Sand in the salt marsh: Contribution of high-energy conditions to salt-marsh accretion

Marine Geology 282 (2011) 240–254

Contents lists available at ScienceDirect

Marine Geology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m a r g e o

Sand in the salt marsh: Contribution of high-energy conditions to salt-marsh accretion Alma V. de Groot a,b,c,⁎, Roos M. Veeneklaas b, Jan P. Bakker b a b c

Kernfysisch Versneller Instituut, University of Groningen, Zernikelaan 25, 9747 AA Groningen, The Netherlands Community and Conservation Ecology Group, Centre for Ecological and Evolutionary Studies, University of Groningen, P.O. Box 11103, 9700 CC, Groningen, The Netherlands Nature Conservation and Plant Ecology Group, and Land Degradation and Development Group, Wageningen University, P.O. Box 47, 6700 AA Wageningen, The Netherlands

a r t i c l e

i n f o

Article history: Received 30 May 2010 Received in revised form 4 February 2011 Accepted 6 March 2011 Available online 13 March 2011 Communicated by J.T. Wells Keywords: Tidal marsh Washover Sedimentation Coastal processes Wadden Sea Schiermonnikoog

a b s t r a c t The environmental dynamics at barrier-island salt marshes are reflected in lateral and vertical textural patterns of the marsh sediment. During normal conditions, fine-grained sediment is deposited, whereas during high-energy conditions also sand accretion can occur. This paper describes the occurrence and importance of sand deposits for the building of salt marshes. The study was carried out in the Wadden Sea on the islands of Schiermonnikoog (NL), Terschelling (NL) and the peninsula of Skallingen (DK). Firstly, we recorded the presence of sand in the sediment representing initial salt-marsh formation. The results indicate that part of the salt marsh developed under conditions that were dynamically enough for sand to be transported. The spatial distribution of these conditions depends on soil elevation and location on the marsh, modified by the presence of artificial sand dikes. Further we recorded the presence and thickness of sand layers within the salt-marsh sediment. Sand layers are found on twenty percent of the marsh area and are partly associated with the local sources of the sand, i.e. marsh creeks, the salt-marsh edge and washovers. In total, sand layers contribute less than ten percent to the volume of marsh deposits on Schiermonnikoog. We dated the layers using the thickness of the deposits and known marsh age. The ages of the layers indicate that for the decadal occurrence of storms capable of depositing sand in the salt marsh, the local hydrodynamics and availability of sand determine whether a site receives sand or not. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Coastal salt marshes are valuable for both biodiversity and coastal protection (e.g. Doody, 2008). They are situated at relatively sheltered places along the coast where sufficient fine-grained sediment is available. Nevertheless, as salt marshes are in direct contact with the sea, they experience a certain degree of dynamics of wind, waves and currents, resulting in variations in marsh extent and vertical growth. Salt marshes may form a protection for the hinterland by buffering these dynamics, but on the other hand if the dynamics become too large (e.g. with changes in climate and sea level) a salt marsh may drown. Knowledge on the natural dynamics of salt marshes is therefore essential for making decisions regarding the conservation of their specific ecosystems and biodiversity, coastal protection and island response to (global) sea-level rise.

⁎ Corresponding author at: Nature Conservation and Plant Ecology Group, and Land Degradation and Development Group, Wageningen University, P.O. Box 47, 6700 AA Wageningen, The Netherlands. Tel.: +31 317 484874, fax +31 317 484845. E-mail addresses: [email protected] (A.V. de Groot), [email protected] (R.M. Veeneklaas), [email protected] (J.P. Bakker). 0025-3227/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.margeo.2011.03.002

The sediment of salt marshes, mainly mud, silt and organic material, reflects the sheltered conditions under which salt marshes generally develop. However, variations in energy conditions, for example from tidal and seasonal periodicity and the occurrence of storms and tsunamis, are reflected in the grain-size distribution of the deposited sediment (e.g. Allen, 2000). Consequently, salt marshes exhibit textural variations in a lateral direction as well as in the vertical stratigraphy (e.g. Wheeler et al., 1999; Allen and Haslett, 2002). During high-energy conditions, coarse-grained material, sand, may be deposited on the salt marsh (e.g. Stumpf, 1983; Ehlers et al., 1993; Roman et al., 1997; Wheeler et al., 1999). Because these events occur infrequently, their deposits form sand layers in the profile, forming records of the dynamics of the salt-marsh environment. Storm-related coarse-grained layers and sand deposits may occur at various locations within a salt marsh, reflecting their sources and depositional processes (Fig. 1). Firstly, sand laminae are observed thinning out from the creek levees (Van Straaten, 1954). During overmarsh tide, the velocity of the flooding waters decreases as the result of lateral spreading from the creek and drag from vegetation on the creek banks. This reduces the sediment-carrying capacity of the floodwaters and the coarsest sediment, sand, is deposited. Secondly, sand layers are observed along the salt-marsh edge, sometimes taking

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

241

Fig. 1. Schematic representation of the expected dominant sand deposits and the responsible processes for the salt marsh of Schiermonnikoog, the Netherlands.

the form of ridges that follow the salt-marsh edge or cliff (Van Straaten, 1954; Ehlers, 1988; Ehlers et al., 1993; Eisma and Dijkema, 1997; Ward et al., 1998; Wheeler et al., 1999). The sand is deposited when the marsh vegetation dampens waves and currents that eroded sand from the intertidal flats (Bartholdy, 1997; Möller et al., 1999; Neumeier and Amos, 2006). As sedimentation rates along the marsh edge are generally high, variations in dynamics will be clearly recorded in the form of distinct layers (Allen, 2000). If a salt marsh advances in cyclic growth, relict edges with their associated layering may remain in the landscape (Wheeler et al., 1999; Van de Koppel et al., 2005; Pedersen and Bartholdy, 2007; Van der Wal et al., 2008; Chauhan, 2009). Thirdly, sand layers on barrier-connected salt marshes may be the result of overwash (e.g. Ehlers et al., 1993; Warren and Niering, 1993; Flemming and Davis, 1994; Oost and De Boer, 1994; Donnelly et al., 2001; Nielsen and Nielsen, 2006; Boldt et al., 2010). During these conditions, the marsh is not only inundated from the back-barrier area, but also from the open sea through gaps in the dunes. The eroded sediment from the beach and dunes is deposited in the form of a washover fan. Finally, aeolian transport may add sand to the salt marsh (French and Spencer, 1993; Neuhaus, 1994; Reineck and Gerdes, 1996). Sources are dunes, washover deposits, beach and beach plains. Additionally, two not-storm related processes that may deposit sand on the salt marsh are ice rafting (Van Proosdij et al., 2006) and tsunamis (e.g. Morton et al., 2007; Komatsubara et al., 2008). Most studies only mention sand layers from one of the above marsh locations, whereas in principle a single salt marsh could contain sand layers from all sources. And even though in allochthonous marshes coarse-grained deposits from high-energy events are less important for salt-marsh building than the deposition of clays and silts (Wheeler et al., 1999), their relative contribution is not known. This leads to the question of what the spatial distribution of sand deposits within a salt marsh looks like, and what this tells about the environmental conditions and their effect on salt-marsh development. And what is exactly the importance of sand for salt-marsh accretion? The ongoing research on salt-marsh ecology at the University of Groningen has produced a large database of soil cores from the Frisian Islands, mainly from the barrier-island salt marsh of Schiermonnikoog, The Netherlands (e.g. Olff et al., 1997; Van Wijnen and Bakker, 2001). This dataset provides a unique opportunity to study the temporal and spatial development of salt marshes and answer questions such as the ones posed above. The purpose of this study is to create an overview of the spatial and temporal aspects of sand deposits in salt marshes, and relate this to the process of salt-marsh development. Firstly, we will record the importance of sand during

the initial stage of salt-marsh formation. Secondly, we describe the occurrence of layers within the marsh deposits and date these deposits using the available data of marsh age and net accretion. We will describe the patterns on the scale of the entire marsh and subsequently zoom in on selected locations to identify the sources of the sand. 2. Methods 2.1. Study sites The main study area is the Dutch barrier island of Schiermonnikoog (Fig. 2). Additional measurements were done on Terschelling (NL) and Skallingen (DK). The studied marshes are all barrier-island salt marshes in the southern North Sea. Such marshes develop when newly-formed dunes shelter an area from flooding from the open sea. The salt marsh itself consists of a vegetated layer of fine-grained sediment, which we call the top layer. It is mainly mineral sediment with a median grain size in the silt or very fine sand fraction. It overlies what once were the intertidal flats, beach plain and lowest part of dunes. This sand surface (the base layer, with the median grain size in the fine sand fraction) generally slopes down from the dune foot towards the intertidal flats, and the marsh surface largely follows this inherited topography (De Leeuw et al., 1993). The barrier island of Schiermonnikoog (53°30′N, 6°10′E) is located in the Dutch Wadden Sea. The tidal range is 2.3 m and Mean High Tide (MHT) is around 1.00 m + NAP (where NAP is the local ordnance datum representing mean sea level). The main salt-marsh complex of Schiermonnikoog is located at the eastern part of the island and is 8 km in length and 0.5–1.5 km in width (Fig. 2A). It is drained by several large creeks, oriented roughly north–south (Fig. 2E). The intertidal flats bordering the salt marsh are dominated by relatively sandy sediment. Some of these creeks are connected to overwash channels (Oost and De Boer, 1994; Ten Haaf and Buijs, 2008). The average net surface elevation change ranges from 0.1 to 0.5 cm a− 1 (with extremes of 1.1 cm a− 1 in the pioneer zone) and depends on marsh age and base elevation (Van Wijnen and Bakker, 2001). The island of Schiermonnikoog has been extending eastward and southward, resulting in gradual new marsh formation. Consequently the marsh exhibits a chronosequence: marsh age decreases from approximately 200 year-old in the middle of the island towards very young at the eastern end. Average top-layer thickness varies accordingly (Olff et al., 1997). In 1959 a sand dike was built that cut off a large part of the beach plain from the North Sea. The sand dike extended 5.5 km from west to east and strongly facilitated the growth

242

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

Fig. 2. Study sites. Panels A–D: measurement locations on Schiermonnikoog (A), Terschelling (C) and Skallingen (D) (see the text for explanation) and the regional setting of the study sites in North-west Europe (B). Panel E: ages of the salt-marsh on Schiermonnikoog, with names of the major creeks (after Kers et al., 1998). In panels A and E, the original location of the artificial sand dike on Schiermonnikoog is given by the dashed line; the part that is still present in dark grey and the by now reworked part in light grey.

of the salt marsh south of it. Part of this area was already vegetated (Van Tooren et al., 1983; De Leeuw et al., 2008). Currently the eastern part of the sand dike is not maintained anymore and has been partly

breached by overwash complexes (Oost and De Boer, 1994). Aerial photographs indicate that since 1970, especially the eastern part of the sand dike has been strongly reshaped by wind and water. By 1989,

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

virtually the entire sand dike east of the 4th creek had been reworked (Fig. 2E). New dunes have developed around the location of the previous artificial sand dike. By now, the salt marsh has extended beyond the previous limit of the sand dike, in the shelter of new natural dunes. Additionally to the eastwards growth, the majority of the marsh is laterally growing southwards. Only the oldest parts are eroding. The second study area is the Dutch Wadden Sea island of Terschelling (53°26′N, 5°28′E, Fig. 2C). The tidal range is 2 m and MHT is 0.87 m + NAP. The salt marsh is located at the eastern part of the island. The majority of the area formed after an artificial sand dike was built between 1931 and 1938 (Roozen and Westhoff, 1985). This part is of more uniform age than the marsh on Schiermonnikoog and the eastern part is presently eroding. Otherwise the salt marsh is comparable to the one on Schiermonnikoog. The third site is the peninsula of Skallingen in Denmark (55°30′N, 8°20′E, Fig. 2D), the barrier spit that forms the northern end of the international Wadden Sea. The tidal range is on average 1.5 m (Bartholdy et al., 2004; Christiansen et al., 2004) and MHT lies around 0.75 m above mean sea level. The salt marsh developed in two phases, starting around 1900 and in the 1950s (Bartholdy and Madsen, 1985; Aagaard et al., 1995). An artificial sand dike was constructed in 1933 close to our measurement locations, sheltering the existing marsh from the North Sea and blocking overwash channels (Aagaard et al., 1995; Nielsen and Nielsen, 2006).

2.2. Measurements 2.2.1. Recording the occurrence of sand The presence of sand within the salt-marsh sediment as a record of dynamic conditions was described in two ways from sediment cores. The cores were taken with a small soil corer of 1 cm diameter and 53 cm length, which is a relatively simple method that can be repeated often. The core was cut in length and examined for grain-size transitions. The thickness of each individual homogeneous layer, if thicker than 0.5 cm, was measured with a ruler to the nearest 0.5 cm (Fig. 3). For simplicity, all sediment where cohesive properties dominated and where most grains were not individually visible was classified as ‘fine-grained’. This encompasses all sediment with considerable amounts of lutum, silt and possibly organic material. Sediments that showed no clear cohesive properties and where most grains were individually visible were classified as ‘sand’. The top layer is defined as the sum of all fine and coarse layers above a continuous layer of sand, the base layer. The salt marsh is defined as the vegetated area with a fine-grained top layer of at least 0.5 cm thick. All measurements without a fine-grained top layer, i.e. creeks, intertidal flats and dunes, were excluded from the analysis. Information on the environmental conditions during initial saltmarsh development was derived from the amount of sand mixed into

Fig. 3. Classification of the marsh top layer from soil cores. The figure is a cross-section of a sediment core, containing fine-grained sediment (dark colours) and sand (light colours).

243

the lower part of the fine-grained marsh sediment. Any fine-grained sediment is considered to be deposited when a location was salt marsh, as such layers are only seldom found on the present intertidal flats and beach plains. For each core, the transition between top and base layer was classified into one of three classes (right panel of Fig. 3). If the transition was abrupt, i.e. there was no mixing of finegrained material and sand, the transition was classified as ‘sharp’. This represents a rather abrupt change from high-energy conditions in which only sand can settle, to low-energy conditions in which sand is not transported and mud is able to settle. Conditions during initial marsh formation were thus calm. If the transition graded along several cm from sand to fines (fining upward, often including all intermediate grain sizes), it was referred to as ‘gradual’. Gradual contacts are the sedimentary result of conditions under which (often very fine) sand as well as fine-grained sediment was transported and deposited. These conditions during initial marsh formation were thus more dynamic. Transitions grading within 1 cm were referred to as ‘intermediate’. For gradual and intermediate transitions, the contact between salt-marsh sediment and base layer was taken as the midpoint of the section. Some of the gradual and intermediate transitions may result from bioturbation: burrowing organisms or plants roots that mix base and top layer. Burrowing organisms are absent from the largest part of the type of salt marsh under consideration (hence the preservation of the layers, Van Straaten, 1954), except for the high marsh (e.g. ants) and in the pioneer zone (organisms that are otherwise common on the intertidal flats). In many of the intermediate and gradual transitions, there was some fining upward observed, grading through the fine sand/coarse silt fraction. We interpreted this as being energy-driven instead of being the result of bioturbation. Mixing by plant roots could in principle occur at the entire salt marsh. However, the good preservation of even small sand layers and sharp transitions in cores where roots were clearly present indicates that this is of minor importance. Sand layers higher in the profile were used as record of past storms. Their thickness was used to identify the contribution of sand to the total salt-marsh sediment budget. The number and thickness of layers within the top layer were determined from the core descriptions (left panel of Fig. 3). Soil elevation was measured with respect to local ordnance datum. In the case of the Netherlands that is Normaal Amsterdams Peil (Dutch Ordnance Level, NAP). For locations in Denmark the Danish equivalent DNN was used (DNN = NAP — 0.16 m, Lassen, 1989). Using local tide conditions, the soil elevations were converted to elevation relative to Mean High Tide (MHT). Two types of levelling equipment were used: a Trimble Spectra Precision laser LL500 and an automatic level. The precision of the measurements was 1 mm and the accuracy was several mm's. Soil elevation minus top-layer thickness gives the elevation of the underlying sand surface, i.e. base elevation. Geographical coordinates were determined from GPS readings (using various Garmin GPS receivers) if measurement spacing was more than 25 m. For smaller distances a combination of measuring tape and GPS recordings was used. The topography of potential sand sources on Schiermonnikoog (salt-marsh edge, creeks, dune foot and current and relict overwash channels) was digitised from georeferenced aerial photographs from 2006 (50 cm horizontal resolution, available through Eurosense bv DKLN-2006) and LIDAR data (5 m horizontal resolution) and a track along the salt-marsh edge in 2004 that was recorded with GPS. The digitising of the creeks was done to much more detail than (for reasons of clarity) is shown in Fig. 2. Local salt-marsh age on Schiermonnikoog was determined from the age of the vegetated surface (Kers et al., 1998, Fig. 2E). This was done for previous studies (e.g. Van Wijnen and Bakker, 2001; Bakker et al., 2003) by tracing vegetated surfaces on aerial photographs and historical maps, with the help of a stereoscope. They used aerial photographs of 1996, 1990, 1979/80, 1959 and 1952 from the Dutch Topografische Dienst in

244

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

Emmen, and maps from 1927, 1888, 1861, 1843 and 1809 in the collection of the ‘Ryksargyf’ in Leeuwarden. This hand-drawn map was subsequently digitised in ArcView. The assigned year of development is the average year between the maps or photographs with and without vegetation. All locations that were classified as being salt marsh during surveys and were not on the age map of 1998 were considered to have developed after the map was produced and were assigned to the year 2000. As the sedimentation of fine-grained material is directly related to the presence of vegetation, the establishment of vegetation is effectively the same as the beginning of marsh development. Marsh age was estimated from island descriptions, maps and aerial photographs for Terschelling (in Roozen and Westhoff, 1985; Schoorl, 2000; Ten Haaf and Buijs, 2008) and Skallingen (in Aagaard et al., 1995; Nielsen and Nielsen, 2006). Data analysis was carried out using MS Excel, Surfer (version 8.00, Golden Software, 1999) and ArcGIS. Aerial photographs of the study area from several years were used for the interpretation of past geomorphology. 2.2.2. Dating the sand layers Storm layers on other European marshes were dated successfully using 137Cs (Ehlers et al., 1993; Wheeler et al., 1999). The large number of cores used in our study made that this method is not feasible for the present work. Therefore, we reconstructed the age of each recorded sand layer in the ‘grid’ and the catchment (see Section 2.3) from its depth within the top layer combined with marsh age. As accretion rates are generally exponentially declining with time (e.g. Allen, 2000), we assumed that the elevation of any point within a core of salt-marsh sediment can be related to the time it was deposited through an exponential function: hðt Þ = hmax + ðhbase − hmax Þe

−ct

ð1Þ

in which h(t) is surface elevation at time t, hmax is the maximum elevation to which the salt-marsh surface grows up, hbase is base elevation and c is a parameter indicating how fast the marsh grows towards hmax. It was assumed that accretion rates are time- and location-dependent but that hmax is the same for every location on the salt marsh. Sea-level rise was not taken into account. The value of hmax was taken as 125 cm + MHT, which is the elevation above which no significant fine-grained deposits were observed in the current dataset. We determined parameter c for all individual cores by fitting Eq. (1) with the core-specific values of local marsh age (t0), base elevation, surface elevation (hsurface) and time of measurement (tmeasurement): c=−

1 tmeasurement − t0

ln

  hsurface − hmax : hbase − hmax

ð2Þ

The year a sand layer was deposited was derived from: tsand = t0 −

  1 h − hmax ln sand : c hbase − hmax

ð3Þ

In case a core contained more than one sand layer, all sand layers were dated individually. The accuracy of tsand is a few years for young parts of the marsh and increases to about 20 year for older areas, related to a number of sources of uncertainty. Firstly, the age of the salt marsh at a particular point is known within a few years for the most recent marshes, but the bandwidth on the older marshes is up to decades. Secondly, the chosen model is a schematisation of marsh accretion that lumps the processes of accretion and compaction, and does not distinguish between the deposition of sand and fines. As a test, we applied the model of Van Wijnen and Bakker (2001) to our data, which includes inundation frequency, accretion and compaction separately and has been calibrated for Schiermonnikoog. The fact that 20% of the sand layers were erroneously dated in the future showed

that this model designed for generalisation and long-term marsh behaviour should not be used for point-by-point hindcasting in small time steps, as marsh-average calibrated values for accretion are not valid on single points. A more sophisticated sedimentation model for Schiermonnikoog is being developed and falls outside the scope of the present study. A third source of uncertainties in calculated age is that fine-grained sediment may be eroded from the surface of the marsh during storms, although this is generally negligible in this type of marshes (Friedrichs and Perry, 2001). Finally, in case the transition from underlying sand surface to salt-marsh deposits of a site is ‘intermediate’ or ‘gradual’, the thickness of the top layer has an uncertainty of 0.5–2 cm, leading to uncertainties of a few years. The cumulative consequence of these uncertainties is that the calculated ages of sand layers from a single event will form a Gaussian or similar distribution when plotted in a histogram. This will be most pronounced for the older layers. We interpret the top of such a distribution as the time that the layers were deposited. Assuming that sand layers are mainly deposited during storms in which the sea level is significantly raised above average, we tried to relate the calculated sand-layer ages to recorded water levels. For this we used the single highest water level recorded per year (Rijkswaterstaat, 2008). Because the time series for Schiermonnikoog did not cover the entire period of interest, additional data from Harlingen are given, which is located 60 km to the west in the Wadden Sea and has comparable water levels to Schiermonnikoog (R2 = 0.74 for water levels from Schiermonnikoog and Harlingen). 2.3. Measurement layout The cores were taken between 2003 and 2007 in a number of measurement layouts that comprise various spatial scales (Fig. 2), the majority of which at Schiermonnikoog. Parts of these layouts were established previous to this study, and earlier data have been reported in various other publications (Olff et al., 1997; Van Wijnen and Bakker, 2001). The current analyses use a total number of 8565 cores of saltmarsh deposits. The layouts, which we describe below, start with measurements at the scale of the entire marsh on Schiermonnikoog, after which we zoom in on a selected catchment (i.e. the area connected by a single creek and its branches), again on Schiermonnikoog. Finally, we zoom further in on locations close to potential sand sources, which are the salt-marsh edge (Schiermonnikoog) and the high-middle marsh (all islands). The presence of a chronosequence on Schiermonnikoog allows comparison between marsh areas of various ages and provides information on the temporal variation in sand deposition. 2.3.1. Salt-marsh landscape Patterns on the scale of the entire salt marsh were described using measurements in a large grid over the salt marsh of Schiermonnikoog that developed after 1900 (‘grid’ in Van Wijnen and Bakker, 2001). It consists of transects 200 m apart, perpendicular to the salt-marsh edge, on which every 50 m a measurement point is located (N = 1747 salt-marsh cores on 590 measurement points, Fig. 2A). At every measurement point, three cores and elevation recordings were taken within 1 m2, which were averaged. The spacing of the grid is relatively large compared to the geomorphology of the marsh, hence the following zooming in on selected locations. 2.3.2. Catchment Patterns on the intermediate scale were described from a young catchment on Schiermonnikoog, which has always been outside the influence of the sand dike. The catchment (approximately 400 m × 500 m) is situated in the east of the salt marsh and started to develop around 1980 (Fig. 2A). A larger creek on the western side of the study area extends into an overwash channel. Cores were taken every 5 m

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254 Table 1 Characteristics of the high-middle marsh measurement transects. Island

Schiermonnikoog Schiermonnikoog Schiermonnikoog Schiermonnikoog Schiermonnikoog Schiermonnikoog Terschelling Terschelling Terschelling Terschelling Skallingen Skallingen Skallingen

Name

SCH_T0 SCH_T1 SCH_T2 SCH_T3 SCH_T5 SCH_T7 TERS_T1 TERS_T2 TERS_T3 TERS_T4 SKAL_T1 SKAL_T2 SKAL_T3

Year of vegetation establishment

Sheltered by artificial sand dike During marsh formation

During further marsh growth

1995 1984 1974 1957 1900 1850 1925–1940 1925–1940 1925–1940 around 1920 1900–1930 1900–1930 1900–1930

– – Yes Yes – – Yes Yes Yes Partly – – –

– – Only Only Only Only Yes Yes Yes Yes Only Only Only

initially initially later later

245

located on comparable elevation of the underlying sand flat, so that their accretionary histories should be comparable. Some transect locations have always been in the shelter of an artificial sand dike, some only during part of their development and some have always been outside the shelter of an artificial sand dike (Table 1). Six transects were measured along the chronosequence of Schiermonnikoog (SCH_T0–T7, Fig. 2A and Table 1, see also e.g. Olff et al., 1997; Van Wijnen and Bakker, 2001), four on Terschelling (TERS_T1–T4, Fig. 2C) and three on Skallingen (SKAL_T1–T3, Fig. 2D). From TERS_T1, TERS_T2, SKAL_T1 and SKAL_T3, only the two outer columns were measured. 3. Results

later later later

along six transects, three perpendicular to the salt-marsh edge and three parallel to it, resulting in a total of 300 data points. 2.3.3. Salt-marsh edge Sand directly originating from the intertidal flats is expected around the salt-marsh edge. Seven transects were established on the old marsh (E1–E7) and three on the young marsh (E8–E10) of Schiermonnikoog, all perpendicular to the salt-marsh edge (Fig. 2A). Part of the transects are included in the catchment layout. Top-layer thickness and soil elevation were measured every 5 or 10 m along transects of between 100 and 350 m long, resulting a total of 499 usable data points. 2.3.4. High-middle marsh Sand from dunes and washover is expected on the high and middle salt marsh. We established transects on all three islands. These consisted of arrays of 10 m wide and 40 to 68 m long, with measurement spacing of 1 m, resulting in 400 to 680 data points per transect. The transects were laid out from the dune foot to the middle marsh and were all

3.1. Dynamics during initial salt-marsh development On the scale of the entire salt marsh of Schiermonnikoog, the proportions in which the sharp, intermediate and gradual transitions occur are approximately equal. Although the three cores taken per measurement point do not always have the same transition type, a large-scale pattern emerges. Sharp transitions, reflecting calm conditions during initial marsh formation, are mostly found at the middle part of the salt marsh and in areas that developed while they were protected by the artificial sand dike (Fig. 4 main panel). These locations are at the low and middle part of the base-elevation gradient (upper panel of Fig. 5). Towards the intertidal flats, coinciding with low base elevations, the sharp transitions give way to gradual transitions that reflect more dynamic conditions during initial marsh growth. Parts of the marsh that have been laterally extending southwards frequently have gradual transitions. At high base elevations along the northern fringe of the marsh, intermediate and gradual transitions dominate. The same overall pattern appears when regarding the transitions as function of distance from the potential sand sources (not shown). The percentage of sharp transitions is highest on marsh areas that developed around the time the artificial sand dike was finished (age classes 1955 and 1964 in lower panel of Fig. 5). Younger marsh areas mostly developed east of the protecting influence of the sand dike or south of already existing marsh area in the western part of the study area and have a larger proportion of gradual transitions.

Fig. 4. Distribution of transitions types. Main panel: transition types on the post-1900 salt marsh at the landscape scale on Schiermonnikoog (‘grid’). The symbols give the dominant transition type from three observations per measurement location (N = 597). The original location of the artificial sand dike is given by the dashed line. Inset: Distribution of transition types at the young catchment on Schiermonnikoog (location see Fig. 2A, N = 300).

246

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

% of observations

100

N= 6

22 54 234 234 224 166 178 174 139 90 64 48 22 15 54

3

80 60 40 20 0

-45 -35 -25 -15 -5

5

15 25 35 45 55 65 75 85 95 105 115 125 135

base elevation (cm+MHT)

% of observations

100

N= 199

216

99

434

225

329

142

64

80 60 40 % sharp % intermediate % gradual

20 0

1913 1939 1955 1964 1974 1986 1993 2000

year of vegetation establishment Fig. 5. Percentage of transition types in cores at the landscape scale (large-scale grid) on Schiermonnikoog, given per base elevation class (upper panel) and per year of vegetation establishment (lower panel). The numbers above the bars represent the number of individual observations in that class.

Within the catchment on the young marsh, sharp transitions dominate and conditions were thus on average calm (inset Fig. 4). Gradual transitions are mainly found around the salt-marsh edge and in depressions with developing creeks. At the transects of the salt-marsh edge, gradual and intermediate transitions dominate (Appendix A). The shape of the cross-sections, or the dominance of lateral accretion or erosion, does not show a relation with the patterns of transition types. At the high-middle salt marsh of Schiermonnikoog, Terschelling and Skallingen, the proportions of sharp, intermediate and gradual transitions vary greatly between locations and the transition types generally form patches of several metres across (Appendix B). The proportions in which they occur differ between the transects. On Terschelling, the conditions during initial marsh formation were calm in the transects that developed in the lee of the sand dike (TERS_T1 and TERS_T2). The conditions were rougher at the transects TERS_T3 and TERS_T4 that developed close to a major creek, a former overwash channel (Ten Haaf and Buijs, 2008). The transects on Skallingen have relatively many sharp transitions. These locations developed relatively far from the salt-marsh edge, in the lee of a large dune area. 3.2. Sand layers 3.2.1. Salt-marsh landscape On the scale of the entire marsh, twenty percent of the measurements contain one, or sometimes more, sand layers (Fig. 6). Not all three points within 1 m2 always have the same number and thickness of layers. The highest number and thickness of sand layers are found along the salt-marsh edge and, to a smaller degree, creek levees. At both locations, often additionally a large number of layers were found that were too thin to be recorded individually, and/or sediment consisting of a mixture of sand and fines. These places are often associated with low base elevation. As a result of that, both the number of layered measurements and the number of layers per core decrease with increasing base elevation (left panel of Fig. 7). At the

high marsh, sand layers are rare or absent. Some of the creeks (the 4th, 5th, 6th, 8th, 11th and 13th creek) are associated with thicker sand deposits than the other creeks (lower panel of Fig. 6). These specific creeks act as overwash channels during storms and most probably all or part of the associated sand layers are washover deposits. Apart from these locations, sand layers are scattered throughout the marsh. When the sand deposits are plotted as a function of distance from the potential sand sources, the pattern is obscured by these scattered locations (Fig. 8). The contribution of the individual sources can consequently not be calculated based on these data. They confirm however that most sand is deposited far from the dunes and overwash channels, around 100 m from the salt-marsh edge. Also within 400 m from overwash channels there is a relative concentration of sand deposits. Single high values are artefacts related to a small number of cores at that distance. A multiple regression of sand-layer thickness with distances from the sources leads to a statistically significant model (p b 0.001), but with an R2 of 0.024 the model is not meaningful. The contribution of the sand layers to the thickness of the top layer is only slightly related to base elevation. Sand contributes mainly to the top layer at marsh area that developed at base elevations lower than 70 cm + MHT (right panel of Fig. 7). The total contribution of sand to the marsh top layer is somewhat under ten percent of the total volume of marsh deposits. Most sand layers are situated relatively deep within the top layer, and were therefore deposited when the marsh at that location was young. Around the salt-marsh edge and creeks, where multiple sand layers occur, the layers are distributed through the entire top layer. Overlaying the transition types with the occurrence of sand layers (compare Fig. 4 with Fig. 6) showed that of all points with gradual transitions, one quarter also has sand layers, and of all points with sand layers, one third has a gradual transition. The ages of the sand layers are given in Fig. 9. The layers were deposited during several individual events, each time at a different location. The oldest layers are situated in the interior marsh between

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

247

% of observations

100

N =1

7 sand layers 6 sand layers 5 sand layers 4 sand layers 3 sand layers 2 sand layers 1 sand layer

80 60 22

40

12 62

239 237

20

186

224

174

178 139

90 22 64 48

0 -60 -40 -20

0

15 4

20 40 60 80 100 120

base elevation (cm + MHT)

contribution to top-layer thickness (%)

Fig. 6. Sand layers within the fine-grained top layer at the landscape scale of Schiermonnikoog. Upper panel: average number of sand layers; lower panel: thickness of sand layers. Crosses indicate layered locations. In the upper panel, the average number of sand layers is only shown if at least two out of three cores on a location contain a sand layer. The lower panel also includes locations where in only one of the cores a relatively thick sand layer was found. The second method leads to a larger number of layered measurement points. The original location of the artificial sand dike on Schiermonnikoog is given by the dashed line.

100 80 60 40

N=1

20

22 12

62

174 239 186 139 237 224 178 90

0 -60 -40 -20

0

64 48 22 15 4

20 40 60 80 100 120

base elevation (cm + MHT)

Fig. 7. Percentage of salt-marsh area on the landscape scale on Schiermonnikoog that contain sand layers within the top layer (left panel) and the volumetric contribution of these layers to the top layer (right panel), as percentage of all observations per base-elevation class. The numbers next to the bars represent the number of observations within each baseelevation class; note that the lowest base-elevation class contains only one observation.

248

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

Fig. 8. Average thickness of sand layers per core as a function of distance from the potential sand sources. Note that the scale on the horizontal axis varies between the figures.

Fig. 9. Ages of sand layers on the landscape scale of Schiermonnikoog. Map: if several layers were present in the profile these were plotted adjacent to the measurement point, with the deepest layers on the left. Inset: the histogram of sand layer ages was corrected for the surface of the salt marsh at the time of deposition. The lines in the graph represent yearly maximum recorded water levels for Schiermonnikoog and Harlingen (60 km west of Schiermonnikoog, Rijkswaterstaat, 2008).

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

the 3rd and 4th creek and date from the time before the artificial sand dike was constructed. These locations did not receive sand afterwards so that the sand layers were most probably deposited by overwash and/or aeolian activity. Later, new marsh developed southward of this area (Fig. 9). The age of these layers decreases with decreasing marsh age, indicating that sand was always deposited within a limited distance from the intertidal flats. Towards the east of the island, the ages of the layers decrease with decreasing age of the marsh. Overall, the frequency of sand deposition increases after 1965. We compared the estimated age of the layers with the records of yearly maximum water levels. Deposition events seemed to have occurred around 1918, 1940–1944, 1966, 1972, 1977, 1987, 1994– 1995 and 2000–2001. Taking into account the accuracy of the estimated ages, these may be related to recorded high tides in 1917, 1936, 1962, 1973, 1976, 1990, 1994 and 2000. From at least the storm in 1976 it is known that sand was deposited on the island during overwash (Ten Haaf and Buijs, 2008). The elsewhere catastrophic flood of 1953 is not reflected in the sand layers. 3.2.2. Catchment Within the young marsh catchment, a third of the area contains sand layers, in most cases only one layer. They are organised in two main patches (outlined in Fig. 10), that appear to be deposited in

249

at least two phases. The older deposit in the northwest of the catchment may be related to one of the storms in 1981 or 1983 and the younger one in the middle of the area in 1994. Whether the sand layers from around 2001 are related to a storm in 2000 is unclear, because assignment of marsh age to the associated locations is uncertain. 3.2.3. Salt-marsh edge A relatively high number of sand layers occur along the salt-marsh edge of Schiermonnikoog, especially where the edge is steep (Appendix A). The sand often forms patches with a size of approximately 100 m perpendicular to the salt-marsh edge. The layers are often associated with a ridge in the underlying sand surface. There is no consistency between the patterns of transition types and sand layers. 3.2.4. High-middle marsh At the high and middle salt marsh of the three islands, sand layers generally occur in patches of several metres across, in between larger areas without layers. In most cases, only one layer is present in the profile. Layer thickness varies mostly between patches. The proportion of the area containing layers and the volumetric contribution of the layers to the top layer varies between the transects (Fig. 11). The volumetric contributions are, except for SCH_T1 and TERS_T1, all below

Fig. 10. Map and histogram of the years of deposition of sand layers at the young catchment on Schiermonnikoog. The outlined areas are discussed in the text. The line in the graph represents yearly maximum recorded water levels (Rijkswaterstaat, 2008).

250

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

Fig. 11. Volumetric contribution of sand layers to the top layer at high-middle marsh transects of Schiermonnikoog, Terschelling and Skallingen.

ten percent. Along the chronosequence of Schiermonnikoog, these proportions are unrelated to marsh age. Transect TERS_T4, close to a large creek, shows relatively thick sand layers over almost half of the transect. The other Terschelling transects contain fewer sand layers. On Skallingen, the high-middle marsh contains hardly any sand layers. 4. Discussion 4.1. Spatial patterns of sand deposition The sand deposits in the salt marsh occur at all the expected locations given in Fig. 1, but also on more interior parts of the marsh. The areas with sand deposits coincide largely with those of the dynamics conditions during marsh beginning, although mostly not on the individual core level. The spatial distribution of the sand is related to the typical zonation of barrier-island salt marshes related to their sloping base layer (De Leeuw et al., 1993). Apart from this pattern, the occurrence of the transition types is patchy on all spatial scales. This suggests that in addition to the general forcing on the landscape scale (waves, currents and wind) there are controls on several (hierarchical) scales that determine the sedimentary result of the environmental conditions. For instance, on the smallest scale this concerns temporarily changing irregularities of the salt-marsh surface, mainly created by individual plants (Van Straaten, 1954; Ehlers et al., 1993; Reineck and Gerdes, 1996; Chang et al., 2008). On the scale of a catchment, vegetation patterns, creek topography and local variations in surface elevation may play a role. The importance of the sand in the form of layers for total salt-marsh sedimentation is for the island of Schiermonnikoog less than ten percent. If we use the small-scale data of the high-middle marsh transects as indication, this figure would be comparable for Terschelling and smaller for Skallingen. Ten percent is not a large proportion of the total sedimentation, though if a marsh would be threatened by sea-level rise also small contributions count. The fact that the sand layers are generally less than a few cm thick means that they do not have a large impact on ‘normal’ sedimentation processes on the studied marsh, such as compaction of underlying sediment (Godfrey and Godfrey, 1974). The percentage of salt marsh area affected by sand deposition is with twenty percent slightly larger. This implies that large parts of the salt marsh do not experience any sand deposition by high-energy waves, currents or wind. The contribution of sand is slightly underestimated as we excluded all sand layers of less than 0.5 cm thickness. However, based on field observations, we estimate that the total volumetric contribution of sand to the top layer will not be much higher, as the omitted layers are very thin, and also the area affected by sand deposition is not much larger.

The local sources of the sand deposits can be partly inferred based on their spatial pattern. If the sand deposits are close to creeks, the saltmarsh edge or overwash channels, the sources of the sand can be assumed to be the creek bed, intertidal flats and dunes plus beach, respectively. It becomes more difficult to identify the sources at locations in the inner marsh, around creeks that connect to overwash channels and at creek mouths. There the sand may have several sources and may be deposited by various processes. This pattern is further affected by the changing locations of the salt-marsh edge, dune ridges and washover mouths during the development of the salt marsh. Consequently there are no clear relations between the topography of the potential sources as it is today and the locations of the sand layers (Fig. 8). Hence, the relative contribution of the sources can only be estimated based on Figs. 6, 8 and 10 and Appendices A and B. The largest proportion of sand deposits seems to be associated with the salt-marsh edge. The distance up to which the layers are deposited into the marsh is around hundred metres, and reaches up to a few hundred metres in case the marsh has been prograding. For outlining the landward extent of washover deposits, the present dataset is not always detailed enough. Based on the present results and aerial photographs it is estimated that substantial deposits within the marsh reach up to approximately 200 m from the dunes and overwash channels. This is typical for barrier islands along the North Sea (Ehlers, 1988; Nielsen and Nielsen, 2006; Ten Haaf and Buijs, 2008). Examples from the USA can be much larger and cover large parts of a barrier island (Donnelly et al., 2006). Possible causes for the smaller extents in the Wadden Sea include the meso- to macrotidal setting that results in relatively short islands and wide inlets, and the absence of hurricanes. Overall, we estimate the area of marsh surface affected by overwash deposits to be around ten percent and the contribution to total marsh sedimentation in the order of a few percent at most. This pattern is of course affected by the building of the sand dike on Schiermonnikoog that has blocked sand transport from the direction of the open sea. As new marsh area developed outside the extent of the dike and more than half of the dike was breached again by overwash, the observed pattern will not deviate too much from a natural situation. The blocking of overwash is thought to have contributed to the loss of marsh area on barrier islands (Godfrey and Godfrey, 1974; Ehlers, 1988; Eleveld, 1999). Based on the present study, it is unclear if the direct contribution from overwash into the marsh is so important for sustaining salt marshes like the one studied here under the present rate of sea-level rise. The direct aeolian input from the dunes into the salt marsh is hard to quantify. Sand transport by the wind is dominated by saltation. Any aeolian sand deposits are, therefore, expected directly adjacent to (formerly) bare sand areas. Part of the layers around overwash channels is likely to be formed by aeolian activity. Further, sand from the sand flat forming the distal end of the island may blow into the young marsh of Schiermonnikoog, where it is captured by the vegetation of the saltmarsh edge (pers. obs.). Deposition from sand by ice rafting may occur over the entire marsh, but will be most prominent around the salt-marsh edge for a mesotidal salt marsh like the one of Schiermonnikoog. It can then not be distinguished from wave- and current-driven processes at salt-marsh edge. Under the temperate climate, the Wadden Sea freezes only occasionally so that this process is probably of minor importance. Sand layers from tsunamis are not expected in this salt marsh. 4.2. Dynamics and sand deposition The transformation from bare sand flat into barrier-island salt marsh is generally considered to take place under calm conditions (e.g. Olff et al., 1997). The presented sediment record indicates that these conditions may be more dynamic than expected, including

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

conditions under which sand transport still occurs. The most dynamic areas during initial marsh formation are the salt-marsh edge, the creek levees and the dune foot. The duration of the process responsible for the gradual or intermediate transitions at a specific location cannot be deducted from the individual cores. Events of sand deposition on the salt-marsh surface occur approximately every decade (Fig. 9). With this given frequency of high-energy events (and perhaps ice rafting), it depends on the topography if a certain site receives sand or not, as the topography determines local hydrodynamics and sand availability. The sand record at Schiermonnikoog suggests that changes in topography that affected sand deposition have taken place, such as lateral marsh growth, creek development and ongoing dune formation. This follows from that firstly most locations with sand layers contain only one layer, deposited when the marsh at that location was young. Secondly, on the core level gradual transitions and sand layers often do not occur together, indicating a change in dynamics. Only along the creeks and salt-marsh edge, the marsh remains open enough for the deposition of new sand layers during the entire lifetime of the marsh. The increased frequency of sand deposition after 1965 (Fig. 9) may be related to the breaching and reworking of the protecting sand dike, a larger proportion of new marsh formation taking place along the generally dynamic salt-marsh edge and changes in the frequency of extreme high tides. Sand layers were often found associated with a ridge in the underlying sand surface. The conditions at such locations are probably so dynamic that when the marsh is young, only sand is deposited and no fine-grained sediment, leading to such a ridge. When a marsh experiences cyclic growth and retreat, relict edges with sand layers may occur (Wheeler et al., 1999; Van de Koppel et al., 2005; Pedersen and Bartholdy, 2007; Van der Wal et al., 2008; Chauhan, 2009). Transects E1, E4, E5, E6 and E7 (Appendix B) contain such relict ridges of various age. However, we also observed sand layers at relict edges where the marsh has always been advancing (transects E5, E6 and E7). Sand layers are therefore not necessarily an indication for a (formerly) retreating marsh edge. The dating of the sand layers was done using a relatively simple approach. This introduces a number of errors in the dating, related to compaction, fluctuating accretion rates due to tidal and weather conditions, surface erosion, the accuracy of measured top-layer thickness, sea-level rise and the difference in accretion rates of finegrained sediment and sand (see also Section 2.2.2). However, when looking at the distribution of the calculated ages (Fig. 9), the dominant influence is the age of the salt marsh derived from the age map of Schiermonnikoog. Without more detailed information on salt-marsh age (which is unlikely to arise as the age map is based on all available aerial photographs), a more sophisticated calculation model for reconstructing age, especially for a large number of individual cores, will hardly give better results. Alternative approaches could be based on models like those of Temmerman et al. (2003) or Bartholdy et al. (2010). Drawbacks of such models are that additional information, for instance suspended sediment concentrations, is necessary and that often marsh-wide values are applied to individual cores which may locally lead to large uncertainties. This study shows that sand deposition on a single salt marsh may occur along various pathways. On a given salt marsh, the frequency, extent, amount and locations of sand deposits will depend on a number of factors. Firstly, the potential of sand transport depends on local dynamics of wind, waves and currents. This is affected by tidal range (affecting how far waves reach into the salt marsh and how much wind setup increases flooding of the marsh; Stumpf, 1983; Friedrichs and Perry, 2001), fetch length, occurrence of hurricanes or cyclones, sheltering from overwash by dunes or human constructions such as seawalls and the occurrence of tsunamis. Secondly, the availability of sand is of importance. Salt marshes bordered by muddy intertidal

251

flats will have a lower potential for sand deposition, although sand layers were observed in muddier environments than the studied marshes (pers. comm. W.E. van Duin). Washover deposits will only be associated with barrier islands and aeolian transport is only possible if there are bare sand areas around. It is not of importance whether the dominant sedimentation is clastic or organic, as sand layers are reported from both type of marshes. Thirdly, the composition and internal stratigraphy of the layers may also vary between locations, and may include shells (Ehlers et al., 1993) and pebbles (in the case of tsunamis, Haslett and Bryant, 2007; Morton et al., 2007; Komatsubara et al., 2008). Finally, the preservation potential of the layers depends on subsequent erosion and bioturbation (Hippensteel, 2008).

5. Summary and conclusions This paper has given an overview of the spatial and temporal aspects of sand deposition on salt marshes, backed up with a large database of soil cores from the North Sea area. These results can be used as a framework for the interpretation of sedimentary records of salt marshes and improving management related to the natural dynamics of salt marshes. Sand may be deposited during the onset of marsh formation, as a considerable part of a salt marsh can initiate under conditions dynamic enough for the transport of sand. During further marsh development, sand layers may be deposited as the result of individual events. On the salt marsh of the present study, the sand layers are not continuous and occur at twenty percent of the salt-marsh area. The sand may have several sources, which are partly reflected in the location of the layers: • Sand taken up from the intertidal flats is deposited on the marsh close to the marsh edge, or transported into the marsh creeks. The sand deposits along relict edges are not necessarily an indication for a formerly retreating marsh edge; • Sand eroded from the marsh creeks is deposited at and around creek levees; • Overwash erodes sediment from the beach and dunes and deposits it onto the salt marsh; • Sand from beach plains, dunes and washover deposits is blown into the salt marsh. The overall contribution of sand layers to total salt-marsh sedimentation in the studied marshes is less than 10%. In general, the volume, frequency and extent of sand deposition on a salt marsh will depend strongly on its geographical and climatological setting.

Acknowledgements Collecting the data for this paper was a joint effort of many people and we wish to thank Nicole Feige, Petra Daniels, Yzaak de Vries and the participants of the Schiermonnikoog field courses 2004–2007 and the Coastal Ecology Expedition 2005. Vereniging Natuurmonumenten is acknowledged for allowing access to the National Park Schiermonnikoog. This work is part of the research programme of the Foundation for Fundamental Research on Matter (FOM). It was financially supported by the Netherlands Organisation for Scientific Research (NWO), through the programme Land-Ocean Interaction in the Coastal Zone (LOICZ no 014.27.005). Dick Visser is thanked for enhancing the graphical work and Harm van Wijnen, Rob de Meijer, Emiel van der Graaf and two anonymous reviewers for discussions and their useful comments on an earlier version of this paper.

252

Appendix A

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

253

Appendix B

Fig. 13. Maps of the distribution of transition type for the high-middle marsh locations on Schiermonnikoog (SCH_T0–T7), Terschelling (TERS_T1–T4) and Skallingen (SKAL_T1–T3). Locations can be found in Fig. 2. The upper part of each figure is located at the dune side. Data are from 2005 (for SCH_T7 from 2007).

Fig. 12. Cross-sections of the salt-marsh edge on Schiermonnikoog with the transition type (upper graphs), number of sand layers and the total thickness of those sand layers. For locations see Fig. 2. The Wadden Sea (South) is to the left of the cross-sections. The black layer represents the top layer on top of the sand base (grey). The marsh edges of the left panels are mainly erosive, those of the right panels mainly accretionary. Data are from 2004.

254

A.V. de Groot et al. / Marine Geology 282 (2011) 240–254

References Aagaard, T., Nielsen, N., Nielsen, J., 1995. Skallingen — Origin and Evolution of a Barrier Spit. Institute of Geography, University of Copenhagen, Copenhagen. 85 pp. Allen, J.R.L., 2000. Morphodynamics of Holocene salt marshes: a review sketch from the Atlantic and Southern North Sea coasts of Europe. Quaternary Science Reviews 19 (12), 1155–1231. Allen, J.R.L., Haslett, S.K., 2002. Buried salt-marsh edges and tide-level cycles in the mid-Holocene of the Caldicot Level (Gwent), South Wales, UK. Holocene 12 (3), 303–324. Bakker, J.P., Bos, D., De Vries, Y., 2003. To graze or not to graze: that is the question. In: Wolff, W.J., Essink, K., Kellerman, A., Van Leeuwe, M.A. (Eds.), Proceedings of the 10th International Scientific Wadden Sea Symposium. Ministery of agriculture, nature management and fisheries, Department of Marine Biology, University of Groningen, Groningen, pp. 67–88. Bartholdy, J., 1997. The backbarrier sediments of the Skallingen peninsula, Denmark. Geografisk Tidsskrift 97, 11–32. Bartholdy, J., Madsen, P.P., 1985. Accumulation of fine-grained material in a Danish tidal area. Marine Geology 67 (1–2), 121–137. Bartholdy, J., Christiansen, C., Kunzendorf, H., 2004. Long term variations in backbarrier salt marsh deposition on the Skallingen peninsula — the Danish Wadden Sea. Marine Geology 203 (1–2), 1–21. Bartholdy, A.T., Bartholdy, J., Kroon, A., 2010. Salt marsh stability and patterns of sedimentation across a backbarrier platform. Marine Geology 278 (1–4), 31–42. Boldt, K.V., Lane, P., Woodruff, J.D., Donnelly, J.P., 2010. Calibrating a sedimentary record of overwash from Southeastern New England using modeled historic hurricane surges. Marine Geology 275 (1–4), 127–139. Chang, E.R., Veeneklaas, R.M., Buitenwerf, R., Bakker, J.P., Bouma, T.J., 2008. To move or not to move: determinants of seed retention in a tidal marsh. Functional Ecology 22 (4), 720–727. Chauhan, P.P.S., 2009. Autocyclic erosion in tidal marshes. Geomorphology 110 (3–4), 45–57. Christiansen, C., Pejrup, M., Kepp, R.R., Nielsen, A., Volund, G., Pedersen, J.B.T., 2004. Tidal and meteorological induced nutrient (N, P) dynamics in the micro-tidal Ho Bugt, Danish Wadden Sea. Danish Jounal of Geography 104 (1), 87–96. De Leeuw, J., De Munck, W., Olff, H., Bakker, J.P., 1993. Does zonation reflect the succession of salt-marsh vegetation — a comparison of an estuarine and a coastal bar island marsh in the Netherlands. Acta Botanica Neerlandica 42 (4), 435–445. De Leeuw, C.C., Grootjans, A.P., Lammerts, E.J., Esselink, P., Stal, L., Stuyfzand, P.J., Van Turnhout, C., Ten Haaf, M.E., Verbeek, S.K., 2008. Ecologische effecten van duinboog- en washoverherstel. University of Groningen, Groningen. 124 pp. Donnelly, J.P., Roll, S., Wengren, M., Butler, J., Lederer, R., Webb III, T., 2001. Sedimentary evidence of intense hurricane strikes from New Jersey. Geology 29 (7), 615–618. Donnelly, C., Kraus, N., Larson, M., 2006. State of knowledge on measurement and modeling of coastal overwash. Journal of Coastal Research 22 (4), 965–991. Doody, J.P., 2008. Saltmarsh Conservation, Management and Restoration. Springer. Ehlers, J., 1988. The Morphodynamics of the Wadden Sea. Balkema, Rotterdam. Ehlers, J., Nagorny, K., Schmidt, P., Stieve, B., Zietlow, K., 1993. Storm-surge deposits in North-Sea salt marshes dated by Cs-134 and Cs-137 determination. Journal of Coastal Research 9 (3), 698–701. Eisma, D., Dijkema, K.S., 1997. The influence of salt marsh vegetation on sedimentation. Intertidal Deposits: River Mouths, Tidal Flats and Coastal Lagoons. CRC Press, Boca Raton, pp. 403–414. Eleveld, M. A., 1999. Exploring coastal morphodynamics of Ameland (the Netherlands) with remote sensing techniques and dynamic modelling in GIS. PhD-thesis, University of Amsterdam. Flemming, B.W., Davis Jr., R.A., 1994. Holocene evolution, morphodynamics and sedimentology of the spiekeroog barrier island system (Southern North Sea). Senckenbergiana Maritima 24, 117–155. French, J.R., Spencer, T., 1993. Dynamics of sedimentation in a tide-dominated backbarrier salt-marsh, Norfolk, Uk. Marine Geology 110 (3–4), 315–331. Friedrichs, C.T., Perry, J.E., 2001. Tidal salt marsh morphodynamics: a synthesis. Journal of Coastal Research SI 27, 7–37. Godfrey, P.J., Godfrey, M.M., 1974. The role of overwash and inlet dynamics in the formation of salt marshes on North Carolina barrier islands. In: Reimold, R.J., Queen, W.H. (Eds.), Ecology of Halophytes. Academic Press, New York, pp. 407–427. Golden Software, I., 1999. Surfer User's Guide; Contouring and 3D Surface Mapping for Scientists and Engineers. Golden Software, Inc., Golden, Colorado. Haslett, S.K., Bryant, E.A., 2007. Reconnaissance of historic (post-AD 1000) high-energy deposits along the Atlantic coast of southwest Britain, Ireland and Brittany, France. Marine Geology 242, 207–220. Hippensteel, S.P., 2008. Preservation potential of storm deposits in South Carolina backbarrier marshes. Journal of Coastal Research 24 (3), 594–601.

Kers, A.S., Van der Brug, S.R., Schoen, L., Bos, D., Bakker, J.P., 1998. Vegetatiekartering Oost-Schiermonnikoog, 1993–1996. Laboratorium voor Plantenoecologie, University of Groningen, Groningen. Komatsubara, J., Fujiwara, O., Takada, K., Sawai, Y., Aung, T.T., Kamataki, T., 2008. Historical tsunamis and storms recorded in a coastal lowland, Shizuoka Prefecture, along the Pacific Coast of Japan. Sedimentology 55 (6), 1703–1716. Lassen, H., 1989. Ortliche und Zeitliche Variationen des Meeresspiegels in der südöstlichen Nordsee. Die Küste. Archiv für Forschung und Technik an der Nord und Ostsee 50, 65–96. Möller, I., Spencer, T., French, J.R., Leggett, D.J., Dixon, M., 1999. Wave transformation over salt marshes: a field and numerical modelling study from north Norfolk, England. Estuarine, Coastal and Shelf Science 49 (3), 411–426. Morton, R.A., Gelfenbaum, G., Jaffe, B.E., 2007. Physical criteria for distinguishing sandy tsunami and storm deposits using modern examples. Sedimentary Geology 200 (3–4), 184–207. Neuhaus, R., 1994. Mobile dunes and eroding salt marshes. Helgoländer Meeresuntersuchungen 48, 343–358. Neumeier, U., Amos, C.L., 2006. The influence of vegetation on turbulence and flow velocities in European salt-marshes. Sedimentology 53 (2), 259–277. Nielsen, J., Nielsen, N., 2006. Washover geomorphology on a barrier spit, Skallingen, Denmark. Journal of Coastal Research SI 39, 678–681. Olff, H., De Leeuw, J., Bakker, J.P., Platerink, R.J., Van Wijnen, H.J., De Munck, W., 1997. Vegetation succession and herbivory in a salt marsh: changes induced by sea level rise and silt deposition along an elevational gradient. Journal of Ecology 85 (6), 799–814. Oost, A.P., De Boer, P.L., 1994. Sedimentology and development of barrier islands, ebb-tidal deltas, inlets and backbarrier areas of the Dutch Wadden Sea. Senckenbergiana Maritima 24, 65–115. Pedersen, J.B.T., Bartholdy, J., 2007. Exposed salt marsh morphodynamics: an example from the Danish Wadden Sea. Geomorphology 90, 115–125. Reineck, H.E., Gerdes, G., 1996. A seaward prograding siliciclastic sequence from upper tidal flats to salt marsh facies (southern North Sea). Facies 34, 209–218. Rijkswaterstaat, 2008. WaterStat. http://www.waterstat.nl/. Roman, C.T., Peck, J.A., Allen, J.R., King, J.W., Appleby, P.G., 1997. Accretion of a New England (USA) salt marsh in response to inlet migration, storms, and sea-level rise. Estuarine, Coastal and Shelf Science 45 (6), 717–727. Roozen, A.J.M., Westhoff, V., 1985. A study on long-term salt-marsh succession using permanent plots. Vegetatio 61 (1–3), 23–32. Schoorl, H., 2000. De Convexe Kustboog; Texel - Vlieland - Terschelling; Bijdragen tot de kennis van het westelijk Waddengebied en de eilanden Texel, Vlieland en Terschelling. Deel 4; De Convexe Kustboog en het eiland Terschelling. Pirola, Schoorl. Stumpf, R.P., 1983. The process of sedimentation on the surface of a salt marsh. Estuarine, Coastal and Shelf Science 17 (5), 495–508. Temmerman, S., Govers, G., Meire, P., Wartel, S., 2003. Modelling long-term tidal marsh growth under changing tidal conditions and suspended sediment concentrations, Scheldt estuary, Belgium. Marine Geology 193 (1–2), 151–169. Ten Haaf, M.E., Buijs, P., 2008. Verdiepende Studies Morfologie: Morfologie, hydraulica en ecologie van washoversystemen. Utrecht University, Utrecht. Van de Koppel, J., Van der Wal, D., Bakker, J.P., Herman, P.M.J., 2005. Self-organization and vegetation collapse in salt marsh ecosystems. The American Naturalist 165 (1), E1–E12. Van der Wal, D., Wielemaker-Van den Dool, A., Herman, P.M.J., 2008. Spatial patterns, rates and mechanisms of saltmarsh cycles (Westerschelde, The Netherlands). Estuarine, Coastal and Shelf Science 76 (2), 357–368. Van Proosdij, D., Ollerhead, J., Davidson-Arnott, R.G.D., 2006. Seasonal and annual variations in the volumetric sediment balance of a macro-tidal salt marsh. Marine Geology 225 (1–4), 103–127. Van Straaten, L.M.J.U., 1954. Composition and structure of recent marine sediments in the Netherlands. Leidse Geologische Mededelingen XIX, 4–96. Van Tooren, B.F., Schat, H., Ter Borg, S.J., 1983. Succession and fluctuation in the vegetation of a Dutch beach plain. Vegetatio 53, 139–151. Van Wijnen, H.J., Bakker, J.P., 2001. Long-term surface elevation change in salt marshes: a prediction of marsh response to future sea-level rise. Estuarine, Coastal and Shelf Science 52 (3), 381–390. Ward, L.G., Kearney, M.S., Stevenson, J.C., 1998. Variations in sedimentary environments and accretionary patterns in estuarine marshes undergoing rapid submergence, Chesapeake Bay. Marine Geology 151 (1–4), 111–134. Warren, R.S., Niering, W.A., 1993. Vegetation change on a northeast tidal marsh: interaction of sea-level rise and marsh accretion. Ecology 74 (1), 96–103. Wheeler, A.J., Orford, J.D., Dardis, O., 1999. Saltmarsh deposition and its relationship to coastal forcing over the last century on the north-west coast of Ireland. Geologie en Mijnbouw- Netherlands Journal of Geosciences 77 (3–4), 295–310.