Spatial dynamic patterns of saltmarsh vegetation in southern Hangzhou Bay:Exotic and native species

Si-long Hung,Yi-ning Chen,*,Yn Li

aSecond Institute of Oceanography,Ministry of Natural Resources,Hangzhou 310012,China

bState Key Laboratory of Satellite Ocean Environment Dynamics,Second Institute of Oceanography,Ministry of Natural Resources,Hangzhou 310012,China

cDongshan Swire Marine Station,Xiamen University,Xiamen 361005,China

Abstract

Keywords:Saltmarsh;Spartina alterniflora;Scirpus mariqueter;Spatial variation;Competition;Exotic species;Native species

1.Introduction

Saltmarshes are found in the transitional zone between marine and terrestrial environments and are inhabited by halophyte communities(Chapman and Chadwick,1974).These are areas where physical actions of tides,waves,and sediments coalesce to form mudflats or peat wetland ecosystems(Adam,1990).Saltmarshes are among the most productive ecosystems worldwide.They provide shelter from storm surges,improve water quality,dissipate waves,capture sediments,and act as habitats for a wide variety of flora and fauna(Teal and Howes,2002).Saltmarshes are found globally,especially in intertidal zones at middle-to-high latitudes,and are among the most common types of coastal wetlands in China(Yang and Chen,1995).However,due to several factors,including land reclamation,the area of China's coastal wetlands decreased by 8×106hm2between 1950 and 2014,which corresponds to 58% of all the wetlands in the 1950s in China(Sun et al.,2015).This loss of wetland areas has already led to serious ecological problems.

Coastal saltmarshes are highly dynamic and complex,but also fragile(Teal and Howes,2002)and extremely vulnerable to exotic species invasion(Wang,2007).Spartina alterniflora(S.alterniflora),which is native to the west coast of the Atlantic Ocean and the Gulf of Mexico,has spread to Europe,the west coast of North America,New Zealand,and China,due to intentional or unintentional introductions mediated by humans(An et al.,2007).In recent years,S.alterniflora has become the dominant invasive species in China's saltmarshes(Li et al.,2009).In China,this species was introduced in 1979 to promote sediment accumulation and to reclaim land(Chung,2006).Although this introduction resulted in a few ecological and economic benefits,it also caused a series of problems.Owing to the ecological tenacity and adaptability of S.alterniflora,this plant easily forms extremely dense and productive monoculture communities,which hinder the survival of other plants(Li et al.,2009)and alter the geomorphological processes(Chen et al.,2018a)occurring in saltmarshes.Studies in Willapa Bay and San Francisco Bay have also shown that S.alterniflora eliminates native vegetation such as Zostera marina,Salicornia virginica,Triglochin maritima,Jaumea carnosa,and Fucus distichus(Scholten and Rozema,1990;Simenstad and Thom,1995;Daehler and Strong,1996).Under certain conditions,S.alterniflora competes for space with native vegetation,occupying the bare tidalflat to form a single dense S.alterniflora community(Callaway and Josselyn,1992).

Scirpus mariqueter(S.mariqueter)is a native saltmarsh plant in China,mainly found in the intertidal zones of the Yangtze River Estuary and Hangzhou Bay.This species contributes to wave action reduction,sediment capture,and biodiversity maintenance(Sun et al.,2001;Chen et al.,2004).Owing to the ever-increasing intensity of land reclamation and invasion of exotic species,the area covered by S.mariqueter communities has decreased dramatically,thus threatening the survival of this species(Sun et al.,2015).Saltmarsh vegetation is an important component of coastal wetland ecosystems,as it contains the main producers in these ecosystems and serves as a basis for biogeochemical cycles and energy flows in the local environment(Adam,1990).Hence,any change in saltmarsh communities will affect the structure and function of the saltmarsh ecosystem(Li et al.,2014).Enhancing our understanding of the spatiotemporal distribution and dynamic evolutionary processes of saltmarsh plant communities is therefore crucial,as this knowledge is key to the protection and remediation of saltmarsh ecosystems.

Most studies have shown that saltmarsh vegetation communities generally present a distinct zonation due to the clear environmental gradient and relatively simple species composition(Emery et al.,2001;Moffett et al.,2010,2012).The spatial distribution pattern of saltmarsh plant communities results from the combination of biological factors(e.g.,interspecific and intraspecific competition)and abiotic factors(e.g.,elevation,waterlogging,and human activities)(Pennings and Callaway,1992;Morris and Haskin,1990;Wang et al.,2010;Schwarz et al.,2011;Marani et al.,2013).The tidalflat elevation is one of the most important environmental factors directly determining the survival and distribution of saltmarsh plants(Silvestri et al.,2005),and it also affects other factors and their importance(Pennings et al.,2005).For example,the influence of tides gradually decreases as the elevation increases.The competitiveness of saltmarsh plants gradually becomes a key factor in plant distribution,as the less competitive species can only inhabit the areas where dominant species cannot survive(Pennings and Callaway,1992).On the other hand,as elevation decreases,the waterlogging time and salt stress become important environmental factors affecting the survival,growth,and distribution of saltmarsh plants(Pennings et al.,2005;Li,2018).The tidal creek network also affects the spatial distribution of the saltmarsh vegetation,as it is an important channel for the continuous exchange of material and energy between the saltmarsh and the tidal system itself as well as the external system,determining its long-term development and evolution under the physical and ecological factors(D'Alpaos et al.,2005;Hughes,2012;Coco et al.,2013;Zhou et al.,2016).The interactions between vegetation growth and geomorphodynamics have been studied in detail in order to provide an understanding of the mechanisms controlling saltmarsh evolution in relation to vegetation distribution and elevation(Murray et al.,2008;D'Alpaos,2011;Chen et al.,2016,2018a).In addition to these fine-scale investigations,observations of the large-scale spatial patterns over a relatively long period also provide insight into new phenomena relating to saltmarsh biogeomorphology.Therefore,in this paper,attempts have been made to report the new phenomena observed in a rapidly developing saltmarsh on the southern bank of Hangzhou Bay,via remote sensing images.

Hangzhou Bay is a typical macrotidal coastal area in China with high sedimentation rates(Li and Xie,1993b;Xia et al.,2004).The construction of a seawall led to the formation of a depositional zone in front of it,which is beneficial to the rapid development of tidal flats and saltmarshes(Wang et al.,2012),and therefore the competition pattern between native and exotic plant species can be studied in this region over a relatively short time scale due to the rapid development of the saltmarsh.In a geostatistical study performed by Gao et al.(2014)on invasive S.alterniflora in the saltmarshes of China,it was found that Hangzhou Bay is a demarcation zone for the environmentalecological changes in this species.In this study,we examined the saltmarshes on the southern bank of Hangzhou Bay(Andong Shoal)because of its unique characteristics regarding the competition between S.alterniflora and the native saltmarsh plants.The spatiotemporal distribution of saltmarsh vegetation in Andong Shoal was analyzed using a time series of highresolution remote sensing images.The objective of this study was therefore to investigate the distribution and spatiotemporal variation of native and invasive plant species in a newly developed saltmarsh after embankment,in an attempt to understand the interspecific competition pattern over a large spatial scale and relatively long temporal scale.

2.Study area

Fig.1.Location of Andong Shoal on southern bank of Hangzhou Bay(units of isobaths:m).

Andong Shoal(Fig.1)is located in Andong,Zhejiang Province(121°05′E to 121°22′E,30°14′N to 30°20′N),and it is an important component of the wetland ecosystem of Hangzhou Bay,which isatypicalfunnel-shaped bay.Hangzhou Bay is the largest macrotidal estuarine bay in China and has a large tidal range and strong currents.This region is dominated by mixed semidiurnal tides with rectilinear tidal currents.The durations of the high and low tides are 6.0 h and 6.4 h,respectively.The mean tidal range is 2.89 m at Dajishan(the mouth of the bay)and 5.54 m at Ganpu(the bayhead),and the maximum tidal range is 8.93 m(Editorial Board of China Bay Survey,1991).The local terrain is highly complex as deep troughs and mudflats coexist within the marine areas of Hangzhou Bay.Andong Shoal is a mudflat on the southern bank of Hangzhou Bay,which was formed in a macrotidal environment.It extends to the mouth of the Qiantang River in the west,is adjacent to the tidal inlet in the east,and expands to form a fan-like shape toward the north.Andong Shoal is also the primary modern tidal flat of Hangzhou Bay(Li and Xie,1993a).Receiving a large amount of sediment from the long shore transport of the Yangtze River,the vertical sedimentation rate in Andong Shoal ranges between 2 cm/year and 4.5 cm/year,which is beneficial to the rapid development of saltmarshes(Li and Xie,1993b).

The saltmarshes of Hangzhou Bay(Fig.2)are mainly found on the tidal flats of Andong Shoal,where the intertidal zone is 7-8 km wide(Song et al.,2014).The rapidly changing rectilinear currents of the intertidal zone affect the saltmarsh vegetation of this area.In October 2018,saltmarsh coverage in the study area was about 3 km wide(Fig.2)and mainly developed in the middle and upper parts of the intertidal zone.The measured slope was only about 4×10-4because of the wide tidal flat,and therefore changes in elevation were not clear at a small scale.Based on field studies,the vegetation growing in the middle and lower saltmarsh is dominated by native S.mariqueter,whereas the uppermost part of the saltmarsh(near the seawall)contains sporadic patches of the common reed Phragmites communis(P.communis)and Suaeda glauca(S.glauca).The remaining areas of the saltmarsh are dominated by the invasive S.alterniflora(Chen et al.,2018b).Because the P.communis and S.glauca patches were relatively small(less than 10 m2)and were rare in comparison with the other two species,the spectral information of these patches was difficult to extract from the multispectral images due to distortion.Therefore,we mainly focused on the spatiotemporal distribution of S.mariqueter and S.alterniflora within the study area and on the ecological succession processes in this area.

Fig.2.Transect profiles for saltmarshes on Andong Shoal(SA1 is the saltmarsh dominated by homogenous coverage of S.alterniflora,and SA2 is the pioneer zone covered by patches of S.alterniflora).

3.Methods

3.1.Data sources

We analyzed remote sensing images of the study area taken by the DMC-3 satellite in the same season over three years.These images were provided by Twenty First Century Aerospace Technology Co.,Ltd.,and their spatial resolutions were 0.8 m in panchromatic images and 3.2 m in multispectral(blue,green,red,and near-infrared)images.We chose to examine the images taken in late summer because saltmarsh plants grow the most during this period(Ouyang et al.,2013).As saltmarsh plants are periodically submerged by tidal water(Silvestri et al.,2005),tides are important when studying the distribution and range of saltmarsh vegetation using remote sensing images.Hence,the selected remote sensing images corresponded to the low tide to increase the reliability and rigor of our findings.Based on the growth cycles of S.alterniflora and S.mariqueter,and on the tide data measured at a nearby tide station(Zhapu Port),we selected three remote sensing images that were compatible with the requirements of this study(Table 1).These images were used to determine the vegetation cover of Andong Shoal.

Table 1Overview of remote sensing images.

3.2.Data processing

The remote sensing images were processed in ENVI 5.3 and ArcGIS 10.2 software packages.The data processing steps included image preprocessing,surface feature classification,and the establishment of a land-use conversion matrix(Fig.3).During the preprocessing step,radiometric calibration and atmospheric corrections were performed on the remote sensing images,followed by geometric correction based on measured ground control points,limiting the errors to 0.5 pixels.The nearest-neighbor diffusion(NNDiffuse)pan-sharpening algorithm was then used to combine the panchromatic and multispectral images,thus increasing the spatial resolution of the images while retaining their spectral information.Finally,a 33.26 km2study area was cropped out of the processed images.

Because the reflectance of S.alterniflora in the nearinfrared band is far greater than that of S.mariqueter(Ouyang et al.,2013),threshold-based segmentation was performed on the images according to the normalized difference vegetation index(NDVI)and simple ratio index(SRI).These were calculated as NDVI=(NIR-R)/(NIR+R)and SRI=NIR/R,where NIR and R represent the reflectance in the near-infrared and red bands,respectively.Using this segmentation method,the areas covered by S.alterniflora and S.mariqueter were accurately distinguished from each other(Ouyang et al.,2013).The surface features of the study area were identified using an expert-knowledge-based decision tree hierarchical classifier.First,the spectra of the various surface features in the processed images were analyzed(Fig.4)to formulate identification thresholds for each surface feature.Five surface features were examined:S.alterniflora(SA),S.mariqueter(SM),artificial building(AB),mudflat(MF),and water(W).A decision tree was then constructed(Fig.5).The preliminary classifications were corrected based on the field survey data(e.g.,samples from the real surface features and their coordinates as measured by RTK-GPS,as well as drone photos)and their visual interpretation.When the overall classification accuracy was greater than 80%,the results of the classification were considered to be reliable(Zheng et al.,2016).According to a confusion matrix,the overall classi fication accuracies in 2016,2017,and 2018 were 91.13%,90.11%,and 92.42%,respectively.Finally,ArcGIS was used to calculate the coverage of each saltmarsh species in the classified results.A 100 m×100 m net covering the study area was built in ArcGIS and superimposed on the classi fication results.The continuous vegetation belt of saltmarsh plants(defined as vegetation coverage larger than 50%)expanding seaward in 2016 and 2018 was extracted as a boundary,and the interannual migration rate of each boundary(Cao et al.,2014)was determined.The Thematic Change workflow tool in ENVI 5.3 was also used to perform conversion matrix analysis on the classified results.

Fig.3.Technical roadmap for remote sensing image processing by ENVI and ArcGIS.

Fig.4.Spectral curves of each surface feature after atmospheric correction.

Fig.5.Decision tree for remote sensing image classification(B means the blue band value of remote sensing images).

4.Results

4.1.Spatiotemporalcharacteristicsofsaltmarshvegetation cover

Fig.6.Results of vegetation classification in 2016,2017,and 2018(red boxes display a plot of pioneer zone of saltmarsh).

Based on the classified time series data,S.mariqueter and S.alterniflora were distributed in distinct strip-like areas(Fig.6).S.mariqueter was mainly found in the middle and lower saltmarsh,while S.alterniflora was mainly found in the upper saltmarsh.S.alterniflora and S.mariqueter distributions showed significant variations according to elevation,consistent with the results of previous studies on saltmarsh zonation(Pennings and Callaway,1992;Marani et al.,2013).Notably,S.alterniflora showed a bimodal distribution in space.In addition to the upper saltmarsh,a large number of sporadic patches of S.alterniflora occurred in the pioneer zone of the saltmarsh(lower saltmarsh),on either bank of the tidal creeks,and among S.mariqueter communities.In the lower saltmarsh,the S.alterniflora patches were parallel to the isobath lines.For the purpose of our discussion,the areas covered by S.alterniflora in the upper saltmarsh and pioneer zone are henceforth denoted as SA1 and SA2,respectively.S.alternifl ora is capable of both sexual and asexual reproduction.Sexual reproduction allows this species to be distributed over long distances,to occupy vacant niches in its home habitat,and to develop in new habitats.Asexual reproduction mostly allows for short-range community expansion(Xiao et al.,2010).We hypothesized that SA1 and SA2 were formed via different biological strategies.While SA1 was formed by the lateral expansion of the S.alterniflora community in the upper saltmarsh via asexual reproduction,SA2 was formed via the dispersion of S.alterniflora seeds and seedlings by tidal currents that resulted in the formation of S.alterniflora patches(distribution peaks)in the furthest area where they remain viable(as determined by the habitat limitations of this species),i.e.,the pioneer zone of the saltmarsh.

Many tidal creeks were observed in the study area.The tidal creek network is an important part of the tidal environment and continuously transports matter and energy to the saltmarsh ecosystem(Hughes,2012;Da Lio et al.,2013).More than 20 tidal creeks with widths larger than 5 m and numerous smaller secondary tidal creeks were observed.The tidal creeks that extended into the saltmarsh often had saltmarsh plants(usually S.mariqueter)growing on their banks.The S.alterniflora patches outside the species-dominant areas were also found along the banks of tidal creeks.The seeds and seedlings of saltmarsh plants might be transported into the tidal creeks by the tides,and then migrate along tidal creeks,and therefore were often observed growing on the tidal creek banks.

4.2.Changes in areas covered by saltmarsh vegetation communities and their boundaries

Saltmarsh dynamics were quantified by analyzing the changes in the coverage of each saltmarsh species from 2016 to 2018.Overall,saltmarsh vegetation coverage in the study area increased from 2016 to 2018(Table 2).Saltmarshes grow by expanding seaward due to a high sedimentation rate(Li and Xie,1993b).In the present study,saltmarsh growth was relatively slow as it only increased by 6.76% per annum(1.07 km2/year)on average.Native species accounted for 78.98%,74.33%,and 62.37% of the saltmarsh vegetation in 2016,2017,and 2018,respectively.Although the native species was still dominant in terms of coverage,its proportion in saltmarsh vegetation decreased over time at a rate of-0.61 km2/year.The areas covered by the exotic species(SA1 and SA2)grew continuously from 2016 to 2018(31.68% from 2016 to 2017 and 54.95% from 2017 to 2018),and its proportion in the saltmarsh vegetation increased significantly during the same period.The total area of S.alterniflora in SA1 maintained a rapid growth rate(1.68 km2/year),but the total area in SA2 varied within a small range,showing no expansion of this species(-0.005 km2/year).

The spatial expansion of each saltmarsh species is also reflected by shifts in its boundaries.In Fig.7,the seaward boundaries of each saltmarsh species in 2016 and 2018 were superimposed on the remote sensing image that was taken in 2018.These boundaries include the saltmarsh vegetation/mudflat boundary(SVMB),the SA1 boundary(SA1B),and the SA2 boundary(SA2B)that separate the S.mariqueter and S.alterniflora communities.The saltmarsh vegetation continuously expanded toward the sea ata rate of 180±15.6 m/year,indicating that the saltmarsh vegetation coverage is growing at a steady rate.Because SA1B wasexpanding at a rate of 287±25.7 m/year,the S.alterniflora community in the upper saltmarsh expanded more rapidly than in the entire saltmarsh and was encroaching on the S.mariquetercommunity.TheexpansionrateofSA2B was 167±9.2 m/year,which was lower than the migration rate of SVMB,indicating that the S.alterniflora patches at the seaward side were not invading the S.mariqueter community.Overall,the boundary migrations of the saltmarsh plants were consistent with the changes in their coverage.This expansion of the entire saltmarsh was mainly caused by the large sediment supply and the consequent high accretion rate of the tidalflat(Li and Xie,1993b),which created favorable conditions for saltmarsh expansion.

Table 2Changes in saltmarsh vegetation areas in 2016,2017,and 2018.

Fig.7.Changes of saltmarsh boundaries between 2016 and 2018 in remote sensing image of 2018.

During a study on the expansion of S.alterniflora saltmarshes in the Jiangsu coast,Zhang et al.(2004)found that the expansion of these saltmarshes could be divided into three stages.In the first stage,the S.alterniflora community expanded at a relatively slow rate,as it mainly relied on the spread of rhizomes to increase plant density.In the second stage,the S.alterniflora community expanded extremely rapidly.This was because the density and thickness of the S.alterniflora community reached their optimal levels,resulting in large-scale releases of seeds.These seeds then drifted along water currents and colonized new areas.In the third stage,the rate of expansion of the S.alterniflora community decreased owing to the restrictions imposed by the elevation of the mudflat face.Based on these classifications,the S.alterniflora community in our study area was likely in its second stage of expansion.The coverage of S.alterniflora increased at an accelerating rate throughout the observation period,and the SA1B in the upper saltmarsh was also rapidly expanding toward the sea.Furthermore,the appearance of SA2 in the pioneer zone of the saltmarsh occurred along isobaths,implying that S.alterniflora seeds were produced and then dispersed over a long distance to establish patches in the pioneer zone.

4.3.Conversion matrix analysis

The conversion matrix,which is based on the Markov model,provides a comprehensive outline of the structural characteristics of the wetland landscape,and of the quantity and direction of the landscape-type conversions.According to the conversion matrix(Fig.8,Tables 3 and 4),the spatial changes of the landscape comprised changes in the coverage of S.mariqueter(SM),S.alterniflora(SA),and mudflat(MF).S.alterniflora mainly established new habitats by replacing S.mariqueter in the middle saltmarsh,while the native species occupied the bare mudflat for tradeoff.During the observation period,the coverage of the S.mariqueter community decreased the most of all landscape types.The replacement of S.mariqueter by S. alterniflora accounted for 68.64% and 67.43% of the decrease in S.mariqueter coverage from 2016 to 2017 and from 2017 to 2018,respectively.This indicated that S.alterniflora was continuously invading and replacing the S.mariqueter community,and this process mainly occurred around SA1 in the upper saltmarsh.Overall,the increase in S. alterniflora coveragewas greater than its decrease, indicating that the S. alterniflora communitieswere expanding continuously.This is consistent with the results of our previous analyses.However,from 2016 to 2018,the SA2 area in the saltmarsh pioneer zone and the S.alterniflora community in the southeastern area of the study area were not stable,as they were frequently replaced by S.mariqueter.From 2016 to 2018,the southeastern corner of this saltmarsh was used for ecological restoration experiments,and as a result,S.alterniflora was replaced by S.mariqueter.However,the SA2 area showed a constant replacement of S.alterniflora by S.mariqueter under natural conditions. This pattern is possibly caused by a decreasing mudflat elevation,which better suits the growth of S.mariqueter.This observation also suggested that changes in elevation can make the competition between S.alterniflora and S.mariqueter reversible.

Fig.8.Results of conversion matrix over two-year periods.

Table 3Conversion matrix for surface features from 2016 to 2017.

Table 4Conversion matrix for surface features from 2017 to 2018.

It was also found through conversion matrix analyses(Fig.8,Tables 3 and 4)that the replacement of mudflat areas by S.mariqueter accounted for 76.89% and 79.12% of the increase in S.mariqueter coverage from 2016 to 2017 and from 2017 to 2018,respectively.Hence,S.mariqueter mainly relied on its expansion toward mudflats to colonize new habitats in the saltmarsh pioneer zone,which is less subject to competitive stress by S.alterniflora.However,the decrease in mudflat coverage,from 7.51% in 2016-2017 to 5.81% in 2017-2018,indicated that the rate at which the mudflat was being replaced by S.mariqueter decreased over time.A detailed examination revealed that the replacement of saltmarsh vegetation coverage by mudflat mainly occurred in the middle saltmarsh and on the sides of tidal creeks.

5.Discussion

5.1.Spatial competition between S.mariqueter and S.alterniflora

S.mariqueter and S.alterniflora are both perennial plants and have similar life cycles.They typically begin to sprout in the first three to four months and grow rapidly in the following six to nine months.The surface of the plant begins to wilt after the eleventh month(Wang,2007).S.mariqueter mainly grows in the middle and lower saltmarsh,while S.alterniflora is able to adapt to a wider range of habitats and conditions,which allows it to grow in the middle and upper saltmarsh(Yan et al.,2007).Thus,the habitats of S.mariqueter and S.alterniflora overlap in the middle saltmarsh,which might lead to spatial competition between these species.In the present study,the distribution of landscape types in the study area exhibited a distinct elevation-dependent pattern,i.e.,mudflat-S.mariqueter-S.alterniflora,with increasing elevation.This is consistent with the typical spatial distribution of plant communities in saltmarshes(Adam,1990;Morris et al.,2002).The lower elevations are dominated by pioneer plants that are resistant to flooding and salinity,which are gradually replaced by mesophytes at higher elevations(Pennings and Callaway,1992).Other than the upper saltmarsh,there are areas in the overlap zone and in the saltmarsh pioneer zone,where S.mariqueter and S.alterniflora coexist,thus indicating that the native and invasive species have overlapping habitats.The conversion matrix showed that interconversions occur between S.mariqueter,S.alterniflora,and mudflat areas.Hence,these areas are engaged in a dynamic competition for space.

Previous studies have shown that S.alterniflora outcompetes S.mariqueter(Chen et al.,2004;Zhang et al.,2004).Compared to intertidal plants like S.mariqueter,S.alterniflora is more tolerant of the abiotic stresses of saltmarshes(e.g.,salinity,flooding,and pH),and therefore highly adaptable to complex and diverse intertidal environments.This adaptability allows S.alterniflora to readily invade other saltmarsh plant communities(Wang et al.,2010;Schwarz et al.,2011).According to our analysis,the coverage of S.mariqueter gradually decreased over the 2016 to 2018 period,and its dominance in the study area also decreased over time.In contrast,S.alterniflora expanded rapidly,as the S.alterniflora community in the upper saltmarsh was able to expand its boundaries via asexual reproduction,and thus invade surrounding S.mariqueter habitats while strengthening its original community.This finding is consistent with those of previous studies(Chen et al.,2004;Zhang et al.,2004),and it shows that S.alterniflora holds a competitive advantage over S.mariqueter for space.

However,we have also found that S.alterniflora does not hold an absolute competitive advantage over S.mariqueter.The present study revealed that S.alterniflora displayed a bimodal spatial distribution,peaking in both the upper saltmarsh and the pioneer zone,although it does not flourish or expand readily in the latter,being restricted to patches.Significant decreases in tidal flat elevation occur in this area due to its large width,which alters environmental conditions such as the degree of flooding,the strength of wind and wave actions,soil conditions,and the degree of soil erosion.Thus,S.alterniflora communities do not hold the advantage in competition with S.mariqueter communities in low-elevation areas such as the pioneer zone although the elevation change is limited in this zone,indicating that macrotidal environments with considerable elevation changes are more beneficial to the growth and expansion of S.mariqueter than S.alterniflora in the saltmarsh pioneer zone.The replacement of S.alterniflora by S.mariqueter also occurred in upper saltmarsh regions,where the elevation decreased due to human disturbance,suggesting that S.mariqueter holds a competitive advantage over S.alterniflora in the low-elevation areas of the tidal flat.Previously,Li and Dissertation(2018)and Zheng et al.(2016)revealed through ecological experiments that the optimum habitat elevation for S.mariqueter was that of lower habitats,and that S.mariqueter had a competitive advantage over S.alterniflora and P.communis when the habitat elevation was less than 2.5 m.In summary,S.mariqueter and S.alterniflora are in spatial competition with each other,but the competitiveness of each species depends on the elevation.Although S.alterniflora dominates the upper saltmarsh most of the time,S.mariqueter still has the potential to overcome S.alterniflora at low elevations.In Andong Shoal,S.alterniflora is more competitive in high-elevation intertidal areas,while S.mariqueter is better suited to the low-elevation areas of the tidalflat.

5.2.Saltmarsh development under human influence

Andong Shoal has an abundant supply of sediment and is always under rapid siltation.The combination of this condition with regular, rectilinear tidal currents has led to the development of stable coastal saltmarshes(Xie et al.,2013).As saltmarsh plant communities tend to be distributed in particular zones along the elevation of the tidal flat,the spatial distribution of plant communities in estuarine saltmarshes often reflects their succession sequence.In most estuarine saltmarshes,theareas at the lowest elevations consist of mudflats,followed by S.mariqueter at higher elevations,and then P.communis.However,after S.alterniflora invasions,this sequence changes to mudflat-S.alterniflora-P.communis in high-salinity areas,and mudflat-S.mariqueter-S.alterniflora-P.communis in low-salinity areas(Wang,2007;Yan et al.,2007).In recent years,the increase in land reclamation projects has led to everincreasing levels of human disturbance,which alter environmental factors such as tidal flat elevation,hydrodynamic conditions,sediment supplies,and sediment characteristics.These effects have disrupted the native mode of saltmarsh succession and have led to the establishment of a unique saltmarsh succession process in Andong Shoal.

The construction of seawalls directly affects the hydrodynamic conditions(currents and waves)near the seawall,thereby altering the sedimentation and transport of silt.These changes disrupt the previously equilibrated profile and generate new erosion and sedimentation trends in the tidal flat(Wang et al.,2012).The elevation of the intertidal zone is relatively low at this point and it is characterized by extremely high levels of salinity and water content,due to frequent tides.Thus,the intertidal environment is extremely harsh and only a few pioneer halophytes(e.g.,S.mariqueter)can grow and flourish in this environment(Sun et al.,2001).Nonetheless,the tidal flat will gradually re-equilibrate and the seawall will modulate the sedimentation dynamics of the tidal flat.This leads to rapid sedimentation in front of the seawall and increased elevation of the tidal flat,reducing soil salinity and the effects of tidal disturbance(Wang et al.,2012).In addition,this tidal flat receives a large sediment supply to create a high accretion rate over a long term(Li and Xie,1993b).The sediment trapping ability of saltmarsh pioneer vegetation further accelerates this accretion rate(Marani et al.,2013;Chen et al.,2018a).As a result,the saltmarsh starts to develop at a relatively rapid rate,as indicated by the rapid expansion of the seaward vegetation edge,but the competition of vegetation further drives the changes within the saltmarsh.

Salt-tolerant plants begins to develop near the constructed seawall,and plant communities begin to be distributed in structurally complex zones,thus increasing plant diversity(Wang,2007)and further driving the spatial dynamics of saltmarsh vegetation.However,the invasion of S.alterniflora in the saltmarshes of Hangzhou Bay has led to sharp reductions in the populations of upper-saltmarsh plants such as P.communis through competitive exclusion,thus heralding a new mode of saltmarsh development(Li et al.,2005).As human activities cause seawalls to advance further toward the sea,only plants like S.mariqueter,which were originally in the pioneer zone of the saltmarsh,can remain.These plants then serve as pioneers for the aforementioned saltmarsh developmental processes.

The outermost seawall in the study area was Dyke No.11,which was constructed in 2010.Therefore,although tidal flat development outside the seawall is strongly affected by human activities,the more recently developed saltmarshes are minimally affected by human disturbance,and thus provide accurate information on the development of natural saltmarshes under rapid sedimentation conditions(Fig.9).In the first stage of saltmarsh development(Fig.9(a)),only a small number of S.mariqueter plants remained outside the seawall.Owing to environmental stresses, such as high salinity and strong currents,S.mariqueter communities expanded toward the mudflat through seeds or corms and gradually strengthened the community(Sun et al.,2001).The coverage of the saltmarsh was approximately 0.5 km at this stage.In the second stage(Fig.9(b)),the formation of a new tidal flat equilibrium and the sediment-capturing effects of S.mariqueter led to rapid siltation-induced rises in the mudflat outside the seawall,thereby establishing suitable conditions for the intrusion of halophytes such as S.alterniflora and P.communis.S.mariqueter was replaced by these plants in the upper saltmarsh, as it is unable to compete in this habitat.Nonetheless,the S.mariqueter community continued to expand toward the lower saltmarsh.The coverage of the saltmarsh was approximately 1.5 km at this stage.In the third stage(Fig.9(c)),the development of the tidal flat reached a stable state of siltation.The S.alterniflora community was expanding rapidly,and it had already occupied the P. communis habitat in the upper saltmarsh and begun invading the S.mariqueter community in the pioneer zone of the saltmarsh,thus creating a bimodal S.alterniflora distribution.Due to competitive exclusion by S.alterniflora,the growth of the S.mariqueter community was restricted and its proportion in the saltmarsh vegetation coverage gradually decreased over time.The saltmarsh coverage was approximately 3 km at this stage.In the fourth stage(for the future),assuming the exponential growth model for the expansion of S.alterniflora in a mudflat(Call away and Josselyn,1992),the distance between the S.alterniflora communities in the upper saltmarsh and pioneer zone will be gradually shortened as the tidal flat will continue to accumulate silts.S.alterniflora is expected to occupy most of the saltmarsh and the S.mariqueter habitat will migrate seaward.If the sediment supply is reduced,the migration of S.mariqueter could be halted,thus turning Andong Shoal into a monospecific community.In summary,the development of saltmarshes under rapid sedimentation in Andong Shoal can be summarized into four stages of succession.The spatial pattern of Andong Shoal in each of these stages(from low to high elevations)is as follows:(a)mudflat-S.mariqueter in the first stage,(b)mudflat-S.mariqueter-S.alterniflora(P.communis)in the middle stage,(c)mudflat-S.mariqueter(S.alterniflora)-S.alterniflora in the middle-to-late stage,and(d)mudflat-S.alterniflora(S.mariqueter)in the final stage,which is likely to take place under a reduced sediment supply.

Fig.9.Evolution of saltmarsh under investigation from 2010 to 2018,through a typical transection profile(units:km).

6.Conclusions

A time series of remote sensing images was processed and analyzed to elucidate spatiotemporal changes in the vegetation coverage of Andong Shoal during 2016-2018,and the competition between native and invasive plant species in this area.The following conclusions were drawn from this study:

(1)Plant communities in the saltmarshes of the study area exhibit zonation patterns.The native S.mariqueter community mainly grows in the lower half of the saltmarsh,while the invasive S.alterniflora species grows in the upper saltmarsh.The distribution of S.alterniflora is bimodal,as it consists of a dominant area in the upper saltmarsh and a patchy area in the saltmarsh pioneer zone.

(2)Between 2016 and 2018,the saltmarshes in the study area expanded seaward at a rapid rate of 1.07 km2/year.S.alterniflora expanded over the upper to middle saltmarsh at a mean rate of 1.68 km2/year.However,the S.alterniflora patches at the seaward edge expanded at a negative rate of-0.005 km2/year,indicating a different pattern in competition.The pattern of boundary changes also implies a similar pattern in interspecific competition.

(3)The interspecies spatial competition between S.mariqueter and S.alterniflora changes according to elevation.S.mariqueter is better-suited to the low-elevation areas of the tidal flat,while S.alterniflora is better-suited to the highelevation areas of the tidal flat.S.mariqueter is thus forced to expand toward the mudflat,while the S.alterniflora community invades and occupies S.mariqueter habitats in its vicinity via asexual reproduction.

(4)The saltmarshes of Andong Shoal are developing rapidly under the influence of human activities. The saltmarsh succession in this area(from low to high elevations)proceeds as follows:(a)mudflat-S.mariqueter in the first stage,(b)mudflat-S.mariqueter-S.alterniflora(P.communis)in the middle stage,(c)mudflat-S.mariqueter(S.alterniflora)-S.alterniflorainthe middle-to-late stage,and(d)mudflat-S.alterniflora(S.mariqueter)in the final stage.

(5)The observations using high-resolution images revealed a bimodal pattern for the competition between native and exotic species,which was found to be related to the elevation.As such,future biogeomorphological models could take this pattern into account when examining vegetation distribution and the associated elevation changes.In particular,vegetation competition results can be reversible when the bedform changes,which will provide a more complicated feedback to the model.

Acknowledgements

The remote sensing images in this study were produced by Twenty First Century Aerospace Technology Co.,Ltd.We would like to thank Mr.Ting-lu Cai and Dr.Xin-kai Wang for technical assistance.