Unprecedented Historical Erosion of US Gulf Coast: A Consequence of Accelerated Sea‐Level Rise?

Most of the US Gulf Coast is composed of barrier islands, peninsulas, chenier plains, and mainland beaches that are the main line of defense for wetlands, estuaries, and urban and industrial centers from rising sea level and severe storms. These wave‐dominated shorelines are currently experiencing widespread erosion. Using newly acquired and existing results from 13 sites spanning south Florida to south Texas, we compare shoreline migration rates during the late Holocene (∼−4000 to 1850 CE) with historical changes since the mid‐19th century. The records show an overall trend of seaward growth during the late Holocene followed by landward migration or a decrease in the rate of growth during historical time. Diminishing offshore sand supply, human alteration of rivers and coastal sand transport, and severe storms have contributed to this change in shoreline trajectory, but their influence has been mostly limited in extent. The most likely cause of this reversal from coastal stability and growth to widespread shoreline retreat is the dramatic historical increase in the rate of sea‐level rise over the past century.

migration (Byrnes et al., 2012;Himmelstoss et al., 2017;Martinez et al., 2005;Morton, 2008;Paine et al., 2021;Parkinson & Wdowinski, 2023).Results from numerical modeling and experimental studies have shown that wave-dominated coastal settings respond to multiple factors, including barrier width and height, offshore gradient, substrate conditions, and changes in sea level, sand supply, wave climate, storm overwash, and dune dynamics (e.g., Lorenzo-Trueba & Ashton, 2014;Reeves et al., 2021;Rodgers & Paola, 2021).Thus, it is not surprising that they are experiencing changes that can also vary spatially and temporally.This variability presents challenges in measuring and projecting long-term coastal change and identifying its causes.Thus, projections like the one shown in Figure 1 may fail to fully capture future changes along wave-dominated coastal settings.
The rate of sea-level rise increased six-fold in historical time, leading to the argument that Gulf Coast environments may shift into a new equilibrium regime similar to that of the early Holocene when sea level was rising at a similar rate, and the coast was undergoing dramatic change (Anderson et al., 2010(Anderson et al., , 2022;;Donoghue, 2011).This raises the question: at what point will sea-level rise again dominate shoreline behavior at the basin scale and has the Gulf Coast crossed that tipping point?

Methods
Unfortunately, observational records of shoreline migration only extend back to the mid-19th century, so the changes lack pre-historic context.Our research has focused on examining the late Holocene record of coastal change (∼−4000 to 1850 CE) to historical changes since the mid-19th century in the northern Gulf of Mexico region to gain context for comparing historical shoreline migration to pre-historic changes and for examining the causes of these changes.This analysis includes 13 coastal barriers and chenier plains spanning nearly 2,000 km of geologically diverse coastline with historical average rates of relative sea-level rise ranging from 2.2 to 6.6 mm yr −1 (Figure 2).The highest rates occur in Louisiana and Texas, due mainly to subsidence driven by the compaction of Holocene sediments that are typically thicker in both areas (Anderson et al., 2022;Törnqvist et al., 2020Törnqvist et al., , 2008) ) in addition to regional hydrocarbon production (Morton et al., 2006).The rate is lower (2.2-3.3 mm yr −1 ) in Florida and closer to the global average.Our 13 sites across the Gulf of Mexico Coast were selected because they have yielded records of late Holocene coastal evolution that are well constrained by radiocarbon (calibrated using Heaton et al., 2020;Stuiver et al., 2021; see Supporting Information) and/or optically stimulated luminescence (OSL) ages.
The offshore stratigraphic record of coastal change in Florida and Alabama is sparse due to erosion and burial of Holocene coastal deposits beneath a large transgressive sand sheet on the continental shelf, the MAFLA Sheet Sand (McBride et al., 1999).We rely on published records obtained from beach ridges to measure average shoreline migration rates in these areas, but there are few onshore locations where land development has not significantly altered the landscape.Exceptions include Sanibel Island (Stapor et al., 1991) and St. Vincent Island (López & Rink, 2008;Rink & López, 2010) in Florida and Morgan Peninsula in Alabama (Blum et al., 2002;Rodriguez & Meyer, 2006) (Figure 2).Results from Sanibel Island and St. Vincent Island show decadal rates of ridge formation, punctuated by episodes of coastal erosion and formation of larger ridges at roughly a 300-yr periodicity.These findings are indicative of ridge formation by severe storms, as previously described (Donnelly & Giosan, 2008;Otvos, 2000;Taylor & Stone, 1996).Observed truncation of ridges indicates that the average rate of progradation derived from beach ridges is a minimal rate.The Morgan Peninsula age constraints are less precise and more variable ridge orientations indicate greater influence by antecedent topography and inlet migration on ridge formation.Thus, these rates are considered less reliable than those from Sanibel and St. Vincent islands.
Mississippi barrier islands are located more than 20 km from the mainland and exposed to strong longshore currents that have resulted in lateral migration of these islands recorded by westward accretion of beach ridges (Morton, 2008).We include results from detailed studies of Petit Bois and Horn Island in this study (Gal et al., 2021;Hollis et al., 2019;Otvos, 1979;Otvos & Giardino, 2004).The evolution of the barrier islands that rim the Mississippi Delta is closely connected to delta lobe abandonment, which occurred at different times during the late Holocene (Otvos, 2018).They are all relatively thin, landward migrating (transgressive) barriers that experience high rates of subsidence and have been severely impacted by hurricanes.Thus, we have excluded these barriers from this analysis.
In western Louisiana and east Texas, the Holocene coastal geomorphology remains largely intact and includes the Calcasieu and Sabine chenier plains, which consist of well-preserved cheniers that have yielded consistent 10.1029/2023EF003676 3 of 16 radiocarbon and OSL ages that decrease in an offshore direction (Gould & McFarlan, 1959;Hijma et al., 2017).Ridge formation has been interpreted as being regulated by sediment delivery from the Mississippi River, but the frequency of ridge formation is not inconsistent with a storm-influenced mechanism (Anderson et al., 2022).The remaining portion of the Texas coast is composed mainly of barrier islands and peninsulas that have, for the most part, remained in their natural state.Texas has a long history of coastal research that has focused on both onshore and offshore deposits and has yielded a detailed record of Holocene coastal evolution that addresses multiple variables, such as antecedent topography, climate change, relative sea level, and sediment supply as drivers of coastal change (Anderson et al., 2022(Anderson et al., , 2014)).A comprehensive USGS analysis of navigational charts and aerial imagery provides estimates of shoreline migration for the Gulf Coast from the late 19th century through 2001 (Himmelstoss et al., 2017).Since 2001, various state agencies have conducted shoreline migration analyses that, when compared to the USGS data set, indicate an increase in landward and alongshore migration of some coastal barriers while other locations show only minor change (Table 1) (Byrnes et al., 2012;Florida Department of Environmental Protection, 2021;Martinez et al., 2005;Morton, 2008;Paine et al., 2021).

Holocene Coastal Evolution
Formation of most Gulf Coast barriers and chenier plains took place during the late Holocene when the sea level was within a few meters of its current elevation and the average rate of rise was ∼0.5 mm yr -1 (Joy, 2019;Milliken et al., 2008;Törnqvist et al., 2004).While these were ideal conditions for barrier formation, the timing and rate of barrier development varied across the Gulf Coast due mainly to regional differences in sand supply (Anderson et al., 2022(Anderson et al., , 2014;;Otvos, 2018).Results from all 13 study locations are summarized in Figure 3.
Results from Sanibel Island, located on the southwest Florida coast (Figure 2), indicate ∼5.1 km of total seaward growth during the past ∼2.5 ka BP (Stapor et al., 1991), yielding an average long-term progradation rate of +2.0 m yr −1 , with shorter intervals ranging from +0.6 to +5.5 m yr −1 (Figure 3).However, several erosional events are indicated by truncation of ridges, suggesting the actual rate was higher.
Results from previous studies indicate a historical increase in coastal erosion along the west Florida coast (Dean & Houston, 2016;Parkinson & Wdowinski, 2023).Sanibel Island's shoreline was migrating seaward at an average rate of +1.1 m yr −1 between 1855 through 2001 CE, but the rate decreased to +0.57 between 2000 and 2020 CE (Table 1).Likewise, St Vincent Island's shoreline was migrating seaward at a rate of +0.5 m yr −1 between 1855 and 2001 CE but was migrating cross-shore at a rate −0.37 m yr −1 between 1998 and 2018 CE.
In Alabama, Morgan Peninsula was in its current location by ∼5.5 ka BP, indicating a long history of relative shoreline stability (Blum et al., 2002;Rodriguez & Meyer, 2006).The peninsula experienced its most rapid progradation of 1.2 m yr −1 between ∼5.5 and 4.0 ka BP, with an average rate of 0.9 m yr −1 (Figure 3).Historical rates of shoreline migration are poorly constrained, but indicate that the shoreline has been stationary in some locations and retreating landward in others, but there are gaps in these records.
The barrier islands off the Mississippi coast experienced their most rapid growth between ∼6 and ∼2 ka BP (Gal et al., 2021;Hollis et al., 2019;Otvos, 1979;Otvos & Giardino, 2004).The timing of island development varied due to differences in the sources of sand and its dispersal along the coast.Petit Bois Island laterally migrated 1.5 km westward from ∼2.1 ka BP to 100 BP (Hollis et al., 2019), while Horn Island laterally migrated 3.5 km westward between ∼5.1 and ∼3.5 ka BP and vertically aggraded and prograded from ∼3.5 ka BP to 100 BP (Gal et al., 2021).From 1847 to 2010 CE, net lateral migration for Petit Bois and Horn islands was −53.0 and −10.9 m yr −1 , respectively (Table 1).Petit Bois and Horn islands are currently experiencing landward migration at rates of −2.4 and −1.5 m yr −1 , respectively.Hollis et al. (2019) concluded that there is significantly more sand eroded at the eastern ends of islands than is deposited on their western ends.
The Calcasieu and Sabine chenier plains are distinctive features of the western Louisiana and east Texas coast that began to form ∼3.0 ka BP (Gould & McFarlan, 1959).They are marked by prominent beach ridges, but only the Calcasieu ridges have yielded sufficient radiocarbon and OSL ages for this analysis (Gould & McFarlan, 1959;Hijma et al., 2017).These results indicate over ∼11 km of progradation over a 2.4 ka period, yielding an average long-term progradation rate of +4.5 m yr −1 , with rates varying between +0.7 and + 17.0 m yr −1 (Figure 3).Historical shoreline migration data for the Calcasieu chenier plain suggests an average rate of growth of +0.8 m yr −1 during the past century.Erosion of younger beach ridges on the eastern portion of the chenier plain reflects a recent history of erosion, with rates ranging from −7.9 to +8.4 m yr −1 during multi-decadal periods between 1884 and 2005 CE (Table 1).Age constraints for the Sabine chenier plain are not as well constrained as the Calcasieu  chenier plain, but indicate a similar history of growth for both features, followed by a historical trend toward erosion (Anderson et al., 2022).
In Texas, combined onshore and offshore stratigraphic records of barrier evolution reveal spatial and temporal differences in regression, transgression, and aggradation (Anderson et al., 2022(Anderson et al., , 2014;;Rodriguez et al., 2004Rodriguez et al., , 2001)).Bolivar Peninsula is a relatively narrow and low barrier that thickens from less than 2 m to approximately 8 m toward the west and over the Trinity River paleo-valley (Figure 2).Two cross-barrier core transects and 30 radiocarbon ages were used to reconstruct the peninsula's evolution (Rodriguez et al., 2004) (Figure 5a).Initial formation of the peninsula dates back to ∼2.5 ka BP, with progradation interrupted by a major transgressive event after ∼1.2 ka BP that manifest as an erosional surface cutting through the barrier.Offshore, this erosional event is recorded in sediment cores as a landward shift in the boundary between the lower shoreface and marine mud.Following this erosional event, the peninsula experienced between 0.5 and 1.1 km of progradation to form the younger beach ridge set.Radiocarbon ages indicate that this recent phase of barrier progradation occurred after ∼0.65 ka BP, yielding a progradation rate of +3.3 m yr −1 .During historical time, Bolivar Peninsula has retreated cross-shore at an average rate of −0.7 m yr −1 (Table 1).The exception is the western 3 km of the peninsula, where progradation has occurred east of a jetty at the entrance to the Houston Ship Channel (Supporting Information).
Galveston Island is up to 8 m thick at its eastern end, where it overlies the Trinity River paleo-valley, and decreases in thickness toward the west and away from the valley (Anderson et al., 2014).Prominent beach ridges occur on the east end of the island and provide a record of progradation in the form of decreasing ridge ages in a seaward direction (Bernard et al., 1970;Rodriguez et al., 2004) (Figure 5b).Radiocarbon ages indicate an early phase of slow progradation between ∼5.5 and 2.5 ka BP (between +0.3 and +0.7 m yr −1 ), followed by faster progradation (between +1.5 and +3.3 m yr −1 ) between ∼2.5 and 1.7 ka BP.Beach ridges younger than ∼1.6 ka BP have been eroded, but offshore cores record the reversal from progradation to transgression manifest as marine mud onlapping shoreface deposits (Figure 5b).A seawall stabilized the eastern portion of Galveston Island since its construction in the early 1900s.Since the wall was constructed, the shoreline has retreated ∼120 m with the rate having increased after the 1950s.Meanwhile, the shoreline of the western portion of the island has experienced a spatially variable historical cross-shore migration rate that averaged −0.6 m yr −1 between 1850 and 2001 CE and −0.9 m yr −1 between the 1930s and 2019 CE (Table 1).for Follets.See Table S1 in Supporting Information S1 for additional information on radiocarbon dates.Note that the scales are different and cores were collected at or near sea level so topography has been removed.Actual core transects correspond to the stratigraphic sections shown in map views, note multiple offshore core locations for each transect.Arrows and numbers at top of each section are rates of shoreline migration.
Follets Island is a small barrier located west of Galveston Island (Figure 2).The history of the island was reconstructed from four cross-island core transects and 21 radiocarbon ages (Odezulu et al., 2018).
The island is composed of less than 2 m of barrier and upper shoreface sand separated from underlying back-barrier deposits by a sharp transgressive surface (Figure 5c).Near vertical facies boundaries separating barrier island sand from adjacent washover and shoreface deposits above the transgressive surface suggest that the island remained fairly stationary after ∼2.8 ka BP, with an estimated cross-shore migration rate of >−0.3 m yr −1 .Historical records indicate that the shoreline was migrating cross-shore at an average rate of −1.2 m yr −1 from 1850 to 2001 CE, but that rate has increased to −1.9 m yr −1 in more recent times (Table 1).
Matagorda Peninsula is separated from Follets Island by the Brazos and Colorado deltaic headlands (Figure 2).The peninsula ranges in thickness from 6 m at its bay side to 1.5 m at the Gulf side and rests on bay deposits (Wilkinson & Basse, 1978).Onshore and offshore sediment cores sampled a transgressive surface marked by a prominent shell layer.A total of 27 radiocarbon ages from this shell layer cluster between ∼2.5 and ∼1.4 ka BP (recalibrated from Yeager et al., 2019).Two radiocarbon ages from the barrier unit above this surface yielded calibrated ages younger than ∼1.4 ka BP.Thus, the barrier has not migrated landward significantly from its current location since at least ∼1.4 ka BP.This indicates a late Holocene cross-shore migration rate of >−0.5 m yr −1 .During historical time, the peninsula has been migrating landward at an average rate between −0.9 and −1.3 m yr −1 (Table 1).
The central Texas Coast includes Matagorda, San Jose, and Mustang islands.These are among the thickest and oldest barrier islands on the Gulf Coast, which is attributed to the relatively steep coastal gradient and their location within a coastal embayment where convergence of longshore currents has nourished the coast with sand from the ancestral Colorado and Rio Grande deltas (Odezulu et al., 2021;Wilkinson, 1975).These islands are composed mainly of tidal deposits, indicating a long history of barrier aggradation and shoreline stability that began during the middle Holocene (Anderson et al., 2022;Simms et al., 2006;Wilkinson, 1975).
Matagorda Island and San Jose Island are joined by the same beach ridge set and separated by a younger storm surge channel, so they formed as a single barrier island.Offshore sediment cores also show similar stratigraphic records (Odezulu et al., 2021;Rodriguez et al., 2001).Radiocarbon ages are lacking for Matagorda Island, but correlation to a well-dated surface that separates bay muds from barrier sands on adjacent Matagorda Peninsula indicates that the island formed before ∼4.0 ka BP (Anderson et al., 2022).Radiocarbon ages from beach ridges on San Jose Island indicate an average rate of progradation of +3.3 m yr −1 between ∼1.4 and ∼1.2 ka BP and +0.2 m yr −1 between 1.2 ka BP and 100 BP.Historical records indicate that the island has experienced negligible shoreline migration in recent decades (Table 1).
Mustang Island is the oldest and thickest coastal barrier in Texas (Simms et al., 2006).It is composed of up to 30 m of sand that is mainly tidal deposits capped by thin (<2 m) barrier sands (Figure 6).A study of offshore cores showed that the island experienced minor transgression between ∼4.9 and ∼0.9 ka BP, when the shoreline migrated landward less than 1 km, followed by over 3 km of shoreface progradation that began ∼0.9 ka BP (Odezulu et al., 2021).In historical time, Mustang Island's shoreline has remained relatively stable, with a small shift toward landward migration over the past several decades (Table 1).
South Padre Island is a transgressive barrier with thin (<2 m) barrier sands resting on back-barrier washover deposits (Anderson et al., 2022).Radiocarbon ages from the barrier and washover deposits indicate that the island was located at its current location by ∼1.9 ka BP.Based on the modern limits of washover deposits behind the island, we estimate that it has had a cross-shore migration rate of ∼>−0.3 m yr −1 .Between 1850 and 2001 CE, South Padre Island was migrating cross-shore at an average rate of −2.7 m yr −1 , and has continued to retreat at a cross-shore rate of −2.5 m yr −1 over the past several decades (Table 1).
The combined results reveal a variable history of barrier island, peninsula, and chenier plain evolution during the late Holocene (Figure 3).Seven of these coastal barriers (Sanibel Island, St. Vincent Island, Morgan Peninsula, Bolivar Peninsula, Galveston Island, Matagorda/San Jose Islands, and Mustang Island) and the Calcasieu Chenier Plain experienced their most rapid growth after ∼3.5 ka BP, but rates of progradation varied spatially due to variable sediment supply, antecedent topography and severe storm impacts.All of these sites experienced either a reduction in their rate of progradation or a reversal from progradation to retrogradation during historical time.
Barrier islands off the Mississippi coast have a late Holocene history of lateral growth and stability followed by faster rates of lateral migration and erosion during historical time.Follets Island, Matagorda Peninsula, and South Padre Island had transgressive histories during the late Holocene, with rates of shoreline migration in the range of −0.3 to −0.5 m yr −1 .During historical time, the rate of landward migration of these coastal barriers accelerated to −1.3 to −2.5 m yr −1 .If these historical rates of shoreline migration are extrapolated back in time to when these barriers initially occupied their current locations, their modern shorelines would be positioned ∼2-∼3 km inland of their current locations.Likewise, Mississippi barrier islands would be located well to the west of their current locations.

Cause(s) of the Reversal in Coastal Barrier Evolution
The widespread extent of a shift from late Holocene stability and growth to historical retreat of coastal barriers and chenier plains implies a mechanism that is impacting the entire Gulf Coast.The potential causes of this change in coastal evolution include a decrease in sand supply to the coast, increased storm activity, direct human impacts, and accelerated sea-level rise.
Many rivers and streams have contributed sand to the Gulf Coast, but most of these flow through valleys flooded during the Holocene, so the sand transported by these rivers has been sequestered in estuaries that occupy these valleys (Anderson & Rodriguez, 2008;Osterman et al., 2009).The Florida, Alabama, and Mississippi coasts receive little direct sediment supply from rivers.The Mississippi, Brazos, and Rio Grande rivers delivered enough sediment to the coast to maintain offshore deltas during the late Holocene, but only the Mississippi River and Brazos River presently deliver significant amounts of sediment to the coast.Most of Louisiana's barrier islands were formed by reworking of delta lobes.Sand supply from the Brazos River is mostly confined to a wave-dominated delta, with progradation of the delta isolated to a few kilometers of the river mouth (Rodriguez et al., 2000).The Rio Grande River nourished a prominent wave-dominated delta during the late Holocene and was a major source of sand for the south Texas coast.It had been reduced to a small wave-dominated delta when the first relatively accurate navigation charts were constructed in the late 1850s, but today the delta no longer exists and the river struggles to maintain access to the Gulf.The historical demise of the delta is attributed to a combination of increasing aridity in the southwestern US, large-scale water usage for municipal and agricultural purposes and construction of the Falcon Dam in the lower river basin (   S1 in Supporting Information S1 for additional information on radiocarbon dates.Cores were collected at or near sea level so topography has been removed.Arrows and numbers at top of each section are rates of shoreline migration based on radiocarbon ages from multiple core transects (Odezulu et al., 2021;Simms et al., 2006).Colors are the same as Figure 5.
Cushing, 2005).Shorelines on either side of the river mouth are currently eroding at rates ranging from −1 to −2 m yr −1 , which is estimated to be roughly six times the rate prior to 1850 CE.
Most of the sand that composes the modern barriers of the Gulf Coast came from reworking of offshore sand sources and this supply has decreased through time as the rate of transgression decreased and offshore sand bodies were draped in marine mud and sand (Anderson et al., 2022(Anderson et al., , 2014;;Odezulu et al., 2021).Erosion of the Apalachicola, Perdido and Escambia river deltas resulted in the formation of the MAFLA sheet sand, which extends across the inner continental shelf from west Florida to Mississippi and is believed to have been the main source of sand for the west Florida and Alabama coast (McBride et al., 1999).The MAFLA surface has sand waves indicative of east to west sand transport, but quantitative information on how much sand is currently delivered to the coast is lacking.
The Mississippi River constructed four large delta lobes offshore Louisiana and Mississippi during the late Holocene (Frazier, 1967).These delta lobes were subsequently reworked to provide sand for the Louisiana coastal barriers (Otvos, 2018).In Mississippi, erosion of offshore valleys and deltas associated with the Biloxi and Pascagoula rivers provided significant sand for the modern Horn and Petit Bois islands (Hollis et al., 2019;Otvos & Giardino, 2004).Lateral shifting of the Mississippi River is believed to have influenced growth of the Calcasieu and Sabine chenier plains (Gould & McFarlan, 1959), but does not appear to have influenced east Texas barriers (Anderson et al., 2022).In Texas, erosion of offshore valleys and deltas of the Mississippi, Trinity, Brazos, Colorado, and Rio Grande rivers yielded most of the sand composing the modern coastal barriers (Anderson et al., 2022(Anderson et al., , 2014;;Odezulu et al., 2021).Those sand sources have been largely buried beneath the Texas Mud Blanket, which experienced significant expansion during the late Holocene (Weight et al., 2011).In summary, sand supply to the coast has decreased during the late Holocene, but a decrease in sand supply during historical time in the sites presented in this study has been mainly restricted to south Texas and does not explain the observed Gulf-wide changes.
Severe storms are known to significantly erode coasts and remove sand from the Gulf littoral system, but their impacts are restricted in geographic extent and are partly mediated during post-storm recovery when some of the sand eroded from the coast is transported back onshore, with the exception of very fine sand (Anderson et al., 2014;Odezulu et al., 2018;Wallace & Anderson, 2013).Over the past ∼5.0ka BP, paleostorm records across the Gulf of Mexico show significant centennial variability of hurricane impacts over the past ∼5.0ka BP (Brandon et al., 2013;Bregy et al., 2018;Lane et al., 2011;Liu & Fearn, 2000a, 2000b;Rodysill et al., 2020;Wallace & Anderson, 2010).However, it has been demonstrated that over multimillennial time periods, intense hurricane impacts have not varied significantly in the Gulf of Mexico (Wallace et al., 2014).It is possible that severe storms contributed to the observed reversal in shoreline migration in certain locations as well as variability between sites over geologic time (i.e., Donnelly & Giosan, 2008).However, this does not explain the observed Gulf-wide shoreline changes during the historic period.
Direct human impacts on coastal evolution include dam construction and alteration of river sediment transport pathways, beach nourishment, urbanization, dredging of navigation channels and construction of associated jetties, and construction of hard structures aimed at combating shoreline erosion.Of these, urbanization is hardest to quantify as it occurs more gradually, but is known to influence sand supply and distribution, in particular by impeding storm washover (Miselis & Lorenzo-Trueba, 2017).Beach nourishment projects, while also challenging to fully quantify due to incomplete and fragmentary record keeping, would simply serve to slow shoreline retreat and thus suggest these modern rates are minimum values (i.e., reduced rates of shoreline retreat, enhanced rates of shoreline progradation; Hapke et al., 2013) (Table 1).We used historical charts and aerial photographs to measure the impacts of river course alterations, channel dredging and construction of hard structures by measuring the extent of coastal change before and after major coastal alteration (Supporting Information S1).The results show that impacts have been minimal on the Florida, Alabama, and Mississippi coasts, although parts of the west Florida coast have experienced unbridled urbanization that has influenced storm washover.The Mississippi barrier islands are part of the Gulf Islands National Seashore and largely protected from human impacts.In Texas, the most significant human impact includes Galveston Island and the west end of Bolivar Peninsula where disruption of longshore sand transport by jetties on either side of the Houston/Galveston Ship Channel has occurred (Figure S1b in Supporting Information S1).The east end of the island has also experienced significant urban development, with construction of the Galveston Seawall having had the most profound impact (Figure S1c in Supporting Information S1).South Padre Island has also experienced a significant impact from human activity, mainly due to increased demand for water from the Rio Grande River and associated reduction in sand supply to the coast.In summary, direct impacts related to alteration of the coast are mostly confined to Bolivar Peninsula, Galveston Island, and South Padre Island in Texas; otherwise human impacts have been relatively restricted in scale (Figures S1 and S2 in Supporting Information S1).
Of the 13 study sites, Sanibel Island, St. Vincent Island, Matagorda-San Jose Island, and the Calcasieu Chenier Plain have yielded the best records of progradation during the late Holocene that has ended in historical time.
None of these areas have experienced significant human alteration or multiple storm impacts that could have caused this change in shoreline trajectory.Instead, our results indicate a Gulf-wide change in shoreline trajectory during historical time and calls for a basin-scale mechanism (Figure 3).
Tide gauge records and satellite measurements show an acceleration in the rate of global sea-level rise since the 20th century from an average of 1.4 mm yr −1 to the current rate of around 4.0 mm yr −1 (Oppenheimer et al., 2019;Wang, Church, Zhang, & Chen, 2021).This rate varies regionally due to a number of factors, including ocean dynamics, subsidence, and glacial-isostatic adjustment (Love et al., 2016;Morton et al., 2006;Wang, Church, Zhang, & Chen, 2021;Wang et al., 2021).The rate along the Gulf Coast was near 10.0 mm yr −1 in the last decade, most likely due to climate variability (Dangendorf et al., 2023;Yin, 2023).
Not since the early-mid Holocene (∼11650 to ∼4000 BP), has the rate of sea-level rise in the Gulf of Mexico been as fast as in recent decades (Figure 7).This was a period of rapid but variable rates of shoreline migration in western Louisiana and Texas (Anderson et al., 2022).By ∼4000 BP, the average rate of sea-level rise decreased to 0.5 mm yr −1, and the shoreline approached its current location.Modern coastal barriers and chenier plains formed during this time, although there continued to be significant variability in shoreline migration due to regional differences in subsidence, sand supply, and antecedent topography (Anderson et al., 2022;Rodriguez et al., 2004).Numerical models indicate that an increase in sea-level rise of a few mm yr −1 can significantly influence shoreline migration on low gradient, sand-starved coasts (Lorenzo-Trueba & Ashton, 2014;Odezulu et al., 2018), but the impacts of sea-level rise may be delayed over multi-decadal timescales due to a lag between increased sea-level rise and changes in offshore gradients (Mariotti & Hein, 2022).A shift in shoreline trajectory from regression or stasis to transgression is commonly associated with erosion and this break in the stratigraphic record may obscure the timing of the regime change.Galveston, TX exemplifies this where ∼1,600 yr of the sedimentary record is missing leading up to the first historical shoreline maps.Other sites have more complete sedimentary records and resolve a decrease in regression rate (Sanibel, FL) or a shift from regression or stasis to transgression (Bolivar, Matagorda, San Jose, Mustang, and North Padre, TX) in the 19th century around the time sea-level rise began to accelerate.The tide-gauge record of sea-level rise in the Gulf of Mexico spans more than a century, with the linear relative mean sea level corrected for vertical land motion of about 1.4-2.0mm yr −1 from 1900 to 2021 (Dangendorf et al., 2023) yielding a total rise of ∼20 cm during this time (Figure 7).Again, numerical models indicate that this is enough to cause the observed shoreline changes (Lorenzo-Trueba & Ashton, 2014;Odezulu et al., 2018).Sea level is expected to rise another ∼30 cm in the Gulf of Mexico by ∼2040 CE, with significant local variability (Fox-Kemper et al., 2021;Garner et al., 2021;Kopp et al., 2023).The impact on deltas and wetlands is well underway and is projected to result in their virtual destruction (Figure 1).Our results suggest that the shift toward accelerated erosion of wave-dominated coastlines is also well underway.These results are consistent with those from a study of Ocean Beach, Tasmania, where an abrupt shift from episodic shoreline erosion and accretion to persistent shoreline recession during the past 70 yr is attributed to accelerated sea-level rise and increasing wind driven wave influence (Sharples et al., 2020).This implies a global mechanism.

Conclusions
Late Holocene geological records from 13 sites that include coastal barriers and chenier plains spanning the US Gulf Coast reveal differences in late Holocene coastal evolution that are mostly related to spatial variations in sand supply, but the overall trend is one of growth or slow shoreline migration.This contrasts with the historical trend since 1850 CE that is marked by either a decrease in the rate of growth or increased rates of landward shoreline migration.This trend results in part from a long-term decrease in sand supply to the coast, but sand sources vary across the Gulf coast and there is no evidence for a Gulf-wide decrease in supply during historical time.Human alteration of rivers and coasts and severe storms contributed to historical changes, but these impacts were localized.The ubiquitous shift toward increased shoreline erosion along the US Gulf Coast during the 20th century is most plausibly due to the historical acceleration of sea-level rise, currently about an order of magnitude faster than during the late Holocene.With the continued increase in the rate of sea-level rise, coastal populations, engineered shorelines, and use of dwindling sand resources along the US Gulf Coast, there is a critical need to adapt management practices to this new state of continuous landward retreat.Given the variable behavior of wave-dominated coasts during the late Holocene and modern times, coastal inundation models are poorly suited for predicting changes in wave-dominated coastal settings in coming decades.These models likely under-estimate the rate and magnitude of change.

Figure 1 .
Figure 1.Model results from NOAA (https://coast.noaa.gov/slr/#/layer/slr)showing portions of the Gulf Coast that are expected to be drowned by a ∼30 cm rise in Gulf of Mexico sea level by the year ∼2040 CE (Fox-Kemper et al., 2021; Garner et al., 2021; Kopp et al., 2023).Dark blue designates the modern extent of coastal water bodies.Light blue designates areas that would be below sea level based on current elevations.Inset figure shows the global sea-level record beginning 1950 CE based on historical sea-level data and AR6 IPCC projections to 2100 CE (modified from Fox-Kemper et al., 2021).

Figure 2 .
Figure 2. Locations of the 13 sites used to compare Holocene versus historical records of shoreline change along the US Gulf of Mexico coast.Also shown are rates of historical sea-level rise based on tide gauge records (numbers in italics from https://tidesandcurrents.noaa.gov/sltrends/).Rates are generally highest in locations where tide gauges are located in old river valleys with relatively thick Holocene sediments.

Figure 3 .
Figure 3. Shoreline movement from the middle Holocene to 2001 CE for 13 sites along the Gulf of Mexico coast.The summary map (a) contrasts shoreline movement during the late Holocene with historical time.Arrows on location map are scaled by the rate of shoreline movement with larger arrows representing higher rates of seaward movement (blue; positive) or landward movement (red; negative).The arrows positioned on land illustrate the most recent late Holocene rate measured just before 1850 CE and the offshore arrows illustrate the average historical rate.Site locations numbered in orange with adjacent white numbers indicating subsidence rates.Bar charts show rate of shoreline movement from the Holocene to present.Study sites are grouped by type regressive (b), aggradational (c) and lateral migration and transgressive (d) and ordered from east to west along the coast.Matagorda and San Jose Island are grouped because they were connected during Late Holocene progradation.Rates of shoreline movement were measured along profiles positioned in a shore-normal orientation, unless otherwise indicated.Historical average rates from USGS denoted by an asterisk (see Table1) and average late Holocene rates denoted by dashed blue lines.Note changes in the scale of the Y-axis for Bolivar Peninsula and Y-axis units labeled in (b) are the same for (c and d).See TableS1in Supporting Information S1 for details.

Figure 4 .
Figure 4. Aerial view of St. Vincent Island, Florida showing beach ridges that record progradation of the island.Red circles are sample locations and black numbers are optically stimulated luminescence ages (in CE and ka).Black arrows show average rates (in m yr −1 ) of shoreline progradation.Larger ridges are white in color.Ages from López and Rink (2008).The 4-band 8-bit image was acquired in 2017 by the National Agriculture Imagery Program and was obtained from the NOAA Office for Coastal Management, https://coast.noaa.gov.

Figure 5 .
Figure5.Box diagrams showing zoomed in stratigraphy for Bolivar Peninsula, Galveston Island, and Follets islands demonstrating the variable timing and types of barrier evolution (modified fromAnderson et al., 2022).(a) Bolivar Peninsula has had a history of progradation that began ∼2.6 ka and was interrupted by a significant erosional event around 0.7 ka.Recent erosion of the peninsula is recorded by absence of younger beach ridges and onlap of marine mud onto shoreface deposits.(b) Beach ridges on Galveston Island record fairly continuous progradation between ∼5.6 and 1.7 ka, followed by retrogradation that marked by erosion of younger beach ridges and onlap of marine mud onto shoreface deposits.(c) Follets Island sits above back-barrier open bay mud and storm washover, documenting a transgressive history.Following initial rapid transgression after ∼2.9 ka, shoreface, barrier, and washover deposits began to stack vertically, indicating a decrease in shoreline migration.Core locations are indicated by vertical bold lines and dotted lines are isochrons in CE (converted to ka) for Bolivar and Galveston, and BP (converted to ka) for Follets.See TableS1in Supporting Information S1 for additional information on radiocarbon dates.Note that the scales are different and cores were collected at or near sea level so topography has been removed.Actual core transects correspond to the stratigraphic sections shown in map views, note multiple offshore core locations for each transect.Arrows and numbers at top of each section are rates of shoreline migration.

Figure 6 .
Figure 6.Stratigraphic section for Mustang Island illustrating aggradational stacking pattern.Core locations are indicated by vertical bold lines and isochrons in BP (converted to ka) by dotted lines.See TableS1in Supporting Information S1 for additional information on radiocarbon dates.Cores were collected at or near sea level so topography has been removed.Arrows and numbers at top of each section are rates of shoreline migration based on radiocarbon ages from multiple core transects(Odezulu et al., 2021;Simms et al., 2006).Colors are the same as Figure5.

Table 1
Rates of Historical Coastal Retreat Along the Northern Gulf of Mexico

Table 1
) and average late Holocene rates denoted by dashed blue lines.Note changes in the scale of the Y-axis for Bolivar Peninsula and Y-axis units labeled in (b) are the same for (c and d).See TableS1in Supporting Information S1 for details.