- Email: [email protected]
Post-storm beach and dune recovery: Implications for barrier island resilience
Geomorphology 234 (2015) 54–63
Contents lists available at ScienceDirect
Geomorphology journal homepage: www.elsevier.com/locate/geomorph
Post-storm beach and dune recovery: Implications for barrier island resilience Chris Houser a,⁎, Phil Wernette a, Elizabeth Rentschlar a, Hannah Jones b, Brianna Hammond a, Sarah Trimble a a b
Department of Geography, College of Geosciences, 810 O&M Building, Texas A&M University, College Station, TX 77843-3147, United States Department of Geology, Carleton College, Northfield, MN 55057, United States
a r t i c l e
i n f o
Article history: Received 17 April 2014 Received in revised form 6 December 2014 Accepted 10 December 2014 Available online 24 January 2015 Keywords: Barrier island Post-storm recovery Resilience Hurricane
a b s t r a c t The ability of beaches and dunes to recover following an extreme storm is a primary control of barrier island response to sea-level rise and changes in the frequency and/or magnitude of storm surges. Whereas erosion of the beach and dune occurs over hours and days, it can be years to decades before the beach and dune are able to recover to their pre-storm state. As a consequence, there are numerous descriptions of near-instantaneous beach and dune erosion due to storms, the immediate onshore transport of sand, and the initial phases of beach and dune recovery following a storm, but a paucity of data on long-term beach and dune recovery. A combination of previously published data from Galveston Island, Texas and new remotely sensed data from Santa Rosa Island, Florida is used in the present study to quantify the rate of dune recovery for dissipative and intermediate beach types, respectively. Recovery of the dune height and volume on Galveston Island was observed within two years following Hurricane Alicia (1983) and was largely complete within six years of the storm, despite extensive washover. In contrast, the dunes on Santa Rosa Island in Northwest Florida began to recover four years after Hurricane Ivan (2004), and only after the profile approached its pre-storm level and the rate of vegetation recovery (regrowth) was at a maximum. Results show that complete recovery of the largest dunes (in height and volume) will take approximately 10 years on Santa Rosa Island, which suggests that these sections of the island are particularly vulnerable to significant change in island morphology if there is also a change in the frequency and magnitude of storm events. In contrast, the areas of the island with the smallest dunes before Hurricane Ivan exhibited a rapid recovery, but no further growth in profile volume and dune height beyond the pre-storm volume and height, despite continued recovery of the largest dunes to their pre-storm height. A change in storm magnitude and/or frequency is a potential threat to barrier island resilience, particularly for those sections of the island where dune recovery has historically taken the longest time. Further study is required to determine how and why dune recovery varies for the dissipative and intermediate beaches of Galveston Island and Santa Rosa Island, respectively. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The vulnerability of a barrier island to extreme storms depends on the elevation of the total water level (tide + storm surge + wave runup) relative to the geometry of the coast, which is largely dependent on the height and alongshore extent of the foredune (Thieler and Young, 1991; Sallenger, 2000; Morton, 2002; Nott, 2006; Houser and Hamilton, 2009). Storm impact can have a range of visible impacts, from minor scarping at the base of the dune to overwash and/or breaching, when dune heights are relatively small compared to the storm surge (Sallenger, 2000; Hesp, 2002). Resiliency of a barrier island, or the ability to return to its previous equilibrium state (Woodroffe, 2007), is dependent on the rate of post-storm dune recovery. This rate is dependent on the transfer of sediment from the nearshore to the beach, which can occur through the landward migration and welding of the innermost nearshore bars, the alongshore migration of sand ⁎ Corresponding author. E-mail address: [email protected] (C. Houser).
http://dx.doi.org/10.1016/j.geomorph.2014.12.044 0169-555X/© 2015 Elsevier B.V. All rights reserved.
waves, the recolonization and expansion of dune-building vegetation, and/or aeolian transfer of sediment from the beach to the backshore and recovering dune. Depending on the magnitude and duration of the storm surge, erosion of the beach and dune occurs over hours and days, whereas recovery of the nearshore, beach and dune can take years to decades (Lee et al., 1998). The differential timescale of erosion and recovery makes barrier island response to sea-level rise dependent on the sequence of storm events and vulnerable to widespread erosion and washover when storms occur in quick succession (see Houser and Hamilton, 2009). The areas of the island that could exhibit significant morphological change and a transition to a new and possibly irreversible equilibrium state in the future are not those with already limited dune development and ineffective dune-building vegetation (low islands; Duran and Moore, 2013), but those sections of the island where the development of large dunes depends on the recovery of beach and the recolonization of dune-building vegetation (high islands; Duran and Moore, 2013). Dune recovery requires the transfer of sediment from nearshore to the beach and ultimately to the dune, assuming that dune-building
C. Houser et al. / Geomorphology 234 (2015) 54–63
vegetation is also present. Based on 10 years of monitoring data, Morton et al. (1994) tracked the beach and dune recovery following Hurricane Alicia (1983) along Galveston Island. Calculating dune heights from the published profiles provides a time series of dune recovery in which peak dune growth is observed ~3 years after the storm and the dune reaches the pre-storm height (of ~ 2 m) 4–5 years after the storm (Fig. 1). Growth of the dune follows a sigmoid curve consistent with the growth models used to quantify vegetation growth (see Hugenholtz and Wolfe, 2005a,b). Specifically, the data presented in Morton et al. (1994) and several survey transects from Priestas and Fagherazzi (2010) all follow the growth model of Verhulst (1838): dN N ¼ rN 1− dt K
ð1Þ
where N is a system attribute (i.e. dune height), r is the growth rate, t is the time elapsed since the last disturbance, and K is the upper boundary (asymptote) of dune growth (Verhulst, 1838). Only those transects from Priestas and Fagherazzi (2010) with pre-storm dune heights of ~2 m, similar to those of Morton et al. (1994), were used in this preliminary analysis. Integration of Eq. (1) gives: Nt ¼
KN o ðK−No Þe−n þ No
ð2Þ
where Nt is the height of the dune at time t, No is the initial height of the dune (t = 0) and e is the base of the natural logarithm. The dunes of Morton et al. (1994) and Priestas and Fagherazzi (2010) appear to follow the same logistic curve, but it is not clear why they exhibit similar recoveries given different environmental conditions. As described by Morton et al. (1994), the first stage of recovery begins immediately after the storm and can last a few weeks or up to a year, depending on the severity of the storm (Sallenger, 2000). This stage is characterized by berm reconstruction and steepening of the beach face. Specifically, sediment is returned to the beachface and the beach undergoes gradual accretion as the innermost bar migrates landward and welds to the beachface, leading to a steep beach ridge in reflective environments or a low-gradient berm in more dissipative environments. Landward migration of nearshore bars is driven by the waves as they shoal across the bar during fair-weather conditions (Elgar et al., 2001; Houser et al., 2006), with recovery and beach welding requiring several years following a large storm or multiple storms in succession (Lee et al., 1998). In lacustrine environments, the width of the backshore is closely tied to water levels, in which low water levels promote dune recovery and progradation, while high water levels allow even moderate storms to erode the foredune and reset the recovery (Saunders and DavidsonArnott, 1990). Following Morton et al. (1994), backbeach aggradation (around year 2 in Fig. 1) is largely dependent on deposition from swash events that
55
exceed the elevation of the beach ridge in intermediate to reflective environments, or through the landward migration of the nearshore bars in dissipative environments. Since storm winds capable of entraining sediment are usually accompanied by elevated water levels (Ruz and Meur-Ferec, 2004; Delgado-Fernandez and Davidson-Arnott, 2011) and precipitation (Keijsers et al., 2012), it is reasonable to assume that sediment only becomes available to the dunes when the backshore expands and allows for the development/recovery of a dune ramp (Christiansen and Davidson-Arnott, 2004). While aeolian transport is possible as soon as the upper-beach and any washover deposits become dry, the expansion of the backshore is required to increase the fetch length, which controls the amount of sediment exchanged from the beach to the dune (Davidson-Arnott, 1988; Davidson-Arnott and Law, 1990, 1996; Bauer and Davidson-Arnott, 2003; Houser, 2009). Sediment emplacement in the backshore and lengthening of the available fetch can either occur between storms, when winds tend to be below the transport threshold, or during storms, through the landward migration of subtidal and intertidal bars (Houser and Greenwood, 2005, 2007), the alongshore migration of sandwaves (Law and Davidson-Arnott, 1990; Davidson-Arnott and Law, 1996) or in response to lake levels in lacustrine environments (Saunders and Davidson-Arnott, 1990). As wind speeds increase above threshold for aeolian transport, it is reasonable to expect that there is a narrow spatial or temporal window in which sediment can be transported to the dune before the storm surge extends into the backshore and the transport system begins to shut down (Delgado-Fernandez and Davidson-Arnott, 2011). In response, Bauer and Davidson-Arnott (2003) present a model to show that oblique winds are largely responsible for transport of sediment from beach to dune, assuming that there is an unlimited supply of sediment in the backshore and across the beach. The greatest potential for aeolian transport occurs on wide, low-angle dissipative beaches, while narrow reflective beaches have limited dune development due to the short and steep fetch that limits aeolian transport (Short and Hesp, 1982). Morton et al. (1994) observed that the narrow beach width on developed beaches limited dune recovery, and that only undeveloped beaches went through all stages of post-storm recovery. Along eroding coasts, the post-storm dune height was unable to reach the pre-storm height due to the lack of new sediment, with only 67% of the prestorm volume recovered. The remaining sediment was transported alongshore and deposited as a spit at the distal end of the island. While Morton et al. (1994) described the final stages of dune recovery as dune formation through sediment capture by vegetation or wrack (see Eamer and Walker, 2010; Ollerhead et al., 2013), followed by dune expansion and vegetation recolonization, it should be noted that dune formation can only be initiated when vegetation is able to colonize the backshore. Depending on the extent that the roots and rhizomes are impacted, post-storm recovery of vegetation can take two to eight years (Brodhead and Godfrey, 1979). Snyder and Boss (2002) found that vegetation recovery on Santa Rosa Island in northwest Florida following
Fig. 1. Recovery of dune height (dune crest elevation–dune base elevation) presented by Morton et al. (1994) and Priestas and Fagherazzi (2010).
56
C. Houser et al. / Geomorphology 234 (2015) 54–63
Hurricane Opal was dependent on the presence of seed banks, vegetation fragments, and rhizomes that survived the storm, while Hesp (1988) observed that the morphology of the incipient foredune is dependent on the mode of beach colonization, plant density and distribution – in addition to the species and morphology available after a storm. The species available for recolonization vary with respect to the beach state; species richness and zonation are greatest and narrowest on reflective beaches, and lowest and widest on dissipative beaches (Hesp, 1988). The rate that the surviving vegetation is able to re-emerge depends on the depth of the washover and the speciesdependent burial toleration. The faster the vegetation emerges, the greater the chance of that vegetation remaining viable (Maun, 2009). Burial-tolerant species are adapted to washover and quickly colonize the surface through creeping and rhizome extension, but are not effective at trapping sand and building dunes. Conversely, dune-building species expend their energy growing upwards as sediment is trapped, but as a result are slow to colonize following a storm. As modeled conceptually in Fig. 2, an increase in the frequency and/or magnitude of storm surge has the potential to promote burial-tolerant species and the development of low dunes and a new equilibrium state (Stallins and Parker, 2003; Duran and Moore, 2013). Low islands have the potential to become high islands given sufficient time for the return of dune-building grasses and a supply of sediment to support the rebuilding of the dune (Duran and Moore, 2013). Compared to studies that document beach and dune erosion during storms, the rates and mechanisms of post-storm dune recovery have received relatively little attention in the literature - with notable exception for Olson (1958), Saunders and Davidson-Arnott (1990), Morton et al. (1994), McLean and Shen (2006), Houser and Hamilton (2009), Priestas and Fagherazzi (2010), Duran and Moore (2013), Wolner et al. (2013), and Lucas and Carter (2013). The majority of studies focus on erosion of the beach and dune, and the immediate recovery of the nearshore and beach profile following storms (Zeigler et al., 1959; Warnke et al., 1966; Katuna, 1991; Sexton and Hayes, 1991). It is reasonable to expect that the lack of field data reflects the disparate timescales of impact and recovery (see Lee et al., 1998) and the difficulty in maintaining and acquiring funding for long-term beach and dune monitoring programs. The purpose of the present study is to document the recovery of Santa Rosa Island in Northwest Florida using LiDAR data acquired between 2004 and 2010 to determine the rates of dune recovery with respect to the Verhulst (1838) model (Eq. (1)), as well as comparing both the rates and mechanisms of dune recovery with the recovery data of Morton et al. (1994; see Fig. 1). The results of this study are also compared to previous studies that describe the geological controls on island response to recent storm activity and over the Holocene (see Houser et al., 2008a, 2008b; Houser and Hamilton, 2009; Houser, 2012).
2. Study site Santa Rosa Island is a narrow sandy Holocene barrier island extending 96 km from East Pass near Destin to Pensacola Pass in the west (Fig. 3). The focus of this study is a 35 km stretch of Santa Rosa Island in northwest Florida that was impacted by hurricanes Ivan (2004), Dennis (2005) and Katrina (2005). Prior to Hurricane Ivan, the dunes east of Pensacola Beach formed a ridge and swale complex, with the primary dune parallel to the shoreline (see Houser et al., 2008a). During the hurricane seasons of 2004 and 2005, the island morphology changed from a discontinuous foredune backed by hummocky backbarrier dunes and maritime forest (at the bayside cuspate headlands) into washover terraces at the cuspate back-barrier headlands and washover corridors that reach Santa Rosa Sound between headlands. The shoreface profile is relatively steep seaward of the washover corridors (which forces the outermost bar closer to the shoreline) and is responsible for a transverse bar and rip morphology with a steep beachface (see Wright and Short, 1984). This beach and nearshore state limits the exchange of sediment from beach to dune, leading to small discontinuous dunes (Houser et al., 2011; Barrett and Houser, 2012). The beaches seaward of the back-barrier cuspate headlands are more gently sloped and fronted by a relatively dissipative nearshore that promotes the development of larger dunes, which limits washover penetration to the back-barrier shoreline (Houser et al., 2008a, 2008b). The pre-storm dune morphology and the impact of Hurricane Ivan exhibited a quasi-regular variation alongshore that corresponded to the alongshore variation in island width and the location of transverse ridges on the inner-shelf (Houser et al., 2008a, 2008b; Houser and Hamilton, 2009; Houser, 2012). The ridges and swales produce bathymetric highs and lows that force an alongshore variation in the surf similarity parameter and beach state (Houser et al., 2011a,b) and is quite visible in the position of the outermost nearshore bar along Santa Rosa Island (see Houser et al., 2008a, 2008b). The largest dunes before Hurricane Ivan are found landward of the transverse ridges, such that washover was limited in these sections of the island and most of the sediment from the beachface and dunes was deposited within the upper shoreface. In contrast, sections of the coast with little to no dune development before Hurricane Ivan are observed in the narrowest portions of the island (between headlands) and landward of the swales on the inner shelf. As a result, washover penetration was greater in these areas and island breaching was common, leaving the surface close to the watertable and covered by a lag of shell and gravel (Houser, 2012). A geological survey by Houser (2012) supported the theory that the ridge and swale topography may be a transgressive surface, representing a multi-scale feedback of the response and recovery of the island to storms. Houser and Hamilton (2009) observed beachface recovery in the Santa Rosa Unit of the island within two years of Hurricane Ivan, and backshore accretion in the Fort Pickens Unit. Similarly, Leadon (1999) observed rapid beach recovery following Hurricane Opal (1994) on Santa Rosa Island in northwest Florida, while Stone et al. (2004) observed limited dune recovery for ~6 years after the storm. In both sections of the island, the recovery exhibited the same alongshore pattern as the pre-storm dune morphology and storm impact, following the ridge and swale pattern (Houser et al., 2008a, 2008b). The purpose of the present study is to describe the recovery of these sections of the island two years after the 2004/2005-hurricane season. It was anticipated that recovery would vary alongshore, coincident with the transverse ridge and swale topography of the inner-shelf and between the Santa Rosa and Fort Pickens Units of the island. 3. Methodology
Fig. 2. Conceptual model of low and high island equilibrium states.
The alongshore variation in island recovery was examined using light detection and ranging (LiDAR), a mapping technology that has transformed coastal change studies (see Stockdon et al., 2007). Airborne
C. Houser et al. / Geomorphology 234 (2015) 54–63
57
Fig. 3. Location of Gulf Islands National Seashore showing Fort Pickens and Santa Rosa Units in addition to storm tracks from 2004 and 2005.
LiDAR involves the collection of spatially dense and accurate topographic data using aircraft-mounted lasers capable of recording elevation measurements with a vertical accuracy of up to 0.15 m and a horizontal accuracy of b1 m. This study uses 5 LiDAR datasets. The LiDAR data from 2004 and 2005 were collected by the US Geological Survey Center for Coastal and Watershed Studies before and (immediately) after Hurricane Ivan in May and September 2004 respectively, with subsequent surveys completed after Hurricane Dennis (July 2005) and Hurricane Katrina (September 2005). Escambia County completed an additional survey in July 2006. The pre-Ivan dataset (May 2004) is a product of the USACE and Optech Inc., which is more commonly known as the Compact Hydrographic Airborne Rapid Total Survey (CHARTS). This system is a combination of a 1000 Hz hydrographic sensor, a 10,000 Hz topographic sensor, and an image sensor. The post-Ivan LiDAR dataset is a product of the NASA sensor known as EAARL (Experimental Advanced Airborne Research LiDAR). The 2010 LiDAR data were collected by the USACE Joint Airborne LiDAR Bathymetry Technical Center of Expertise (JALBTCX) using the Hawkeye system. This system operates at 1064 nm (red wavelength) for topographic sensing and 532 nm (green wavelength) for bathymetric sensing. Positional accuracy was assessed using post-processed KGPS methods, and the horizontal and vertical accuracy (root mean squared error, or RMSE) of the 2010 LiDAR is ~1.5 m and ~0.15 m, respectively. The point spacing for the pre-Ivan data was ~0.3 m, but only ~1 m in the subsequent surveys. Following Young and Ashford (2006), all data were first converted from raw xyz data into ArcSDE feature classes. The raw point data were interpolated using inverse distance weighting (IDW) into pre-Ivan, post-Ivan, post-Katrina, and recovery elevation models. The IDW power was set to four to negate distant points, and the search radius was limited to the closest eight points, again to negate distant neighbors. The RMSE was 0.34 m, 0.54 m, 0.67 m and 0.62 m for
the pre-Ivan, post-Ivan, post-Katrina, and July 2006 models, respectively. The 2010 LiDAR data was best approximated by a spherical semivariogram model with nugget 0.00, sill 1.00, and range 23 m. The modeled semivariogram was used in an ordinary kriging algorithm to interpolate a 1 m DEM. As described in Rentschlar (2014), the alongshore variation in the aerial cover of vegetation and the recovery of the vegetation cover following Hurricane Ivan was examined using remotely sensed imagery acquired through the USGS EarthExplorer, National Oceanic and Atmospheric Administration (NOAA), Florida Department of transportation (FDOT) and the National Agricultural Imagery Program (NAIP). The distribution of vegetation was quantified using a RGB object oriented classification. Object-oriented classifiers were developed to help overcome the problems of classifying high spatial resolution images (Cleve et al., 2008; Li et al., 2013). Higher resolution images are more likely to have a variety of spectral responses that represent the same object. When using traditional per-pixel classifiers, this high spectral variability increases the chance of salt and pepper classification noise. Objectoriented analysis overcomes this by incorporating the spatial relationship of pixels as well as the spectral relationship when determining the class an object (Myint et al., 2011). Cross-shore transects spaced at 35-m intervals were extracted from the LiDAR data for a total of 302 transects in the Santa Rosa Unit of Gulf Islands National Seashore. Each transect starts from a straight line that bisects the island (point 0 on the profiles) and extends towards the Gulf of Mexico at 2-m intervals to the point where the LiDAR signal is unable to penetrate the water column; the average length of a transect is ~375 m. 4. Results LiDAR imagery of the Santa Rosa Unit from May to September 2004, 2005, 2006 and 2010 are presented in Fig. 4. The alongshore variability
58
C. Houser et al. / Geomorphology 234 (2015) 54–63
Fig. 4. LiDAR imagery from (A) May and (B) September 2004; (C) 2005, (D) 2006 and (E) 2010.
in the beach and dune morphology described by Houser et al. (2008a, 2008b) is evident in the May 2004 LiDAR imagery (Fig. 4A). This morphology is characterized by the island's largest dunes located seaward of back-barrier cuspate headlands and smaller dunes in the narrow sections of the island between headlands. Washover penetration during Hurricane Ivan follows a similar pattern, with the washover terrace extending further landward and into the back-barrier sound in the narrow sections of the island, but relatively limited washover penetration in the wider sections of the island. This variation is also evident in the alongshore variation in washover penetration following Hurricane Ivan (September 2004; Fig. 4B) and the presence of remnant dunes in the widest sections of the island where washover penetration is limited. The influence of the transverse ridge and swale bathymetry on the berm and dune elevation was further examined using cross-shore transects sampled every 35 m alongshore (Fig. 5). The alongshore variability is relatively muted in the 2005 and 2006 profile data, suggesting further erosion of the largest dunes and infilling of the washover and breach areas. Recovery along the entire length of the Santa Rosa Unit is evident in the 2010 data, with the areas of greatest recovery located where the dunes were largest pre-Ivan, and limited recovery present in the narrow sections of the island where pre-storm dune heights were smaller. As a consequence, the alongshore variation in dune height associated with the ridge and swale bathymetry is again evident 5 years after Hurricane Ivan (Fig. 4D).
The alongshore variation in island morphology is also evident in the representative cross-shore profiles in Fig. 6. Recovery in the narrow sections of the island (Profiles 129 and 249) is characterized by berm reconstruction and steepening of the beachface with a near-uniform increase in elevation through storm washover (Fig. 6). Recovery in the wider sections (Profiles 150 and 294) is characterized by berm reconstruction and steepening of the beachface but also includes incipient dune development in the backshore and sporadic dune development further landward (Fig. 6A, D). Dune recovery in the widest sections of the island is associated with the seaward accretion of the beachface through the landward migration of the nearshore bars, while seaward development of the beachface is limited in the narrowest sections of the island. The average maximum island elevation (i.e. the elevation of the berm or dune if present) was calculated for each year to develop a recovery curve similar to Galveston and St. John's Island (see Fig. 1). In areas where there was further erosion of the dune during hurricanes Dennis and Katrina (e.g. transect 150) the recovery was assumed to start after 2006. In the majority of areas, however, there was elevation recovery during hurricanes Dennis and Katrina through berm development. In other areas data were not included for years in which an artificial dune was not yet constructed when the access road was rebuilt (e.g. transects 129 and 151). The alongshore-average recovery is presented in Fig. 7 using the May 2004 (pre-Ivan) dune heights as the predicted height after 10 years of recovery since Hurricane Opal
Fig. 5. Alongshore variation in dune and/or berm height extracted from the LiDAR data from May and September 2004, 2005, 2006 and 2010.
C. Houser et al. / Geomorphology 234 (2015) 54–63
59
Fig. 6. Representative cross-shore profiles from areas with large pre-Ivan dunes (A, D), and from areas with small pre-Ivan dunes (B, C). Shown are profiles (A) 150, (B) 129, (C) 249 and (D) 294.
(1994). The recovery appears to follow the growth model of Verhulst (1838) but with a slower rate of recovery than that seen in the data of Morton et al. (1994) and Priestas and Fagherazzi (2010). For an average pre-storm dune height of ~2.1 m, the growth rate (r) is ~0.25 m yr−1, meaning it would take 7 years for the dunes to reach their pre-Ivan height following either hurricanes Ivan or Katrina. The alongshoreaverage rate of recovery in dune height is compared to the recovery of the vegetation cover (from Rentschlar, 2014) and profile volume in
Fig. 8. On average, the island volume reaches a peak rate of recovery ~4 years after the disturbance, while vegetation and dune height both reach a peak rate of recovery ~5 years after the disturbance. While the vegetation and dune height exhibit peak recovery at the same time, the vegetation recovery initiates more quickly after the storm. The rate of dune recovery remains small until ~3 years following the storm. The rate of dune recovery varies considerably, with a weak correspondence to the height of the pre-storm dune (Fig. 9). Specifically,
60
C. Houser et al. / Geomorphology 234 (2015) 54–63
Fig. 7. Recovery of dune height (dune crest elevation- dune base elevation) from the present study and presented by Morton et al. (1994) and Priestas and Fagherazzi (2010).
the rate of recovery is fastest (r N 0.4) where the pre-storm dunes were the smallest, and slowest (r b 0.2) where the pre-storm dunes were the largest. Similarly, the growth rate of the vegetation (from Rentschlar, 2014) and profile volume are inversely related to the pre-storm dune height. Recovery of both vegetation and profile volume are relatively fast where the pre-storm dune heights were small, and are slowest where the pre-storm dune heights were the largest. A comparison between transects 249 (fast recovery of small dunes) and 294 (slow recovery of larger dunes) suggests that the profile volume recovers rapidly where the pre-storm dunes are the smallest and is complete within two years of the storm (Fig. 10). In contrast, recovery of the profile volume continues for at least seven years following the storm, where the pre-storm dune heights were the largest (transect 294). While recovery of the profile volume is fastest where the pre-storm dunes are the smallest, the ultimate height of those dunes is limited by the availability of sediment from the nearshore. 5. Discussion The rate and extent of post-storm dune recovery determines the vulnerability of barrier islands to future storm events, which in turn control how the island will respond to accelerated sea-level rise and a change in the frequency and/or magnitude of storm events. Compared to studies describing the impact of storms on beaches and dunes, there is a paucity of field and remotely sensed data to quantify the rate of post-storm beach and dune recovery. It is reasonable to assume that the scarcity of field data is due to the disparate timescales of storm response and post-storm recovery. Results presented by Stone et al. (2004) suggest that dune recovery on Santa Rosa Island took up to a decade between hurricanes Opal (1994) and Ivan (2004). Using the May 2004 (pre-Ivan) dune heights as the predicted height after 10 years of recovery,
results of the present study also suggest that the dunes along Santa Rosa Island were eroded and overwashed over the course of several hours during Hurricane Ivan and could take up to a decade to return to their pre-storm height. While it is not clear why the dunes recovered faster on Galveston Island (Morton et al., 1994; Fig. 1), the results of the present study suggest that beach and dune recovery can be modeled using a sigmoid curve (Eq. (1)). A sigmoid recovery curve is consistent with the recovery curves for vegetation following a major disturbance (Hugenholtz and Wolfe, 2005; Rentschlar, 2014) and allows for models to predict the response of barrier islands to changes in storm activity and sea-level rise (Duran and Moore, 2013). Results from Olson (1958) and Saunders and Davidson-Arnott (1990) suggest that the differing alongshore rates of recovery between Santa Rosa Island and Galveston Island are dependent on whether the dunes were primarily scarped or overwashed (see also Sallenger, 2000; Hesp, 2002). However, it is also possible that the faster recovery from Galveston Island (Morton et al., 1994) is a result of the dissipative beachface (see Hesp, 1982, 1988; Short and Hesp, 1982; Bauer and Sherman, 1999) and wider fetch (Bauer and Davidson-Arnott, 2003), but we would have then expected the dune recovery on St. George Island (Priestas and Fagherazzi, 2010) to be similar to Santa Rosa Island. Further study is required to determine how recovery varies along and among different barrier island environments and with respect to beach morphology and whether the dune was scarped or overwashed (see Olson, 1958; Hesp, 2002). The rate of post-storm recovery on Santa Rosa Island varies alongshore in a manner consistent with the morphology of the barrier island. As reported by Houser et al. (2008a, 2008b), the largest pre-storm dunes are seaward of the back-barrier cuspate headlands, while the smallest pre-storm dunes are in the narrowest sections of the island between the back-barrier headlands. This alongshore variation is forced by
Fig. 8. Rate of change in profile volume, vegetation area (from Rentschlar, 2014) and dune height.
C. Houser et al. / Geomorphology 234 (2015) 54–63
61
Fig. 9. Alongshore variation in the May 2004 dune height and the calculated growth rate (r; Eq. (1)).
the transverse ridge and swale bathymetry of the inner-shelf that in turn formed as the island transgressed over the Holocene (Houser, 2012). An intermediate to dissipative nearshore profile develops landward of the transverse ridges, while a reflective to intermediate nearshore profile develops landward of the swales in the narrowest sections of the island (Barrett and Houser, 2012). As argued by Houser and Hamilton (2009) the dunes landward of the transverse ridges are supply limited, while the dunes landward of the swales are (aeolian) transport limited. Results of the present study suggest that this may have been an oversimplification and that the alongshore variation was reinforced by the alongshore variation in washover potential, which is in turn dependent on the rate of dune recovery following storms. Washover in areas with small and/or discontinuous dunes causes sediment to be lost to the back of the island, which limits the supply of sediment that can be returned to the dunes during the recovery period in the absence of seaward winds (Leatherman and Zaremba, 1987). As reported by Houser et al. (2008a, 2008b) the offshore winds are relatively weak and represent an insignificant supply of sediment to the recovering dunes. Areas with larger and continuous dunes have limited washover, but also lose more sediment offshore. The dunes are taller because that sediment can be returned to the beachface through the landward migration of the nearshore bars. In the same way, the dunes in the narrow sections of Santa Rosa Island are the not the smallest because they take longer to recover but rather because there is a limited supply of sediment for recovery. Results of the present study suggest that the landward migration of the nearshore bars and recovery of the profile volume are the primary control of dune recovery on Santa Rosa Island. Limited dune development in the narrow sections of the island reflects the limited supply of sediment available in the nearshore, and profile volume recovery was observed within ~2 years following Hurricane Ivan (see Fig. 10). In comparison, profile recovery continues for at least six years following
Hurricane Ivan in areas where the pre-storm dunes were the largest in May 2004. The vegetation recovery in these same locations reaches its maximum at the same time that the growth of the profile volume begins to slow, with the vegetation growth slowing 10 to 15 years after the disturbance (Rentschlar, 2014). Recovery of the vegetation closely follows the recovery of the profile volume, which may reflect the dependency of vegetation growth on sediment deposition (Zhang and Maun, 1992). As a consequence, dune recovery is relatively fast in areas where the prestorm dunes were small and discontinuous alongshore, but dune heights at Pensacola Beach are ultimately limited by the availability of sediment returned to the beach and backshore by landward migration of the innermost nearshore bar. In contrast, the largest dunes on Santa Rosa Island experience a longer recovery due to the longer time required for the sediment to be transferred from the nearshore bars to the beach and to the dune. The dependency of dune recovery on the availability of sediment in the nearshore (versus in the washover deposits) suggests that the development of small discontinuous dunes is not reversible without an influx of new sediment from an offshore or alongshore source, which in turn requires a cessation of storm activity and the return of dune-building vegetation. However, it also suggests that an inability of the largest dunes to recover due to some anthropogenic interference (Saunders and Davidson-Arnott, 1990; Houser, 2009; Houser et al., 2012), or a rapid succession of storms (Houser and Hamilton, 2009) can force the dune system into an irreversible small dune (i.e. low island) state. It is possible that the longer recovery of the dunes on Santa Rosa Island is a result of the limited erosion and washover that occurred during hurricanes Dennis and Katrina in 2005 (see Houser and Hamilton, 2009). While these storms did not reduce the island elevation, and in some places increased island elevation through washover (Fig. 5), the storms kept the nearshore bars offshore and limited the recovery of the beachface. This suggests, as previously noted, that the
Fig. 10. Comparison of nearshore profile recovery for a representative profile with small dunes (Transect 249) and a representative profile with large dunes (Transect 294).
62
C. Houser et al. / Geomorphology 234 (2015) 54–63
initial recovery of the dune on Santa Rosa Island is limited by the availability of sediment returned to the beachface from the landward migration of the nearshore bars, and that vegetation recovery is a reaction to the frequency of washover that results from the limited dune development. This is consistent with arguments made by Stallins and Parker (2003) and Duran and Moore (2013), but the authors of those studies believed that the transition between low- and high-island states is a consequence of a change in vegetation from effective to ineffective dune builders. Based on these previous studies, a ‘high’ island will develop if the frequency of storm events permits the recovery of dune building vegetation, while a low island will develop if frequent storms limit dune building vegetation (Stallins and Parker, 2003; Duran and Moore, 2013). The vegetation type is a response to changes in the amount of sediment partitioned between the nearshore and overwash, which suggests that a change in the frequency of storms relative to the rate of sediment exchange from the nearshore to the beachface (and ultimately the dune) may lead to a self-reinforcing cycle that ends with a ‘low’ island dominated by maintainer species that preserve flow flat topography (see Wolner et al., 2013). The maximum dune height can be predicted from the transient form ratio (TFR) of Brunsden and Thornes (1979). Through this model dune height depends on the frequency of disturbance (Fi) relative to the recovery time (Ri): T FR ¼
Ri : Ti
ð3Þ
When TFR N 1 the maximum dune height is limited by the frequency of storm events capable of eroding the dune, while sediment supply, beachface slope, or the relative abundance of effective and ineffective dune building vegetation are the primary limit when the TFR b 1. Comparison of the recovery curves with the storm surge annual maximum water levels for Santa Rosa Island (Xu and Huang, 2008), suggests that storm surges capable of eroding the dune base (~ 1 m asl) occur on average every 7 years. While this means that there is a high probability that the largest dunes will have largely recovered before the next erosive storm, there is the potential for a transition to a new equilibrium state if storms occur in close succession, and/or there is an increase in the frequency and magnitude of storm events. The resulting change towards a low island would signal a new equilibrium response and create the potential for rapid island transgression and potential drowning. In this respect, it is the areas with largest dunes on Santa Rosa Island that should be of greatest concern, and be offered the greatest protection, rather than the small-dune areas that are overwash and breached. Without intervention, the larger dunes could remain in a new, lower dune state, and regularly be the site of washover and breaching - a new and possibly irreversible equilibrium state in the future. 6. Conclusions If a barrier island is faced with a change in the frequency and/or magnitude of extreme storms, its resilience will be dependent on the ability of the beach and dune to recover following extreme storms. Results suggest that the recovery of the beach and dune to their prestorm morphology can take up to a decade, and that the rate and amount of recovery is dependent on the pre-storm dune height. On Santa Rosa Island in northwest Florida, recovery varies alongshore. It is fast where the pre-storm dune heights were small and there appears to be no further recovery since most of the sediment was lost to overwash. The largest dunes take longer to recover, with the rate of recovery dependent on the landward migration and welding of the nearshore bars to the beachface and the recolonization by dune-building vegetation. The rate of recovery for Santa Rosa Island in northwest Florida is therefore susceptible to a significant loss of height overall, and the
island has the potential to become a ‘low’ island if there is an increase in the frequency and/or magnitude of storm events. Acknowledgments This study was partially funded through support from the National Science Foundation (BCS-1001126; BCS-1262711). References Barrett, G., Houser, C., 2012. Identifying hot spots of rip current activity using wavelet analysis at Pensacola Beach, Florida. Phys. Geogr. 33, 32–49. Bauer, B.O., Davidson-Arnott, R.G.D., 2003. A general framework for modeling sediment supply to coastal dunes including wind angle, beach geometry, and fetch effects. Geomorphology 49, 89–108. Bauer, B., Sherman, D., 1999. Coastal dune dynamics: problem and prospects. In: Goudre, A.S., Livingston, I., Stokes, J. (Eds.), Aeolian Environments, Sediments And Landforms. J. Wiley and Sons Ltd., London. Brodhead, J.M., Godfrey, P.J., 1979. The Effects Of Off-Road Vehicles On Coastal Dune Vegetation In The Province Lands, Cape Cod National Seashore. Environmental Institute, Massachusetts. Brunsden, D., Thornes, J.B., 1979. Landscape sensitivity and change. Trans. Inst. Br. Geogr. 4, 463–484. Christiansen, M.B., Davidson-Arnott, R., 2004. Rates of landward sand transport over the foredune at Skallingen, Denmark and the role of dune ramp. Dan. J. Geogr. 104, 31–43. Cleve, C., Kelly, M., Kearns, F.R., Moritz, M., 2008. Classification of the wildland–urban interface: a comparison of pixel-and object-based classifications using high-resolution aerial photography. Comput. Environ. Urban. Syst. 32, 317–326. Davidson-Arnott, R.G.D., 1988. Temporal and spatial controls on beach–dune interaction, Long Point, Lake Erie. In: Psuty, N.P. (Ed.), J. Coast. Res. Spec. Issue vol. 3. C.E.R.F., Florida, pp. 131–136. Davidson-Arnott, R.G.D., Law, M.N., 1990. Seasonal patterns and controls on sediment supply to coastal foredunes, Long Point, Lake Erie. In: Nordstrom, K.F., Psuty, N.P., Carter, R.W.G. (Eds.), Coastal Dunes: Form And Process. Wiley, London, pp. 177–200. Davidson-Arnott, R.G.D., Law, M.N., 1996. Measurement and prediction of long-term sediment supply to Coastal Foredunes. J. Coast. Res. 12, 654–663. Delgado-Fernandez, I., Davidson-Arnott, R., 2011. Meso-scale aeolian sediment input to coastal dunes: the nature of aeolian transport events. Geomorphology 126, 217–232. Duran, O., Moore, L.J., 2013. Vegetation controls on the maximum size of coastal dunes. Proc. Natl. Acad. Sci. U. S. A. 110, 17217–17222. Eamer, J.B.R., Walker, I.J., 2010. Quantifying sand storage capacity of large woody debris on beaches using LiDAR. Geomorphology 118, 33–47. Elgar, S., Gallagher, E.L., Guza, R.T., 2001. Nearshore sandbar migration. J. Geophys. Res. 106, 11623–11627. Hesp, P.A., 1982. Morphology And Dynamics Of Foredunes In S.E. Australia. (Unpubl. PhD Thesis), Dept. Geography, University of Sydney (397 pp.). Hesp, P.A., 1988. Surfzone, beach and foredune interactions on the Australian southeast coast. In: Psuty, N.P. (Ed.), J. Coast. Res. Spec. Issue vol. 3. C.E.R.F., Florida, pp. 15–25. Hesp, P., 2002. Foredunes and blowouts: initiation, geomorphology and dynamics. Geomorphology 48, 245–268. Houser, C., 2009. Synchronization of transport and supply in beach-dune interaction. Prog. Phys. Geogr. 33, 733–746. Houser, C., 2012. Role of ridge and swale bathymetry in barrier island transgression. Geomorphology 173 (174), 1–16. Houser, C., Greenwood, B., 2005. Hydrodynamics and sediment transport within the inner surf zone of a lacustrine multiple-barred nearshore. Mar. Geol. 218, 37–63. Houser, C., Greenwood, B., 2007. Onshore migration of a swash bar. J. Coast. Res. 23, 1–14. Houser, C., Hamilton, S., 2009. Sensitivity of post-hurricane beach and dune recovery to event frequency. Earth Surf. Process. Landf. 34, 613–628. Houser, C., Greenwood, B., Aagaard, T., 2006. Divergent response of an intertidal swash bar. Earth Surf. Process. Landf. 31, 1775–1791. Houser, C., Hapke, C., Hamilton, S., 2008a. Controls on coastal dune morphology, shoreline erosion and barrier island response to extreme storms. Geomorphology 100, 223–240. Houser, C., Hobbs, C., Saari, B., 2008b. Post-hurricane airflow and sediment transport over a recovering dune. J. Coast. Res. 24, 944–953. Houser, C., Barrett, G., Labude, D., 2011. Alongshore variation in the rip current hazard at Pensacola Beach, Florida. Nat. Hazards 57, 501–523. Houser, C., Labude, B., Haider, L., Weymer, B., 2012. Impacts of driving on the beach: case studies from Assateague Island and Padre Island national seashores. Ocean Coast. Manag. 71, 33–45. Hugenholtz, C., Wolfe, S., 2005. Biogeomorphic model of dunefield activation and stabilization on the northern Great Plains. Geomorphology 70, 53–70. Katuna, M.P., 1991. The effects of Hurricane Hugo on the Isle of Palms, South Carolina: from destruction to recovery. In: Finkl, C.W., JR, Pilkey, O.H., JR (Eds.), Impacts Of Hurricane Hugo, September 10–22, 1989. J. Coast. Res. 8, pp. 263–273. Keijsers, J.G.S., Poortinga, A., Riksen, M.J.P.M., Groot de, A.V., 2012. Connecting aeolian sediment transport with foredune development. NCK-Days 2012: Crossing Borders In Coastal Research, 13 March 2012–16 March 2012, Enschede, The Netherlands. Law, M.N., Davidson-Arnott, R.G.D., 1990. Seasonal controls on aeolian processes on the beach and foredune. Proc. Symposium On Coastal Sand Dunes. NRC, Ottawa, pp. 49–67.
C. Houser et al. / Geomorphology 234 (2015) 54–63 Leadon, M.E., 1999. Beach, dune and offshore profile response to a severe storm event. Coastal Sediments (1999). ASCE, pp. 2534–2549. Leatherman, S.P., Zaremba, R.E., 1987. Overwash and aeolian processes on a United States Northeast coast. Sed. Geol. 52 (3–4), 183–206. Lee, G., Nicholls, R.J., Birkemeier, W.A., 1998. Storm-driven variability of the beachnearshore profi le at Duck, North Carolina, USA, 1981–1991. Mar. Geol. 148, 163–177. Li, J., Wan, J., Lu, Y., Chen, J., Fu, Y., 2013. A new object-oriented approach towards GIS seamless spatio-temporal data model construction. Pervasive Computing And The Networked World. Springer, pp. 318–331. Lucas, K.L., Carter, G.A., 2013. Change in distribution and composition of vegetated habitats on Horn Island, Mississippi, norther Gulf of Mexico in the initial five years following Hurricane Katrina. Geomorphology 199, 129–137. Maun, M.A., 2009. The Biology Of Coastal Sand Dunes. Oxford University Press (265 pp.). McLean, R., Shen, J.S., 2006. From foreshore to foredune: foredune development over the last 30 years at Moruya Beach, New South Wales, Australia. J. Coast. Res. 22, 28–36. Morton, R.A., 2002. Factors controlling storm impacts on coastal barriers and beaches — a preliminary basis for near real-time forecasting. J. Coast. Res. 18, 486–501. Morton, R.A., Paine, J.G., Gibeaut, J.C., 1994. Stages and durations of post-storm beach recovery, southeastern Texas coast, USA. J. Coast. Res. 884–908. Myint, S.W., Gober, P., Brazel, A., Grossman-Clarke, S., Weng, Q., 2011. Per-pixel vs. objectbased classification of urban land cover extraction using high spatial resolution imagery. Remote Sens. Environ. 115, 1145–1161. Nott, J., 2006. Tropical cyclones and the evolution of the sedimentary coast of northern Australia. J. Coast. Res. 22, 49–62. Ollerhead, J., Davidson-Arnott, R., Walker, I.J., Mathew, S., 2013. Annual to decadal morphodynamics of the foredune system at Greenwich Dunes, Prince Edward Island, Canada. Earth Surf. Process. Landf. 38, 284–298. Olson, J.S., 1958. Lake Michigan dune development 3: lake level, beach and dune oscillations. J. Geol. 66, 473–483. Priestas, A.M., Fagherazzi, S., 2010. Morphological barrier island changes and recovery of dunes after Hurricane Dennis, St. George Island, Florida. Geomorphology 114, 614–625. Rentschlar, E., 2014. Quantifying Vegetation Recovery On Santa Rosa Island. (MSc Thesis), Department of Geography, Texas A&M University (105 pp.). Ruz, M., Meur-Ferec, C., 2004. Influence of high water levels on aeolian sand transport: upper beach/dune evolution on a macrotidal coasta, Wissant Bay, northern France. Geomorphology 60, 73–87. Sallenger Jr., A.H., 2000. Storm impact scale for barrier islands. J. Coast. Res. 890–895. Saunders, K.E., Davidson-Arnott, R.G.D., 1990. Coastal dune response to natural disturbances. Proc. Symposium On Coastal Sand Dunes. NRC, Ottawa, pp. 321–345.
63
Sexton, W.J., Hayes, M.O., 1991. The geologic impact of Hurricane Hugo and post-storm shoreline recovery along the undeveloped coastline of South Carolina, Dewees Island to the Santee Delta. In: Finkl, C.W., JR, Pilkey, O.H., JR (Eds.), Impacts Of Hurricane Hugo, September 10–22, 1989. J. Coast. Res. 8, pp. 275–290. Short, A., Hesp, P.A., 1982. Wave–beach–dune interaction in southeast Australia. Mar. Geol. 48, 259–284. Snyder, R., Boss, C., 2002. Recovery and stability in barrier island plant communities. J. Coast. Res. 530–536. Stallins, J.A., Parker, A.J., 2003. The influence of complex systems interactions on barrier island dune vegetation pattern and process. Ann. Assoc. Am. Geogr. 93, 13–29. Stockdon, H.F., Sallenger Jr., A.H., Holman, R.A., Howd, P.A., 2007. A simple model for the spatially-variable coastal response to hurricanes. Mar. Geol. 238 (1), 1–20. Stone, G.W., Liu, B., Pepper, D.A., Wang, P., 2004. The importance of extratropical and tropical cyclones on the short-term evolution of barrier islands along the northern Gulf of Mexico, USA. Mar. Geol. 210, 63–78. Thieler, E.R., Young, R.S., 1991. Quantitative evaluation of coastal geomorphological changes in South Carolina after Hurricane Hugo. J. Coast. Res. 8, 187–200. Verhulst, P.F., 1838. Notice sur la loi que la population suit dans son accroissement. Corresp. Math. Phys. Publiée A Quetelet 10, 113–121. Warnke, D.A., Goldsmith, V., Grose, P., Holt, J.J., 1966. Drastic beach changes in a lowenergy environment caused by Hurricane Betsy. J. Geophys. Res. 71, 2013–2016. Wolner, C.W.V., Moore, L.J., Young, D.R., Brantley, S.T., Bissett, S.N., McBride, R.A., 2013. Ecomorphodynamic feedbacks and barrier island response to disturbance: insights from the Virginia Barrier Islands, Mid-Atlantic Bight, USA. Geomorphology 199, 115–128. Woodroffe, C.D., 2007. 2007. The natural resilience of coastal systems: primary concepts. In: McFadden, L., Penning-Rowsell, E., Nicholls, R.J. (Eds.), Managing Coastal Vulnerability. Elsevier, Amsterdam, pp. 45–60. Wright, L.D., Short, A.D., 1984. Morphodynamic variability of beaches and surfzones: A synthesis. Mar. Geol. 56, 92–118. Xu, S., Huang, W., 2008. Frequency analysis for predicting 1% annual maximum water levels along Florida coast, US. Hydrol. Process. 22, 4507–4518. Young, A.P., Ashford, S.A., 2006. Application of airborne LIDAR for seacliff volumetric change and beach-sediment budget contributions. J. Coast. Res. 22, 307–318. Zeigler, J.M., Hayes, C.R., Tuttle, S.D., 1959. Beach changes during storms on outer Cape Cod, Massachusetts. J. Geol. 67, 318–336. Zhang, J., Maun, M., 1992. Effects of burial in sand on the growth and reproduction of Cakile edentula. Ecography 15, 296–302.
Contents lists available at ScienceDirect
Geomorphology journal homepage: www.elsevier.com/locate/geomorph
Post-storm beach and dune recovery: Implications for barrier island resilience Chris Houser a,⁎, Phil Wernette a, Elizabeth Rentschlar a, Hannah Jones b, Brianna Hammond a, Sarah Trimble a a b
Department of Geography, College of Geosciences, 810 O&M Building, Texas A&M University, College Station, TX 77843-3147, United States Department of Geology, Carleton College, Northfield, MN 55057, United States
a r t i c l e
i n f o
Article history: Received 17 April 2014 Received in revised form 6 December 2014 Accepted 10 December 2014 Available online 24 January 2015 Keywords: Barrier island Post-storm recovery Resilience Hurricane
a b s t r a c t The ability of beaches and dunes to recover following an extreme storm is a primary control of barrier island response to sea-level rise and changes in the frequency and/or magnitude of storm surges. Whereas erosion of the beach and dune occurs over hours and days, it can be years to decades before the beach and dune are able to recover to their pre-storm state. As a consequence, there are numerous descriptions of near-instantaneous beach and dune erosion due to storms, the immediate onshore transport of sand, and the initial phases of beach and dune recovery following a storm, but a paucity of data on long-term beach and dune recovery. A combination of previously published data from Galveston Island, Texas and new remotely sensed data from Santa Rosa Island, Florida is used in the present study to quantify the rate of dune recovery for dissipative and intermediate beach types, respectively. Recovery of the dune height and volume on Galveston Island was observed within two years following Hurricane Alicia (1983) and was largely complete within six years of the storm, despite extensive washover. In contrast, the dunes on Santa Rosa Island in Northwest Florida began to recover four years after Hurricane Ivan (2004), and only after the profile approached its pre-storm level and the rate of vegetation recovery (regrowth) was at a maximum. Results show that complete recovery of the largest dunes (in height and volume) will take approximately 10 years on Santa Rosa Island, which suggests that these sections of the island are particularly vulnerable to significant change in island morphology if there is also a change in the frequency and magnitude of storm events. In contrast, the areas of the island with the smallest dunes before Hurricane Ivan exhibited a rapid recovery, but no further growth in profile volume and dune height beyond the pre-storm volume and height, despite continued recovery of the largest dunes to their pre-storm height. A change in storm magnitude and/or frequency is a potential threat to barrier island resilience, particularly for those sections of the island where dune recovery has historically taken the longest time. Further study is required to determine how and why dune recovery varies for the dissipative and intermediate beaches of Galveston Island and Santa Rosa Island, respectively. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The vulnerability of a barrier island to extreme storms depends on the elevation of the total water level (tide + storm surge + wave runup) relative to the geometry of the coast, which is largely dependent on the height and alongshore extent of the foredune (Thieler and Young, 1991; Sallenger, 2000; Morton, 2002; Nott, 2006; Houser and Hamilton, 2009). Storm impact can have a range of visible impacts, from minor scarping at the base of the dune to overwash and/or breaching, when dune heights are relatively small compared to the storm surge (Sallenger, 2000; Hesp, 2002). Resiliency of a barrier island, or the ability to return to its previous equilibrium state (Woodroffe, 2007), is dependent on the rate of post-storm dune recovery. This rate is dependent on the transfer of sediment from the nearshore to the beach, which can occur through the landward migration and welding of the innermost nearshore bars, the alongshore migration of sand ⁎ Corresponding author. E-mail address: [email protected] (C. Houser).
http://dx.doi.org/10.1016/j.geomorph.2014.12.044 0169-555X/© 2015 Elsevier B.V. All rights reserved.
waves, the recolonization and expansion of dune-building vegetation, and/or aeolian transfer of sediment from the beach to the backshore and recovering dune. Depending on the magnitude and duration of the storm surge, erosion of the beach and dune occurs over hours and days, whereas recovery of the nearshore, beach and dune can take years to decades (Lee et al., 1998). The differential timescale of erosion and recovery makes barrier island response to sea-level rise dependent on the sequence of storm events and vulnerable to widespread erosion and washover when storms occur in quick succession (see Houser and Hamilton, 2009). The areas of the island that could exhibit significant morphological change and a transition to a new and possibly irreversible equilibrium state in the future are not those with already limited dune development and ineffective dune-building vegetation (low islands; Duran and Moore, 2013), but those sections of the island where the development of large dunes depends on the recovery of beach and the recolonization of dune-building vegetation (high islands; Duran and Moore, 2013). Dune recovery requires the transfer of sediment from nearshore to the beach and ultimately to the dune, assuming that dune-building
C. Houser et al. / Geomorphology 234 (2015) 54–63
vegetation is also present. Based on 10 years of monitoring data, Morton et al. (1994) tracked the beach and dune recovery following Hurricane Alicia (1983) along Galveston Island. Calculating dune heights from the published profiles provides a time series of dune recovery in which peak dune growth is observed ~3 years after the storm and the dune reaches the pre-storm height (of ~ 2 m) 4–5 years after the storm (Fig. 1). Growth of the dune follows a sigmoid curve consistent with the growth models used to quantify vegetation growth (see Hugenholtz and Wolfe, 2005a,b). Specifically, the data presented in Morton et al. (1994) and several survey transects from Priestas and Fagherazzi (2010) all follow the growth model of Verhulst (1838): dN N ¼ rN 1− dt K
ð1Þ
where N is a system attribute (i.e. dune height), r is the growth rate, t is the time elapsed since the last disturbance, and K is the upper boundary (asymptote) of dune growth (Verhulst, 1838). Only those transects from Priestas and Fagherazzi (2010) with pre-storm dune heights of ~2 m, similar to those of Morton et al. (1994), were used in this preliminary analysis. Integration of Eq. (1) gives: Nt ¼
KN o ðK−No Þe−n þ No
ð2Þ
where Nt is the height of the dune at time t, No is the initial height of the dune (t = 0) and e is the base of the natural logarithm. The dunes of Morton et al. (1994) and Priestas and Fagherazzi (2010) appear to follow the same logistic curve, but it is not clear why they exhibit similar recoveries given different environmental conditions. As described by Morton et al. (1994), the first stage of recovery begins immediately after the storm and can last a few weeks or up to a year, depending on the severity of the storm (Sallenger, 2000). This stage is characterized by berm reconstruction and steepening of the beach face. Specifically, sediment is returned to the beachface and the beach undergoes gradual accretion as the innermost bar migrates landward and welds to the beachface, leading to a steep beach ridge in reflective environments or a low-gradient berm in more dissipative environments. Landward migration of nearshore bars is driven by the waves as they shoal across the bar during fair-weather conditions (Elgar et al., 2001; Houser et al., 2006), with recovery and beach welding requiring several years following a large storm or multiple storms in succession (Lee et al., 1998). In lacustrine environments, the width of the backshore is closely tied to water levels, in which low water levels promote dune recovery and progradation, while high water levels allow even moderate storms to erode the foredune and reset the recovery (Saunders and DavidsonArnott, 1990). Following Morton et al. (1994), backbeach aggradation (around year 2 in Fig. 1) is largely dependent on deposition from swash events that
55
exceed the elevation of the beach ridge in intermediate to reflective environments, or through the landward migration of the nearshore bars in dissipative environments. Since storm winds capable of entraining sediment are usually accompanied by elevated water levels (Ruz and Meur-Ferec, 2004; Delgado-Fernandez and Davidson-Arnott, 2011) and precipitation (Keijsers et al., 2012), it is reasonable to assume that sediment only becomes available to the dunes when the backshore expands and allows for the development/recovery of a dune ramp (Christiansen and Davidson-Arnott, 2004). While aeolian transport is possible as soon as the upper-beach and any washover deposits become dry, the expansion of the backshore is required to increase the fetch length, which controls the amount of sediment exchanged from the beach to the dune (Davidson-Arnott, 1988; Davidson-Arnott and Law, 1990, 1996; Bauer and Davidson-Arnott, 2003; Houser, 2009). Sediment emplacement in the backshore and lengthening of the available fetch can either occur between storms, when winds tend to be below the transport threshold, or during storms, through the landward migration of subtidal and intertidal bars (Houser and Greenwood, 2005, 2007), the alongshore migration of sandwaves (Law and Davidson-Arnott, 1990; Davidson-Arnott and Law, 1996) or in response to lake levels in lacustrine environments (Saunders and Davidson-Arnott, 1990). As wind speeds increase above threshold for aeolian transport, it is reasonable to expect that there is a narrow spatial or temporal window in which sediment can be transported to the dune before the storm surge extends into the backshore and the transport system begins to shut down (Delgado-Fernandez and Davidson-Arnott, 2011). In response, Bauer and Davidson-Arnott (2003) present a model to show that oblique winds are largely responsible for transport of sediment from beach to dune, assuming that there is an unlimited supply of sediment in the backshore and across the beach. The greatest potential for aeolian transport occurs on wide, low-angle dissipative beaches, while narrow reflective beaches have limited dune development due to the short and steep fetch that limits aeolian transport (Short and Hesp, 1982). Morton et al. (1994) observed that the narrow beach width on developed beaches limited dune recovery, and that only undeveloped beaches went through all stages of post-storm recovery. Along eroding coasts, the post-storm dune height was unable to reach the pre-storm height due to the lack of new sediment, with only 67% of the prestorm volume recovered. The remaining sediment was transported alongshore and deposited as a spit at the distal end of the island. While Morton et al. (1994) described the final stages of dune recovery as dune formation through sediment capture by vegetation or wrack (see Eamer and Walker, 2010; Ollerhead et al., 2013), followed by dune expansion and vegetation recolonization, it should be noted that dune formation can only be initiated when vegetation is able to colonize the backshore. Depending on the extent that the roots and rhizomes are impacted, post-storm recovery of vegetation can take two to eight years (Brodhead and Godfrey, 1979). Snyder and Boss (2002) found that vegetation recovery on Santa Rosa Island in northwest Florida following
Fig. 1. Recovery of dune height (dune crest elevation–dune base elevation) presented by Morton et al. (1994) and Priestas and Fagherazzi (2010).
56
C. Houser et al. / Geomorphology 234 (2015) 54–63
Hurricane Opal was dependent on the presence of seed banks, vegetation fragments, and rhizomes that survived the storm, while Hesp (1988) observed that the morphology of the incipient foredune is dependent on the mode of beach colonization, plant density and distribution – in addition to the species and morphology available after a storm. The species available for recolonization vary with respect to the beach state; species richness and zonation are greatest and narrowest on reflective beaches, and lowest and widest on dissipative beaches (Hesp, 1988). The rate that the surviving vegetation is able to re-emerge depends on the depth of the washover and the speciesdependent burial toleration. The faster the vegetation emerges, the greater the chance of that vegetation remaining viable (Maun, 2009). Burial-tolerant species are adapted to washover and quickly colonize the surface through creeping and rhizome extension, but are not effective at trapping sand and building dunes. Conversely, dune-building species expend their energy growing upwards as sediment is trapped, but as a result are slow to colonize following a storm. As modeled conceptually in Fig. 2, an increase in the frequency and/or magnitude of storm surge has the potential to promote burial-tolerant species and the development of low dunes and a new equilibrium state (Stallins and Parker, 2003; Duran and Moore, 2013). Low islands have the potential to become high islands given sufficient time for the return of dune-building grasses and a supply of sediment to support the rebuilding of the dune (Duran and Moore, 2013). Compared to studies that document beach and dune erosion during storms, the rates and mechanisms of post-storm dune recovery have received relatively little attention in the literature - with notable exception for Olson (1958), Saunders and Davidson-Arnott (1990), Morton et al. (1994), McLean and Shen (2006), Houser and Hamilton (2009), Priestas and Fagherazzi (2010), Duran and Moore (2013), Wolner et al. (2013), and Lucas and Carter (2013). The majority of studies focus on erosion of the beach and dune, and the immediate recovery of the nearshore and beach profile following storms (Zeigler et al., 1959; Warnke et al., 1966; Katuna, 1991; Sexton and Hayes, 1991). It is reasonable to expect that the lack of field data reflects the disparate timescales of impact and recovery (see Lee et al., 1998) and the difficulty in maintaining and acquiring funding for long-term beach and dune monitoring programs. The purpose of the present study is to document the recovery of Santa Rosa Island in Northwest Florida using LiDAR data acquired between 2004 and 2010 to determine the rates of dune recovery with respect to the Verhulst (1838) model (Eq. (1)), as well as comparing both the rates and mechanisms of dune recovery with the recovery data of Morton et al. (1994; see Fig. 1). The results of this study are also compared to previous studies that describe the geological controls on island response to recent storm activity and over the Holocene (see Houser et al., 2008a, 2008b; Houser and Hamilton, 2009; Houser, 2012).
2. Study site Santa Rosa Island is a narrow sandy Holocene barrier island extending 96 km from East Pass near Destin to Pensacola Pass in the west (Fig. 3). The focus of this study is a 35 km stretch of Santa Rosa Island in northwest Florida that was impacted by hurricanes Ivan (2004), Dennis (2005) and Katrina (2005). Prior to Hurricane Ivan, the dunes east of Pensacola Beach formed a ridge and swale complex, with the primary dune parallel to the shoreline (see Houser et al., 2008a). During the hurricane seasons of 2004 and 2005, the island morphology changed from a discontinuous foredune backed by hummocky backbarrier dunes and maritime forest (at the bayside cuspate headlands) into washover terraces at the cuspate back-barrier headlands and washover corridors that reach Santa Rosa Sound between headlands. The shoreface profile is relatively steep seaward of the washover corridors (which forces the outermost bar closer to the shoreline) and is responsible for a transverse bar and rip morphology with a steep beachface (see Wright and Short, 1984). This beach and nearshore state limits the exchange of sediment from beach to dune, leading to small discontinuous dunes (Houser et al., 2011; Barrett and Houser, 2012). The beaches seaward of the back-barrier cuspate headlands are more gently sloped and fronted by a relatively dissipative nearshore that promotes the development of larger dunes, which limits washover penetration to the back-barrier shoreline (Houser et al., 2008a, 2008b). The pre-storm dune morphology and the impact of Hurricane Ivan exhibited a quasi-regular variation alongshore that corresponded to the alongshore variation in island width and the location of transverse ridges on the inner-shelf (Houser et al., 2008a, 2008b; Houser and Hamilton, 2009; Houser, 2012). The ridges and swales produce bathymetric highs and lows that force an alongshore variation in the surf similarity parameter and beach state (Houser et al., 2011a,b) and is quite visible in the position of the outermost nearshore bar along Santa Rosa Island (see Houser et al., 2008a, 2008b). The largest dunes before Hurricane Ivan are found landward of the transverse ridges, such that washover was limited in these sections of the island and most of the sediment from the beachface and dunes was deposited within the upper shoreface. In contrast, sections of the coast with little to no dune development before Hurricane Ivan are observed in the narrowest portions of the island (between headlands) and landward of the swales on the inner shelf. As a result, washover penetration was greater in these areas and island breaching was common, leaving the surface close to the watertable and covered by a lag of shell and gravel (Houser, 2012). A geological survey by Houser (2012) supported the theory that the ridge and swale topography may be a transgressive surface, representing a multi-scale feedback of the response and recovery of the island to storms. Houser and Hamilton (2009) observed beachface recovery in the Santa Rosa Unit of the island within two years of Hurricane Ivan, and backshore accretion in the Fort Pickens Unit. Similarly, Leadon (1999) observed rapid beach recovery following Hurricane Opal (1994) on Santa Rosa Island in northwest Florida, while Stone et al. (2004) observed limited dune recovery for ~6 years after the storm. In both sections of the island, the recovery exhibited the same alongshore pattern as the pre-storm dune morphology and storm impact, following the ridge and swale pattern (Houser et al., 2008a, 2008b). The purpose of the present study is to describe the recovery of these sections of the island two years after the 2004/2005-hurricane season. It was anticipated that recovery would vary alongshore, coincident with the transverse ridge and swale topography of the inner-shelf and between the Santa Rosa and Fort Pickens Units of the island. 3. Methodology
Fig. 2. Conceptual model of low and high island equilibrium states.
The alongshore variation in island recovery was examined using light detection and ranging (LiDAR), a mapping technology that has transformed coastal change studies (see Stockdon et al., 2007). Airborne
C. Houser et al. / Geomorphology 234 (2015) 54–63
57
Fig. 3. Location of Gulf Islands National Seashore showing Fort Pickens and Santa Rosa Units in addition to storm tracks from 2004 and 2005.
LiDAR involves the collection of spatially dense and accurate topographic data using aircraft-mounted lasers capable of recording elevation measurements with a vertical accuracy of up to 0.15 m and a horizontal accuracy of b1 m. This study uses 5 LiDAR datasets. The LiDAR data from 2004 and 2005 were collected by the US Geological Survey Center for Coastal and Watershed Studies before and (immediately) after Hurricane Ivan in May and September 2004 respectively, with subsequent surveys completed after Hurricane Dennis (July 2005) and Hurricane Katrina (September 2005). Escambia County completed an additional survey in July 2006. The pre-Ivan dataset (May 2004) is a product of the USACE and Optech Inc., which is more commonly known as the Compact Hydrographic Airborne Rapid Total Survey (CHARTS). This system is a combination of a 1000 Hz hydrographic sensor, a 10,000 Hz topographic sensor, and an image sensor. The post-Ivan LiDAR dataset is a product of the NASA sensor known as EAARL (Experimental Advanced Airborne Research LiDAR). The 2010 LiDAR data were collected by the USACE Joint Airborne LiDAR Bathymetry Technical Center of Expertise (JALBTCX) using the Hawkeye system. This system operates at 1064 nm (red wavelength) for topographic sensing and 532 nm (green wavelength) for bathymetric sensing. Positional accuracy was assessed using post-processed KGPS methods, and the horizontal and vertical accuracy (root mean squared error, or RMSE) of the 2010 LiDAR is ~1.5 m and ~0.15 m, respectively. The point spacing for the pre-Ivan data was ~0.3 m, but only ~1 m in the subsequent surveys. Following Young and Ashford (2006), all data were first converted from raw xyz data into ArcSDE feature classes. The raw point data were interpolated using inverse distance weighting (IDW) into pre-Ivan, post-Ivan, post-Katrina, and recovery elevation models. The IDW power was set to four to negate distant points, and the search radius was limited to the closest eight points, again to negate distant neighbors. The RMSE was 0.34 m, 0.54 m, 0.67 m and 0.62 m for
the pre-Ivan, post-Ivan, post-Katrina, and July 2006 models, respectively. The 2010 LiDAR data was best approximated by a spherical semivariogram model with nugget 0.00, sill 1.00, and range 23 m. The modeled semivariogram was used in an ordinary kriging algorithm to interpolate a 1 m DEM. As described in Rentschlar (2014), the alongshore variation in the aerial cover of vegetation and the recovery of the vegetation cover following Hurricane Ivan was examined using remotely sensed imagery acquired through the USGS EarthExplorer, National Oceanic and Atmospheric Administration (NOAA), Florida Department of transportation (FDOT) and the National Agricultural Imagery Program (NAIP). The distribution of vegetation was quantified using a RGB object oriented classification. Object-oriented classifiers were developed to help overcome the problems of classifying high spatial resolution images (Cleve et al., 2008; Li et al., 2013). Higher resolution images are more likely to have a variety of spectral responses that represent the same object. When using traditional per-pixel classifiers, this high spectral variability increases the chance of salt and pepper classification noise. Objectoriented analysis overcomes this by incorporating the spatial relationship of pixels as well as the spectral relationship when determining the class an object (Myint et al., 2011). Cross-shore transects spaced at 35-m intervals were extracted from the LiDAR data for a total of 302 transects in the Santa Rosa Unit of Gulf Islands National Seashore. Each transect starts from a straight line that bisects the island (point 0 on the profiles) and extends towards the Gulf of Mexico at 2-m intervals to the point where the LiDAR signal is unable to penetrate the water column; the average length of a transect is ~375 m. 4. Results LiDAR imagery of the Santa Rosa Unit from May to September 2004, 2005, 2006 and 2010 are presented in Fig. 4. The alongshore variability
58
C. Houser et al. / Geomorphology 234 (2015) 54–63
Fig. 4. LiDAR imagery from (A) May and (B) September 2004; (C) 2005, (D) 2006 and (E) 2010.
in the beach and dune morphology described by Houser et al. (2008a, 2008b) is evident in the May 2004 LiDAR imagery (Fig. 4A). This morphology is characterized by the island's largest dunes located seaward of back-barrier cuspate headlands and smaller dunes in the narrow sections of the island between headlands. Washover penetration during Hurricane Ivan follows a similar pattern, with the washover terrace extending further landward and into the back-barrier sound in the narrow sections of the island, but relatively limited washover penetration in the wider sections of the island. This variation is also evident in the alongshore variation in washover penetration following Hurricane Ivan (September 2004; Fig. 4B) and the presence of remnant dunes in the widest sections of the island where washover penetration is limited. The influence of the transverse ridge and swale bathymetry on the berm and dune elevation was further examined using cross-shore transects sampled every 35 m alongshore (Fig. 5). The alongshore variability is relatively muted in the 2005 and 2006 profile data, suggesting further erosion of the largest dunes and infilling of the washover and breach areas. Recovery along the entire length of the Santa Rosa Unit is evident in the 2010 data, with the areas of greatest recovery located where the dunes were largest pre-Ivan, and limited recovery present in the narrow sections of the island where pre-storm dune heights were smaller. As a consequence, the alongshore variation in dune height associated with the ridge and swale bathymetry is again evident 5 years after Hurricane Ivan (Fig. 4D).
The alongshore variation in island morphology is also evident in the representative cross-shore profiles in Fig. 6. Recovery in the narrow sections of the island (Profiles 129 and 249) is characterized by berm reconstruction and steepening of the beachface with a near-uniform increase in elevation through storm washover (Fig. 6). Recovery in the wider sections (Profiles 150 and 294) is characterized by berm reconstruction and steepening of the beachface but also includes incipient dune development in the backshore and sporadic dune development further landward (Fig. 6A, D). Dune recovery in the widest sections of the island is associated with the seaward accretion of the beachface through the landward migration of the nearshore bars, while seaward development of the beachface is limited in the narrowest sections of the island. The average maximum island elevation (i.e. the elevation of the berm or dune if present) was calculated for each year to develop a recovery curve similar to Galveston and St. John's Island (see Fig. 1). In areas where there was further erosion of the dune during hurricanes Dennis and Katrina (e.g. transect 150) the recovery was assumed to start after 2006. In the majority of areas, however, there was elevation recovery during hurricanes Dennis and Katrina through berm development. In other areas data were not included for years in which an artificial dune was not yet constructed when the access road was rebuilt (e.g. transects 129 and 151). The alongshore-average recovery is presented in Fig. 7 using the May 2004 (pre-Ivan) dune heights as the predicted height after 10 years of recovery since Hurricane Opal
Fig. 5. Alongshore variation in dune and/or berm height extracted from the LiDAR data from May and September 2004, 2005, 2006 and 2010.
C. Houser et al. / Geomorphology 234 (2015) 54–63
59
Fig. 6. Representative cross-shore profiles from areas with large pre-Ivan dunes (A, D), and from areas with small pre-Ivan dunes (B, C). Shown are profiles (A) 150, (B) 129, (C) 249 and (D) 294.
(1994). The recovery appears to follow the growth model of Verhulst (1838) but with a slower rate of recovery than that seen in the data of Morton et al. (1994) and Priestas and Fagherazzi (2010). For an average pre-storm dune height of ~2.1 m, the growth rate (r) is ~0.25 m yr−1, meaning it would take 7 years for the dunes to reach their pre-Ivan height following either hurricanes Ivan or Katrina. The alongshoreaverage rate of recovery in dune height is compared to the recovery of the vegetation cover (from Rentschlar, 2014) and profile volume in
Fig. 8. On average, the island volume reaches a peak rate of recovery ~4 years after the disturbance, while vegetation and dune height both reach a peak rate of recovery ~5 years after the disturbance. While the vegetation and dune height exhibit peak recovery at the same time, the vegetation recovery initiates more quickly after the storm. The rate of dune recovery remains small until ~3 years following the storm. The rate of dune recovery varies considerably, with a weak correspondence to the height of the pre-storm dune (Fig. 9). Specifically,
60
C. Houser et al. / Geomorphology 234 (2015) 54–63
Fig. 7. Recovery of dune height (dune crest elevation- dune base elevation) from the present study and presented by Morton et al. (1994) and Priestas and Fagherazzi (2010).
the rate of recovery is fastest (r N 0.4) where the pre-storm dunes were the smallest, and slowest (r b 0.2) where the pre-storm dunes were the largest. Similarly, the growth rate of the vegetation (from Rentschlar, 2014) and profile volume are inversely related to the pre-storm dune height. Recovery of both vegetation and profile volume are relatively fast where the pre-storm dune heights were small, and are slowest where the pre-storm dune heights were the largest. A comparison between transects 249 (fast recovery of small dunes) and 294 (slow recovery of larger dunes) suggests that the profile volume recovers rapidly where the pre-storm dunes are the smallest and is complete within two years of the storm (Fig. 10). In contrast, recovery of the profile volume continues for at least seven years following the storm, where the pre-storm dune heights were the largest (transect 294). While recovery of the profile volume is fastest where the pre-storm dunes are the smallest, the ultimate height of those dunes is limited by the availability of sediment from the nearshore. 5. Discussion The rate and extent of post-storm dune recovery determines the vulnerability of barrier islands to future storm events, which in turn control how the island will respond to accelerated sea-level rise and a change in the frequency and/or magnitude of storm events. Compared to studies describing the impact of storms on beaches and dunes, there is a paucity of field and remotely sensed data to quantify the rate of post-storm beach and dune recovery. It is reasonable to assume that the scarcity of field data is due to the disparate timescales of storm response and post-storm recovery. Results presented by Stone et al. (2004) suggest that dune recovery on Santa Rosa Island took up to a decade between hurricanes Opal (1994) and Ivan (2004). Using the May 2004 (pre-Ivan) dune heights as the predicted height after 10 years of recovery,
results of the present study also suggest that the dunes along Santa Rosa Island were eroded and overwashed over the course of several hours during Hurricane Ivan and could take up to a decade to return to their pre-storm height. While it is not clear why the dunes recovered faster on Galveston Island (Morton et al., 1994; Fig. 1), the results of the present study suggest that beach and dune recovery can be modeled using a sigmoid curve (Eq. (1)). A sigmoid recovery curve is consistent with the recovery curves for vegetation following a major disturbance (Hugenholtz and Wolfe, 2005; Rentschlar, 2014) and allows for models to predict the response of barrier islands to changes in storm activity and sea-level rise (Duran and Moore, 2013). Results from Olson (1958) and Saunders and Davidson-Arnott (1990) suggest that the differing alongshore rates of recovery between Santa Rosa Island and Galveston Island are dependent on whether the dunes were primarily scarped or overwashed (see also Sallenger, 2000; Hesp, 2002). However, it is also possible that the faster recovery from Galveston Island (Morton et al., 1994) is a result of the dissipative beachface (see Hesp, 1982, 1988; Short and Hesp, 1982; Bauer and Sherman, 1999) and wider fetch (Bauer and Davidson-Arnott, 2003), but we would have then expected the dune recovery on St. George Island (Priestas and Fagherazzi, 2010) to be similar to Santa Rosa Island. Further study is required to determine how recovery varies along and among different barrier island environments and with respect to beach morphology and whether the dune was scarped or overwashed (see Olson, 1958; Hesp, 2002). The rate of post-storm recovery on Santa Rosa Island varies alongshore in a manner consistent with the morphology of the barrier island. As reported by Houser et al. (2008a, 2008b), the largest pre-storm dunes are seaward of the back-barrier cuspate headlands, while the smallest pre-storm dunes are in the narrowest sections of the island between the back-barrier headlands. This alongshore variation is forced by
Fig. 8. Rate of change in profile volume, vegetation area (from Rentschlar, 2014) and dune height.
C. Houser et al. / Geomorphology 234 (2015) 54–63
61
Fig. 9. Alongshore variation in the May 2004 dune height and the calculated growth rate (r; Eq. (1)).
the transverse ridge and swale bathymetry of the inner-shelf that in turn formed as the island transgressed over the Holocene (Houser, 2012). An intermediate to dissipative nearshore profile develops landward of the transverse ridges, while a reflective to intermediate nearshore profile develops landward of the swales in the narrowest sections of the island (Barrett and Houser, 2012). As argued by Houser and Hamilton (2009) the dunes landward of the transverse ridges are supply limited, while the dunes landward of the swales are (aeolian) transport limited. Results of the present study suggest that this may have been an oversimplification and that the alongshore variation was reinforced by the alongshore variation in washover potential, which is in turn dependent on the rate of dune recovery following storms. Washover in areas with small and/or discontinuous dunes causes sediment to be lost to the back of the island, which limits the supply of sediment that can be returned to the dunes during the recovery period in the absence of seaward winds (Leatherman and Zaremba, 1987). As reported by Houser et al. (2008a, 2008b) the offshore winds are relatively weak and represent an insignificant supply of sediment to the recovering dunes. Areas with larger and continuous dunes have limited washover, but also lose more sediment offshore. The dunes are taller because that sediment can be returned to the beachface through the landward migration of the nearshore bars. In the same way, the dunes in the narrow sections of Santa Rosa Island are the not the smallest because they take longer to recover but rather because there is a limited supply of sediment for recovery. Results of the present study suggest that the landward migration of the nearshore bars and recovery of the profile volume are the primary control of dune recovery on Santa Rosa Island. Limited dune development in the narrow sections of the island reflects the limited supply of sediment available in the nearshore, and profile volume recovery was observed within ~2 years following Hurricane Ivan (see Fig. 10). In comparison, profile recovery continues for at least six years following
Hurricane Ivan in areas where the pre-storm dunes were the largest in May 2004. The vegetation recovery in these same locations reaches its maximum at the same time that the growth of the profile volume begins to slow, with the vegetation growth slowing 10 to 15 years after the disturbance (Rentschlar, 2014). Recovery of the vegetation closely follows the recovery of the profile volume, which may reflect the dependency of vegetation growth on sediment deposition (Zhang and Maun, 1992). As a consequence, dune recovery is relatively fast in areas where the prestorm dunes were small and discontinuous alongshore, but dune heights at Pensacola Beach are ultimately limited by the availability of sediment returned to the beach and backshore by landward migration of the innermost nearshore bar. In contrast, the largest dunes on Santa Rosa Island experience a longer recovery due to the longer time required for the sediment to be transferred from the nearshore bars to the beach and to the dune. The dependency of dune recovery on the availability of sediment in the nearshore (versus in the washover deposits) suggests that the development of small discontinuous dunes is not reversible without an influx of new sediment from an offshore or alongshore source, which in turn requires a cessation of storm activity and the return of dune-building vegetation. However, it also suggests that an inability of the largest dunes to recover due to some anthropogenic interference (Saunders and Davidson-Arnott, 1990; Houser, 2009; Houser et al., 2012), or a rapid succession of storms (Houser and Hamilton, 2009) can force the dune system into an irreversible small dune (i.e. low island) state. It is possible that the longer recovery of the dunes on Santa Rosa Island is a result of the limited erosion and washover that occurred during hurricanes Dennis and Katrina in 2005 (see Houser and Hamilton, 2009). While these storms did not reduce the island elevation, and in some places increased island elevation through washover (Fig. 5), the storms kept the nearshore bars offshore and limited the recovery of the beachface. This suggests, as previously noted, that the
Fig. 10. Comparison of nearshore profile recovery for a representative profile with small dunes (Transect 249) and a representative profile with large dunes (Transect 294).
62
C. Houser et al. / Geomorphology 234 (2015) 54–63
initial recovery of the dune on Santa Rosa Island is limited by the availability of sediment returned to the beachface from the landward migration of the nearshore bars, and that vegetation recovery is a reaction to the frequency of washover that results from the limited dune development. This is consistent with arguments made by Stallins and Parker (2003) and Duran and Moore (2013), but the authors of those studies believed that the transition between low- and high-island states is a consequence of a change in vegetation from effective to ineffective dune builders. Based on these previous studies, a ‘high’ island will develop if the frequency of storm events permits the recovery of dune building vegetation, while a low island will develop if frequent storms limit dune building vegetation (Stallins and Parker, 2003; Duran and Moore, 2013). The vegetation type is a response to changes in the amount of sediment partitioned between the nearshore and overwash, which suggests that a change in the frequency of storms relative to the rate of sediment exchange from the nearshore to the beachface (and ultimately the dune) may lead to a self-reinforcing cycle that ends with a ‘low’ island dominated by maintainer species that preserve flow flat topography (see Wolner et al., 2013). The maximum dune height can be predicted from the transient form ratio (TFR) of Brunsden and Thornes (1979). Through this model dune height depends on the frequency of disturbance (Fi) relative to the recovery time (Ri): T FR ¼
Ri : Ti
ð3Þ
When TFR N 1 the maximum dune height is limited by the frequency of storm events capable of eroding the dune, while sediment supply, beachface slope, or the relative abundance of effective and ineffective dune building vegetation are the primary limit when the TFR b 1. Comparison of the recovery curves with the storm surge annual maximum water levels for Santa Rosa Island (Xu and Huang, 2008), suggests that storm surges capable of eroding the dune base (~ 1 m asl) occur on average every 7 years. While this means that there is a high probability that the largest dunes will have largely recovered before the next erosive storm, there is the potential for a transition to a new equilibrium state if storms occur in close succession, and/or there is an increase in the frequency and magnitude of storm events. The resulting change towards a low island would signal a new equilibrium response and create the potential for rapid island transgression and potential drowning. In this respect, it is the areas with largest dunes on Santa Rosa Island that should be of greatest concern, and be offered the greatest protection, rather than the small-dune areas that are overwash and breached. Without intervention, the larger dunes could remain in a new, lower dune state, and regularly be the site of washover and breaching - a new and possibly irreversible equilibrium state in the future. 6. Conclusions If a barrier island is faced with a change in the frequency and/or magnitude of extreme storms, its resilience will be dependent on the ability of the beach and dune to recover following extreme storms. Results suggest that the recovery of the beach and dune to their prestorm morphology can take up to a decade, and that the rate and amount of recovery is dependent on the pre-storm dune height. On Santa Rosa Island in northwest Florida, recovery varies alongshore. It is fast where the pre-storm dune heights were small and there appears to be no further recovery since most of the sediment was lost to overwash. The largest dunes take longer to recover, with the rate of recovery dependent on the landward migration and welding of the nearshore bars to the beachface and the recolonization by dune-building vegetation. The rate of recovery for Santa Rosa Island in northwest Florida is therefore susceptible to a significant loss of height overall, and the
island has the potential to become a ‘low’ island if there is an increase in the frequency and/or magnitude of storm events. Acknowledgments This study was partially funded through support from the National Science Foundation (BCS-1001126; BCS-1262711). References Barrett, G., Houser, C., 2012. Identifying hot spots of rip current activity using wavelet analysis at Pensacola Beach, Florida. Phys. Geogr. 33, 32–49. Bauer, B.O., Davidson-Arnott, R.G.D., 2003. A general framework for modeling sediment supply to coastal dunes including wind angle, beach geometry, and fetch effects. Geomorphology 49, 89–108. Bauer, B., Sherman, D., 1999. Coastal dune dynamics: problem and prospects. In: Goudre, A.S., Livingston, I., Stokes, J. (Eds.), Aeolian Environments, Sediments And Landforms. J. Wiley and Sons Ltd., London. Brodhead, J.M., Godfrey, P.J., 1979. The Effects Of Off-Road Vehicles On Coastal Dune Vegetation In The Province Lands, Cape Cod National Seashore. Environmental Institute, Massachusetts. Brunsden, D., Thornes, J.B., 1979. Landscape sensitivity and change. Trans. Inst. Br. Geogr. 4, 463–484. Christiansen, M.B., Davidson-Arnott, R., 2004. Rates of landward sand transport over the foredune at Skallingen, Denmark and the role of dune ramp. Dan. J. Geogr. 104, 31–43. Cleve, C., Kelly, M., Kearns, F.R., Moritz, M., 2008. Classification of the wildland–urban interface: a comparison of pixel-and object-based classifications using high-resolution aerial photography. Comput. Environ. Urban. Syst. 32, 317–326. Davidson-Arnott, R.G.D., 1988. Temporal and spatial controls on beach–dune interaction, Long Point, Lake Erie. In: Psuty, N.P. (Ed.), J. Coast. Res. Spec. Issue vol. 3. C.E.R.F., Florida, pp. 131–136. Davidson-Arnott, R.G.D., Law, M.N., 1990. Seasonal patterns and controls on sediment supply to coastal foredunes, Long Point, Lake Erie. In: Nordstrom, K.F., Psuty, N.P., Carter, R.W.G. (Eds.), Coastal Dunes: Form And Process. Wiley, London, pp. 177–200. Davidson-Arnott, R.G.D., Law, M.N., 1996. Measurement and prediction of long-term sediment supply to Coastal Foredunes. J. Coast. Res. 12, 654–663. Delgado-Fernandez, I., Davidson-Arnott, R., 2011. Meso-scale aeolian sediment input to coastal dunes: the nature of aeolian transport events. Geomorphology 126, 217–232. Duran, O., Moore, L.J., 2013. Vegetation controls on the maximum size of coastal dunes. Proc. Natl. Acad. Sci. U. S. A. 110, 17217–17222. Eamer, J.B.R., Walker, I.J., 2010. Quantifying sand storage capacity of large woody debris on beaches using LiDAR. Geomorphology 118, 33–47. Elgar, S., Gallagher, E.L., Guza, R.T., 2001. Nearshore sandbar migration. J. Geophys. Res. 106, 11623–11627. Hesp, P.A., 1982. Morphology And Dynamics Of Foredunes In S.E. Australia. (Unpubl. PhD Thesis), Dept. Geography, University of Sydney (397 pp.). Hesp, P.A., 1988. Surfzone, beach and foredune interactions on the Australian southeast coast. In: Psuty, N.P. (Ed.), J. Coast. Res. Spec. Issue vol. 3. C.E.R.F., Florida, pp. 15–25. Hesp, P., 2002. Foredunes and blowouts: initiation, geomorphology and dynamics. Geomorphology 48, 245–268. Houser, C., 2009. Synchronization of transport and supply in beach-dune interaction. Prog. Phys. Geogr. 33, 733–746. Houser, C., 2012. Role of ridge and swale bathymetry in barrier island transgression. Geomorphology 173 (174), 1–16. Houser, C., Greenwood, B., 2005. Hydrodynamics and sediment transport within the inner surf zone of a lacustrine multiple-barred nearshore. Mar. Geol. 218, 37–63. Houser, C., Greenwood, B., 2007. Onshore migration of a swash bar. J. Coast. Res. 23, 1–14. Houser, C., Hamilton, S., 2009. Sensitivity of post-hurricane beach and dune recovery to event frequency. Earth Surf. Process. Landf. 34, 613–628. Houser, C., Greenwood, B., Aagaard, T., 2006. Divergent response of an intertidal swash bar. Earth Surf. Process. Landf. 31, 1775–1791. Houser, C., Hapke, C., Hamilton, S., 2008a. Controls on coastal dune morphology, shoreline erosion and barrier island response to extreme storms. Geomorphology 100, 223–240. Houser, C., Hobbs, C., Saari, B., 2008b. Post-hurricane airflow and sediment transport over a recovering dune. J. Coast. Res. 24, 944–953. Houser, C., Barrett, G., Labude, D., 2011. Alongshore variation in the rip current hazard at Pensacola Beach, Florida. Nat. Hazards 57, 501–523. Houser, C., Labude, B., Haider, L., Weymer, B., 2012. Impacts of driving on the beach: case studies from Assateague Island and Padre Island national seashores. Ocean Coast. Manag. 71, 33–45. Hugenholtz, C., Wolfe, S., 2005. Biogeomorphic model of dunefield activation and stabilization on the northern Great Plains. Geomorphology 70, 53–70. Katuna, M.P., 1991. The effects of Hurricane Hugo on the Isle of Palms, South Carolina: from destruction to recovery. In: Finkl, C.W., JR, Pilkey, O.H., JR (Eds.), Impacts Of Hurricane Hugo, September 10–22, 1989. J. Coast. Res. 8, pp. 263–273. Keijsers, J.G.S., Poortinga, A., Riksen, M.J.P.M., Groot de, A.V., 2012. Connecting aeolian sediment transport with foredune development. NCK-Days 2012: Crossing Borders In Coastal Research, 13 March 2012–16 March 2012, Enschede, The Netherlands. Law, M.N., Davidson-Arnott, R.G.D., 1990. Seasonal controls on aeolian processes on the beach and foredune. Proc. Symposium On Coastal Sand Dunes. NRC, Ottawa, pp. 49–67.
C. Houser et al. / Geomorphology 234 (2015) 54–63 Leadon, M.E., 1999. Beach, dune and offshore profile response to a severe storm event. Coastal Sediments (1999). ASCE, pp. 2534–2549. Leatherman, S.P., Zaremba, R.E., 1987. Overwash and aeolian processes on a United States Northeast coast. Sed. Geol. 52 (3–4), 183–206. Lee, G., Nicholls, R.J., Birkemeier, W.A., 1998. Storm-driven variability of the beachnearshore profi le at Duck, North Carolina, USA, 1981–1991. Mar. Geol. 148, 163–177. Li, J., Wan, J., Lu, Y., Chen, J., Fu, Y., 2013. A new object-oriented approach towards GIS seamless spatio-temporal data model construction. Pervasive Computing And The Networked World. Springer, pp. 318–331. Lucas, K.L., Carter, G.A., 2013. Change in distribution and composition of vegetated habitats on Horn Island, Mississippi, norther Gulf of Mexico in the initial five years following Hurricane Katrina. Geomorphology 199, 129–137. Maun, M.A., 2009. The Biology Of Coastal Sand Dunes. Oxford University Press (265 pp.). McLean, R., Shen, J.S., 2006. From foreshore to foredune: foredune development over the last 30 years at Moruya Beach, New South Wales, Australia. J. Coast. Res. 22, 28–36. Morton, R.A., 2002. Factors controlling storm impacts on coastal barriers and beaches — a preliminary basis for near real-time forecasting. J. Coast. Res. 18, 486–501. Morton, R.A., Paine, J.G., Gibeaut, J.C., 1994. Stages and durations of post-storm beach recovery, southeastern Texas coast, USA. J. Coast. Res. 884–908. Myint, S.W., Gober, P., Brazel, A., Grossman-Clarke, S., Weng, Q., 2011. Per-pixel vs. objectbased classification of urban land cover extraction using high spatial resolution imagery. Remote Sens. Environ. 115, 1145–1161. Nott, J., 2006. Tropical cyclones and the evolution of the sedimentary coast of northern Australia. J. Coast. Res. 22, 49–62. Ollerhead, J., Davidson-Arnott, R., Walker, I.J., Mathew, S., 2013. Annual to decadal morphodynamics of the foredune system at Greenwich Dunes, Prince Edward Island, Canada. Earth Surf. Process. Landf. 38, 284–298. Olson, J.S., 1958. Lake Michigan dune development 3: lake level, beach and dune oscillations. J. Geol. 66, 473–483. Priestas, A.M., Fagherazzi, S., 2010. Morphological barrier island changes and recovery of dunes after Hurricane Dennis, St. George Island, Florida. Geomorphology 114, 614–625. Rentschlar, E., 2014. Quantifying Vegetation Recovery On Santa Rosa Island. (MSc Thesis), Department of Geography, Texas A&M University (105 pp.). Ruz, M., Meur-Ferec, C., 2004. Influence of high water levels on aeolian sand transport: upper beach/dune evolution on a macrotidal coasta, Wissant Bay, northern France. Geomorphology 60, 73–87. Sallenger Jr., A.H., 2000. Storm impact scale for barrier islands. J. Coast. Res. 890–895. Saunders, K.E., Davidson-Arnott, R.G.D., 1990. Coastal dune response to natural disturbances. Proc. Symposium On Coastal Sand Dunes. NRC, Ottawa, pp. 321–345.
63
Sexton, W.J., Hayes, M.O., 1991. The geologic impact of Hurricane Hugo and post-storm shoreline recovery along the undeveloped coastline of South Carolina, Dewees Island to the Santee Delta. In: Finkl, C.W., JR, Pilkey, O.H., JR (Eds.), Impacts Of Hurricane Hugo, September 10–22, 1989. J. Coast. Res. 8, pp. 275–290. Short, A., Hesp, P.A., 1982. Wave–beach–dune interaction in southeast Australia. Mar. Geol. 48, 259–284. Snyder, R., Boss, C., 2002. Recovery and stability in barrier island plant communities. J. Coast. Res. 530–536. Stallins, J.A., Parker, A.J., 2003. The influence of complex systems interactions on barrier island dune vegetation pattern and process. Ann. Assoc. Am. Geogr. 93, 13–29. Stockdon, H.F., Sallenger Jr., A.H., Holman, R.A., Howd, P.A., 2007. A simple model for the spatially-variable coastal response to hurricanes. Mar. Geol. 238 (1), 1–20. Stone, G.W., Liu, B., Pepper, D.A., Wang, P., 2004. The importance of extratropical and tropical cyclones on the short-term evolution of barrier islands along the northern Gulf of Mexico, USA. Mar. Geol. 210, 63–78. Thieler, E.R., Young, R.S., 1991. Quantitative evaluation of coastal geomorphological changes in South Carolina after Hurricane Hugo. J. Coast. Res. 8, 187–200. Verhulst, P.F., 1838. Notice sur la loi que la population suit dans son accroissement. Corresp. Math. Phys. Publiée A Quetelet 10, 113–121. Warnke, D.A., Goldsmith, V., Grose, P., Holt, J.J., 1966. Drastic beach changes in a lowenergy environment caused by Hurricane Betsy. J. Geophys. Res. 71, 2013–2016. Wolner, C.W.V., Moore, L.J., Young, D.R., Brantley, S.T., Bissett, S.N., McBride, R.A., 2013. Ecomorphodynamic feedbacks and barrier island response to disturbance: insights from the Virginia Barrier Islands, Mid-Atlantic Bight, USA. Geomorphology 199, 115–128. Woodroffe, C.D., 2007. 2007. The natural resilience of coastal systems: primary concepts. In: McFadden, L., Penning-Rowsell, E., Nicholls, R.J. (Eds.), Managing Coastal Vulnerability. Elsevier, Amsterdam, pp. 45–60. Wright, L.D., Short, A.D., 1984. Morphodynamic variability of beaches and surfzones: A synthesis. Mar. Geol. 56, 92–118. Xu, S., Huang, W., 2008. Frequency analysis for predicting 1% annual maximum water levels along Florida coast, US. Hydrol. Process. 22, 4507–4518. Young, A.P., Ashford, S.A., 2006. Application of airborne LIDAR for seacliff volumetric change and beach-sediment budget contributions. J. Coast. Res. 22, 307–318. Zeigler, J.M., Hayes, C.R., Tuttle, S.D., 1959. Beach changes during storms on outer Cape Cod, Massachusetts. J. Geol. 67, 318–336. Zhang, J., Maun, M., 1992. Effects of burial in sand on the growth and reproduction of Cakile edentula. Ecography 15, 296–302.