Page  202



DennisA. Riege


Beavers(Castor canadensis andC. fibre)aresignificantmodifiersofplantcommunitiesbutstudies arelackingonthe indirect effectsofbeaveractivityontheunderstoryvegetationofnon-riparian forests. Beaver cutting of saplings during 2011–2016 altered the vegetation in part of a permanent plotestablishedin2007attheHuronMountainClubReserveinaPinus strobus–Picea glauca–Acer rubrum standwithanunderstoryofAcer saccharum saplings.Temporarypondingchangedthecommunitycompositionofanothersectionoftheplot. Thisstudyillustratestheimportanceofpermanent plots,wherebeforeandafterdataareavailabletoexaminetheeffectsofdisturbancesonsuccession. To examine beaver effects, a0.34-ha study area thatincluded asection cutbybeaver, as well as an adjacent uncut section was established within the permanent plot. Within the study area, beaver felled342stemsof1-9cmDBH(diameteratbreastheight),mostlyA. saccharum.Diametergrowth of saplings in the cut area was 1.8 times greater than in the uncut area. Acer saccharum or Acer rubrum sproutedfrom the majority of cut stumps andgrew upwardin spite of setbacks andforking fromdeerbrowsing.TheA. saccharum saplingthicketthatexistedpriortocuttingappearedtobereproducing itself, in contrast to other studies that reported that beaver cutting redirected succession from hardwoods to conifers. Frequencies of groundcover species were recorded annually during 2013-2019 in permanent belt transects, replicating data collected prior to beaver activity. Ground- cover richness has increased in the cut area, along with an influx of Rubus strigosus, Rubus parviflorus, and other species associated withgaps. Impatiens capensis has colonizedboth cut and uncut areas, but its frequency has decreased since 2017. The most-abundant species prior to disturbance (Dryopteris carthusiana, Maianthemum canadense, Trientalis borealis)have retainedhighfrequencies and should retain their dominance as the saplingthicket recovers. By contrast, in an area inundated during 2012-2015, these upland groundcover species have been replaced in dominance by denseR. strigosus.Whilesuccessionofthecutareamayshowaresiliencetobeaverdisturbance,that of the flooded area may be entering a recalcitrant understoryphase, dominatedby R. strigosus and resistanttotreeestablishment.PeriodicbeaverharvestingintheuplandforestborderingFisherCreek maybe maintainingtheA. saccharum saplingthicketin acyclicalunderstory succession.

KEYWORDS:Beaver, HuronMountains,disturbance, succession, groundvegetation


Beavers (Castor canadensis and C. fibre)arethe only animals other than humans that are able to cut andfelltrees, hence they can be asignificant modifier ofplantcommunities(Wrightetal.2002;Roselletal.2005;NummiandKuuluvainen 2013). Effects of beaver foraging on woody vegetation and succession have been investigated (Barnes and Dibble 1988; Johnston and Naiman 1990; Donkor andFryxell1999;Barnes andMallik2001). Donker andFryxell(1999) and Barnes and Mallik (2001) concluded that beaver preference for hardwood species wouldfavor succession toward conifers attheir mixed study sites.


Studies are lacking on the effect ofbeaver cutting on vegetation ofthe forest floor. Beavers mostlyforage on woody vegetation and aquatic herbs (Northcutt 1971; Allen 1983). Though beavers are known to ingest sedges and grasses (Allen 1983), the direct effects of beaver herbivory on upland groundcover are likely minimal. However, indirect effects caused by opening the subcanopy to increased light and by the disturbance created by cutting activity may be expectedtoalter ground species composition.

Long-termpermanentplotstudiesareneededtoexamineempiricalprocesses and test theoretical patterns of succession (Bakker et al. 1996; Johnson and Miyanshi2008;Riege2012).Apermanentplotwasestablishedin2007inalatesuccessional Pinus strobus–Picea glauca–Acer rubrum forest at the Huron MountainClubReserveinUpperMichigan(referredtoastheFCSplotinRiege 2011,2013).During2012–2016,beavercuttingofsaplingsopenedupareasofa thick Acer saccharum subcanopyin the plot.Withthe availability ofpre-disturbancedata, Iinitiated along-termstudyto investigate direct andindirecteffects ofbeaver cutting on vegetation ofthe forestfloor.

Part ofthe FCSpermanentplot was floodedby shallow beaver pondingduring2012– 2015.Upontherecessionofthewatersin2016,Ibegantoexaminethe effectsongroundcoverofthisepisodeofflooding.Althoughchangesinriparian vegetation in abandoned beaver ponds have been reported (Barnes and Mallik 2001;McMasterandMcMaster2001;Wrightetal.2002),researchislackingon changesingroundcoverinanuplandforestaftertemporaryinundation.Hyvönen and Nummi (2008) investigated post-inundation changes in woody vegetation, but notgroundherbs,in the uplands bordering abandonedponds.

Thispaperaddressesthequestions:(1)howdoesbeavercuttingaffectunderstory vegetation in a mature mixed forest? (2) how does an episode of beaver ponding affect groundcover? (3) what are the implications of these beaver disturbancesforfuturesuccessionofthestand? Previousstudiesofsuccessionafter beaver disturbance used achronosequence approach, comparing sites thatdiffer inyearssincebeaverabandonment(BarnesandMallik2001;McMasterandMc- Master 2001; Hyvönen and Nummi 2008). This beaver-effects study applied a temporal approach, tracking changes over many years via replicated data in a permanentplot.


Study area

The2500-haHuronMountainClubReserveinUpperMichigancontainsextensiveforestsofoldgrowthmesichemlockandhardwoods( Braun1950;Woods2000).During2007–2008,Iestablished twoadjacentpermanentplotstotaling2.4habyFisherCreekinthereserve,whichweredescribedin Riege(2011,2013).The0.82-hasouthplot(FisherCreekSouth,orFCS)containedaPinus strobus– Picea glauca–Acer rubrum stand with athick sapling understory of A. saccharum. During October 2011, considerable flooding along Fisher Creek was noted and beaver cutting of saplings along its bankwasevident,butnocutting was yetobserved on theFCSplot.However,byMay2012,beaver cuttinghad created openings inthemaple saplingthicket ofFCS.Acanalhadbeen constructedthat reached some 25 minto the plot. BySeptember 2012, floodinginundatedthe south end ofthe FCS plot.

During2013,Iestablishedastudyarea(N46.858°,W87.882°)of0.34hawithinFCStoexamine effectsofbeavercuttingonthevegetation.Thestudyareaincludeda0.1-hasubplot(H1),whichhad

Page  204 204 THE GREAT LAKES BOTANIST Vol. 59

TABLE 1. The basal area in m2/ha of individual species and the totals in the Huron Mountain Club Reserve FCS plot in 2011 and 2016.

2011 2016

Pinus strobus 30.4 28.8 Picea glauca 5.8 4.8 Acer rubrum 5.5 4.8 Acer saccharum 1.3 0.7 Tsuga canadensis 1.2 1.0 Betula alleghaniensis 0.9 0.9 Abies balsamea 0.5 0.1 Ostrya virginiana 0.0 0.0 Acer pensylvanicum 0.0 0.0 Totals 45.5 40.8

beenembeddedinFCStomonitorthegrowthandsurvivalofsmallsaplings.Thestudyareaboundaries were delineatedto include all area within 10 mofpreviously establishedtransectlines in FCS thatwerenotflooded.Areasofbeavercuttingweredistinctfromuncutareas.Foranalysisofbeaver effects,thestudyareawasdividedintoacutarea(0.20ha),wherebeaverhadfelledsaplings,andan uncut area (0.14 ha). The cut area was delineated to extend 1 m beyond the outermost line of cut stems.

By 2016, waters had receded from the south end of FCS. Changes in the permanent transects within the flooded zone were addedto this projectto examine effects of temporarybeaver ponding on the vegetation. Mosttrees (stems ≥10 cm DBH)diedin the flooded area by2016.Table 1illustrates adecrease in basal area oftree species in the FCSplotfrom 2011to 2016. (Beaver effects on tree composition oftheplotarenotincludedin thispaper.)

Field Methods

Saplings aredefined as alltrees with stems between 1.0 and10.0 cm DBH. For purposesofthis study,thesearedividedintosmallsaplings(stems1.0–4.9cmDBH)andlargesaplings(stems5–9.9 cmDBH). DuringAugust2013, all small saplings that occurred within 10 mfrom the transectlines weremappedanddiametersmeasured,aswereallcutstems≥1cmdiameterat15cminheight.Diametermeasurementswereat1- cmscale,roundingdown.Largesaplingshadalreadybeenmeasured in 2011. Each autumn thereafter, allnew cuts in the studyplotwererecorded. Duringautumn 2018, Iremeasureddiameters of all saplingsfrom1to9cm DBHinthe beaver study area.

InordertocalculatearegressionformulatoestimatetheDBHofcutstemspriortocutting(with theexceptionofcutstemsinsubplotH1whereDBHhadbeenmeasuredin2011),nineteensaplings were measuredin2013 atboth15 cm height and atDBH(i.e., at about1.4 mheight).The resultant formula of DBH = (0.95 * base diameter) – 0.36 cm was used to estimate DBH prior to cutting. Using the formula, stem bases that measured 5–9 cm in diameter were adjusted for DBH by subtracting 1 cm, since diameters were recorded in whole numbers in the plots.Asample of 19 stems wasconsideredsufficienttomakeadjustments atthis 1-cmscale.

Beginning in July 2016, the tallest living height was measured on sprouts from the cut stumps thatwerewithin1mfromthetransectmidlines.During2016–2018,thesemeasurementswereaninformaladdition to theground sampling, and somesprouts were missed. DuringJuly2019, in aseparatededicatedsurvey, allcutstumpswithin1.5mofthemidlinesweretallied,andallsproutswere mappedandtheirheightsmeasured.

Beginning in summer 2013, then annually through 2019, the presence by species of tree seedlings, shrubs, and herbs within the study area was recorded in continuous transects of 2 ¥ 2-m quadrats (that is, all species occurring within 1 m of a tape laid along the transect lines were recorded).Treeseedlingsweredefinedasstemslessthan1cmDBH(diameteratbreastheight).This replicatedthe samplingprotocol of2011for these permanentplottransects, whichhadbeen carried outpriortothebeaveractivity.Atotalof96quadratsweresampledwithinthebeaverstudyarea:65 in areasof cutting,31inuncut areas.


As part ofthe 2016 remeasurementprotocol, species composition ofgroundcover was recorded in the 32 south FCS transects that were inundated for most of 2012–2015, but in which the waters had recededby2016.Thereafter, the groundcover in the 32quadrats was sampled annuallythrough 2019.


Diameter growth rates of saplings within the study area and within the subplot H1 were comparedbythe nonparametricWilcoxonrank sum test, as the rates were not normallydistributed.The Chi-squaretestwasusedtoexamineanybeaverpreferenceforsizeofsaplingtocut.Statisticaltests wereconsideredsignificantifp<0.05.

Quadrat species richness is defined as the number of different species found in a quadrat. The specieslistedinthispaperareonlythosefoundinthequadratsandthereforedonotrepresentacomplete flora ofthe study area. Individuals in the genera Carex, Juncus,and Solidago were notidentifiedtospecies. However,itwasevidentwhenmorethanoneCarex specieswaspresentinaquadrat. Thus,forpurposesofdeterminingspeciesrichness,thenumberofCarex speciesineachquadratwas recorded.NomenclaturefollowsVossandReznicek(2012)forseedplantsandMICHIGANFLORA ONLINE(2011)forpteridophytes.

With the groundcover sample quadrats, the two-sample t-test was applied to determine if mean quadratspeciesrichnessdifferedbetweenthecutanduncutareasduringeachyearofthestudy.With the 32 inundated quadrats, species richness before and after flooding was also examined by the pairedt-test. Statistix9.0. software was usedforstatisticalanalyses.


Effects of beaver cutting on saplings

ByOctober 2016, beavers had cut342 saplings of1–9 cm DBHin the 0.34ha study area. No new cuts were found after 2016. Most of the cutting (267 stems)hadoccurredbySeptember2013.Nostemsgreaterthan9cmDBHwere felled. Beavers downed82% ofthe saplings in the 0.20-ha cut area, therebydecreasing the density from 2090 to 380 saplings/ha. Beaver cut 85% of the saplings 1–5 cm DBH (n = 398), but only 23% of saplings 6–9 cm DBH (n = 20).However,nosignificantdifferencewasfoundinbeaverpreferenceforavailablestemsizesbytheChi- squaretest(χ2=12.2,p=0.14,df=8).Frompre-cut stem maps or sprouts from stumps, I was able to identify 68% of the cuts to species; of these 94% were Acer saccharum. Also identified among the cuts were A. rubrum (n =10), A. pensylvanicum (n = 2), and Picea glauca (n =1), which wasconsistent withtheir small numbers on the plot.

Since the beaver disturbance, the diameter growth of saplings has been greaterincutthan uncutareas.WithinsubplotH1, which wasmeasuredin2011 prior to beaver activity,growth of all saplings from 2011to 2018 averaged0.91 cm versus 0.50 cm in the uncut area (p = 0.03, Wilcoxon rank sum test). The smallsaplingsinthestudyareaoutsideofH1werefirstmeasuredin2013.Their 5-year diameter growth rate wasalso greater incutthan inuncut areas (0.63 cm versus0.35cm),butnotatasignificantlevel(p=0.15,Wilcoxonranksumtest), althoughtheir meanannualgrowth rate (ca. 0.06cm/year)was similar tothatin H1.Within the entire beaver study area, diameter growth oflarge saplings from 2011to 2018 was also greater in cut than in uncut areas (mean of1.62 cm versus 0.92 cm,p =0.04,Wilcoxon rank sum test).

Page  206 206 THE GREAT LAKES BOTANIST Vol. 59

FIGURE1. Beaver-cutopeninginathicketofmaplesaplings. June2013. Photo byDennis Riege.

FIGURE 2. The same view as Figure 1: Maple saplings reclaiming the opening. September 2019. PhotobyKeithNelson.


Throughoutthecutarea,mapleseedlingshavesproutedfromthestumps.The Acer saccharum saplingthicketthatwasfelledappearedtobereproducingitself (Figures 1and2). DuringJuly2019, atotal of64 cut stumps were found within

1.5 m from the transect midlines. Maple sprouts (40 Acer saccharum,3 Acer rubrum)have emergedfrom 67% ofthe stumps.Althoughthe median height of these sprouted seedlings was 140 cm, eleven A. saccharum sprouts have grown toexceed200cminheightby2019,whichislikelyabovethedeerbrowsezone. The tallest A. saccharum sproutin the transects reached 400 cm in seven years of growth. However, most spouts have experienced browse back during this time. Of 15 sprouts with four years of measurements, ten have experienced at least one annual decrease in height (Figure 3), with sign of browsing by deer. This has resulted in multi-branched growth with forking of the many stems sprouting from the cut stumps (Figure 4). Although some were repeatedly browsed,onlytwoofthe43monitoredsproutswerenotedtohavediedby2019 (Figure 5). Effects of beaver cutting on the groundlayer

The groundlayer species that were dominant in 2011 maintained high frequenciesthrough2019, in bothcut and uncutareas(Table2: Dryopteris carthusiana, Maianthemum canadense, Trientalis borealis, seedlings of Acer saccharum andAcer rubrum).Thegreatestchangeinthevegetationwastheinvasionof the annual Impatiens capensis, a native species that colonized both the cut and the uncut areas (Table 2; Figure 6). In 2014, Impatiens capensis was not recordedinthequadrats.In2015itwaswidespreadinthecut(32% ofquadrats) and uncut areas (26%). After a maximum of 52% in 2017, Impatiens capensis frequencies have decreased (Figure 5). Rubus strigosus and Rubus parviflorus, species that associate with gaps, have increased from 2–5% to 20–29% frequencyinthecutareas. Severalspeciesappearedinthecut-areaquadratsforthe firsttimein2014(e.g., Scutellaria laterifolia, Scuttelaria galericulata, Fallopia convolvulus, Circaea alpina, Potentilla norvegica, Ribes glandulosum), although the latter three were absent by 2019. Two forbs, Viola blanda and Osmorhiza claytonii, which were common prior to cutting, were absent from the uncut area by 2014 and from the cut area by 2017 (Figure 6). Acer saccharum and Acer rubrum dominated the tree seedlings in the quadrats (Table 2). Seedlings of the two most dominant trees in FCS, Pinus strobus and Picea glauca, were rare.The cut areas exhibitedgreatertree seedling richness,includingfourspecies( Alnus incana,Fraxinus nigra,Prunus serotina,Quercus rubra) that were notfoundinthe uncutquadrats.

Effects of temporary beaver ponding on the groundlayer

Whereasdominantgroundspeciesinthemainstudyareamaintainedhighfrequencies after beaver cutting, the ground vegetation in the flood zone quadrats shifted dramatically from upland forest species (e.g., Dryopteris carthusiana, Maianthemum canadense, Acer saccharum seedlings) in 2011 to riparian and gap species (e.g., Rubus strigosus, Impatiens capensis, Carex spp.) in 2016–

Page  208 208 THE GREAT LAKES BOTANIST Vol. 59

FIGURE3.Changesinheightof15mapleseedlingsfrom2016to2019thathad sproutedfromstumpscutbybeaverin2013or2014,showingageneralupward trendbutwithsetbacksdue tobrowsing.Eachlinerepresentsoneseedling.

FIGURE 4. Asprout of Acer rubrum from a cut stump, illustrating forking of the multiple stems.PhotobyDennisRiege.


FIGURE 5.Asprout of Acer rubrum from a cut stump that died in 2019 after much browsing. Photo by Keith Nelson.

2019 (Table 3). Almost all trees and saplings died after inundation, creating a largegap.Impatiens capensis reached100% frequencyin2017,buthassincedecreased, paralleling its trend in the main study area.As in the cut areas, Rubus strigosus hasincreased,althoughmoreextensivelyinthefloodzone,to94% frequencyin 2019.Adense thicketdominatedby R. strigosus has grown to ca. 1.4 m in height (Figure 7). Many species that were not observed in 2011 appeared after thewaters receded,mostnotably Calamagrostis canadensis, Fallopia convolvulus, and Scutellaria galericulata (Table3).

Effects of beaver disturbance on species richness

Mean groundlayer species richness in the cut area increased from 5.3 to 6.3 speciesper4m2quadratfrom2011to2019,whileitdecreasedintheuncutarea from 5.9to 5.3. Since 2015, mean richness has been significantlygreater in the cut area (Table 4). Species richness declinedin the flood zone from 5.2 species perquadratin2011to4.4in2016afterflooding,butnotsignificantly(p=0.10,

Page  210 210 THE GREAT LAKES BOTANIST Vol. 59

TABLE2. Frequencies ofground species in the cut area andthe uncut area in selectedyears, as measured by the percentage of quadrats in each area in which each species is present. The 2011 column reflects data prior to beaver cutting. Species with all zero values were present in years other than listed.



2011 2016 2019 2011 2016 2019 Tree Seedlings Abies balsamea L. 3 0 0 3 0 0 Acer pensylvanicum L. Acer rubrum L. 14 20 14 48 15 60 13 39 16 42 19 84 Acer saccharum Marshall 84 71 75 80 55 65 Alnus incana (L.)Moench Betula alleghaniensis Britton Fraxinus nigra Marshall Ostrya virginiana (Mill)K.Koch Picea glauca (Moench)Voss Pinus strobus L. 0 2 0 9 5 5 2 6 0 9 6 0 0 12 2 3 6 2 0 0 0 13 0 0 0 0 0 13 0 0 6 0 0 6 0 0 Prunus virginiana L. Prunus serotina Ehrh. 0 0 3 0 2 0 0 0 3 0 0 0 Quercus rubra L. 0 3 3 0 0 6 Tsuga canadensis L. 3 8 2 3 0 0 Herbs and Shrubs Aralia nudicaulis L. 22 28 35 23 16 23 Brachyelytrum aristosum Michx Carex spp Circaea alpina L. Clintonia borealis (Aiton) Raf. Coptis trifolia (L.)Salisb. Cornus canadensis L. 12 26 0 11 3 3 11 39 0 0 6 6 17 35 0 0 3 6 19 16 0 19 10 13 16 6 0 6 6 6 13 19 0 6 6 13 Corylus cornuta Marshall Dryopteris carthusiana (Vill.)H. P.Fuchs Eurybia macrophylla (L.)Cass Fallopia convolvulus (L.)A. Love Galium triflorum Michx. 6 97 0 0 5 5 95 0 6 2 6 94 0 2 0 0 94 0 0 0 0 84 0 0 0 0 90 0 0 0 Huperzia lucidula (Michx.)R.Trevis Impatiens capensis Meerb. Juncus spp. Lonicera canadensis Marshall 6 0 0 9 0 35 3 6 2 18 2 5 0 0 0 16 0 36 0 16 0 23 0 16 Maianthemum canadense Desf. 78 74 78 87 74 74 Oryzopsis asperifolia Michx. Osmorhiza claytonii (Mich.) C.B. Clarke Polygonatum pubescens (Willd.)Pursh. Potentilla norvegica L. Pteridium aquilinum (L.)Kuhn Pyrola elliptica Nutt. Ribes glandulosum Grauer Ribes lacustre (Pers.)Poir. Rubus strigosus Michx, Rubus parviflorus Nutt. Scutellaria galericulata L. Scutellaria lateriflora L. 6 11 0 0 5 0 0 2 2 5 0 0 8 0 3 2 6 0 0 0 20 20 2 2 5 0 2 0 3 2 0 0 29 20 2 2 6 23 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 Streptopus lanceolatus (Aiton)Reveal Symphyotrichum lanceolatus Willd. Trientalis borealis Raf. 6 0 58 0 2 59 0 0 74 16 0 65 0 0 55 0 0 58 Trillium cernuum L. 2 3 2 3 0 3 Vaccinium angustifolium Aiton Veronica officinalis L. 0 0 2 0 2 0 0 0 0 0 0 0 Viola blanda Willd. 28 6 0 23 0 0


FIGURE6.Frequenciesovertimeoffivespecieswithinthebeaver-cutareaas measuredbythe percentage ofquadrats in which each species is present.The 2011 data was recorded prior to beaver cutting. Vibl = Viola blanda; Oscl = Osmorhiza claytonii; Imca = Impatiens capensis; Rust = Rubus strigosus; Rupa=Rubus parviflorus.

df=31,paired t-test). However,by2017 meanrichness perquadrat was similar to that in 2011 (Table 4), although species composition had radically changed (Table 3).


Acer saccharum sprouts and seedlings are beginning to grow rapidly in the cut areas, especiallyfrom the remnant stumps. Bythis trajectory, A. saccharum willbethemostsuccessfulspeciesinreclaimingtheclearingscreatedbybeaver. Thebeaver disturbance willlikelyhave theeffect of acyclical succession in the understory back to an A. saccharum sapling thicket similar to the one that existedin2011. Thispatterncontrastswithstudiesthatconcludedthatbeaverpreference for hardwood species would alter succession toward conifers (Donkor andFryxell1999;BarnesandMallik2001).Althoughconiferspeciesaccounted for 85% of the basal area of the FCS stand (Table 1), conifer saplings (Riege 2011) and seedlings (Tables 2 and 3) were very rare, perhaps due to the dense shade of the thicket. Hence, A. saccharum should face little competition from conifers inthe understory.

Similar to this study, Jacobs (1969) found considerable episodic height reduction and forking in deer-browsed Acer saccharum seedlings. In spite ofthis browse back, he reported that seedlings in favorable, open areas grew out of

Page  212 212 THE GREAT LAKES BOTANIST Vol. 59

TABLE 3. Frequencies of ground species in the transects that were inundated for most of 2012–2015, as measured by the percentage of quadrats in which each species is present. The 2011 column reflects data prior to inundation, and the 2016 and 2019 columns reflect post-flooding data. Species that were present in 2017 or 2018 are listed with all zero values for 2017 and 2018.

2011 2016 2019 Tree Seedlings Abies balsamea 0 6 0 Acer pensylvanicum 13 0 0 Acer rubrum 19 9 25 Acer saccharum 75 13 16 Alnus incana 0 0 6 Betula alleghaniensis 0 0 3 Ostrya virginiana 3 0 0 Picea glauca 0 0 13 Prunus virginiana 0 0 0 Quercus rubra 0 0 3 Herbs and Shrubs Aralia nudicaulis 19 3 6 Athyrium felix-femina (L.) Roth 0 0 0 Brachyelytrum aristosum 13 3 19 Calamagrostis canadensis (Michx.)P. Beauv. 0 19 25 Carex spp 47 88 69 Chelone glabra L. 0 0 6 Cirsium muticum Michx. 0 3 0 Dryopteris carthusiana 88 38 63 Eqisetum sylvaticum L. 3 3 3 Eurybia macrophyllum 6 0 0 Fallopia convolvulus 0 9 25 Galium triflorum 6 3 3 Impatiens capensis 0 94 50 Iris versicolor L. 0 3 3 Juncus spp. 0 19 6 Lonicera canadensis 6 0 0 Maianthemum canadense 84 16 16 Onoclea sensibilis L. 6 9 9 Oryzopsis asperifolia 3 0 0 Osmunda cinnamomea (L.)C. Presl 0 0 3 Osmorhiza claytonii 38 0 0 Potentilla norvegica 0 3 0 Pteridium aquilinum 3 0 0 Pyrola elliptica 3 0 0 Rubus strigosus 0 63 91 Rubus parviflorus 6 3 3 Sambucus racemosa L. 3 0 0 Scutellaria galericulata 0 31 16 Solidago spp. 0 6 3 Trientalis borealis 44 9 19 Trillium cernuum 3 0 0 Verbascum thaspus L. 0 0 0 Viola blanda 16 0 0 Viola cucullata Aiton 13 0 0


FIGURE 7. Dense Rubus strigosus thicket in 2018 that developed in an area inundated during 2012–2015. The ground cover was dominated by upland herbs in 2011 prior to flooding. Photo by KeithNelson.

reach of deer. This also seems to be the case at Fisher Creek, where deer may slow but not stop saplinggrowthin A. saccharum.Fei andSteiner (2009) noted thatA. rubrum canregainthegrowingspaceitoccupiedinsevenyearsafterharvestingbystumpsproutingalone. Acer saccharum isalessprolificsprouterthan Acer rubrum (SolomonandBlum1967),butappearstobewellonitswaytoregainingits space atFisher Creek.

Diameter growth of saplings in the cut area was 1.8 times greater than the

TABLE4.Speciesrichness(meannumberofspeciesperquadrat)±standarderrorperyearofground plantsinthecut,uncut,andfloodedareas.The2011datawasrecordedpriortobeaveractivity. The flooded area was inundated during the period 2012–2015. A single asterisk (*) indicates data for whichp<0.05; adoubleasterisk(**)indicates data forwhichp<0.01,indicatingthatthecutarea differsfromtheuncut areabythe twosamplet-test.

Year Uncut Cut Flooded 2011 5.9±0.3 5.3±0.4 5.2 ± 0.3 2013 4.9±0.5 5.0±0.3 2014 5.0±0.4 5.9±0.3 2015 4.8±0.4 6.8±0.4** 2016 4.6±0.4 6.1±0.3** 4.4 ± 0.3 2017 4.6±0.4 5.8±0.3* 5.0 ± 0.3 2018 4.6±0.4 5.9±0.3** 5.0±0.3

Page  214 214 THE GREAT LAKES BOTANIST Vol. 59

uncutarea.TheuncutarearemainedshadedbyadenseA. saccharum saplingunderstory. The cut area also was closer to the open corridor ofFisher Creek, further increasinglight availability.The increasedgrowth ratewillfacilitate there- production ofthe A. saccharum saplingthicketin the cut area.This leads me to speculate that past episodes of beaver cutting may have assisted in limiting A. saccharum to a dense subcanopy under the Pinus strobus–Picea glauca–Acer rubrum forest.McGinleyandWhitham(1985)reportedthatselectivebeaverforaging on Populus fremontii keptitina perpetualjuvenile condition.

Most stems cut in this study were 2–5 cm DBH, and the largest was 9 cm DBH. Raffel et al. (2009) found a preference for 2–6 cm DBH stems, and that treesgreaterthan9cmDBHwereavoidedbybeavers.Beaverareknowntofell much larger trunks. Donkor and Fryxell (1999) reported an average cut size of 15cmDBHinlowlandborealforests.JohnstonandNaiman(1990)foundmeans of14 cm and10 cm at two sites, with a maximum cut of43.5 cm DBH. When establishing atFisher Creekin 2011,beaver encountered an abundance of small saplings. The overwhelming abundance of small Acer saccharum saplings limitedsizeandspecieschoice. DonkorandFryxell(1999)didnotethatA. saccharum and Acer rubrum are preferred species.

The retention ofhighfrequencies ofthe dominant species (Table 2) suggests resilience of the groundlayer community at Fisher Creek to disturbance by beaver cutting. Rubus strigosus and Rubus parviflorus have increased with the opening of the A. saccharum sapling thicket, but will likely decrease as the saplingsreproduceshadycover.DonosoandNyland(2006)foundthatR. strigosus abundance declines in clearcuts after about5years, as trees grow above the raspberry layer. This especially occurs in stands with advance regeneration, whichmaybe the case withthe stumpspouting atFisherCreek.Archambault et al. (1998) and Donoso and Nyland (2006) reported that most individuals of R. strigosus die within 10years after logging.

Many species that first appeared in the cut-area quadrats in 2014 and 2015 (e.g., Impatiens capensis, Ribes glandulosum, Scutellaria laterifolia, S. galericulata, Potentilla norvegica, Fallopia convolvulus, Circaea alpina)are generally associated with moist, streambank, or disturbed habitats (Curtis 1959; Chadde 2013). With the exception of I. capensis, these species were rare and might be expectedtodeclinewithclosureofthesaplingsubcanopy.Impatiens capensis is an annualplantthatis able to germinate earlyin the springin disturbed riparian areasanddevelopadensepopulationwithacontinuouscanopythatexertshabitat dominance over the herbaceous layer (Winsor 1983). Early germination allowsittopersistmanyyearsafteradisturbance, althoughitsrangewasdecreasing five years after beaver cutting at Fisher Creek. Interestingly, Impatiens capensis colonizedtheuncutareaatasimilarrateasitdidinthecutarea(Table 3). Most of its invasion of the uncut area occurred in a heavily shaded sapling thicket withlow ground cover. Impatiens capensis maywellpersistin theuncut areaslonger,asitisfacingdensecompetitivegrowthinthecutarea,particularly bytheRubus shrubs.TheforestherbsViola blanda andOsmorhiza claytonii exhibitedthe greatestdecrease, with neither species notedin cut or uncutquadrats after 2017 (Table 3). This may not be related to beaver disturbance, as both speciesalsodeclinedatotherHuronMountainReserveplotsin2016(Riege,un


publisheddata). Wiegmann andWaller (2006) reportedboth species as “losers†in 50years of change in regionalforests andbelieveddeer herbivory akeyfactor. Shelton et al. (2014) reportedthatdeer reduced abundance of O. claytonii.

While composition of the groundlayer in the main study area suggests resilience after beaver disturbance, vegetation in the flooded zone has been radically altered.Almost all trees have died. By2019, dense Rubus strigosus filled the opened area (Figure 7). Donoso and Nyland (2006) reviewed examples where R. strigosus inhibited tree establishment beyond 15 years, in sites that werepoorlydrainedorlackedadvanceregeneration.Intheirreviewof125studies, Royo and Carson (2006) included Rubus spp. among plants that can form what they termed a “recalcitrant understory layer†that can alter the rate or direction of succession.

Studies of succession post-inundation on non-riparian forests arerare. Hyvönen and Nummi (2008) reported that deciduous trees may be favored over conifersinthisenvironmentafterbeaverflooding.TerwilligerandPastor(1999) foundthatflooding maykillthe ectomycorrhizae necessaryfor conifer seedling establishment. Acer saccharum, Acer rubrum,and Picea glauca seedlings were present in the flood zone quadrats at Fisher Creek in 2019 (Table 3) but were small in size and numbers—far from an advance regeneration. Whether tree seedlings can develop through a potential recalcitrant R. strigosus layer is an open question. Isuspectthatthe flooded area ofFCS will not return to a Pinus strobus–Picea glauca–A. rubrum standwithanA. saccharum subcanopy,unless perhaps if undisturbedfor severaldecades.

Groundlayer species richness in the cut area atFisher Creekhas increased as expected with colonization ofgap species. Continuation ofthis study will allow testing of a hypothesis that species richness will peak and then decrease in the cut area as the A. saccharum sapling subcanopy redevelops and inhibits shade- intolerant species. This result would be in accord with the intermediate disturbance hypothesis (Connell 1978), which proposes that diversity increases to a maximum following adisturbance thendeclines.

This study illustrates the value of long-term permanent plots, where vegetationdataarecollectedbeforeandtrackedafteradisturbance. Forestsaresubject to many disturbances (e.g., disease, species invasions, wind-throw, climate change)thatinfluence succession. Long-term plots in place prior to disturbance events provide invaluable opportunities for a cause-and-effect examination of vegetation dynamics (Bakker et al. 1996). Continuation of this project will test hypothesesthatthetrajectoryofthebeaver-cutareawilltrendtoreproductionof pre-existing vegetation, while that ofthe beaver-flood area will stallin amultidecadalrecalcitrant understorylayer.


The Huron Mountain Wildlife Foundation has generously supported my research at the Huron MountainClubReserve.Iespeciallythankitsdirector,KerryWoods,foradvicefromthestartofmy long-term studies.The HanesTrust andthe Michigan BotanicalFoundation have been instrumental inprovidingfundstocontinuethisbeaver-effectsstudy.KeithNelson,PaulBaumann,BarryRosett, andAlison Paulson assisted in the field data collection over the years. Thanks are due to Michael Rotter, BilAlverson, Katie Frerker, Sarah Johnson, and Don Waller for help in species identifica

Page  216 216 THE GREAT LAKES BOTANIST Vol. 59



Allen,A.W.(1983).Habitatsuitabilityindexmodels:Beaver.WesternEnergyandLandUseTeam, DivisionofBiologicalService,ResearchandDevelopment,FishandWildlifeService,U.S.Dept. ofthe Interior,Washington, D.C.

Archambault, L., J. Morrissette, and M. Bernier-Cardou. (1998). Forest succession over a 20-year period following clearcutting in balsam fir-yellow birch ecosystems of eastern Quebec, Canada. ForestEcology andManagement102:61–74.

Bakker,J .P.,H.Olff,J. H.Willems,andM.Zobel.(1996).Whydoweneedpermanentplots inthe study oflong-termvegetationdynamics? Journal ofVegetation Science7:147–155.

Barnes,W. J., andE.Dibble. (1988).Theeffectsofbeaver in riverbankforest succession. Canadian Journal ofBotany66:40–44

Barnes,W. J.,andA. U.Mallik. (2001). Effects ofbeaver, Castor canadensis, herbivory on stream- side vegetation inanorthernOntariowatershed.CanadianField-Naturalist115:9–21.

Braun,E.L. (1950).Deciduousforestsof easternNorthAmerica. Macmillan,NewYork,N.Y.

Chadde,S.W.(2013).Wisconsinflora:AnillustratedguidetothevascularplantsofWisconsin.CreateSpaceIndependentPublishingPlatform. Connell, J. H.(1978). Diversityintropicalrainforestsandcoralreefs.Science 199:1302–1310.

Curtis,J.T.(1959).Vegetation ofWisconsin.UniversityofWisconsin Press, Madison.

Donkor, N. T., andJ. M. Fryxell. (1999). Impact of beaver foraging on structure oflowland boreal forestofAlgonquinProvincialPark,Ontario.ForestEcology andManagement118:83–92.

Donoso,P.J.,andR.D.Nyland.(2006).InterferencetohardwoodregenerationinnortheasternNorth America:Theeffectsofraspberries(Rubus spp.)followingclearcuttingandshelterwoodmethods. NorthernJournal ofAmericanForestry23:288–296.

Fei,S.,andK.C.Steiner.(2009).Rapidcaptureofgrowingspacebyredmaple.CanadianJournalof ForestResearch39:1444–1452.

Hyvönen, T., and P. Nummi. (2008). Habitat dynamics of beaver Castor canadensis at two spatial scales.Wildlife Biology14:302–308.

Jacobs, R. D. (1969) Growth and development of deer-browsed sugar maple seedlings. Journal of Forestry67:870–874.

Johnson,E.A.,andK.Miyanshi.(2008).Testingtheassumptionsofchronosequencesinsuccession. EcologyLetters11:419–431.

Johnston, C. A., and R. J . Naiman. (1990). Browse selection by beaver: Effects on riparian forest composition. Canadian JournalofForestResearch20:1063–1043.

McGinley, M.A., andT. G. Whitham. (1985), Central place foraging by beavers (Castor canadensis): Atest offoragingpredictions andimpact of selective feeding on the growthform of cottonwoods( Populus fremontii). Oecologia 66:558-562.

McMaster, R. T., andN. D. McMaster. (2001). Composition, structure, anddynamics of vegetation infifteenbeaver-impactedwetlands inwestern Massachusetts. Rhodora 103:293–320.

MICHIGANFLORAONLINE.A.A.Reznicek,E.G.Voss,andB.S.Walters.(2011).Universityof Michigan.Availableat (AccessedApril15,2020).

Northcutt,T.H.(1971).Feedinghabits ofbeaver in Newfoundland.Oikos 22:407–410.

Nummi, P.,andT.Kuuluvainen.(2013)Forestdisturbancebyanecosystemengineer:Beaverinborealforestlandscape. BorealEnvironmentResearch18(A):13–24.

Raffel, T. R., N. Smith, C. Cortwright, andA. J. Gatz. (2009). Central place foraging by beavers (Castor canadensis)in acomplexlakehabitat.TheAmericanMidlandNaturalist162:62–73.

Riege, D.A. (2011).Demography of old-growth white pine standsatthe Huron Mountain ClubReserveandEstivantPines inUpper Michigan.TheMichiganBotanist50:107–117.

Riege, D. A. (2012). Surge in regeneration of Pinus strobus L. in three Wisconsin forests not projectedbypastdemography. JournaloftheTorreyBotanicalSociety.139:299–310.

Riege,D.A.(2013).Groundvegetationofold-growthwhitepinestandsattheHuronMountainClub ReserveandEstivantPines in Upper Michigan.TheMichigan Botanist52:80–92.

Rosell,F.,O.Bozser,P.Collen,andH.Parker.(2005).EcologicalimpactofbeaversCastor fibre and Castor canadensis andtheir abilityto modifyecosystems.MammalReview35:248–276.


Royo,A.A.,andW.P.Carson.(2006).Ontheformationofdenseunderstorylayersinforestsworldwide: Consequencesandimplicationsforforestdynamics,biodiversity,andsuccession.Canadian Journal ofForestResearch36:1345–1362.

Shelton,A.L., J.A. Henning, P. Schultz, andK. Clay. (2014). Effectsof abundant white-taileddeer onvegetation,animals,mycorrhizalfungi,andsoils.ForestEcologyandManagement320:39–49.

Solomon,D. S., andB. M.Blum. (1967). Stumpsprouting offournorthern hardwoods. USDAForestService, ResearchPaper NE-59. Northeastern ForestExperimentStation, UpperDarby, Pennsylvania.

Terwilliger, J., and J. Pastor (1999). Small mammals, ectomycorrhizae, and conifer succession in beavermeadows. Oikos85:83–94.

Voss, E. G., andA.A. Reznicek. (2012). Field manual ofMichigan flora. The University ofMichigan Press,AnnArbor.

Wiegmann, S. M., andD. M.Waller. (2006). Fiftyyears of change in northern uplandforest under- stories: Identity and traits of “winner†and “loser†plant species. Biological Conservation 126: 109-123.

Winsor, J. (1983). Persistence by habitat dominance in the annual Impatiens capensis (Balsaminaceae). JournalofEcology71:451–466.

Woods, K. D. (2000). Dynamics in late-successionalhemlock-hardwoodforests over three decades. Ecology81:110–126.

Wright, J. P., C. G. Jones, andA. S. Flecker. (2002).An ecosystem engineer, the beaver, increases speciesrichness atthelandscapescale.Oecologia 132:96–101.