Timing of disturbance affects biomass and flowering of a saltmarshplant and attack by stem-boring herbivores

SHANZE LI AND STEVEN C. PENNINGS �

Department of Biology and Biochemistry, University of Houston, 3455 Cullen Blvd, Suite 342, Houston, Texas 77204 USA

Citation: Li, S., and S. C. Pennings. 2017. Timing of disturbance affects biomass and flowering of a saltmarsh plant andattack by stem-boring herbivores. Ecosphere 8(2):e01675. 10.1002/ecs2.1675

Abstract. In salt marshes, disturbance by wrack (floating mats of dead vegetation) is common andaffects plant productivity and species composition, but little is known about how the timing of disturbancemediates these effects, nor how it interacts with herbivory. Using a field experiment on the Georgia coast,we simulated the effects of wrack disturbance at different times of the year on the marsh grass Spartinaalterniflora and its stem-boring herbivorous insects. The timing of disturbance throughout the growing sea-son strongly affected fall biomass, stem height, the proportion of stems flowering, and the proportion ofstems colonized by stem-boring herbivorous insects. End-of-season biomass in plots disturbed in Marchdid not differ from undisturbed controls, but biomass was reduced by 50% in plots disturbed in May, andby over 90% in plots disturbed in September. Disturbance in March and May stimulated flowering, but dis-turbance later in the growing season suppressed it. Plots disturbed late in the growing season had a lowfrequency of stem-boring herbivores. Stems containing stem borers rarely flowered. These results indicatethat the timing of disturbance matters in coastal salt marshes. Late-season disturbances had the strongesteffects on S. alterniflora and its herbivores. Disturbances early in the growing season did not affect end-of-season biomass, and stimulated flowering, suggesting parallels between fire disturbance in grasslandsand wrack disturbance in salt marshes. Late-season disturbance did reduce herbivory by stem-boringinsects, but not enough to compensate for the direct effects of disturbance on the plants. Future studies ofdisturbance in salt marshes should consider how the timing of experimental disturbance treatments relatesto the timing of natural disturbances.

Key words: disturbance; flowering; phenology; plant population and community dynamics; plant–herbivoreinteractions; primary production; stem borer; wrack.

Received 12 December 2016; accepted 19 December 2016. Corresponding Editor: Debra P. C. Peters.Copyright: © 2017 Li and Pennings. This is an open access article under the terms of the Creative Commons AttributionLicense, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.� E-mail: spennings@uh.edu

INTRODUCTION

Natural disturbances play important roles inmany ecosystems (Pickett and White 1985), andthe outcome of disturbances is often affected bytheir timing (Crawley 2004). In some cases, aswith fire, the intensity of the disturbance mayvary seasonally (Whelan 1995). In other cases,the timing of the disturbance affects which life-history stages of species in the community areaffected (Fill et al. 2012), and the time remainingin the growing season for species to recover

(Shepherd et al. 2012). In terrestrial plant com-munities, the effects of fire and drought ongrowth and flowering often vary with the time ofyear in which the disturbances occur (Waldropet al. 1992, Craine et al. 2012, Shepherd et al.2012). Similarly, hydrological disturbances toaquatic communities have different effects at dif-ferent times (Peterson et al. 1990, Peterson andStevenson 1992).In coastal habitats, one of the most common

natural disturbances is from wrack, dead plantmaterial that is deposited in the intertidal by

❖ www.esajournals.org 1 February 2017 ❖ Volume 8(2) ❖ Article e01675

waves and tides (Pennings and Bertness 2001, Orret al. 2005, Li and Pennings 2016). Wrack patchesdeposited low in the intertidal may be moveddaily by high tides; patches deposited high in theintertidal or into the supra-tidal may remain inplace indefinitely, decomposing over a period ofmonths. In salt marshes, wrack disturbance maycontribute to creekbank erosion (Lottig and Fox2007), and can kill plants and alter plant commu-nity composition (Bertness and Ellison 1987).Wrack can also benefit plants by increasing soilnutrients and reducing porewater salinities (Pen-nings and Richards 1998). Wrack is generatedmost profusely as plants senesce in the fall andwinter, and so disturbances are most common inthe spring; however, wrack patches are periodi-cally moved about by the tides, and so new dis-turbances can occur at any time during thegrowing season (Bertness and Ellison 1987, Valielaand Rietsma 1995). Whether the timing of distur-bance matters has not been studied.

Disturbance also affects herbivores (Dennoet al. 1980). Depending on whether herbivoresare more or less affected by disturbance thanplants, disturbance can interact with herbivory indifferent ways (Menge and Sutherland 1987,Hunter and Price 1992). Salt marshes are proneto drought-induced die-back of vegetation, anddrought stress appears to increase the vulnerabil-ity of some plants to herbivory by snails andstem-boring insects (Moon and Stiling 2002,Goranson et al. 2004, Silliman et al. 2005, Gaetaand Kornis 2011), but effects of wrack on her-bivory have not been examined. Moderate wrackdisturbance is unlikely to affect snails, but stem-boring insects that have limited mobility arelikely to die if the plant stems that they inhabitare killed by wrack. Moreover, whether newstems that grow following a disturbance are colo-nized by stem-boring insects will depend onoviposition preferences of female insects and thetime of year at which eggs are laid.

We conducted a field experiment in creekbankhabitat in coastal Georgia to explore the impor-tance of the timing of wrack disturbance on end-of-season plant biomass, plant flowering, and theprevalence of stem-boring herbivores. We hypo-thesized that plants would be least affected bydisturbances occurring early in the growing sea-son because the plants would not yet be flower-ing and would have more time to regrow than if

they experienced disturbances late in the grow-ing season. We expected stem-boring herbivoresto also be affected by disturbance, but we knewtoo little about their life cycles or preferences forlarger (older) vs. smaller (younger) stems to pre-dict in what way.

MATERIALS AND METHODS

We worked in a salt marsh along Dean Creek atthe southern end of Sapelo Island in Georgia, Uni-ted States (31.39° N, 81.28° W). The dominantplant in low and middle marsh elevations was theC4 grass Spartina alterniflora. In Georgia, this grassgrows year-round, but above-ground biomass ishighest between June and November (Gallagheret al. 1980). Flowering peaks around October(S. C. Pennings, personal observation). During thegrowing season of the study year (2015), therewere no dramatic changes in sea level, droughts,or extremely cold days (Appendix S1: Tables S1–S2, Fig. S1), and average conditions were not unu-sual compared to the previous decades (Wiezskiand Pennings 2014).Spartina alterniflora is attacked by several stem-

boring insects (Pfeiffer and Wiegert 1981, Stilingand Strong 1983), but because of their crypticlifestyle they are often overlooked in ecologicalstudies. We focused on stem-boring herbivores inthis manuscript because their presence and abun-dance over an extended period can be easilyquantified by a one-time collection, and becausetheir intimate associate with plant stems makesthem likely to be affected by disturbances thatharm plants (Armitage et al. 2013).The Georgia Coastal Ecosystems (GCE) Long-

Term Ecological Research Program has docu-mented wrack disturbance to eight sites incoastal Georgia dominated by S. alterniflora since2001 (Li and Pennings 2016). We conducted afield experiment at one of these sites, GCE 6(Appendix S1: Fig. S2), in 2015. We established72 plots, each 1 9 1 m, in S. alterniflora alongDean Creek at GCE 6. We simulated wrack dis-turbance on four different dates by coveringselected plots with 10 cm of wrack collectednearby on 18 March, 27 May, 15 July, and 7 Sep-tember 2015. The wrack bent over tall plantstems, typically breaking them near the base.Because the plots were smaller than naturalwrack patches, the wrack was fixed to the marsh

❖ www.esajournals.org 2 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

surface with string so that it would not washaway. We also simulated the mechanical distur-bance created by wrack, but without the input ofallogenic organic matter (Pennings and Richards1998) by clipping selected plots 1 cm above thesoil surface on the same dates. This created atotal of nine treatments (4 dates 9 2 disturbancesplus a control) with eight replicates each. Wemeasured plant heights in a 50 9 50 cm quadratcentered in each plot before the first treatmentswere conducted in March, again on 9 May, andevery 2 weeks thereafter until plots were har-vested in October. We also noted whether eachstem was flowering, and used this information toestimate stem biomass, using an allometricregression equation (Wiezski and Pennings 2014)for the creekbank zone of GCE site 6, which isimmediately adjacent to our experimental site:Ln (Biomass (g)) = �6.095 + 1.760 9 Ln (height(cm)) + 0.212 9 (flower (1 = flowering; 0 = notflowering)). For creekbank S. alterniflora, thisallometric relationship correlates strongly (R2 =0.97) with plot-level biomass (S. C. Pennings,unpublished data).

The center 50 9 50 cm of each plot washarvested by clipping stems at the soil surface on5–6 October 2015. We measured the height ofeach stem, noting whether it was flowering, andcounted the number of stem-boring insects pre-sent in each stem by splitting the stem open. Insome cases, an emergence hole and hollow inte-rior showed that a stem borer had been presentbut had already emerged, in which case the stemwas scored as “bored,” but the number of stemborers was not assigned. Representative larvaewere reared to adulthood to identify stem borertaxa. Stems were dried for 3 d at 60°C andweighed. Biomass in the different treatments wascompared using one-way analysis of variance(ANOVA), with post hoc Tukey’s HSD (honestsignificant difference) tests to compare individ-ual treatments. Percent data (% of stems boredor flowering) were converted to proportions,arcsine-square-root-transformed, and comparedwith one-way ANOVA, with post hoc Tukey’sHSD tests to compare individual treatments.Experimental data are available online (Li 2016).

To determine how stem borers affected plantend-of-season biomass, at the beginning of Octo-ber we collected 50 S. alterniflora shoots withstem borers and another 50 without from the

area surrounding the experimental plots. Whe-ther or not shoots have stem borers can usuallybe determined by examining the growing tip ofthe plant (Appendix S1: Fig. S3). We measuredthe height and recorded the flowering status ofeach shoot, split it open, and counted the numberof stem-boring larvae present. Shoots were driedfor 3 d at 60°C and weighed. Because most of thestems with stem borers were not flowering, wecompared the effect of stem borers on theheight–mass relationship of non-flowering stems(n = 40 with stem borers and 40 without) usinganalysis of covariance (ANCOVA) with stem bor-ers as the treatment and height as the covariate.

RESULTS

Wrack and clipping treatments reduced esti-mated aboveground biomass to zero on eachtreatment date (Fig. 1A). Otherwise, estimatedbiomass increased steadily with time, except fora slight decline in August, which may have beendue to stem borers or to high densities of grass-hoppers that inflicted obvious damage to theplants at that time. (Both herbivores can affectmeasurements of plant height: stem borers bykilling the meristem and grasshoppers by con-suming parts of the upper leaves.) Plots dis-turbed in March converged with control plots byOctober, but plots disturbed later did not fullyrecover in biomass by October.October standing biomass differed among treat-

ments (Fig. 1B). Plots disturbed in March did notdiffer from undisturbed controls. Plots disturbedin May had only 50% of the biomass of controlplots, and plots disturbed in July and Septemberhad even less biomass.The percent of stems attacked by stem-boring

insects depended on when plots were disturbedby wrack (Fig. 2A). Plots disturbed in March hadthe same frequency of stem borers as controlplots, but plots disturbed in July and Septemberhad very low stem borer attack rates (Fig. 2A).Stems shorter than 40 cm rarely housed stem-boring insects (Appendix S1: Fig. S6).The percent of stems that flowered also dep-

ended on when plots were disturbed by wrack(Fig. 2B). In control plots, about 5% of stems flow-ered. In contrast, in plots disturbed in March andMay, up to 20% of the stems flowered, althoughthe pairwise comparisons with the control plots

❖ www.esajournals.org 3 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

were not always statistically significant. In plotsdisturbed in July and September, none of thestems flowered. Stems shorter than 60 cm rarelyflowered (Appendix S1: Fig. S7).

Disturbance altered the height–frequency dis-tribution of the stems in the plots (Appendix S1:Fig. S4). Plots disturbed in March and May had

height–frequency distributions qualitatively sim-ilar to control plots with the most frequent stemheights in the 60–90 cm size range. Plots dis-turbed in July had a modal height of 20–60 cm,and plots disturbed in September had a modalheight of 10–40 cm. Control plots lacked anystems shorter than 30 cm, but all the disturbedplots had some stems in this size range.Stems containing stem borers rarely flowered

(1 out of 50 vs. 10 out of 50 stems lacking stemborers) and were heavier at a given heightthan stems lacking stem-boring insects (Fig. 3).Almost half (42%) of the bored stems no longercontained a stem-boring insect in October, butthe history of stem boring was revealed by a hol-low stem interior and an emergence hole. Theremaining bored stems contained stem borers,mostly fly larvae (38%, usually in the upper partof the stem) or moth larvae (22%, usually in the

Fig. 1. (A) Spartina alterniflora biomass in differenttreatments varied over time. Plots were sampled on16 March (day 0), 18 March (day 2), 27 May (day 72),10 June (day 86), 25 June (day 101), 9 July (day 115),23 July (day 129), 7 August (day 144), 21 August (day158), 6 September (day 174), 21 September (day 189),and 6 October (day 204). Biomass was estimated usingallometric relationships between height and mass.Data are means � SE (n = 8). (B) Comparison of finalstanding live biomass in different treatments. Data aremeans � SE (n = 8). Bars sharing a letter are not sig-nificantly different from one another. Dashed horizon-tal lines indicate 100% and 50% of the average biomassin the control plots.

Fig. 2. Percent of stems bored and flowering in dif-ferent treatments. Data are means � SE (n = 8). Barssharing a letter are not significantly different from oneanother.

❖ www.esajournals.org 4 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

bottom part of the stem). A few stems (6%) con-tained another stem-boring species that we didnot identify (likely Calamomyia alterniflorae [Dip-tera: Cecidomyiidae]). Most of the stems hadonly one larva inside, but we found one stemwith a fly larva and a moth exuvia, and anotherwith 16 small moth larvae.

We identified the two most common stem borerspecies by raising representative individuals toadulthood (Appendix S1: Fig. S5). The fly, Chaetop-sis apicalis, can be distinguished by the single darkband at the apex of the wing (Appendix S1:Fig. S5C). The moth was in the family Pyralidae(Appendix S1: Fig. S5F), which contains a numberof other stem-boring species. The moth larvaewere approximately two to four times longer thanthe fly larvae, and consumed much more of thestem core than did the fly larvae.

DISCUSSION

In our experiment, the timing of disturbanceaffected how Spartina alterniflora and its stem-boring insects responded. Wrack disturbancereduced fall biomass and stem height, altered theproportion of stems flowering in the fall, and

reduced the frequency of attack by stem borers,but these effects strongly depended on the timingof disturbance.Wrack disturbance significantly and strongly

reduced plant biomass and modal stem height.The effect of wrack disturbance on fall standingbiomass depended on the timing of the distur-bance. Plants recovered from wrack disturbancethat occurred in March, with biomass and stemheights comparable to the control treatment bythe time we harvested the experiment in October.In contrast, plants did not fully recover from dis-turbance that occurred later in the growing sea-son. In nature, wrack is produced by the death ofS. alterniflora stems, which happens primarilyover the winter, and so wrack abundance isgreatest earlier in the growing season and decli-nes into the fall (Bertness and Ellison 1987,Valiela and Rietsma 1995). As a result, distur-bances are likely to be more common earlier inthe growing season. Patches of wrack are movedabout by high tides; however, especially if theydo not lodge high in the intertidal, so new distur-bances can happen at any time of the year.Recovery of vegetation from wrack disturbance

is a function of several variables. The thickness ofthe wrack mat determines whether or not plantscan grow through it or cannot recover until thewrack either washes away or decomposes (Bert-ness and Ellison 1987, Valiela and Rietsma 1995,Brewer et al. 1998). How long the wrack remainsin place determines whether some plants survivethe disturbance or all are killed (Bertness and Elli-son 1987, Valiela and Rietsma 1995). The size ofthe wrack patch and the level of abiotic stress inthe patch determines how long it takes plants tofill a disturbance-generated unvegetated area bygrowing in from the sides (Shumway and Bert-ness 1994, Brewer and Bertness 1996, Angeliniand Silliman 2012). The timing of the disturbancerelative to plant phenology (Gallagher et al. 1980)affects the severity of the disturbance, becauseplants translocate resources from belowgroundrhizomes to aboveground shoots at the start ofthe growing season. Thus, a disturbance in Julyremoves a greater percentage of the plant’s reso-urces than a disturbance in March. Finally, thetiming of the disturbance also affects how muchtime remains in the growing season during whichplants can recover (Alber et al. 2013). In ourexperiment, wrack patches were relatively small

Fig. 3. Relationship between biomass and height ofSpartina alterniflora plants with and without stem bor-ers. With stem borer: Biomass = 0.19 9 Height � 5.81(R2 = 0.43); without stem borer: Biomass = 0.19 9

Height � 8.03 (R2 = 0.51). Compared the two withand without stem borer using ANCOVA. They weresignificantly different from each other (F = 8.250,P < 0.01, R2 = 0.51).

❖ www.esajournals.org 5 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

and thin, and we observed that vegetation wasable to recover by growing through the wrack.Although our experimental manipulation mim-icked the most common type of wrack distur-bance at our study site, we also observed larger,thicker mats of wrack that completely killed vege-tation and produced unvegetated areas in theS. alterniflora zone that appeared to take a fewyears to recover. Exploring the interactive effectsof the timing and size of wrack disturbance wasoutside the scope of this project, which focusedon the timing of relatively modest wrack distur-bances, but deserves attention in the future.

Wrack represents not only a disturbance butalso a source of nutrients and protection fromphysical stress (Pennings and Richards 1998).These positive effects of wrack may be moreimportant in the more-stressful high marsh thanin the lower marsh elevations dominated byS. alterniflora (Pennings and Richards 1998). Wefound no major differences between the wrackand clip treatments, indicating that the primaryway that wrack affected S. alterniflora biomassand height in this experiment was by killingstems, rather than by providing a nutrient sub-sidy or ameliorating physical stress. There weretrends suggesting that the May wrack and cliptreatments had different effects on stem borersand flowering (Fig. 2), but these trends were notstatistically significant.

Wrack disturbance early in the growing seasonappeared to stimulate flowering, although few ofthe individual means comparisons with the con-trol treatment were statistically significant due tomoderate replication and high variability. Wrackand clipped treatments both responded similarly,indicating that the increased flowering was aresponse to disturbance rather than a nutrientsubsidy. Other studies have also found that dis-turbance and stress can increase flowering inS. alterniflora (Shumway and Bertness 1994, Daleoet al. 2015, Liu et al. 2016). In terrestrial grass-lands, fire at the right phenological stage oftenstimulates flowering (Fill et al. 2012, Shepherdet al. 2012). Flowering in response to disturbanceis likely to be adaptive if the disturbance frees upresources and increases performance of seedlings,but plants may only be able to increase effort tosexual reproduction if the disturbance occurs atthe right phenological stages. Alternatively, dis-turbance or stress can also be an indicator of a

suboptimal habitat that causes organisms to shiftresources away from vegetative growth intosexual reproduction in order to disperse (Hamil-ton et al. 1987, van Kleunen et al. 2002, Crosbyet al. 2015). Disturbances that occurred later inthe growing season led to almost no flowering,likely because shoots simply did not grow largeenough to flower (Appendix S1: Fig. S7).Consumers play important roles in controlling

the production of plants in coastal ecosystems(Silliman and Bortolus 2003). We expected thatstems with stem borers would have less biomassthan un-bored stems of the same height, due toconsumption of plant material by the stem borer,but the data indicated that they were insteadheavier. The most likely explanation for thisresult is that stem borers killed the upper, thinnerparts of the stem, leaving only the thicker basalregion to be measured. An alternative possibilityis that the stem borers preferentially selectedstems with larger diameters that might providemore food or better protection from parasitoids.How stem borers affect S. alterniflora productionat the stand level remains unexplored; however,the lack of stem borers in plots disturbed in Julyand September did not allow plants to compen-sate for wrack disturbance.Consistent with past reports, we observed that

S. alterniflora stems containing stem borers usu-ally did not flower (Grevstad et al. 2004). In caseswhere flowering did occur, it is likely that thestem borer had colonized the stem too recently todisrupt the development of the flower. As a result,we hypothesize that widespread attack by stemborers might significantly reduce sexual repro-duction in S. alterniflora. Because stem borers alsolimit the growth of the stem, they may contributeto loss of vegetation in areas where S. alterniflorais already stressed by other factors (Gaeta andKornis 2011). In general, because stem borers aredifficult to directly observe, they have been over-looked in studies of salt marsh plant ecology (butsee Stiling and Strong 1983), and their role inmediating plant reproduction and vegetativegrowth deserves further attention.Natural disturbances can interact with the

mobility or phenology of consumers to affect theabundance of consumers (Armitage et al. 2013).By killing plant stems, wrack disturbance killedany stem borers present at the time of thedisturbance. Everything else being equal, older

❖ www.esajournals.org 6 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

stems will have had more time to accumulatestem-boring larvae than will the younger shootsthat emerged from the disturbed plots. Thisseems unlikely to explain why plants disturbedin May had low levels of stem borers in October,4 months later. More likely, the re-growing stemswere too short (Appendix S1: Fig. S6) to beattractive to egg-laying females during the timethat females were laying eggs. Armitage et al.(2013) also found that clipping of abovegroundS. alterniflora biomass in a Texas marsh early inthe growing season (April) reduced the abun-dance of larval stem borers. Similarly, in agricul-tural and forest systems, burning disturbancecan kill 95% of the stem borers (Adesiyun andAjayi 1980), and disturbance by snow may alsopromote the mortality of some stem borers(Ayres and Lombardero 2000).

A major gap in our understanding of how stemborers affect S. alterniflora is that we do not knowhow stem borers affect clonal integration amongshoots (Pennings and Callaway 2000). Herbivoryby stem borers (Utsumi and Ohgushi 2008, For-noni 2011, Stephens and Westoby 2015) or bychewing herbivores (Cain et al. 1991, Piqueras1999) may weaken other shoots on the same cloneif the attacked stem becomes a resource sink;alternatively, plants may be able to withdrawresources from attacked stems and divert them tosupport growth and flowering of other shoots onthe same clone. As a consequence, we cannotextrapolate from the performance of individualbored stems to estimate the net effect of stem-bor-ing insects on the clone as a whole.

Stem borers in forest and agricultural systemshave been well studied (Adesiyun and Ajayi 1980,Ayres and Lombardero 2000, Stephens and Wes-toby 2015), but fewer studies have focused onstem borers in grasslands. We found that attackby stem borers interacts with disturbance, suchthat the timing of disturbance affects the probabil-ity of attack by stem borers, and that both distur-bance and stem borers affect end-of-seasonstanding biomass. Because end-of-season stand-ing biomass is often used to estimate annual netprimary productivity in grasslands, our estimatesof ANPP could be improved by a better under-standing of both disturbance and herbivory.

Previous experimental studies of wrack distur-bance in salt marshes have typically applied thewrack at a single date that sometimes did and

sometimes did not correspond to the peak timeof natural wrack disturbance (Bertness and Elli-son 1987, Brewer et al. 1998, Pennings andRichards 1998). Future studies should considerthe natural seasonal pattern of wrack disturbancein designing their studies, either initiating treat-ments when natural disturbance peaks, or incor-porating multiple disturbance treatments torepresent the range of times when disturbancecan occur.

ACKNOWLEDGMENTS

S. C. P. was supported by the National Science Foun-dation (OCE12-37140) and S. L. by the China Scholar-ship Council. We thank Fan Li, Leslie Wu, CarolineReddy, and Qiang He for help in the field. This work isa contribution of the Georgia Coastal Ecosystems LTERprogram, and contribution number 1057 of the Univer-sity of Georgia Marine Institute. S. L. and S. P. con-ceived the ideas and designed methodology; S. L.collected and analyzed the data; S. L. led the writing ofthe manuscript; S. L. and S. P. contributed critically todrafts and gave final approval for publication.

LITERATURE CITED

Adesiyun, A. A., and O. Ajayi. 1980. Control of the sor-ghum stem borer, Busseola fusca, by partial burningof the stalks. International Journal of Pest Manage-ment 26:113–117.

Alber, M. A., K. Bulski, and Z. Cress. 2013. Wrack dis-turbance in salt marsh communities. Final report tothe coastal resources division, Georgia Dept. ofNatural Resources, October 2013. Department ofMarine Sciences, University of Georgia, Athens,Georgia, USA.

Angelini, C., and B. R. Silliman. 2012. Patch size-dependent community recovery after massive dis-turbance. Ecology 93:101–110.

Armitage, A. R., C.-K. Ho, and A. Quigg. 2013. Theinteractive effects of pulsed grazing disturbanceand patch size vary among wetland arthropodguilds. PLoS ONE 8:e76672.

Ayres, M. P., and M. J. Lombardero. 2000. Assessingthe consequences of global change for forest distur-bance from herbivores and pathogens. Science ofthe Total Environment 262:263–286.

Bertness, M. D., and A. M. Ellison. 1987. Determinantsof pattern in a New England salt marsh plant com-munity. Ecological Monographs 57:129–147.

Brewer, J. S., and M. D. Bertness. 1996. Disturbanceand intraspecific variation in the clonal morphol-ogy of salt marsh perennials. Oikos 77:107–116.

❖ www.esajournals.org 7 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

Brewer, J. S., J. M. Levine, and M. D. Bertness. 1998.Interactive effects of elevation and burial withwrack on plant community structure in someRhode Island salt marshes. Journal of Ecology 86:125–136.

Cain, M. L., W. P. Carson, and R. B. Root. 1991. Long-term suppression of insect herbivores increases theproduction and growth of Solidago altissima rhi-zomes. Oecologia 88:251–257.

Craine, J. M., J. B. Nippert, A. J. Elmore, A. M. Skibbe,S. L. Hutchinson, and N. A. Brunsell. 2012. Timingof climate variability and grassland productivity.Proceedings of the National Academy of ScienceUSA 109:3401–3405.

Crawley, M. J. 2004. Timing of disturbance and coexis-tence in a species-rich ruderal plant community.Ecology 85:3277–3288.

Crosby, S. C., M. Ivens-Duran, M. D. Bertness,E. Davey, L. A. Deegan, and H. M. Leslie. 2015.Flowering and biomass allocation in U.S. AtlanticCoast Spartina alterniflora. American Journal ofBotany 102:669–676.

Daleo, P., J. Alberti, C. M. Bruschetti, J. Pascual,O. Iribarne, and B. R. Silliman. 2015. Physical stressmodifies top-down and bottom-up forcing on plantgrowth and reproduction in a coastal ecosystem.Ecology 96:2147–2156.

Denno, R. F., M. J. Raupp, D. W. Tallamy, and C. F.Reichelderfer. 1980. Migration in heterogeneousenvironments: differences in habitat selectionbetween the wing forms of the dimorphic plan-thopper, Prokelisia marginata (Homoptera: Del-phacidae). Ecology 61:859–867.

Fill, J. M., S. M. Welch, J. L. Waldron, and T. A.Mousseau. 2012. The reproductive response of anendemic bunchgrass indicates historical timing of akeystone process. Ecosphere 3:61.

Fornoni, J. 2011. Ecological and evolutionary implica-tions of plant tolerance to herbivory. FunctionalEcology 25:399–407.

Gaeta, J. W., and M. S. Kornis. 2011. Stem borer fre-quency and composition in healthy Spartina alterni-flora (smooth cordgrass) and dieback zones in asouthern Atlantic coast salt marsh. Estuaries andCoasts 34:1078–1083.

Gallagher, J. L., R. J. Reimold, R. A. Linthurst, andW. J. Pfeiffer. 1980. Aerial production, mortality,and mineral accumulation-export dynamics inSpartina alterniflora and Juncus roemerianus plantstands in a Georgia salt marsh. Ecology 61:303–312.

Goranson, C. E., C.-K. Ho, and S. C. Pennings. 2004.Environmental gradients and herbivore feedingpreferences in coastal salt marshes. Oecologia 140:591–600.

Grevstad, F. S., M. S. Wecker, and D. R. Strong. 2004.Biological control of Spartina. Pages 267–272 inThird International Conference on Invasive Spar-tina. San Francisco Estuary Invasive Spartina Pro-ject of the California State Coastal Conservancy,San Francisco, California, USA.

Hamilton, N. S., B. Schmid, and J. Harper. 1987. Life-history concepts and the population biology ofclonal organisms. Proceedings of the Royal Societyof London B: Biological Sciences 232:35–57.

Hunter, M. D., and P. W. Price. 1992. Playing chutesand ladders: heterogeneity and the relative roles ofbottom-up and top-down forces in natural commu-nities. Ecology 73:724–732.

Li, S. 2016. Experimental wrack disturbance at fourtimes during the growing season. Georgia CoastalEcosystems LTER Project, Long Term EcologicalResearch Network, University of Georgia. https://doi.org/10.6073/pasta/0efb520e3772f0b97b560325853ebf32

Li, S., and S. C. Pennings. 2016. Disturbance in Georgiasalt marshes: variation across space and time. Eco-sphere 7:e01487.

Liu, W., K. Maung-Douglas, D. R. Strong, S. C.Pennings, and Y. Zhang. 2016. Geographical varia-tion in vegetative growth and sexual reproductionof the invasive Spartina alterniflora in China. Journalof Ecology 104:173–181.

Lottig, N. R., and J. M. Fox. 2007. A potential mecha-nism for disturbance-mediated channel migrationin a southeastern United States salt marsh. Geo-morphology 86:525–528.

Menge, B. A., and J. P. Sutherland. 1987. Communityregulation: variation in disturbance, competition,and predation in relation to environmentalstress and recruitment. American Naturalist 130:730–757.

Moon, D. C., and P. Stiling. 2002. Top-down, bottom-up, or side to side? Within-trophic-level interac-tions modify trophic dynamics of a salt marshherbivore. Oikos 98:480–490.

Orr, M., M. Zimmer, D. E. Jelinski, and M. Mews.2005. Wrack deposition on different beach types:spatial and temporal variation in the pattern ofsubsidy. Ecology 86:1496–1507.

Pennings, S. C., and M. D. Bertness. 2001. Salt marshcommunities. Pages 289–316 in M. D. Bertness,S. D. Gaines, and M. E. Hay, editors. Marine com-munity ecology. Sinauer Associates, Sunderland,Massachusetts, USA.

Pennings, S. C., and R. M. Callaway. 2000. The advan-tages of clonal integration under different ecologi-cal conditions: a community-wide test. Ecology 81:709–716.

❖ www.esajournals.org 8 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS

Pennings, S. C., and C. L. Richards. 1998. Effects ofwrack burial in salt-stressed habitats: Batis maritimain a southwest Atlantic salt marsh. Ecography 21:630–638.

Peterson, C. G., K. D. Hoagland, and R. J. Stevens.1990. Timing of wave disturbance and the resis-tance and recovery of a freshwater epilithicmicroalgal community. Journal of the North Amer-ican Benthological Society 9:54–67.

Peterson, C. G., and R. J. Stevenson. 1992. Resistanceand resilience of lotic algal communities: impor-tance of disturbance timing and current. Ecology73:1445–1461.

Pfeiffer, W. J., and R. G. Wiegert. 1981. Grazers onSpartina and their predators. Pages 87–112 in L. R.Pomeroy and R. G. Wiegert, editors. The ecology ofa salt marsh. Springer-Verlag, New York, NewYork, USA.

Pickett, S. T. A., and P. S. White. 1985. The ecology ofnatural disturbance and patch dynamics. Aca-demic Press, Orlando, Florida, USA.

Piqueras, J. 1999. Herbivory and ramet performance inthe clonal herb Trientalis europaea L. Journal of Ecol-ogy 87:450–460.

Shepherd, B., D. L. Miller, and M. Thetford. 2012. Fireseason effects on flowering characteristics and ger-mination of longleaf pine (Pinus palustris) savannagrasses. Restoration Ecology 20:268–276.

Shumway, S. W., and M. D. Bertness. 1994. Patch sizeeffects on marsh plant secondary succession mech-anisms. Ecology 75:564–568.

Silliman, B. R., and A. Bortolus. 2003. Underestimationof Spartina productivity in western Atlantic marshes:

Marsh invertebrates eat more than just detritus.Oikos 101:549–554.

Silliman, B. R., J. Van de Koppel, M. D. Bertness, L. E.Stanton, and I. A. Mendelssohn. 2005. Drought,snails, and large-scale die-off of southern U.S. saltmarshes. Science 310:1803–1806.

Stephens, A. E., and M. Westoby. 2015. Effects ofinsect attack to stems on plant survival, growth,reproduction and photosynthesis. Oikos 124:266–273.

Stiling, P. D., and D. R. Strong. 1983. Weak competitionamong Spartina stem borers by means of murder.Ecology 64:770–778.

Utsumi, S., and T. Ohgushi. 2008. Host plant variationin plant-mediated indirect effects: moth boring-induced susceptibility of willows to a specialist leafbeetle. Ecological Entomology 33:250–260.

Valiela, I., and C. S. Rietsma. 1995. Disturbance of saltmarsh vegetation by wrack mats in GreatSippewissett Marsh. Oecologia 102:106–112.

van Kleunen, M., M. Fischer, and B. Schmid. 2002.Experimental life-history evolution: selection onthe allocation to sexual reproduction and its plas-ticity in a clonal plant. Evolution 56:2168–2177.

Waldrop, T. A., D. L. White, and S. M. Jones. 1992. Fireregimes for pine-grassland communities in thesoutheastern United States. Forest Ecology andManagement 47:195–210.

Whelan, R. J. 1995. The ecology of fire. CambridgeUniversity Press, Cambridge, UK.

Wiezski, K., and S. C. Pennings. 2014. Climate driversof Spartina alterniflora saltmarsh production inGeorgia, USA. Ecosystems 17:473–484.

SUPPORTING INFORMATION

Additional Supporting Information may be found online at: http://onlinelibrary.wiley.com/doi/10.1002/ecs2.1675/full

❖ www.esajournals.org 9 February 2017 ❖ Volume 8(2) ❖ Article e01675

LI AND PENNINGS