed onto the beach during each nourishment moved alongshore and out of the section, the repeated sand placement has not fundamentally changed the process, e.g., patterns of wave shoaling and breaking. At the end of the recent three nourishments, on average between 60% and 80% of the sand that was added to the dry beach was eroded (Figure 19). In the middle of this section at profile R-61, over 90% of the placed sand was eroded at the end of each nourishment. The present nourishment design has been successful in maintaining a roughly 10 m dry beach. The fact that the beach retreated to similar positions 4 years after each nourishment indicates that the nourishment did not alter the long-term trend of sand deficit. However, without consistent renourishment, the existing beach and dune will be completely eroded to prenourishment conditions (before 1998), re-exposing the existing seawall and riprap (Figure 5A). This may happen in 4–6 years, or one nourishment cycle.

The erosional hotspot at the headland is caused by higher nearshore waves and stronger longshore current as compared to the beaches along the two flanks. This explains why Hurricane Hermine caused more erosion, particularly on the dry beach, along this section than along adjacent sections. No significant alongshore gradient in wave height and longshore current velocity is simulated by the model along this section. The beach profiles at the end of each cycle retreated to similar locations (Figure 4D). Unlike the North Sand Key and Sunset Beach sections, a substantial nearshore bar exists (Figures 4D and 7C). The sandbar tends to move onshore during the summer and offshore during the winter (Roberts and Wang, 2012). The offshore sandbar movement can also be driven by energetic events (Cheng and Wang, 2018; Cheng, Wang, and Smith, 2015). A substantial amount of the sand placed on the beach moved offshore and onshore and stayed within this section. The beaches to the north and south of the headland have been repeatedly nourished, providing alongshore sand supply to this section. The headland experiences considerable longshore variations in terms of sediment losses (Figures 6D and 8). This may be caused by spatial variations of offshore bathymetry. Similar to the North Sand Key erosional hotspot, the repeated nourishment did not fundamentally alter the sand deficit.

The Sunset Beach erosional hotspot is caused by a persistent southward-increasing wave height and longshore current under both northerly and southwesterly incident waves. Under northerly approaching waves, the wave shadowing caused by John’s Pass ebb shoal resulted in a substantial southward-increasing wave height and longshore current. Under southwesterly approaching waves, due to significant wave refraction and a southwest-facing shoreline (Figure 11), the longshore current is still directed to the south with a mild southward-increasing gradient (Figure 18C,D). The erosion by the longshore transport gradient explains the landward shift of the entire profile while the profile shape remains consistent (Figures 4E and 7D).

Since 2006, the dry beach width decreased progressively after each nourishment, even though the shoreline was extended further seaward during each nourishment (Figure 4E), suggesting an increasing erosion rate. At the end of each nourishment cycle, on average, about 85% of the sand that was added to the dry beach was eroded (Figure 20). In the southern section, almost 100% of the placed sand was lost. At the south end of this section, a terminal groin exists between profiles R140 and R141 (Figure 21). This groin has limited influence on the beach nourishment performance (Figure 20). Alongshore spreading of this short project caused some variation at the northern end, i.e. R-137 (Figure 20). Overall, the current design of the Sunset Beach nourishment is not adequate to maintain a stable dry beach. More sand and a wider beach are required for future nourishment.

The Belleair Shores segment is located between the nourished North Sand Key section to the north and the Indian Rocks section to the south. Substantial amounts of sand from the North Sand Key nourishment were transported alongshore, to both the north and south. The northward-moving sand is impounded by the long south jetty at Clearwater Pass, resulting in a wide accretionary beach. However, the southward longshore transport did not result in a dry beach gain at Belleair Shores. Some sand was gained in the subtidal zone, resulting in a wider and shallower nearshore platform. The dry beach, backed by a seawall, has been less than 10 m wide over the past 25 years (Figures 4B and 5B). The beach remains consistently narrow even directly adjacent to the nourishment project at both ends. The beach tends to be eroded to the seawall after energetic events and then recovers after each event. There has not been an overwhelming storm during the 17 year study period since 2006 that significantly overtopped the seawall. A more direct hit by a strong storm, such as Hurricane Michael in 2018 (Wang et al., 2020) or Hurricane Ian in 2022 (Wang, Royer, and Gutierrez, 2024), would induce catastrophic damage. After the first nourishment, the beach gained up to 5 m of dry beach width, which is within the range of seasonal variations. During the following two nourishment cycles, relatively small dry beach width and profile-volume loss were measured.

Overall, Belleair Shores did not directly benefit from adjacent beach nourishments, in terms of the dry beach width. There is no mechanism in place to impound sand that is being transported from the longshore spreading of the North Sand Key and Indian Rocks nourishment. Therefore, for a beach to benefit from adjacent beach fill, a transport gradient with more sand entering than leaving the section must exist. Such a gradient does not exist along Belleair Shores.

For the Indian Rocks and South Sand Key sections, all beach profiles retreated to similar locations at the end of each nourishment. A substantial nearshore bar exists. The sand bar varies seasonally and tends to move onshore during the summer season and offshore during the winter season. The offshore sandbar movement can also be driven by energetic events (Cheng and Wang, 2018; Cheng, Wang, and Smith, 2015; Roberts, Wang, and Puleo, 2013). These sections did not show an apparent and persistent longshore sediment transport gradient. A substantial amount of the placed sand moved offshore and onshore and stayed within the stretch of the beach.

A dry beach width between 5 and 25 m was maintained by the nourishment cycles. Significant longshore variation occurred along the South Sand Key section caused by the breakwater at R-101. The beach profile downdrift of the breakwater experienced the largest dry beach loss (Figure 6E). Overall, these two sections of the beach were successfully maintained by the nourishment. The current nourishment design is adequate to compensate for the sand deficit, although the sand deficit would remain if nourishment were to be discontinued.

In summary, along the 12 km Sand Key project span, a relatively stable beach and dune system was maintained by repeated nourishment. The existing beach and dune system is significantly wider than the beach before the repeated nourishments (Figure 5). The 1.8 km gap at Belleair Shores between North Sand Key and Indian Rocks did not gain a significant amount of sand. There is no mechanism there to impound the sand transported through longshore spreading of the adjacent nourishment projects. Along Sunset Beach on Treasure Island, the beach retreated further landward during each nourishment, even though the shoreline was constructed further seaward. This suggests that the erosion rate is increasing over time, although the reason for this is not clear, and it is beyond the scope of this study.

For the North Sand Key and Sunset Beach erosional hotspots, the alongshore variation of wave height is caused by ebb shoals. For both cases, the ebb shoals are located to the north of the erosional hotspot. The shadowing of northerly approaching waves results in a southward-increasing wave height (Figure 18). Furthermore, wave refraction around the ebb shoal combined with the specific shoreline orientation result in a southward-increasing breaker angle. This southward-increasing wave height and breaking-wave angle result in a significant negative gradient in longshore currents at both the North Sand Key and Sunset Beach erosional hotspots. However, for southwesterly incident waves, the two hotspots responded differently. For North Sand Key, a mild northward-increasing longshore current was computed. The sediment transport and deposition pattern were also influenced by factors other than just wave and current conditions. A long inlet jetty exists along the northern boundary of Sand Key, leading to significant impoundment of the northward-moving sediment. South of the North Sand Key erosional hotspot, the non-nourished Belleair Shores section is severely depleted of sand. Therefore, the longshore current gradient becomes irrelevant due to the lack of sediment supply.

For Sunset Beach, the southwesterly approaching waves still drive a southward longshore current due to wave refraction and the southwest-facing shoreline orientation, but at a reduced magnitude as compared to the northerly approaching wave case. The beach to the north is quite wide due to active sand bypassing from John’s Pass ebb shoal. Therefore, there is an abundance of sand supply for the southward longshore sediment transport. This sand supply did not have significant influence on the behavior of the aggressive erosional hotspot, suggesting that the negative gradient in longshore transport is the dominant factor. In summary, the North Sand Key erosional hotspot is controlled by both a negative transport gradient and a depleted sediment supply from the south, while the Sunset Beach hotspot is controlled by a persistent negative longshore transport gradient despite abundant sediment supply.

Around the broad headland, the incident wave energy and longshore current velocity are greater than those at the adjacent sections. However, no significant alongshore gradient in wave height and alongshore current was simulated by the model. In addition, adequate sand supply appears to be available from adjacent beaches on both sides. Compared to North Sand Key and Sunset Beach, the headland erosional hotspot is not as aggressive. Therefore, elevated wave energy alone does not necessarily cause more erosion.

The following conclusions can be drawn from this study:

The repeated nourishment successfully maintained a minimum 10 m dry beach width along almost all nourished segments over nearly four decades, including the erosional hotspots.

The gap section in the Sand Key nourishment did not gain any significant dry beach width due to a lack of mechanism for retaining the sand from the longshore spreading from the adjacent nourishment projects.

The present nourishment cycles successfully compensated for the existing sand deficit but did not fundamentally alter the processes that are causing the deficit. The repeated nourishment served as a maintenance strategy.

The three erosional hotspots are caused by different processes. At North Sand Key, an aggressive transport gradient is caused by wave refraction around the Clearwater Pass ebb shoal under northerly approaching waves, while under southerly approaching waves, the transport gradient can be attributed to a depleted sand supply from the south. At the protruding Sand Key headland, the elevated erosion is caused by relatively higher waves as compared to adjacent stretches. At Sunset Beach, the southward-increasing wave height and longshore current occur under both northerly and southerly incident waves. The wave shadowing by the shallow John’s Pass ebb shoal results in a southward-increasing transport under northerly incident waves, while under southerly incident waves, a mild southward-increasing transport is caused by significant wave refraction and a southwest-facing shoreline. Sunset Beach has an abundance of sand supply from the north, but this did not fundamentally change the trend of erosion.

Generally, the formation of erosional hotspots along barrier island beaches is caused by alterations of wave and current fields by both natural and anthropogenic features. Natural influences include those caused by adjacent inlets, particularly ebb shoals, headland features, or certain bathymetry characteristics. Anthropogenic features include shoreline hardening, inlet structures, or nearshore dredging.

This study was funded by Pinellas County, Florida. We acknowledge numerous graduate students at USF-CRL for their assistance in field data collection.