2009 NORTHEASTERN NATURALIST 16(2):183–208 Multi-decadal Changes in Salt Marshes of Cape Cod, MA: Photographic Analyses of Vegetation Loss, Species Shifts, and Geomorphic Change Stephen M. Smith* Abstract - Salt marsh ecosystems on Cape Cod, MA have exhibited substantial changes within the last 60+ years. Analyses of aerial photographs dating back to 1947 reveals that extensive marsh area loss and alterations in tidal creek structure have occurred where vegetation along the edges of tidal creeks and mosquito ditches in the low marsh has declined or disappeared. Where edge vegetation has not been lost, and where major changes in tidal inlet size have not resulted in flows that cause erosion and bank slumping, marsh area and creek structure has remained very stable. The extent of high-marsh vegetation in virtually all systems has diminished greatly, particularly since the 1980s, with the seaward edge of this zone rapidly retreating in a landward direction. In several systems, this has resulted in high marsh being replaced by barren mudflat. In others, low-marsh advancement has been able to keep pace with high-marsh retreat. These processes are discussed within the context of various biotic and abiotic factors that are the likely agents of change. Introduction The ecological and socio-economic benefits of salt marsh ecosystems to coastal communities are numerous and have been well documented (Fagherazzi et al. 2005 and references therein; Valiela et al. 2002), but they have suffered from human activities. Many have been directly lost or degraded by dredging, filling, and diking. Currently, they face the threat of submergence from accelerated sea-level rise combined with dwindling opportunities for upslope migration due to extensive development of coastal uplands. A relatively new concern is the disappearance of marsh vegetation from US coastal areas along the Gulf of Mexico and Atlantic Ocean (McKee et al. 2004, Ogburn and Alber 2006). Since the late 1990s, this phenomenon has been reported from Louisiana to Maine. Termed “sudden wetland dieback” (SWD), it has been the subject of much discussion among wetland scientists. In New England, salt marsh vegetation losses were first reported in 2002 from along the south shore of Cape Cod (Ron Rozsa, Connecticut Department of Environmental Protection, Hartford, CT, and Scott Warren, Connecticut College, New London, CT, pers. comm.; http://wetland.neers. org). Subsequently, many sites with vegetation losses presumed to represent SWD were documented in all coastal New England states, with the most widespread losses occurring on Cape Cod. The disappearing vegetation was initially thought to be a relatively recent occurrence, with more or less *National Park Service, Cape Cod National Seashore, 99 Marconi Site Road, Wellfleet, MA 02667; stephen_m_smith@nps.gov. 184 Northeastern Naturalist Vol. 16, No. 2 simultaneous losses in many different marshes beginning around or shortly before 2002—possibly as a result of a pathogen or drought (or both) (Adamowicz 2006, Adamowicz and Wagner 2005; http://wetland.neers.org/). However, this timeline was never substantiated since no long-term field monitoring had ever been conducted at these sites and reconstruction of history using aerial photography had not been done. Recent work by Holdredge et al. (in press) has helped to explain the disappearance of the dominant low-marsh species, Spartina alterniflora Loisel. (Smooth Cordgrass), on Cape Cod. Intensive field monitoring and controlled manipulative experiments have revealed that plants are being consumed and eventually killed from intense, continuous grazing by a species of nocturnal, herbivorous crab—Sesarma reticulatum Say (Purple Marsh Crab). The impact of the Purple Marsh Crab is occurring throughout the extent of the low-marsh zone, but is most concentrated along creekbanks. A similar phenomenon has been documented in South America where the herbivorous crab Chasmagnathus granulata Dana (Southwest Atlantic Burrowing Crab) has been linked with the decline of S. densiflora Brongn. (Denseflower Cordgrass) (Bortolus and Iribarne 1999). In the Arctic, overgrazing by geese has created a mosaic of marsh vegetation and bare ground (McLaren and Jefferies 2004). Although a native species, S. reticulatum have reached very high densities on Cape Cod and are significantly impacting vegetation, but the time period over which this population growth has occurred has not been determined. Regardless, crab herbivory has had obvious consequences on marsh structure. Without the binding capacity of living plants, large sections of denuded creek banks are being eroded away, resulting in significant structural changes to creek networks and marsh edge retreat. Upslope of S. alterniflora, large areas of high-marsh vegetation also have vanished. There, it has been mostly Spartina patens (Aiton) Muhl. (Salt Marsh Hay)—the predominant high-marsh species in this region— that has disappeared, although losses of Distichlis spicata (L.) Greene (Saltgrass) and Juncus gerardii Loisel. (Saltmeadow Rush) have also occurred. Unlike S. alterniflora, high-marsh plants are dying with fewer symptoms of herbivory. Even though S. reticulatum can be seen feeding on S. patens (S.M. Smith, pers. observ.), there is frequently a large amount of non-grazed, dead foliage left on the plants. Eventually, only bare ground remains, which is peculiar since these areas are neither waterlogged nor hypersaline like typical salt marsh pannes (Ewanchuk and Bertness 2004). In contrast to the low marsh, where plant losses have no relationship with elevation, the deterioration of the high marsh has consistently occurred along its seaward edge, suggesting a link with hydrology since the lower limit of S. patens is regulated by elevation of mean high tide (Bertness 1991, Nixon 1982). While reductions in the extent of high marsh have been observed elsewhere (Brinson and Christian 1999, Donnelly and Bertness 2001, Orson and Howes 1992, Warren and Niering 1993), to this author’s knowledge the formation of large unvegetated mudflats in the wake of its retreat has not been previously described. 2009 S.M. Smith 185 Collectively, salt marshes constitute critical intertidal habitat along Cape Cod’s coastline, and understanding the ways in which these systems are changing through time is a key component in managing the resource. The intent of this paper is to describe and show examples of the kinds of geomorphologic and vegetation changes that have occurred in salt marsh landscapes across outer Cape Cod during the last ≈60 years. Particular attention has been paid to areas with known vegetation losses (originally labeled as SWD sites) to provide some perspective on timelines and historic vegetation structure. Patterns and temporal trends have been spatially characterized in both quantitative and qualitative ways using aerial and ground-level photography. This synthesis of information can serve as a tool for interpreting current and future field observations and subsequent imagery of Cape Cod salt marshes. Methods Photo acquisition and GIS analysis The analysis of salt marsh change was conducted using methods similar to those of Civco et al. (1986) and Higinbottham et al. (2004). A variety of aerial photographs of outer Cape Cod (defined here as everything east of Barnstable Marsh) taken between 1938 and 2005 were analyzed in this study (Table 1). Because each photo series did not include all the areas of interest in each year, some of the analyses span slightly different ranges of time. The photos were scanned at 600 dpi and saved as TIFF image files. Geo-rectification was done in ARCGIS 9.1 based on 2001 aerial photo mosaics of Cape Cod (provided by the Massachusetts Geographic Information System, Office of Geographic and Environmental Information) and projected in UTM (NAD83) coordinates (zone 19N). The number of points used for rectification ranged between 8–30, and the total root mean square error during the process was <5 m. Spatial resolution of the images was between 0.5 and 1 m. All aerial photographs were shot around the time of low tide when water levels are well below creekbank edges. Figure 1 shows all the marshes that were included in this study. Limited spatial coverage of certain areas precluded analysis of some marshes (e.g., Blackfish Creek, Wellfleet; Allen’s Harbor, Harwich). Some very small systems (<10 acres in size) and marshes that were, or currently are, tidally restricted also were not included; Table 1. Metadata for aerial photos used in salt marsh change analysis. Date Type Pixel res. Source December 1938 Black and white 1.0 Cape Cod National Seashore September 1947 Black and white 0.6 Cape Cod National Seashore July 1952 Black and white 0.9 NRCS (Barnstable) September 1977 Black and white 0.5 Cape Cod National Seashore September 1984 Color IR 1.0 NRCS (Barnstable) October 2000 Color IR 0.5 Cape Cod National Seashore April 2001 Multi-spectral 0.5 MassGIS April 2005 Multi-spectral 0.5 MassGIS 186 Northeastern Naturalist Vol. 16, No. 2 however, the sites that were analyzed comprise most of the salt marsh habitat east of Barnstable. Interpretation of photo signatures All photographs, including the 1940s–1970s black-and-white series, were of high quality. In general, the seaward limit of creekbank S. alterniflora and the high-marsh/low-marsh boundaries, which are very abrupt in New England (Bertness 1991), were obvious. Spartina alterniflora was clearly distinguishable from bare ground and high-marsh vegetation based on color (or shades of grey in the black-and-white photos) and texture. Texture is a very useful diagnostic tool for mapping S. patens as this species has a characteristic “cow-licked” surface. Color IR photography provides an excellent template for analyzing salt marsh landscapes (Van der Wal et al. 2008) as vegetation is even more visually distinct, with bare ground appearing as shades of blue to grey, high marsh (S. patens and D. spicata) Figure 1. Map of Cape Cod showing individual marsh systems and major water bodies. 2009 S.M. Smith 187 as light pink, and low marsh (S. alterniflora) as bright to dark red. Further confirmation of photo signatures was possible using ground-level salt marsh monitoring data collected between 1997 and 2003 (Smith 2004) and lowaltitude oblique-angle photography taken between 1991 and 2007, and by doing current spot checks of locations using a hand-held GPS unit. Delineations of marsh-edge positions in selected areas of interest were done by hand in ARCGIS 9.1. Because the photographs were taken close to the time of low tide, the seaward edges of vegetated creekbanks and islands are quite apparent, especially in September images when plant foliage is still intact and at maximum biomass. Photographs from 1947 (black and white), 1984 (color IR), and 2000 (color IR) were used extensively because of the consistency in timing (all taken in September) and ease of delineating vegetation in these images. Where color IR photographs from 2000 were not available, April 2001 (true color) and April 2005 (true color) photographs were used. Relative position (proximity to open water or upland features), elevation data, and ground-level photography were occasionally used to inform decisions on where to draw boundaries. Change analysis Quantification of change in selected segments of marshes was done using GIS overlay analysis. Smaller portions of marshes were analyzed rather than entire systems due to the complexity of delineating large, spatially complex marshes. Quantitative analysis of entire systems would require hand-delineation of many thousands of polygons, which was beyond the scope of this study. The marsh areas shown in the various figures below are simply good examples of the changes that have occurred. However, qualitative visual assessments of how representative these changes are within the larger systems also were done (e.g., extensive or limited). These assessments were based on both field observations and aerial photography (see Table 2). Changes in the width of tidal creeks were calculated by randomly choosing 5–10 (number based on segment length) point locations (using ARCGIS randomization tool) and then measuring distances across the creek between the two edges at those points over time. Marsh area loss was also estimated by drawing polygons around selected marsh islands or specific vegetation zones (i.e., high marsh) by year and then calculating the area of each using ARCGIS tools. In addition to these quantitative estimates on selected portions of marshes, the total extent of high-marsh loss (entire system) was estimated visually based on broad cover categories of <10% loss, 10–30% loss, and >30% loss for entire systems since the earliest date of photography (1947–1952). Whether losses were more extensive prior to or after 1984 was also assessed (Table 2). Survey of vegetation losses from crab herbivory All marshes discussed below were surveyed during the summers of 2006 or 2007 for evidence of crab herbivory. In contrast to goose grazing, where 188 Northeastern Naturalist Vol. 16, No. 2 Table 2. Summary of characteristics of various Cape Cod salt marshes as determined by ground-level surveys conducted in 2006–2007 and GIS analysis of historic aerial photography (marsh names followed by an asterisk [*] indicate sites that were thought to have experienced sudden wetland dieback [SWD]). S = symptoms of Sesarma grazing present, L = low-marsh edge vegetation losses, C = creek widening/change in structure, T = tidal inlet widening, HD = high-marsh dieback with mudflat formation, HL = high-marsh loss (since 1947 or 1952), and HR = high-marsh retreat more rapid after 1984. Y = yes, E = extensive (observed along most marsh edges, creek segments), L = limited (occurs only in a few creeks/edges), BR = broad (distance between S. alterniflora and S. patens exceeds >5 m), NR = narrow (distance between S. alterniflora and S. patens is <5 m) , and Inc = inconclusive. Marsh Town Type S L C T HD HL HR West End Provincetown Back barrier <10% Hatches Harbor Provincetown Back barrier <10% Pamet Harbor* Truro Back barrier Y E E <10% + riverine The Gut* Wellfleet Back barrier Y L L BR >30% Y Middle Meadow* Wellfleet Back barrier Y L L BR >30% Y Jeremy Marsh Wellfleet Back barrier Y L L BR >30% Y Indian Neck* Wellfleet Back barrier Y E E BR <10% Lt. Island* Wellfleet Back barrier Y E E BR 10-30% Y Audubon Sanctuary* Wellfleet Back barrier Y E E BR 10-30% Y Herring River* Eastham Back barrier Y L L NR >30% Y Boat Meadow* Eastham Back barrier Y L L NR >30% Y Nauset Marsh Eastham Back barrier >30% Y Namskaket Brewster Riverine L Y >30% Y Paine’s Creek Brewster Riverine <10% Y Quivett Creek Brewster Riverine L Y 10-30% Y Pleasant Bay Orleans/ Back barrier L Y >30% Y Chatham Morris Island Chatham Back barrier Y 10-30% Oyster River* Chatham Riverine Y L L Y 10-30% Hardings Beach* Chatham Back barrier Y E E NR 10-30% Y Cockle Cove* Chatham Back barrier Y L L >30% + riverine Red River* Chatham Back barrier Y L L 10-30% Y + riverine Eel Creek/Taylor Chatham Back barrier Y E E 10-30% Pond* + riverine Saquatucket* Harwich Riverine Y E E Y 10-30% Herring River* Harwich Riverine Y E E NR 10-30% Swan River* Dennis Riverine Y E E inc West Dennis Beach* Dennis Back barrier Y E E NR >30% Bass River* Dennis Riverine Y E E Inc Inc Sesuit Harbor Dennis Riverine Y 10-30% Parker River* Yarmouth Riverine Y E E Inc Inc Bass Hole Yarmouth Back barrier 10-30% Y + riverine Barnstable Marsh Barnstable Back barrier 10-30% Y Percentages of total 65 65 74 23 35 100 55 2009 S.M. Smith 189 leaves are generally nipped off cleanly at about mid-height, S. reticulatum grazing results in leaves being shredded, ripped, and torn near the base of the plant. The remaining foliage is very short and has a tattered appearance. As the growing season progresses, the remaining shoot stubble turns brown and dies with no re-growth occurring. Results Five main patterns of change are evident in salt marshes across Cape Cod. They are: i) tidal creek widening, creek structural changes, and marsh area reductions associated with edge vegetation losses; ii) tidal creek widening and creek structural changes associated with increases in the width of tidal inlets; iii) marsh edge/area stability; iv) high-marsh losses (landward retreat) with replacement by un-vegetated mudflats; and v) high-marsh losses balanced by low-marsh encroachment. These categories are described in further detail below. Tidal creek widening, creek structural changes, and marsh area reductions associated with edge vegetation losses Time-series aerial photography shows substantial creek widening and corresponding marsh-area loss over the course of 2 to 5 decades in many outer Cape Cod marshes. The structural changes in creek networks and loss of marsh area discussed in this section are not related to any changes in tidal inlet geomorphology. They are, however, all associated with losses of S. alterniflora along marsh edges that continue today as a result of herbivory (Holdredge et al., in press). A good example is the Swan River (Dennis), which showed progressive increases in channel width and deterioration of marsh peninsulas between 1952 and 2005. As shown in Figure 2, one peninsula was actually transformed into an island during this period as the oxbow was cut to form a new channel. This shift has occurred without any corresponding changes in the size or structure of the tidal inlet on Nantucket sound, which was hardened with rock jetties sometime before 1947. The creekbanks along this segment of Swan River have suffered major vegetation losses dating back at least 2 decades as the characteristic signatures of denuded marsh (blue-grey) are present in 1984 color IR photos. A similar process has occurred in other marshes along the south shore of Cape Cod. In the nearby West Dennis Beach area, up to 50% of marsh has disappeared since 1952 (Fig. 3). Today, only relic root and shoot stubble remains along most of the creek banks and island edges in both systems. Similar losses are found along Parker River (Yarmouth), Bass River (Yarmouth/Dennis), Herring River (Harwich), and Saquatucket Harbor (Harwich) (see Fig. 1 for locations). Saquatucket Harbor, in particular, has undergone remarkable changes over the last several decades. In contrast to the systems discussed above, the northern portion of this marsh was removed and the main channel widened in 1969 to create a harbor. Since 1971, the western portion of the fringing marsh has lost an estimated 63% of its original area (data not shown). Oblique-angle aerial photos show heavily denuded marsh edges by 1991, which is the earliest 190 Northeastern Naturalist Vol. 16, No. 2 Figure 2. Changes in the Swan River (Dennis) between 1952 and 2005. Note the peninsula within the circled area that eventually becomes an island. The right side of the large photo shows this segment of marsh in April 2005 with the 1952 channel widths demarcated as black lines (quantitative changes are listed in the embedded table). A current, ground-level view of the island is shown in the inset photo (view is to the southeast). Figure 5 (opposite page, bottom). Staggered times for the onset of S. alterniflora losses: color IR photos of a marsh along the Oyster River (Chatham) (first two photos, top left) showing the loss of edge vegetation (bright red signature indicated by arrows) between 1984 and 2000 and of a Bass River marsh (Dennis) (top right) where edge vegetation was already absent by 1984. The bottom photos show ground-level views of the presence (left) and absence (right) of robust creekbank S. alterniflora that can be detected in high-level aerial photographs. 2009 S.M. Smith 191 Figure 3. Changes in an island marsh near West Dennis Beach (Dennis) between 1952 and 2005 with numerical estimates of loss (white polygon demarcates outer perimeter of marsh in 1952). Figure 4. Vertical and oblique aerial photographs of West Dennis Beach marsh in 1984 (left) and Saquatucket marsh in 1991 (right) (Saquatucket photo by E.R. Lilley). Vegetation losses are indicated by the blue-gray signatures in the former while the obvious bare ground (brown) is visible the latter (indicated by white arrows). date that low-level photos have been found for any location originally labeled as a “SWD site” along the south shore of Cape Cod (Fig. 4). 192 Northeastern Naturalist Vol. 16, No. 2 The vegetation losses in Saquatucket Harbor do not seem to be a consequence of harbor creation. For instance, the loss of marsh there was much greater between 1984 and 2000—long after the harbor was dug—than it was immediately following it (between 1971 and 1984). Moreover, harbors have been carved out of other marshes on Cape Cod without having this kind of effect on bank vegetation (e.g., Sesuit Harbor, Dennis). At Saquatucket Harbor, the vegetation disappeared well before any of the peat or sediment had eroded away and there was no undercutting or bank slumping, all of which suggest that increased tidal flows were not the cause of deterioration. In fact, the denuded marsh is identical to those that have not undergone any such engineering but have suffered from overgrazing. That said, the change in tidal regime resulting from harbor creation may allow greater erosion along already denuded creek banks. The photographs of the West Dennis Beach and Saquatucket Harbor marshes in Figure 4 are highly informative because they clearly show that vegetation losses began well before the presumed onset of SWD on Cape Cod. In fact, color IR signatures of tall, vigorously growing S. alterniflora normally found along creekbanks are absent in images of a number of marshes dating back as far as 1984, indicating that losses occurred prior to this date. At sites where there was a distinct absence of creekbank vegetation in 1984, there is typically little to none today—only relic peat remains and the area of denuded marsh have generally persisted or increased. Among different systems, however, the onset of vegetation loss appears to be staggered in time. Some marshes did, in fact, bear the signatures of vigorous creekbank S. alterniflora in 1984, but lost it by 2000. Oyster River and Harding’s Beach (Chatham) marshes follow this timeline (Fig. 5). The Herring River (Harwich) is another system that did not begin losing its creekbank S. alterniflora in large amounts until sometime after 1984 (i.e., the signature of creekbank vegetation was generally present in 1984). It is also an excellent example of how changes in creek geomorphology are, in part, linked with the timing of vegetation loss. Compared to the Swan River (Fig. 2), there was little alteration in the shape and size of the marsh peninsula shown in Figure 6 for over 30 years (between 1952 and 1984). This stasis is contrasted by accelerated deterioration thereafter. The more complex architecture of marshes around Wellfleet and Truro facing Cape Cod Bay has undergone more complex patterns of deterioration. Pamet Harbor (Truro) and Lt. Island (Wellfleet) are two such areas where large portions of contiguous marsh have broken apart to form arrays of tidal channels, mudflats, and shrinking islands (Fig. 7). The time period over which these marshes have deteriorated is roughly similar to that of the Herring River in Harwich. In this regard, there was a much higher degree of geomorphic stability between 1947 and 1984, after which there was more obvious decline. In certain locations around Lt. Island, tidal creeks showed a >100% increase in width between 1984 and 2000, having grown by as much as 20 m (Fig. 8). Similar to the evidence for Saquatucket Harbor and West Dennis Beach, the 1995 photos of denuded creekbanks in Figure 8 show that vegetation losses began well before the presumed onset of SWD. 2009 S.M. Smith 193 Figure 6. Changes in a bend in the Herring River (Harwich) between 1952 and 2005. The signature of edge vegetation was present throughout this marsh in 1984, but decreased greatly between 1984 and 2001. Note how channel widening and deterioration of the peninsula accelerated during the latter period (horizontal bar in 1952 photo indicates where peninsula width was measured in all photographs). Tidal creek widening and creek structural changes associated with increases in the width of tidal inlets In two riverine systems, Quivett and Namskaket creeks (Brewster, Cape Cod Bay), where there is no detectable crab herbivory, changes in creek morphology also have occurred. However, creek changes here have been vastly different from those discussed above in that they are the result of undercutting and subsequent bank slumping. Intact vegetation still grows on the slumping or severed blocks of marsh peat that becomes separated from the bank—i.e., there are no denuded edges like those found in places with intense herbivory. Moreover, after an initial change in response to tidal inlet widening prior to 1984, changes in creek widths and structure greatly decreased to the point where they were essentially stable. The inlet width at Quivett Creek (Brewster) increased approximately threefold between 1938 and 1984, but only by ≈13% (15 m) between 1984 and 2001. The same is true for Namskaket Creek (Brewster) where the tidal inlet width doubled between 1947 and 1984, but increased by only 25% between 1984 and 2000. In both systems, there was much more change in creek structure prior to 1984 than after. In fact, marsh geomorphology has shown very little evolution since 1984 (Fig. 9). The cause of inlet widening is unknown, but is assumed to be natural as these areas did not serve as harbors and no human engineering of the inlets was evident. Regardless, the lack of major change in creek widths or patterning after 1984 is in direct contrast to those systems described above, where deterioration accelerated during this period. 194 Northeastern Naturalist Vol. 16, No. 2 Figure 7. Deterioration of Lt. Island between 1947 and 2000 (black line represents edge of marsh in 1947) and a recent photo of heavily grazed S. alterniflora at the beginning of the growing season (early June, 2007). Note the rapid disintegration between 1984 and 2000. 2009 S.M. Smith 195 Marsh edge/area stability Where marsh edges were not denuded of vegetation, and where widening tidal inlets did not play a physical role in re-shaping tidal creeks, long-term stability was observed. Examples include Paine’s Creek (Brewster; Fig. 9), Nauset Marsh (Eastham; Fig. 10), Bass Hole area (Yarmouth; Fig. 10), and Barnstable Marsh (Barnstable). In these systems, there have been no losses of edge vegetation and the widths of the main tidal inlets have remained more or less unchanged. Consequently, changes in the patterns of tidal channels and/or marsh shrinkage were not perceptible. This stasis is something that has been previously noted in Barnstable Marsh (Redfield 1972). Crab Figure 8 (opposite page, bottom). Examples of creek widening and island shrinkage in a Lt. Island marsh between 1984 and 2000, and ground-level photos showing denuded creekbanks (white arrows) in 1995 (cropped versions of photos taken by K. Rosenthal). Note the calving of peat (middle right-side photo), which results in permanent creek widening. Figure 9. Photo series of Namskaket (top), Quivett (middle), and Paine’s (bottom) creeks. Note the major change in tidal inlet widths (horizontal bars represent initial widths) of Namskaket and Quivett creeks prior to 1984 that resulted in creek widening and re-patterning (see circled areas for examples of oxbow cutting in the Namskaket series; black lines in Quivett series delineate channel edges in 1947). This period of transformation was followed by relative stability as the inlet changed very little between 1984 and 2000/2001. In contrast, the Paine’s Creek inlet remained virtually unchanged throughout this whole time period, which corresponds with creek stability. 196 Northeastern Naturalist Vol. 16, No. 2 herbivory was nonexistent in the abovementioned sites since S. reticulatum was either extremely rare or totally absent. In fact, no crab burrows of any kind were recorded in a 2003 survey of Nauset Marsh (Smith 2004). Pleasant Bay marshes fall into both this and the previous category. There have been no significant changes in marsh area or creek patterning since 1947 (i.e., no oxbow cutting, increased meandering, etc). Furthermore, the widths of the major flow channels through the system have remained virtually unchanged (Fig. 10). However, there are several secondary creeks that exhibited widening without any corresponding changes in architecture. This widening has occurred in the absence of crab herbivory, as S. reticulatum was not present in this system (Smith 2004). The few areas showing such change converted from low marsh in 1947 to high marsh in 1984 and back to low marsh by 2000—the result of inlet migration that dampened tides between 1947 and 1984, followed by a break in the barrier beach in 1987 that immediately increased tidal amplitude by ≈40 cm (Wilhelm 1989). Creek widening in these few instances may be related to vegetation instability caused by rapid shifts in species composition. In general, from the standpoint of marsh geomorphology, the system has been remarkably robust to major fluctuations in hydrology. High-marsh losses and the development of un-vegetated mudflats Reductions in the extent of high-marsh vegetation are conspicuous in aerial photography of virtually every marsh on outer Cape Cod. In only a Figure 10. Marsh stability in Nauset Marsh (top series), Bass Hole area (middle series), and Pleasant Bay (bottom series) between 1947 and 2000, where no detectable crab herbivory on S. alterniflora has occurred (note the stability of creek architecture throughout this period; white lines in the bottom series delineates the channel edges in 1947 and shows the lack of change over time). 2009 S.M. Smith 197 handful of systems, where the amount of high marsh was very limited even in the earliest photos, is it difficult to recognize this trend. In some cases, transgression of the low/high marsh boundary has resulted in the formation of extensive areas of bare ground between the two zones. In general, while high-marsh retreat occured to a small extent between 1947 and 1984, it was much more rapid thereafter. Perhaps the best example of this occurs in the Gut (CCNS, Wellfleet), where the low/high marsh boundary shifted >100 m upslope between 1984 and 2000. In this marsh, a reduction of approximately 46% in high-marsh area has occurred since 1947, with most of the loss (35%) occurring between 1984 and 2000 (Fig. 11). In the wake of the rapid retreat observed in recent decades, large areas of bare ground have formed. Middle Meadow (CCNS, Wellfleet) has followed a similar trend, losing approximately 37% of the high marsh originally present in 1947, with 30% of the loss occurring after 1984 (Fig. 12). In this system, some stands of J. gerardii within the high-marsh area have also vanished. Where large mudflats have replaced high-marsh vegetation (primarily in Wellfleet marshes), it is unclear why S. alterniflora has not been able to shift landward, since the upslope advance of this species is thought only to be constrained by competition with S. patens (Bertness and Ellison 1987). Moreover, annual forbs (Salicornia spp., Suaeda spp.) also have failed to colonize these areas. Because high-marsh losses occur anywhere from a few to tens of meters distant from marsh edges, they have had no obvious influence on erosional losses and decreases in marsh platform area. High-marsh losses balanced by low marsh advancement The majority of marshes on Cape Cod have exhibited variable reductions in high-marsh area, but without the development of unvegetated mudflat during the process. In fact, there have been some very rapid shifts in vegetation that demonstrate the ability of S. alterniflora to keep pace with retreating S. patens. In the northeastern portion of Nauset Marsh (Eastham), for example, the small amount of high marsh that existed in 1947 remained more or less unchanged until after 1984, when much of it was replaced by S. alterniflora (Fig. 13). In Hatches Harbor and West End marshes (CCNS, Provincetown), a similar process has occurred, although temporal trends in the structure and composition of vegetation in these marshes is somewhat confounded by rampant off-road vehicle activity through the 1980s as well as changes in tidal inlet geomorphology, overwashes, and sand deposition from nearby dunes. In Pleasant Bay, large areas of high marsh converted to low marsh between 1984 and 2000. This area has been greatly affected by alterations in tidal amplitude caused by barrier-beach migration and breaks. As mentioned previously, a 1987 storm produced a new inlet that greatly increased tidal amplitude. Subsequently, the majority of high marsh that proliferated between 1947 and 1984 (due to decreasing tidal amplitude from southward migration of the tidal inlet and the establishment of additional mosquito ditches) virtually disappeared. In some places, the low/high marsh boundary shifted landward by as much as 650 m between 1984 and 2000 (Fig. 6). 198 Northeastern Naturalist Vol. 16, No. 2 Despite this rapid reduction in high-marsh area, no bare gaps were detectable in aerial photography and none are present in this area today. Thus, S. Figure 11. Retreat of the high marsh (the black line in the top photos approximates the seaward boundary of high-marsh vegetation) in the Gut (Cape Cod National Seashore) between 1947 and 2000 with the subsequent development of barren mudflat. Note the increase in the rate of loss after 1984. The bottom left photo is a lowaltitude, oblique-angle image of barren mudflat that had developed in the Gut by Nov 1991 (J. Ingoldsby, Landscape Mosaics, Marshfield, MA). The bottom right photo was acquired in June 2007 (S.M. Smith, NPS). 2009 S.M. Smith 199 alterniflora was able to replace high-marsh species at a fast enough rate to prevent the development of unvegetated mudflat. Summary of general trends among outer Cape Cod marshes Table 2 summarizes key characteristics of outer Cape Cod salt marshes and the kinds of changes they have undergone over the last 50+ years. Of all the marshes analyzed in this study, 65% had symptoms of S. reticulatum grazing and every one of these exhibited some degree of edge vegetation loss. In contrast, where crab herbivory was not observed, there was no detectable loss of edge vegetation and no significant changes in marsh structure over time, with the exception only of systems where there were large increases in tidal inlet widths. In the latter, however, there were no edge vegetation losses akin to those described in areas where gazing was evident. Instead, creek banks with healthy vegetation were lost to undercutting, slumping, and subsequent erosion. High-marsh dieback associated with the development of bare mudflat occurred at approximately one third of the sites, while landward advance of the low marsh was evident in all but three locations where there was so little high marsh to begin with that such trends could not be analyzed with a reasonable level of accuracy. Figure 12 (opposite page, bottom). Retreat of the high marsh in Middle Meadow (Cape Cod National Seashore) between 1947 and 2000 with the subsequent development of barren mudflat. Note the increase in the rate of high-marsh loss after 1984. The lowaltitude photo at the bottom of the figure shows the boxed area and was captured from an oblique angle in June 2007 (S.M. Smith, NPS). Black arrows point upslope to show direction of elevation gradients, and white polygon shows extent of high marsh in 2007. Figure 13. Photos showing high-marsh losses in Nauset Marsh (top photos) and Pleasant Bay (bottom photos) between 1947 and 2000. Note the rapid reduction in high-marsh area after 1984 without any development of unvegetated mudflat (the black line in the top photos approximates the seaward boundary of high marsh vegetation in 1947; the structure across the tidal creek is a boardwalk that does not restrict tidal flow). 200 Northeastern Naturalist Vol. 16, No. 2 Discussion The analyses presented above shed light on a number of changes taking place in Cape Cod salt marshes. With respect to the timeline of vegetation losses, color IR photography shows that both low and high-marsh vegetation losses on Cape Cod have been occurring since at least 1984, and possibly for several decades. While actual bare ground can be seen in the 1984 color IR photos where there was once vegetation, low-marsh losses prior to this year are inferred by major changes in creek structure that have occurred in the absence of tidal inlet changes. This inference is reasonable given that it has long been recognized that under normal tidal conditions with no significant loss of vegetation, salt marsh creek systems are very stable (Gabet 1998, Garofalo 1980). The photography also shows that vegetation losses have not occurred simultaneously; they began at different times for different marshes. This temporal staggering suggests that no single climatic event (e.g., extreme drought) forced the disappearance of vegetation all at once, either independently or in conjunction with consumer pressure and/or pathogens, as is reportedly the case with SWD in Georgia (Ogburn and Alber 2006, Silliman et al. 2005) and Louisiana (McKee et al. 2004). The photo series also show that while smaller patches of vegetation may disappear or recover during the course of a single growing season, the losses are generally cumulative and progressive over many years when viewed from a system-wide perspective. This information, along with the work of Holdredge et al. (in press), indicates that salt marsh vegetation losses on Cape Cod are distinctly different than SWD events from other areas of the country. There has been significant geomorphologic change where vegetation along creekbank edges has declined. The role of S. reticulatum grazing in driving this decline is an important, recent finding that correlates exceptionally well with spatial patterns of loss. Where there are no S. reticulatum, or where their densities are so low that herbivory is not detectable, there has been no incidence of SWD. The structural stability of systems like Barnstable Marsh and Nauset Marsh, where S. reticulatum and denuded creekbanks are both absent, suggests an intimate link between changes in creek architecture and herbivory. Tidal creeks in NY (Hartig et al. 2002) and southeastern England (Van der Wal and Pyle 2004) also have widened and lengthened over the last 50 years. While specific causes have not been conclusively determined, these changes may also be the result of disturbance to vegetation—by biogeochemical processes in the former and by the worm Nereis diversicolor O.F. Müller (Ragworm) in the latter (Paramor and Hughs 2004). Similarly, experiments in British Columbia have showed that marsh geomorphology can be altered by geese herbivory (Kirwan et al. 2008). Twentieth-century marsh loss and tidal creek widening in other systems such as Chesapeake Bay have been attributed to decreased sediment supply coincident with sea level rise (Kearny et al. 1988). Structural changes have also been apparent in some Cape Cod marshes in the absence of crab 2009 S.M. Smith 201 herbivory (e.g., Namskaket and Quivett Creek), particularly in the form of creek widening and/or oxbow cutting. However, these changes have followed major alterations in the width and structure of tidal inlets. Furthermore, these changes were the result of physical rather than biological processes, as evidenced by the severe undercutting and slumping of creekbanks with healthy, vigorous vegetation along the edges. It is noteworthy that Nauset Marsh, and to a large extent Pleasant Bay, have not shown these kinds of changes, despite tidal inlet migration and breaks in the barrier beaches that alter tidal regimes. Their setting as large, back-barrier rather than riverine systems may have played a role in this stability in that the ratio of open water to marsh was much higher. Thus, changes in tidal flows were experienced over a much broader area and effectively dampened. In addition to herbivory and physical processes related to shoreline change, sea-level rise has undoubtedly had some influence on Cape Cod marshes. Sea level has risen at an average rate of about 2.65 millimeters/ year (0.87 feet/century) with a standard error of 0.1 mm/yr based on monthly mean sea-level data from 1921 to 1999. However, the steepest slope in any 10-year period between 1938 and present occurred from 1989 to 1998 (source: NOAA; http://tidesandcurrents.noaa.gov). In England, Baily and Pearson (2007) have reported >50% losses of Spartina marsh between 1971 and 2001, a portion of which they attribute to increased erosion from sea-level rise. On Cape Cod, transgression of S. alterniflora should be occurring at the lowest elevations where vegetation becomes short and sparse and eventually grades into mud or sand flats. However, the position of these edges can be difficult to determine accurately in high-altitude aerial photographs. Further complicating the matter is that these boundaries can exhibit large and rapid changes in response to storm-related accretion events or sediment displacement. Creekbank vegetation exists at much higher elevations, is generally thicker and taller, and has a well-defined edge; thus, changes there are easier to see. Also, flooding stress in S. alterniflora should manifest itself as a gradual stunting in plant heights (Anderson and Treshow 1980, Howes et al. 1986) rather than the rapid disappearance of discrete patches within otherwise tall, healthy vegetation as is the case with herbivory. The former is not conspicuous in aerial photographs (as illustrated in above figures), whereas the latter is. Within relatively short time frames, however, rising sea levels could lead to more severe erosion that alters marsh architecture, particularly following disturbances like those generated by herbivory. Even temporary vegetation losses may result in permanent physical changes to tidal marshes (Kirwan and Murray 2007), and this progression certainly seems to be the case in Cape Cod. A more obvious manifestation of sea-level rise may be in the landward retreat of the high marsh. Rates of retreat are variable depending on marsh topography, and large shifts in position occur where elevation gradient is low, while small to no shifts are observed where gradients are 202 Northeastern Naturalist Vol. 16, No. 2 steep. Heterogeneity in sediment properties, including nutrients, may also influence these patterns. Sediment accretion and other aspects of marsh elevation dynamics undoubtedly play a role as well. Unfortunately, there are few data available on these variables. Nonetheless, loss of high-marsh vegetation was apparent across all marsh systems. Some of this loss was related to changes in tidal regimes from inlet widening (artificial or natural). However, there are many systems that have not been altered in this way that similarly exhibited diminishing high-marsh vegetation. The most plausible explanation for this is sea-level rise—i.e., that tidal inundation is exceeding the flood tolerance of the species that grow there (Burdick 1989, Gleason and Zieman 1981). About half of the high-marsh zones on Cape Cod retreated relatively slowly between 1947 and 1984, after which the rates accelerated. It is noteworthy that this time period overlaps with the 1989–1998 period of increased sea-level rise. An increase in the number and size of waterlogged pannes should be another visible symptom of sea-level rise (DeLaune et al. 1994, Hartig et al. 2002, Orson et al. 1985, Warren and Niering 1993). This trend does appear to be happening to a limited extent in some marshes, but there are no ground-level data to provide conclusive evidence of change. Moreover, it is difficult to accurately delineate these features in aerial photographs, particularly if the marshes are vegetated and without standing water at the time of the photo. In addition, many Cape Cod salt marshes are relatively young, have sandy inorganic substrates, and have been extensively ditched, all of which tend to suppress extensive panne formation. Unditched marshes (mostly within CCNS on the outer Cape) that provide better opportunities for panne formation have developed behind barrier beaches where tidal inlet migration and breaks in barrier beaches alter tidal regimes. Separating the effects of these shorter-term events from background sea-level rise is problematic. The process by which unvegetated mudflats form in the wake of highmarsh dieback and why this occurs in some marshes but not in others is still unclear. In other words, why doesn’t the low marsh fill in the gaps left behind? Although competitive interactions with S. patens play an important role in determining the upper limit of S. alterniflora (Bertness 1991), edaphic conditions may also be a factor (Broome et al. 1974, Hartman 1988, Proffitt et al. 2003). Areas vacated by S. patens could experience periodic stresses (e.g., elevated salinities) that might slow the landward encroachment of S. alterniflora. Another possibility is that crab bioturbatation by both Uca spp. (Fiddler Crabs) and S. reticulatum prevents the establishment of new seedlings, particularly annual forbs belonging to the genera Salicornia (pickleweed) and Suaeda (seepweed) that have colonized areas between S. alterniflora and S. patens in other marshes. There is also the question of whether direct grazing on S. patens may be contributing to its retreat. While S. reticulatum evidently prefers S. alterniflora, it will feed on S. patens as well. However, S. patens can tolerate 2009 S.M. Smith 203 extreme levels of defoliation under good growing conditions as evidenced by the fact that humans have harvested S. patens (haying) for many decades. Moreover, high-marsh dieback areas are characterized by the presence of intact, standing dead foliage and frequently exhibit a hummocking pattern of growth. The former indicates that at least some mortality is occurring in the absence of grazing (i.e., the foliage is not consumed). In many cases, the development of hummocks is thought to be a response to flooding stress (Buck 2001, DeLaune et al. 1994). However, high-marsh dieback on Cape Cod may be the result of multiple factors. In this regard, flooding stress may be exacerbating the effects of herbivory or other disturbances such as wrack smothering. It has been shown previously that marsh vegetation subjected to hydrologic stress has a reduced ability to withstand or recover from additional pressures (Baldwin and Mendelssohn 1998, Gough and Grace 1998, Miller et al. 2001, Tolley and Christian 1999). A related idea is that high-marsh dieback is indirectly associated with crab activity. Bioturbation and burrow construction increase soil aeration, which leads to reduced organic matter in the soil as it is oxidized away. This, in turn, can result in lower soil porosity and percolation rates of water (Vidal-Beaudet and Charpentier 2000). In addition, the loss of adjacent S. alterniflora can result in poorer edaphic conditions as rhizosphere oxidation Figure 14. The onset of high-marsh dieback (black arrows indicate direction of retreat) adjacent to healthy S. alterniflora (bright red signature) as seen in color IR photography from 2000 (top left). The larger, low-altitude image taken in 2007 (bottom right; S.M. Smith) shows continued loss of S. patens in locations non-adjacent to, and not previously occupied by, S. alterniflora. 204 Northeastern Naturalist Vol. 16, No. 2 is decreased (Bortolus and Iribarne 1999, Hacker and Bertness 1995, Howes et al. 1981). However, there are numerous examples of high-marsh losses on Cape Cod that are not preceded by adjacent low-marsh losses. Figure 14 illustrates high-marsh dieback occurring in an area that is non-adjacent to S. alterniflora. The dieback pattern here follows topography (i.e., it has died in the lowest spots), rather than any changes in S. alterniflora, which is vigorously growing along its landward border. In general, if low-marsh losses were initiating high-marsh dieback, the expectation would be a weaker relationship with elevation. The dying edge should theoretically be more distorted, with dieback extending further upslope where it is adjacent to heavily grazed patches of S. alterniflora. This spatial pattern, however, has not been observed in the field. In light of such contradictions between known ecological processes and conditions on the ground, much more extensive characterization of soil properties and manipulative experiments with salt marsh plants and fauna are needed to address various hypotheses on the cascading effects of vegetation loss. Through stratigraphy, it is possible to reconstruct the long-term history of salt marsh development, including major vegetation transitions (Allen and Haslett 2002, Orson et al. 1987). This technique is particularly good at capturing large events like hurricanes (Donnelly and Webb 2004; Donnelly et al. 1999, 2004; Van de Plassche et al. 2006). However, such methods generally cannot resolve changes at decadal time scales (Donnelly 2006). Thus, whether the kinds of changes described above have occurred previously on Cape Cod is unknown. In southern Maine, Jacobson (1988) concluded from soil-core analysis that the widths of tidal channels in the Wells Marsh increased dramatically between 1794 and 1872, and attributes this to sea-level rise. Regardless of causes and past history, significant changes in the relative proportions of high vs. low marsh and the geomorphology of Cape Cod salt marshes have occurred since the mid-1900s, and these alterations may be influencing a number of ecosystem processes. The changes have altered habitat structure in obvious ways, but it is unknown how other flora or fauna have responded. Loss of edge vegetation may influence the dynamics of sediment transport. If the quantity of eroded materials in near-shore waters is increasing, benthic organisms including shellfish and sea grasses could be impacted. In human terms, the loss of salt marsh equates to a reduction in the buffering of wave energy along the coastline. In fact, a positive feedback loop may be in place whereby the loss of vegetation increases the rate of erosion, which allows more wave energy to enter the system, which results in more erosion. The same process could occur with sedimentation as vegetation loss decreases the probability that suspended sediment will settle onto the marsh surface, leading to reduced vertical accretion, gradual submergence eliciting flooding stress, and the eventual loss of more vegetation (Nyman et al. 1993). These potential impacts, along with the inability of marshes to creep upslope due to human development of adjacent uplands, 2009 S.M. Smith 205 mean that substantial amounts of salt marsh on Cape Cod could be lost during the next 60 years. Acknowledgments I would like to thank Donald Liptack (NRCS, Barnstable) for graciously providing numerous aerial photographs of Cape Cod and Russ Hume, who piloted the aircraft from which NPS staff have taken numerous low-altitude oblique photos over the past several years. Dr. Mark Bertness and Christine Holdredge (Brown University) provided critical information on crab herbivory. I am also grateful to Mark Adams (CCNS), John Portnoy (CCNS), David Burdick (University of New Hampshire), Ivan Valiela (Marine Biological Laboratory), and Merryl Alber (University of Georgia) for their reviews of early drafts of the manuscript. This work was supported by the National Park Service, Cape Cod National Seashore. Literature Cited Adamowicz, S.C. 2006. Sudden wetland dieback in the Northeast. National Wetland Newsletter 28:4. Adamowicz, S.C., and L. Wagner. 2005. Workshop proceedings and recommendations: Northeast sudden tidal wetland dieback workshop, April 1, 2005, Parker River National Wildlife Refuge. Rachel Carson National Wildlife Refuge, Wells, ME. Allen, J.R.L., and K. Haslett. 2002. Buried salt-marsh edges and tide-level cycles in the mid-Holocene of the Caldicot Level (Gwent), South Wales, UK. The Holocene 12:303–324. Anderson, C.M., and M. Treshow. 1980. A review of environmental and genetic factors that affect height in Spartina alterniflora Loisel (Salt Marsh Cordgrass). Estuaries 3:168–176. Baily, B., and A.W. Pearson. 2007. Change-detection mapping and analysis of salt marsh areas of central southern England from Hurst Castle spit to Pagham Harbour. Journal of Coastal Research 23:1549–1564. Baldwin, A.H., and I.A. Mendelssohn. 1998. Effects of salinity and water level on coastal marshes: An experimental test of disturbance as a catalyst for vegetation change. Aquatic Botany 61:255–268. Bertness, M.D. 1991. Zonation of Spartina patens and Spartina alterniflora in a New England salt marsh. Ecology 72:138–148. Bertness, M.D., and A.M. Ellison. 1987. Determinants of pattern in a New England salt marsh plant community. Ecological Monographs 57:129–147. Bortolus, A., and O. Iribarne. 1999. Effects of the SW Atlantic burrowing crab Chasmagnathus granulata on a Spartina salt marsh. Marine Ecology Progress Series 178:78–88. Brinson, M.M., and R.R. Christian. 1999. Stability of Juncus roemerianus patches in a salt marsh. Wetlands 19:65–70. Broome, S.W., W.W. Woodhouse, Jr., and E.D. Seneca. 1974. Propagation of Smooth Cordgrass, Spartina alterniflora, from seed in North Carolina. Chesapeake Science 15:214–221. Buck, T.L. 2001. High-marsh plant community response to sea-level induced highmarsh subsidence and ecosystem state change. M.Sc. Thesis. Department of Biology, University of Virginia, Charlottesville, VA. 206 Northeastern Naturalist Vol. 16, No. 2 Burdick, D.M. 1989. Root aerenchyma development in Spartina patens in response to flooding. American Journal of Botany 76:777–780. Civco, D.L., W.C. Kennard, and M.W. Lefor. 1986. Changes in Connecticut saltmarsh vegetation as revealed by historical aerial photographs and computerassisted cartographics. Journal Environmental Management 10:229–239. DeLaune, R.D., J.A. Nyman, and W.H. Patrick, Jr. 1994. Peat collapse, ponding, and wetland loss in a rapidly submerging coastal marsh. Journal of Coastal Research 10:1021–1030. Donnelly, J.P. 2006. A revised Late Holocene sea-level record for northern Massachusetts, USA. Journal of Coastal Research 22:1051–1061. Donnelly, J.P., and M.D. Bertness. 2001. Rapid shoreward encroachment of salt marsh cordgrass in response to accelerated sea-level rise: Proceedings of the National Academy of Sciences 98:14218–14223. Donnelly, J.P., and T. Webb III. 2004. Backbarrier sedimentary records of intense hurricane landfalls in the northeastern United States. Pp. 58–96, In R. Murnane and K. Liu (Eds.). Hurricanes and Typhoons: Past Present and Potential. Columbia University Press, New York, NY. Donnelly, J.P., T. Webb III, and W.L. Prell. 1999. The influence of accelerated sealevel rise, human modification, and storms on a New England salt marsh. Current Topics in Wetland Biogeochemistry 3:152–160. Donnelly, J.P., J. Butler, S. Roll, M. Wengren, and T. Webb III. 2004. A backbarrier overwash record of intense storms from Brigantine, New Jersey. Marine Geology 210:107–121. Ewanchuk, P.J., and M.D. Bertness. 2004. The role of waterlogging in maintaining forb pannes in northern New England salt marshes. Ecology 85:1568–1574. Fagherazzi, S., R. Torres, C. Hopkinson, and D. van Proosdij. 2005. Salt marsh geomorphology: Physical and ecological effects on landform. EOS 86:57-58. Gabet, E.J. 1998. Lateral migration and bank erosion in a salt-marsh tidal channel in San Francisco Bay, California. Estuaries 21:745–753. Garofalo, D. 1980. The influence of wetland vegetation on tidal stream channel migration and morphology. Estuaries 3:258–270. Gleason, M.L., and J.C. Zieman. 1981. Influence of tidal inundation on internal oxygen supply of Spartina alterniflora and Spartina patens. Estuarine, Coastal, and Shelf Science 13:47–57. Gough, L., and J.G. Grace. 1998. Effects of flooding, salinity, and herbivory on coastal plant communities, Louisiana, United States. Oecologia 117:527–535. Hacker, S.D., and M.D. Bertness. 1995. Morphological and physiological consequences of a positive plant interaction. Ecology 76:2165–2175. Hartig, E.K., V. Gornitz, A. Kolker, F. Mushacke, and D. Fallon. 2002. Anthropogenic and climate change impacts on saltmarshes of Jamaica Bay, New York City. Wetlands 22:71–89. Hartman, J.M. 1988. Recolonization of small disturbance patches in a New England salt marsh. American Journal of Botany 75:1625–1631. Higinbottham, C.B., M. Alber, and A.G. Chalmers. 2004. Analysis of tidal marsh vegetation patterns in two Georgia estuaries using aerial photography and GIS. Estuaries 27:670–683. Holdredge, C., M.D. Bertness, and A.H. Altieri. In press. Crab herbivory-driven dieoff of New England salt marshes. Conservation Biology. Howes, B.L., R.W. Howarth, J.M. Teal, and I. Valiela. 1981. Oxidation-reduction potential in a salt marsh: Spatial patterns and interactions with primary production. Limnology and Oceanography 26:350–360. 2009 S.M. Smith 207 Howes, B.L., J.W.H. Dacey, and D.D. Goehringer. 1986. Factors controlling the growth form of Spartina alterniflora: feedbacks between above-ground production, sediment oxidation, nitrogen, and salinity. Journal of Ecology 74:881–898. Jacobson, H.A. 1988. Historical development of the saltmarsh at Wells, Maine. Earth Surface Processes and Landforms 13:475–486. Kearney, M.S., R.E. Grace, and J.C. Stevenson. 1988. Marsh loss in Nanticoke estuary, Chesapeake Bay. Geographical Review 78(2):205–220 Kirwan, M.L., and A.B. Murray. 2007. An integrated biologic and physical model of tidal marsh evolution. Proceedings of the National Academy of Science 104:6118–6122. Kirwan, M.L., A.B. Murray, and W.S. Boyd. 2008. Temporary vegetation disturbance as an explanation for permanent loss of tidal wetlands. Geophysical Research Letters 35:L05403. McKee, K.L., I.A. Mendelssohn, and M.D. Materne. 2004. Acute saltmarsh dieback in the Mississippi River deltaic plain: A drought-induced phenomenon? Global Ecology and Biogeography 13:65–73. McLaren, J.R., and R.L. Jefferies. 2004. Initiation and maintenance of vegetation mosaics in an Arctic salt marsh. Journal of Ecology 92:648–660. Miller, W.D., S.C. Neubauer, and I.C. Anderson. 2001. Effects of sea level induced disturbances on high salt marsh metabolism. Estuaries 24:357–367. Nixon, S.W. 1982. The ecology of New England high salt marshes: A community profile. FFWS/OBS-81/55. US Fish and Wildlife Service, Washington, DC. Nyman, J.A., R.D. DeLaune, H.H. Roberts, and W.H. Patrick, Jr. 1993. Relationship between vegetation and soil formation in a rapidly submerging coastal marsh. Marine Ecology Progress Series 96:269–279. Ogburn, M.B., and M. Alber. 2006. An investigation of salt marsh dieback in Georgia using field transplants. Estuaries and Coasts 29:54–62. Orson, R.A., and B.L. Howes. 1992. Salt marsh development studies at Waquoit Bay, Massachusetts: Influence of geomorphology on long-term plant community structure. Estuarine, Coastal, and Shelf Science, 35:453–471. Orson, R.A., W. Panageotou, and S.P. Leatherman. 1985. Response of tidal salt marshes of the US Atlantic and Gulf coasts to rising sea levels. Journal of Coastal Research 1:29–37. Orson, R.A., W.A. Niering, and R.S. Warren. 1987. The development of a New England river valley tidal marsh. Estuaries 10:20–27. Paramor, O.A.L., and R.G. Hughes. 2004. The effects of bioturbation and herbivory by the polychaete Nereis diversicolor on loss of saltmarsh in southeast England. Journal of Applied Ecology 41:449–463. Proffitt, C.E., S.E. Travis, and K.R. Edwards. 2003. Genotype and elevation influence Spartina alterniflora colonization and growth in a created salt marsh. Ecological Applications 13:180–192. Redfield, A.C. 1972. Development of a New England salt marsh. Ecological Monographs 42:201–237. Silliman, B.R., J. van de Koppel, M.D. Bertness, L. Stanton, and I. Mendelsohn. 2005. Drought, snails, and large-scale die-off of southern US salt marshes. Science 310:1803–1806. Smith, S.M. 2004. 2003. Salt marsh vegetation monitoring report, Cape Cod National Seashore. National Park Service Report: Cape Cod National Seashore, Wellfleet, MA. 208 Northeastern Naturalist Vol. 16, No. 2 Tolley, P.M., and R.R. Christian. 1999. Effects of increased inundation and wrack deposition on a high salt marsh plant community. Estuaries 22:944–954. Valiela, I., M.L. Cole, J. Mcclelland, J. Hauxwell, J.Cebrian, and S.B. Joye. 2002. Role of salt marshes as part of coastal landscapes. Pp. 23–36, In M.P. Weinstein and D.A. Kreeger (Eds.). Concepts and Controversies in Tidal Marsh Ecology. Kluwer Academic Publishers, Dordrecht, The Netherlands. Van de Plassche, O., G. Erkens, F. Van Vliet, J. Brandsma, R.J. Van der Borg, and A.F.M. De Jong. 2006. Salt marsh erosion associated with hurricane landfall in southern New England in the fifteenth and seventeenth centuries. Geology 34:829–832. Van der Wal, D., and K. Pyle. 2004. Patterns, rates, and possible causes of saltmarsh erosion in the Greater Thames area (UK). Geomorphology 61:373–391. Van der Wal, D., A. Wielemaker-Van den Dool, and P.M.J. Herman. 2008. Spatial patterns, rates, and mechanisms of salt marsh cycles (Westerschelde, The Netherlands). Estuarine, Coastal, and Shelf Science 76:357–368. Vidal-Beaudet, L., and S. Charpentier. 2000. Percolation theory and hydrodynamics of soil-peat mixtures. Soil Science Society of America Journal 64:827–835. Warren, R.S., and W.A. Niering. 1993. Vegetation change on a northeast tidal marsh: Interaction of sea level rise and marsh accretion. Ecology 74:96–103. Wilhelm, R.W. 1989. Report on the Pleasant Bay water quality study, Cape Cod, Massachusetts. Friends of Pleasant Bay, Inc. Chatham, MA.