Harrowing and Seed Addition for Sandplain Grassland and

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, researchlibraries, and research funders in the common goal of maximizing access to critical research.

Harrowing and Seed Addition for Sandplain Grassland and Heathland RestorationAuthor(s): Kelly A. Omand, Jennifer M. Karberg, Danielle I. O'Dell and Karen C. BeattieSource: Natural Areas Journal, 38(5):356-369.Published By: Natural Areas Associationhttps://doi.org/10.3375/043.038.0505URL: http://www.bioone.org/doi/full/10.3375/043.038.0505

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, andenvironmental sciences. BioOne provides a sustainable online platform for over 170 journals and books publishedby nonprofit societies, associations, museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance ofBioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiriesor rights and permissions requests should be directed to the individual publisher as copyright holder.

https://doi.org/10.3375/043.038.0505

http://www.bioone.org/doi/full/10.3375/043.038.0505

http://www.bioone.org

http://www.bioone.org/page/terms_of_use

356 Natural Areas Journal Volume 38 (5), 2018

ABSTRACT: Sandplain grassland and coastal heathland are globally rare communities threatened by succession, development, and shoreline change. Restricted mainly to coastal plains of the Northeast, they provide vital habitat for many state and regionally rare species, highlighting their importance for conservation and restoration. Brush-cutting and prescribed fire have been successful in maintaining these communities but have proven ineffective in converting overgrown shrubland to an earlier successional state. Previous research has established that the soil seed bank in overgrown shrubland lacks key dominant grassland species, and that while small-scale soil disturbance leads to short-term increases in grasses and forbs, woody species rapidly resume dominance. We examined whether applying larger-scale soil disturbance in an area previously treated with annual brush-cutting could more effectively impede woody regrowth while providing an extensive mineral soil seed bed to recruit desired vegetation. We evaluated the effectiveness of harrowing, harrowing plus seeding, and seeding alone, compared to untreated plots and to existing sandplain grassland and coastal heathland reference sites. Harrowing stimulated seed bank germination of varied native disturbance-associated species, including four Massachusetts state-listed rare species that recruited only within the harrowed area, and indicated that the seed bank lacked key grassland dominants. Comparison to reference grassland and heathland sites suggested that harrowing plus seeding more closely approached the desired restoration outcome than any other treatment. Our results suggest that larger-scale soil disturbance can enhance biodiversity by stimulating germination of native early successional plants from the seed bank while providing germination sites for sown target species.

Index terms: grassland, harrowing, seed addition, seed bank, soil disturbance

INTRODUCTION

Globally rare sandplain grassland and coastal heathland, ranked S1 (Critically Im-periled) in Massachusetts, provide hotspots for regionally rare and endangered early successional species (Carlson et al. 1991; Leahy 1993; Dunwiddie et al. 1996; Sorrie and Dunwiddie 1996; Barbour et al. 1999; Swain and Kearsley 2001). It is estimated that globally >98% of sandplain grassland and 85–98% of coastal heathland (hereafter “grassland” and “heathland”) have been lost (Noss et al. 1995), primarily since the 1850s. Declines are attributed to agricul-tural abandonment and fire suppression, resulting in the expansion of native shrub-lands and pitch pine barrens (Godfrey and Alpert 1985; Tiffney and Eveleigh 1985; Barbour et al. 1999). Similar successional trends have been observed in grasslands in other regions of North America (Briggs et al. 2005) and in Europe (Dzwonko and Loster 2008) with reductions in plant species diversity (Ratajczak et al. 2012), resulting in a shift from grassland to shru-bland in many areas (Briggs et al. 2005; D’Odorico et al. 2012). Without distur-bance-based management, woody species encroachment reduces available habitat for rare obligates, placing them at risk of extirpation (Dunwiddie 1989; Dunwiddie and Caljouw 1990; Barbour et al. 1999; Dzwonko and Loster 2008). Nantucket land management has included prescribed

fire, brush-cutting, and sheep grazing treatments (or combinations thereof) to maintain grasslands and heathlands, or to interrupt succession in shrublands (Dun-widdie 1998; Omand et al. 2014); similar techniques have also been widely used for grassland restoration and maintenance in other regions (Rowe 2010).

On Nantucket Island, Massachusetts, remaining grassland and heathland are of high conservation priority. The most recent island-wide vegetation survey in-dicated that the island retained only 251 ha of grassland and 1121 ha of heathland, compared to approximately 2642 ha of scrub oak shrubland (hereafter “shru-bland”) (TNC 1998). Shrubland expansion and development have further restricted these areas, and currently grasslands and heathlands are concentrated primarily on the southern outwash plain of the island where wind and salt spray limit the growth of shrubs and trees and favor grassland and heathland species. These were once much more widespread on the island, including the harrowing and seed addition site (Tiff-ney and Eveleigh 1985; Dunwiddie et al. 1996; Griffiths 2006). Nantucket’s south shore experiences dramatic erosion losses, up to 3.65 m per year (Bernd-Cohen and Gordon 1999), a rate expected to increase in the coming decades (Parris et al. 2012; MA Office of Coastal Zone Planning 2013). In spite of these threats, Nantucket’s large

Natural Areas Journal 38:356–369

R E S E A R C H A R T I C L E

2 Corresponding author: [email protected]; (508) 228-2884

•

Harrowing and Seed Addition for

Sandplain Grassland and Heathland

Restoration

Kelly A. Omand1,2

Nantucket Conservation FoundationScience and Stewardship Department

P.O. Box 13Nantucket, MA 02554

Jennifer M. Karberg1

Danielle I. O’Dell1

Karen C. Beattie1

•

Associate editor: Noel Pavlovic

Volume 38 (5), 2018 Natural Areas Journal 357

expanses of permanently conserved land (43.8% of the island, approximately 5554 ha) offer opportunities to expand these rare communities farther inland, where they will be buffered from projected shoreline changes (BioMap2 2012).

While brush-cutting and prescribed fire have been useful in maintaining grass-lands and heathlands, these techniques have proven less effective at converting shrublands back to grasslands or heathlands (Dunwiddie 1998; Lezberg et al. 2006; Beattie et al. 2008). After almost 20 y of annual dormant season brush-cutting in a 93-ha scrub oak shrubland on Nan-tucket, the vegetation remains dominated by native woody species, lacking many grassland/heathland species (Beattie et al. 2008; Omand et al. 2014). Research on soil disturbance effects in narrow disc harrowed strips within this brush-cut area documented increased graminoid and forb cover, but many desired grassland and heathland species remained absent. Woody shrubs such as scrub oak (Quercus ilicifolia Wangenh.), dwarf chestnut oak (Quercus prinoides Willd.), and black huckleberry (Gaylussacia baccata [Wangenh.] K. Koch) rapidly recolonized from the edges via rhizomes, as did Pennsylvania sedge (Carex pensylvanica Lam.) (O’Dell 2014; botanical nomenclature follows Haines 2011). Omand et al. (2014) documented a depauperate seed bank lacking key grassland dominants such as little bluestem grass (Schizachyrium scoparium [Michx.] Nash) in shrubland, compared to the seed banks of intact Nantucket grassland and heathland sites. This parallels research conducted on nearby Martha’s Vineyard, where the soil seed bank also appeared to be lacking key sandplain species that only recruited with seed addition during restoration (Lezberg et al. 2006; Neill et al. 2015). Seed limitation and microsite suitability represent important limiting factors for seedling recruitment, partic-ularly for rare or disturbance-associated species in sandplain habitats (Farnsworth 2006; Clark et al. 2007). In the absence of a substantial soil seed bank, researchers in other early successional systems have often recommended seed addition to enhance restoration efforts (Lunt 1997; Laughlin 2003; von Blanckenhagen and Poschlod

2005; Lezberg et al. 2006; Lang and Halpern 2007; Dzwonko and Loster 2008; Valko et al. 2011).

Prescribed fire has been employed to reduce surface accumulation of duff and litter and create mineral seed beds to facilitate germination of disturbance-dependent spe-cies. However, this method is not always effective with low-intensity burns or in areas with deep duff layers, and manual soil scarification has been proposed as a complement to prescribed fire in sandplain grassland restoration (Neill et al. 2007). Tilling is often the first stage in many grassland restoration projects to reduce competition from existing vegetation and create germination sites for sown species (Hofmann and Isselstein 2004; Luzuriaga et al. 2005; Edwards et al. 2007). Past human activity such as Native American agriculture and burning, followed by Eu-ropean colonists’ intensive sheep grazing (Dunwiddie 1989; Motzkin and Foster 2002) and soil disturbance associated with cultivation (Foster and Motzkin 2003), have long been recognized as factors contribut-ing to unusually large areas of grassland and heathland on Nantucket, including within the treatment and reference areas of the current project.

In light of past research and the island’s historical context, we wished to evaluate the effectiveness of combining larger-scale soil disturbance with native seed addition to restore grassland and/or heathland communities in shrubland. We anticipated that soil disturbance over a broader area would more aggressively impact the root systems of woody shrubs, limiting their regrowth and encouraging establishment of early successional species from the seed bank. We hypothesized that seed addition following harrowing would result in higher species diversity and recruitment of the desired grassland/heathland plants, providing a better restoration outcome than annual brush-cutting or seed addition alone.

METHODS

Study Location

This research took place on Nantucket Is-

land, Massachusetts, located 44 km south of Cape Cod (41°15′24″N, 70°3′35″W; Figure 1). Nantucket Island is approximately 116 km2 in land area; elevation ranges from sea level to 33 m. Nantucket receives average annual precipitation of 953 mm (US Climate Data, n.d.) and is located in USDA Plant Hardiness Zone 7b (USDA 2012). The island’s surficial geology con-sists mainly of glacial end moraine and outwash plain deposits of differing age (Oldale 1985). Sandplain grasslands and coastal heathlands dominate the southern outwash plains, while scrub oak shrublands predominate farther inland on the glacial moraine (TNC 1998).

Study sites were located in the Middle Moors in the island’s interior within the outwash plain (Oldale 1985). Two Research Areas (Area 1 = 1.16 ha; Area 2 = 1.25 ha) were delineated in ArcGIS 9.3 (ESRI, Red-lands, California) using aerial photographs, NRCS Soils layers (Langlois 1979), and information on past management (Figure 2) to ensure that the Research Areas were rel-atively homogenous. Both Research Areas were sited within the Evesboro sand with a 0–3% slope soil classification (Langlois 1979) and received identical management with annual dormant season brush-cutting since 1998 (Beattie et al. 2008). Pretreat-ment vegetation consisted of a mosaic of scrub oaks and ericaceous shrubs, such as black huckleberry, common lowbush blueberry (Vaccinium angustifolium [Ait.]), and hillside blueberry (Vaccinium pallidum [Ait.]), which attain a maximum height of 0.5–1.0 m during the growing season and are then reduced to 10–15 cm each year by dormant season brush-cutting.

The study design consisted of a randomized complete block design with an additional nested treatment; the Research Areas were designated as the overall treatment block. Research Area 1 was randomly assigned as the Unharrowed treatment block and Research Area 2 as the Harrowed treat-ment block. These treatment blocks were unreplicated due to their size and lack of available habitat (Figure 2). Harrowing took place in late autumn of 2010, using a tractor-drawn disc harrow (Model 127, Athens Plow Co., Athens, Tennessee), and annual brush-cutting continued in


both areas over the course of the study. The nested treatment consisted of 1-m2 plots located within each Research Area block that were designated as Unseeded or Seeded; the Hawth’s Tools extension in ArcGIS 9.3 randomly selected sixty 1-m2 plot locations within each Research Area, with a 5-m buffer between points to maintain plot independence. The seed treatment was administered within the plot area (defined by a 1-m2 PVC quadrat frame) for each of the Seeded plots. In summary, the full randomized complete block design consisted of the Seeded and Unseeded treatments nested within the two Research Areas (Harrowed and Unharrowed; Figure

2). Hereafter we will refer to the four possible treatment combinations as Har-rowed, Harrowed+Seeded, Unharrowed, and Seeded.

All Seeded and Harrowed+Seeded plots received the treatment of seed addition in May of 2011 and in May of 2012 (seed sowing was repeated due to an unusually dry growing season and brush-cutting in 2011, which may have interfered with recruitment). The seed addition treatment consisted of seven common native grass-land/heathland species absent from both the seed bank and the surrounding vegetation of the research site (O’Dell 2014; Omand

et al. 2014). We harvested all seed from island populations and apportioned it in the following amounts within each seeded plot: little bluestem grass: 78 g; switchgrass (Panicum virgatum L.): 15 g; sickle-leaved golden aster (Pityopsis falcata [Pursh] Nutt.): 0.8 g; toothed white-topped aster (Sericocarpus asteroides [L.] B.S.P): 0.06 g; stiff aster (Ionactis linariifolia [L.] Greene): 0.05 g; yellow wild indigo (Baptisia tinctoria [L.] R. Br. ex Ait. f.): 2 g; wild goat’s rue (Tephrosia virginiana [L.] Pers.): 2.8g. Seeds were cold stratified over the winter in a refrigerator prior to sowing; quantities reflect the total com-bined mass for both seed addition dates.

Figure 1. Map depicting Middle Moors Restoration Site (where harrowing and seed addition were undertaken), Grassland and Heathland Reference Sites, and island-wide conservation lands, Nantucket Island, Massachusetts.


Stiff aster was only added in 2011, and toothed white-topped aster only in 2012, due to lack of availability in different years.Pretreatment vegetation plot sampling occurred across all treatments 29 Septem-ber–7 October 2010. We visually estimated percent cover of the greatest spread of foliage for each individual species, as well as for each growth form category (woody, forb, graminoid, and ericaceous: huckle-berry and blueberry species combined) and substrate category (bare ground, litter, or duff) within every 1-m2 plot. Percent cover categories were visually estimated by a pair of observers to reduce individual bias using modified Daubenmire cover classes: <1%,

1–2%, 2–5%, 5–10%, 10–25%, 25–50%, 50–75%, and 75–100% (Daubenmire 1959). Initial posttreatment vegetation sam-pling occurred 10 August–20 September 2012 and the final posttreatment sampling took place 22 August–18 September 2013.

We randomly selected a subset of vegeta-tion sampling plots within existing areas of reference sandplain grasslands (n = 30) and areas of coastal heathlands (n = 30) from a prior vegetation study on Nantucket at the Head of the Plains (located in the outwash plain along the southwest shore of the island) that were sampled in 2005 and 2006 (Karberg 2014) for comparison

with the vegetation plots in the current study (Figure 1).

Analysis

Percent cover class estimates of both indi-vidual plant species and growth forms were converted to midpoints for analysis. We identified several “Species of Management Interest” that are locally common, aggres-sive and/or clonal species that have been observed to impact the success of early successional management projects: black huckleberry, common lowbush blueberry, scrub oak, dwarf chinquapin oak, Pennsyl-

Figure 2. Map of the Middle Moors Restoration Site depicting plot locations and blocking levels of design; Research Area 1 (Unharrowed) and Research Area 2 (Harrowed), each with Seeded and Unseeded nested plots, Nantucket Island, Massachusetts, 2010–2013.


vania sedge, and bristly dewberry (Rubus hispidus L.) (Karberg 2014; O’Dell 2014).

We used generalized linear mixed models (GLMM) to evaluate species and growth form cover prior to management treatments (2010) and in the final year of the study (2013) to assess the effectiveness of two levels of treatment: Harrowed or Unhar-rowed as the overall Research Area level treatment, and Seeded or Unseeded plots nested within the Research Area level. We used GLMM to analyze this data as it allowed a full analysis of the blocked and nested study design given the lack of repetition of the Harrowed treatment and

also allowed the inclusion of both binary and continuous variables while being more sensitive to unbalanced data and outliers (McCulloch and Searle 2001; Manning 2007; Bolker et al. 2008). The proportional nature of the percent cover response vari-able required a binomial error distribution and logit-link. Additionally, we included random effects of Seeded Treatment nested within Research Area (which represented the Harrowed Treatment). When necessary due to outliers and complete separation of some response variables, we applied penalized maximum likelihood priors and optimizations on the fixed effects. All GLMMs were fit in program R using either

the glmer or blgmer functions in package lme4 or blme, respectively (R Core Team 2015). GLMMs were performed for each response variable to evaluate cover before treatment (2010) and posttreatment (2013). The GLMM provides parameter estimates analyzed for significance using Wald’s z statistic (Table 1), with the Intercept in all models representing the combination of Unharrowed and Unseeded treatment categories.

Plots were sampled in a repeated design over three years but lack of repetition at the overall Research Area level reduced our ability to build analysis models that

Table 1. Fixed effects results of generalized linear mixed model (GLMM) analyses of Growth Form and Species of Management Interest percent cover in 2010 and 2013 as a result of nested treatments. Due to the design of the study, not all measured cover categories could be modeled; presented are the models that could be resolved. The intercept represents the base levels of each treatment: Unharrowed, without harrowing or seeding.

Percent cover (2013)

Treatment levels EstimatesTest statistics (Wald's |z|) Treatment levels Estimates

Test statistics (Wald's |z|)

(intercept) 2.51780 3.959*** (intercept) 6.79700 3.963***TrtLev1: Harrowed −0.5753 0.756 TrtLev1: Harrowed −2.635 2.031*TrtLev2: Seeded 1.13040 1.346 TrtLev2: Seeded −2.639 2.041*

(intercept) −4.09518 3.297*** (intercept) −1.9974 4.367***TrtLev1: Harrowed 0.03506 0.025 TrtLev1: Harrowed 2.46100 5.098***TrtLev2: Seeded 0.03505 0.025 TrtLev2: Seeded 1.51010 3.128**

(intercept) −1.51762 3.732*** (intercept) −0.9264 1.173TrtLev1: Harrowed 0.25301 0.551 TrtLev1: Harrowed −3.533 2.365*TrtLev2: Seeded 0.04704 0.103 TrtLev2: Seeded 0.11130 0.140

Carex pensylvanica (Pennsylvania sedge)(intercept) −4.09522 3.297*** (intercept) −0.9984 2.613**TrtLev1: Harrowed 0.03517 0.025 TrtLev1: Harrowed −0.8959 1.669 .TrtLev2: Seeded 0.03498 0.025 TrtLev2: Seeded −0.6297 1.204

(intercept) −6.20535 2.980** (intercept) −4.861 2.860**TrtLev1: Harrowed 1.25600 0.716 TrtLev1: Harrowed −1.421 0.794TrtLev2: Seeded 1.06300 0.611 TrtLev2: Seeded 1.01000 0.567

(intercept) −1.9509 3.837*** (intercept) −0.6205 1.743TrtLev1: Harrowed −1.1712 1.684 . TrtLev1: Harrowed 1.92160 4.442***TrtLev2: Seeded 0.40400 0.646 TrtLev2: Seeded −0.4737 1.096

(intercept) −2.1332 −4.157*** (intercept) −2.257 0.963TrtLev1: Harrowed 0.20910 0.355 TrtLev1: Harrowed −1.963 0.485TrtLev2: Seeded −0.1316 0.223 TrtLev2: Seeded −0.556 0.441

Vaccinium angustifolium (common lowbush blueberry)

. P < 0.1; * P < 0.05; ** P < 0.01; *** P < 0.001

Percent cover (2010)

Woody cover

Graminoid cover

Ericaceous cover

Gaylussacia baccata (black huckleberry)

Rubus hispidus (bristly dewberry)


included repeated measures. The dramatic initial vegetation cover change created by the Harrowed treatment itself made re-peated measures comparison of vegetation plots problematic in any statistical model. We instead chose to assess pretreatment vegetation composition across all treatment levels and relate these results to a separate analysis of posttreatment vegetation com-position across all treatment levels at the end of the study.

The success of recruitment of seeded spe-cies was analyzed using a nonparametric independent samples Kruskal–Wallis test to compare Seeded and Harrowed+Seeded treatments for each sown species (IBM 2012).

We conducted a nonmetric multidimen-sional scaling analysis (nMDS) based on a Bray–Curtis dissimilarity matrix to assess how resultant vegetation com-munities within the Research Area level and Seeded level treatments compared to existing desired target vegetation commu-nity datasets (Reference Grassland and Reference Heathland areas at the Head of the Plains) (R Core Team 2015). An analysis of similarity (ANOSIM) using the Bray–Curtis index of similarity was conducted to examine similarity in overall vegetation composition among all sites. A SIMPER (similarity percentages) analysis was then conducted to examine the average contribution of individual species to the average dissimilarity among treatments and reference communities (Harrowed, Harrowed+Seeded, Seeded, Unharrowed, Reference Grassland, and Reference Heath-land) using Primer 7 (Plymouth Routines in Multivariate Ecological Research; Clarke et al. 2015). We also evaluated species richness in each treatment over time, calculating species diversity as Shannon’s diversity index (H); H was also calculated for the Reference Grassland and Heathland (R Core Team 2015).

RESULTS

Treatments significantly influenced the percent cover of some growth forms and some individual species two years after the final treatment application (2013). The lack of repetition at the Research Area

level treatment and the nested nature of the Seeded treatment made it difficult to model the response of all growth forms and individual species to treatments. Low percent cover of particular species or growth forms (e.g., forbs), or uniform cover of growth forms (e.g., total cover), made the analysis models either unable to converge or caused them to provide unre-liable results. Results are presented below for data that could be analyzed within the constraints of the study model.

Shifts in Growth Form Cover

Prior to treatment, there were significant differences in the cover of growth forms between the Unharrowed plots and all other treatments (Tables 1 and 2). Unharrowed plots had significantly higher woody cov-er and significantly lower graminoid and ericaceous cover than other treatments, indicating an initial difference in vegetation composition.

At the end of the study in 2013, woody cover remained significantly higher at the Unharrowed plots and was significantly lower in the Harrowed and Seeded plots. The Harrowed plots contained significantly lower cover of ericaceous plants. Finally, the Seeded plots had significantly less graminoid cover, while the Harrowed+-Seeded plots contained significantly more graminoid cover (Tables 1 and 2).

Species of Management Interest

Of the Species of Management Interest that the model was able to analyze, the Unharrowed plots contained significantly lower cover of Pennsylvania sedge, black huckleberry, bristly dewberry, and common lowbush blueberry prior to the application of any treatment in 2010.

In the final year of monitoring, it was possible to model some significant dif-ferences for the Species of Management Interest among treatments. Pennsylvania sedge remained significantly lower in the Unharrowed plots and cover was lower in the Harrowed plots, but not significantly (P < 0.1). Black huckleberry was significantly lower in the Unharrowed plots and bristly

dewberry cover was significantly higher in the Harrowed plots. The response of all other Species of Management Interest could not be significantly modeled among treatments.

Response of Sown Species

Of the sown species, only white-topped toothed aster (three Harrowed plots) and yellow wild indigo (one Harrowed plot) appeared in unseeded plots in 2013. Switchgrass did not establish successfully in any plots. With the exception of white-topped toothed aster, all of the sown spe-cies established with significantly higher percent cover in the Harrowed+Seeded treatment than in the Seeded treatment: little bluestem (H = 45.645, df = 1, P < 0.001), yellow wild indigo (H = 9.771, df = 1, P = 0.001), goat’s rue (H = 26.101, df = 1, P < 0.001), sickle-leaf golden aster (H = 9.710, df = 1, P = 0.002), and stiff aster (H = 5.720, df = 1, P = 0.017) (independent samples Kruskal–Wallis tests). White-topped toothed aster cover in 2013 did not differ significantly among the two seeding treatments (H = 3.331, df = 1, P = 0.068; independent samples Kruskal–Wallis tests). Frequency of establishment was highest for little bluestem, which established in all Harrowed+Seeded plots, compared to only one-third of Seeded plots.

Comparison to Reference Vegetation Communities

The nMDS ordination strongly differ-entiated the treatments in this study and the Reference Grassland and Reference Heathland habitats based on species composition, producing two convergent solutions with a stress value of 8.364749 × 10−5 (R = 0.5417, P < 0.001; Figure 3). Species composition for Unharrowed and Seeded treatments was very similar, and both were strongly differentiated from Har-rowed and Harrowed+Seeded treatments. The Harrowed+Seeded treatment was most similar to the Reference Grassland. ANOSIM confirmed that all treatments and reference sites were significantly differentiated from each other except the Unharrowed and Seeded treatments (r = 0.558, P = 0.001). The SIMPER analy-


Tab

le 2

. Mea

n pe

rcen

t co

ver

and

stan

dard

err

ors

(SE

) fo

r G

row

th F

orm

, Spe

cies

of

Man

agem

ent

Inte

rest

, and

See

ded

Spec

ies

by t

reat

men

t le

vels

(20

10–2

013)

.

Perc

ent c

over

ca

tego

ries

Gro

wth

form

20

1020

1320

1020

1320

1020

1320

1020

13

Tota

l cov

er82

.50

± 2.

2187

.50

± 0.

0085

.71

± 1.

2487

.50

± 0.

0081

.67

± 1.

9687

.50

± 0.

0085

.00

± 1.

3987

.50

± 0.

00

Woo

dy

75.3

3 ±

3.75

83.3

3 ±

1.73

81.2

5 ±

2.77

65.1

8 ±

3.48

76.6

7 ±

3.10

85.8

3 ±

1.16

82.5

0 ±

1.85

87.5

0 ±

0.00

Er

icac

eous

32.0

7 ±

4.60

5.23

± 0

.90

29.7

1 ±

4.72

2.98

± 0

.50

26.3

3 ±

4.18

38.8

7 ±

5.60

23.9

7 ±

4.95

39.7

7 ±

5.08

G

ram

inoi

d8.

77 ±

2.6

654

.50

± 4.

764.

71 ±

1.9

183

.04

± 1.

844.

45 ±

1.4

528

.67

± 5.

736.

50 ±

2.4

431

.42

± 5.

27

Forb

0.10

± 0

.04

5.68

± 1

.90

0.05

± 0

.30

5.34

± 1

.09

0.03

± 0

.02

0.43

± 0

.27

0.15

± 0

.12

0.72

± 0

.30

B

are

grou

nd0.

03 ±

0.0

20.

80 ±

0.2

60.

07 ±

0.0

60.

91 ±

0.2

00.

13 ±

0.0

40.

70 ±

0.2

00.

13 ±

0.0

40.

25 ±

0.0

7Sp

ecie

s of M

anag

emen

t Int

eres

t

Car

ex p

ensy

lvan

ica

7.85

± 2

.88

27.9

3 ±

4.99

2.43

± 1

.38

9.91

± 2

.68

2.67

± 1

.37

20.6

5 ±

5.43

5.52

± 2

.44

18.7

0 ±

5.25

G

aylu

ssac

ia b

acca

ta4.

70 ±

1.4

12.

13 ±

0.6

58.

54 ±

3.3

51.

54 ±

0.4

01.

73 ±

0.6

26.

95 ±

1.8

52.

32 ±

0.6

77.

73 ±

2.4

2

Que

rcus

ilic

ifolia

13.4

8 ±

3.11

2.27

± 1

.26

21.9

6 ±

5.04

2.88

± 1

.45

24.7

8 ±

3.56

37.6

5 ±

5.01

17.6

7 ±

2.74

40.3

0 ±

4.77

Q

uerc

us p

rino

ides

11.2

2 ±

3.66

1.28

± 0

.81

12.5

5 ±

3.89

0.54

± 0

.31

11.1

5 ±

3.85

18.9

8 ±

4.21

15.0

3 ±

3.66

12.9

3 ±

3.18

Ru

bus h

ispi

dus

13.7

2 ±

3.49

75.3

3 ±

3.75

5.71

± 2

.34

55.5

4 ±

5.16

16.3

2 ±

4.67

28.8

5 ±

5.94

28.5

7 ±

5.24

36.5

5 ±

6.37

Va

ccin

ium

angu

stifo

lium

18.9

7 ±

4.30

2.53

± 0

.65

14.4

1 ±

4.00

1.52

± 0

.31

17.1

2 ±

3.94

23.6

7 ±

5.8

211

.02

± 4.

1317

.32

± 4.

60Se

eded

spec

ies

Ba

ptis

ia ti

ncto

ria

0.00

± 0

.00

0.02

± 0

.02

0.00

± 0

.00

0.14

± 0

.04

0.00

± 0

.00

0.00

± 0

.00

0.00

± 0

.00

0.00

± 0

.00

Io

nact

is li

nari

ifolia

0.00

± 0

.00

0.00

± 0

.00

0.00

± 0

.00

0.34

± 0

.17

0.00

± 0

.00

0.00

± 0

.00

0.00

± 0

.00

0.02

± 0

.02

Pi

tyop

sis f

alca

ta0.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

00.

86 ±

0.3

40.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

0

Schi

zach

yriu

m

scop

ariu

m0.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

062

.50

± 3.

150.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

01.

37 ±

0.6

3

Seri

coca

rpus

as

tero

ides

0.00

± 0

.00

0.05

± 0

.03

0.00

± 0

.00

0.05

± 0

.03

0.00

± 0

.00

0.00

± 0

.00

0.00

± 0

.00

0.00

± 0

.00

Te

phro

sia

virg

inia

na0.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

00.

38 ±

0.0

60.

00 ±

0.0

00.

00 ±

0.0

00.

00 ±

0.0

00.

17 ±

0.1

7

Har

row

edH

arro

wed

+See

ded

Unh

arro

wed

Seed

ed


sis revealed the individual plant species responsible for dissimilarity among the treatments and reference communities (Table 3). The Unharrowed and Seeded treatments showed a lower dissimilarity to one another (61.23%) as predicted by the ANOSIM. All other treatments were significantly dissimilar from each other but varied in the percent average dissimi-larity and the species responsible for that dissimilarity (Table 3). The Harrowed and Harrowed+Seeded treatments were 57.38% dissimilar to one another, largely due to the presence of little bluestem in the seeded plots. Comparing the treatment and reference sites, the lowest dissimilarity occurred between the Harrowed+Seeded and Reference Grassland (69.56% average dissimilarity); differences were primarily due to higher amounts of bristly dewberry and little bluestem and lower amounts of Pennsylvania sedge and black huckleber-

ry in Harrowed+Seeded compared to the Reference Grassland (Table 3).

Species Richness and Diversity Response

Species diversity calculations indicate that pretreatment species diversity (Shannon H) was very similar among all treatments, ranging from 2.944 to 3.045 (Figure 4) but lower than that calculated for the Refer-ence Grassland and Reference Heathland. Both Unharrowed and Seeded treatments exhibited a minor increase in species di-versity by the conclusion of the study. In contrast, species diversity in both the Har-rowed and Harrowed+Seeded treatments increased substantially (Figure 4). In 2012 and 2013, calculated species diversity in both the Harrowed and Harrowed+Seeded treatments was higher than the baseline cal-culated species diversity of the Reference

Grassland (H = 3.5835) and the Reference Heathland (H = 3.4965) (Figure 4).

Rare Species Emergence

We recorded the emergence and flowering of four state-listed rare plant species from the soil seed bank that were not documented at the site prior to treatment; all are con-sidered early successional associates or obligates of rare New England sandplain vegetation communities. Regionally rare sandplain blue-eyed grass (Sisyrinchi-um fuscatum Bickn.), listed as “Special Concern” (S3) in Massachusetts (MA NH&ESP, n.d.), emerged in the greatest numbers (736 individuals) in 2012; that year we also documented a small popu-lation (16 individuals) of Massachusetts Endangered (S2) papillose nutsedge (Scle-ria pauciflora Muhl. ex. Willd.). In 2013, we documented six flowering individuals

Figure 3. The nMDS ordination depicting the species composition of treatments compared to the Reference Grassland and Reference Heathland sites at the Head of the Plains. Individual species key in defining the vegetation communities are listed.


Table 3. Table depicting SIMPER Results, showing main species contributing to dissimilarity between treatments and Reference Grassland and Heath-land sites at Head of the Plains.

SpeciesContribution

(%)Cumulative

(%)Avg.

dissimilarity(a) Avg. dissimilarity = 73.26 Unharrowed Harrowed Rubus hispidus 29.05 29.05 25.92 74.75 21.28 Quercus ilicifolia 16.35 45.40 33.48 2.27 11.98 Carex pensylvanica 14.66 60.06 20.77 24.43 10.74 Vaccinium angustifolium 11.37 71.44 23.08 2.53 8.33(b) Avg. dissimilarity = 80.80 Unharrowed Harrowed+Seeded Schizachyrium scoparium 28.04 28.04 0.05 62.50 22.66 Rubus hispidus 18.38 46.42 25.92 55.54 14.85 Quercus ilicifolia 13.83 60.24 33.48 2.88 11.17 Carex pensylvanica 9.76 70.01 20.77 9.91 7.89(c) Avg. dissimilarity = 57.38 Harrowed Harrowed+Seeded Schizachyrium scoparium 39.76 39.76 0.00 62.50 22.81 Rubus hispidus 20.62 60.38 74.75 55.54 11.84 Carex pensylvanica 13.74 74.13 24.43 9.91 7.89(d) Avg. dissimilarity = 61.23 Unharrowed Seeded Rubus hispidus 20.72 20.72 25.92 34.58 12.68 Quercus ilicifolia 18.92 39.63 33.48 40.30 11.58 Carex pensylvanica 16.04 55.68 20.77 18.58 9.82 Vaccinium angustifolium 15.75 71.43 23.08 17.32 9.65(e) Avg. dissimilarity = 69.78 Harrowed Seeded Rubus hispidus 27.07 27.07 74.75 34.58 18.89 Quercus ilicifolia 20.92 47.99 2.27 40.30 14.60 Carex pensylvanica 14.31 62.30 24.43 18.58 9.99 Vaccinium angustifolium 8.07 70.37 2.53 17.32 5.63(f) Avg. dissimilarity = 77.35 Harrowed+Seeded Seeded Schizachyrium scoparium 28.03 28.03 62.50 1.37 21.68 Rubus hispidus 18.15 46.18 55.54 34.58 14.04 Quercus ilicifolia 17.49 63.67 2.88 40.30 13.53 Carex pensylvanica 8.89 72.56 9.91 18.58 6.88(g) Avg. dissimilarity = 86.73 Unharrowed Ref. Grassland Carex pensylvanica 15.73 15.73 20.77 37.79 13.64 Quercus ilicifolia 14.15 29.88 33.48 0.00 12.27 Schizachyarium scoparium 14.08 43.96 0.05 31.34 12.21 Rubus hispidus 11.78 55.74 25.92 10.21 10.22 Vaccinium angustifolium 10.56 66.30 23.08 3.93 9.16 Quercus prinoides 8.02 74.32 18.40 1.61 6.96(h) Avg. dissimilarity = 82.72 Harrowed Ref. Grassland Rubus hispidus 31.94 31.94 74.75 10.21 26.41 Carex pensylvanica 15.45 47.38 24.43 37.79 12.78 Schizachyrium scoparium 14.84 62.22 0.00 31.34 12.28 Gaylussacia baccata 5.18 67.41 2.13 10.34 4.29 Dichanthelium spp. 5.06 72.47 10.78 0.32 4.18(i) Avg. dissimilarity = 69.56 Harrowed+Seeded Ref. Grassland Rubus hispidus 26.04 26.04 55.54 10.21 18.11 Schizachyrium scoparium 20.56 46.60 62.50 31.34 14.30 Carex pensylvanica 17.18 63.78 9.91 37.79 11.95 Gaylussacia baccata 5.62 69.40 1.54 10.34 3.91 Arctostaphylos uva-ursi 4.34 73.74 0.00 9.32 3.02

Continued

Avg. abundance

SpeciesContribution

(%)Cumulative

(%)Avg.

dissimilarity(a) Avg. dissimilarity = 73.26 Unharrowed Harrowed Rubus hispidus 29.05 29.05 25.92 74.75 21.28 Quercus ilicifolia 16.35 45.40 33.48 2.27 11.98 Carex pensylvanica 14.66 60.06 20.77 24.43 10.74 Vaccinium angustifolium 11.37 71.44 23.08 2.53 8.33(b) Avg. dissimilarity = 80.80 Unharrowed Harrowed+Seeded Schizachyrium scoparium 28.04 28.04 0.05 62.50 22.66 Rubus hispidus 18.38 46.42 25.92 55.54 14.85 Quercus ilicifolia 13.83 60.24 33.48 2.88 11.17 Carex pensylvanica 9.76 70.01 20.77 9.91 7.89(c) Avg. dissimilarity = 57.38 Harrowed Harrowed+Seeded Schizachyrium scoparium 39.76 39.76 0.00 62.50 22.81 Rubus hispidus 20.62 60.38 74.75 55.54 11.84 Carex pensylvanica 13.74 74.13 24.43 9.91 7.89(d) Avg. dissimilarity = 61.23 Unharrowed Seeded Rubus hispidus 20.72 20.72 25.92 34.58 12.68 Quercus ilicifolia 18.92 39.63 33.48 40.30 11.58 Carex pensylvanica 16.04 55.68 20.77 18.58 9.82 Vaccinium angustifolium 15.75 71.43 23.08 17.32 9.65(e) Avg. dissimilarity = 69.78 Harrowed Seeded Rubus hispidus 27.07 27.07 74.75 34.58 18.89 Quercus ilicifolia 20.92 47.99 2.27 40.30 14.60 Carex pensylvanica 14.31 62.30 24.43 18.58 9.99 Vaccinium angustifolium 8.07 70.37 2.53 17.32 5.63(f) Avg. dissimilarity = 77.35 Harrowed+Seeded Seeded Schizachyrium scoparium 28.03 28.03 62.50 1.37 21.68 Rubus hispidus 18.15 46.18 55.54 34.58 14.04 Quercus ilicifolia 17.49 63.67 2.88 40.30 13.53 Carex pensylvanica 8.89 72.56 9.91 18.58 6.88(g) Avg. dissimilarity = 86.73 Unharrowed Ref. Grassland Carex pensylvanica 15.73 15.73 20.77 37.79 13.64 Quercus ilicifolia 14.15 29.88 33.48 0.00 12.27 Schizachyarium scoparium 14.08 43.96 0.05 31.34 12.21 Rubus hispidus 11.78 55.74 25.92 10.21 10.22 Vaccinium angustifolium 10.56 66.30 23.08 3.93 9.16 Quercus prinoides 8.02 74.32 18.40 1.61 6.96(h) Avg. dissimilarity = 82.72 Harrowed Ref. Grassland Rubus hispidus 31.94 31.94 74.75 10.21 26.41 Carex pensylvanica 15.45 47.38 24.43 37.79 12.78 Schizachyrium scoparium 14.84 62.22 0.00 31.34 12.28 Gaylussacia baccata 5.18 67.41 2.13 10.34 4.29 Dichanthelium spp. 5.06 72.47 10.78 0.32 4.18(i) Avg. dissimilarity = 69.56 Harrowed+Seeded Ref. Grassland Rubus hispidus 26.04 26.04 55.54 10.21 18.11 Schizachyrium scoparium 20.56 46.60 62.50 31.34 14.30 Carex pensylvanica 17.18 63.78 9.91 37.79 11.95 Gaylussacia baccata 5.62 69.40 1.54 10.34 3.91 Arctostaphylos uva-ursi 4.34 73.74 0.00 9.32 3.02

Continued

Avg. abundance


of Massachusetts Endangered (S1) tall nutsedge (Scleria triglomerata Michx; MA NH&ESP, n.d.). During the study we also observed several flowering individuals of tall bush-clover (Lespedeza stuevei Nutt.), which was in 2017 listed as “Historic” in Massachusetts and is now under review by the Massachusetts Natural Heritage and Endangered Species Program due to the

recent documentation of a handful of small populations in the state (R. Wernerehl, Mas-sachusetts State Botanist, pers. comm.). In all cases these plants only recruited within areas that received harrowing.

Nonnative Species

We observed nonnative fescues (Festuca

spp. L.), bentgrasses (Agrostis spp. L.), velvetgrass (Holcus lanatus L.), and tall rye grass (Schedonorus arundinaceus [Schreb.] Dumort.), primarily in the harrowed area. Fescues and bentgrasses are naturalized in the existing Reference Grassland and Heathland habitats, while tall rye grass and velvetgrass are common in pasture and road edge habitats. Cover of these species

Table 3. (Cont’d.)

SpeciesContribution

(%)Cumulative

(%)Avg.

dissimilarity(j) Avg. dissimilarity = 87.23 Seeded Ref. Grassland Quercus ilicifolia 17.38 17.38 40.30 0.00 15.16 Carex pensylvanica 15.27 32.65 18.58 37.79 13.32 Rubus hispidus 13.83 46.49 34.58 10.21 12.07 Schizachyrium scoparium 13.40 59.88 1.37 31.34 11.68 Vaccinium angustifolium 7.82 67.70 17.32 3.93 6.82 Gaylussacia baccata 6.12 73.82 7.73 10.34 5.34(k) Avg. dissimilarity = 91.58 Unharrowed Ref. Heathland Gaylussacia baccata 16.64 16.64 6.90 42.55 15.24 Quercus ilicifolia 13.61 30.25 33.48 0.00 12.46 Arctostaphylos uva-ursi 10.72 40.96 0.00 27.95 9.81 Rubus hispidus 10.34 51.30 25.92 2.06 9.47 Carex pensylvanica 9.43 60.73 20.77 7.31 8.64 Vaccinium angustifolium 9.41 70.14 23.08 1.73 8.62(l) Avg. dissimilarity = 93.81 Harrowed Ref. Heathland Rubus hispidus 30.67 30.67 74.75 2.06 28.78 Gaylussacia baccata 17.11 47.78 2.13 42.55 16.05 Arctostaphylos uva-ursi 10.53 58.31 0.02 27.95 9.88 Carex pensylvanica 9.62 67.93 24.43 7.31 9.02 Schizachyrium scoparium 8.43 76.36 0.00 22.29 7.90(m) Avg. dissimilarity = 82.51 Harrowed+Seeded Ref. Heathland Rubus hispidus 23.22 23.22 55.54 2.06 19.16 Schizachyrium scoparium 21.27 44.49 62.50 22.29 17.55 Gaylussacia baccata 18.34 62.83 1.54 42.55 15.13 Arctostaphylos uva-ursi 11.28 74.11 0.00 27.95 9.31(n) Avg. dissimilarity = 90.96 Seeded Ref. Heathland Quercus ilicifolia 16.97 16.97 40.30 0.00 15.44 Gaylussacia baccata 16.28 33.26 7.73 42.55 14.81 Rubus hispidus 12.99 46.24 34.58 2.06 11.81 Arctostaphylos uva-ursi 10.52 56.76 0.00 27.95 9.57 Schizachyrium scoparium 8.33 65.09 1.37 22.29 7.58 Carex pensylvanica 7.98 73.07 18.58 7.31 7.25(o) Avg. dissimilarity = 78.64 Ref. Grassland Ref. Heathland Gaylussacia baccata 19.69 19.69 10.34 42.55 15.48 Carex pensylvanica 17.69 37.37 37.79 7.31 13.91 Schizachyrium scoparium 15.99 53.36 31.34 22.29 12.57 Arctostaphylos uva-ursi 13.87 67.23 9.32 27.95 10.91 Morella caroliniensis 5.65 72.89 8.25 5.47 4.45

Avg. abundance


was under 5% in monitoring plots where they were present.

DISCUSSION

The results of this research indicate that large-scale soil disturbance provided by disc harrowing can be an effective method for reducing cover of some woody species while encouraging recruitment of desired sandplain grassland and heathland species and increasing early successional species diversity. Soil disturbance initially led to a dramatic reduction of woody species cover. Woody species cover overall remained significantly reduced by soil disturbance at the end of the study, with the excep-tion of vigorous regrowth of one species, bristly dewberry, a desirable low-growing component of grasslands and heathlands. Additionally, disc harrowing broke up root systems of clonal woody shrubs and the sod-forming native graminoid, Pennsyl-vania sedge, which provided mineral soil germination sites for sown little bluestem and early successional grasses and forbs contained in the seed bank. Plots lacking

soil disturbance or seed addition had lower graminoid cover in the final year of the study than other treatments.

While harrowing alone was effective in boosting species diversity and graminoid cover, key grassland and heathland species such as little bluestem did not recruit spon-taneously from the seed bank. Harrowing followed by seed addition shifted vegeta-tion composition more strongly toward that of the Reference Grassland than harrowing alone, mainly by recruiting little bluestem, which was not present at the site prior to treatments. Rapid growth of little bluestem in Harrowed+Seeded plots appeared to inhibit Pennsylvania sedge recolonization, but may have also limited the establishment success of seeded forbs. Plots without soil disturbance changed little in diversity over time, regardless of seed addition. This mirrors conclusions from earlier research (Beattie et al. 2008) that the litter/duff surface layer and competition with woody vegetation in brush-cut shrublands may inhibit germination and establishment of many forb and graminoid seedlings, and that key sandplain species are seed limited

(Lezberg et al. 2006; Wheeler et al. 2015). There have been similar findings of seed and microsite limitation in grassland res-toration efforts in other regions (Laughlin 2003; Clark et al. 2007), suggesting that combined soil disturbance and seed addi-tion may be beneficial in other grassland restoration efforts.

Our observations support other research-ers’ conclusions that soil disturbance has dramatic positive impacts on grassland restoration when used wisely (Hobbs and Huenneke 1992; Hoffman and Isselstein 2004; Edwards et al. 2007). Key benefits of soil disturbance at our site appeared to include mixing the top layer of accumulated duff and leaf litter into the soil profile and creating large surface expanses of mineral soil. The unexpectedly high establishment of a variety of common early successional graminoids and forbs, as well as several rare species, suggests that disc harrowing may have brought long-buried seeds to the surface, while providing an appropriate substrate for germination with temporarily low competition from aggressive woody species such as scrub oak and blueberry, as well as clonal Pennsylvania sedge.

However, seed addition proved crucial for establishing desired native little bluestem and forbs that were absent from the seed bank. Combined harrowing and seeding enhanced the establishment success of all of the sown species except white-topped toothed aster. As with little bluestem in our study, establishment of more aggressive grasses has been found to limit recruitment of forbs when sown together in experi-mental restorations (Dickson and Busby 2009). Seeding with a lower ratio of little bluestem seed or patch-sowing forbs at higher seeding rates separately from little bluestem are methods that we believe may allow managers to maximize estab-lishment of seeded species. Patch-sowing within an unseeded matrix may provide multiple benefits: enhancing vegetation heterogeneity across a site by allowing desirable early successional native species in the seed bank to recruit in some areas, meanwhile making limited seed resources spread farther.

Key woody species of sandplain grass-

Figure 4. Changes in Species Diversity (Shannon H) in treatments over time (2010–2013); Reference Grassland and Reference Heathland diversity levels are presented for comparison purposes (data col-lected 2005 and 2006 at the Head of the Plains).


lands, and especially coastal heathlands, were lacking or underrepresented within the restoration treatments. These include bayberry (Morella caroliniensis [Mill.] Small), bearberry (Arctostaphylos uva-ur-si [L.] Spreng.), and pine barren false heather (Hudsonia ericoides L.) (Sorrie and Dunwiddie 1996; Swain and Kearsley 2001). The absence or scarcity of these species in our sampling plots contributed to vegetation composition differences between the treatments and the Reference Heathland and Grassland in our ordination. It was beyond the scope of this study to determine whether these species are able to rapidly colonize a restoration site unaided in a short time frame (plausible for bird-dispersed species in particular), or if seed addition would be necessary to more closely approximate the diversity of existing grassland and heathland.

The recruitment of four rare early succes-sional sandplain species from the seed bank following harrowing was considered an unexpected benefit. These plants are clas-sified in Massachusetts as characteristic but infrequent components of intact sandplain communities and are known to be strongly associated with soil disturbance (Swain and Kearsley 2001; Clark 2004; Zaremba 2004; Farnsworth 2006). That none of these species were previously observed in the standing vegetation or in the nearby seed bank study site highlights the value (but also the unpredictability) of soil seed banks in conserving rare plant species.

Rare and spatially limited species may be underrepresented or missed entirely in seed bank studies and random vegetation sam-pling designs (Baskin and Baskin 1989). In contrast to the smaller treatment area (O’Dell 2014) and seed bank sample size (Omand et al. 2014) in previous research, larger-scale harrowing in the current study likely increased the probability of observing these rare plants. Establishment of these species may have increased the resemblance of our restoration area to desired grassland and heathland compo-sition at the landscape level. However, due to their rarity in both the reference and treatment sites, this effect was not detectable in our ordination. Comparison of a whole-site species list may be a useful

tool for managers assessing a restored area with respect to a reference community site, although spatial unpredictability and disparities in the size of restoration and reference sites could make quantitative comparisons challenging.

The low proportion of aggressive non-native or invasive species in surrounding vegetation and in the soil seed bank likely enhanced our results. Other researchers have been plagued by extensive germi-nation of invasive species from the seed bank, which hampers restoration efforts (Wilson et al. 2004; Wheeler et al. 2015). Prior evaluation of the soil seed bank and pretreatment vegetation at the restoration site suggested that there would be little germination of troublesome species with large-scale soil disturbance (Omand et al. 2014). Many of the nonnative species observed in the treatment area are also common in the Reference Grassland and Heathland sites, while others are associ-ated with old field and road-edge habitats and likely appeared due to the close proximity of a main road and bike path. Soil disturbance has been correlated with higher levels of invasive species in other restoration studies (Hobbs and Huenneke 1992; Von Holle and Motzkin 2007), and we caution against employing this tech-nique without prior assessment of invasive seed sources at restoration sites. Likewise, many grassland restorations take place on former agricultural lands where soils and their biota have been dramatically altered by amendments. Soils at our study site had remained acidic, sandy, and low in nutrients for many years, which likely enhanced the establishment of grassland species while minimizing the success of pasture weeds such as velvetgrass and tall rye grass.

CONCLUSION

Further declines in sandplain grassland and coastal heathland are projected as a result of ecological succession, develop-ment, and coastal erosion. Maintenance and expansion of these rare communities remains a high priority in order to retain many vulnerable species during a period of rapid ecological change. While scrub oak and pitch pine barrens are clearly ecologically important plant communities

and merit protection themselves (Goldstein 1997; Wagner et al. 2003), the comparative rarity and continued decline of grassland and heathland communities underscores their vulnerability and the need for highly effective management techniques in order to retain them in the coastal landscape. The results of this study demonstrate that combining soil disturbance and seed ad-dition can be a useful step in converting areas of overgrown native shrubland to higher diversity grassland and heathland. These methods may be applicable for early successional restoration in other regions overgrown with native woody shrubs and vines, which have not been extensively invaded by nonnative species or experi-enced edaphic changes. Following shru-bland treatment with harrowing and seed addition, woody species management will be required such as prescribed fire and/or brush-cutting to maintain structural and species diversity over time.

ACKNOWLEDGMENTS

We would like to thank the seasonal field crew members who assisted with field work and data entry for this project: Gwen Kozlowski, Jin Hong, John Krapek, Nora Harkness Kelly, Emily West, Amanda Swaller, Tyler Refsland, Mara Plato, Iris Clearwater, and Cyndi Park.

Kelly Omand (MS Antioch University New England) is an Ecologist/Field Su-pervisor for the Science and Stewardship Department of the Nantucket Conservation Foundation. Her research interests include restoration of native plant communities, invasive plant ecology and management, and island ecology.

Jennifer Karberg (PhD Michigan Tech-nological University) is the Research Su-pervisor for the Science and Stewardship Department of the Nantucket Conservation Foundation. Her research interests include ecosystem restoration (particularly grass-lands and wetlands), rare plant ecology and carnivorous plants.

Danielle O’Dell (MS University of Arizona) is an Ecologist/Field Supervisor for the Nantucket Conservation Foundation. Her


research interests include wildlife ecology, monitoring and protection of rare species, and the impacts of land use and manage-ment on wildlife populations.

Karen Beattie (MS University of Massa-chusetts at Amherst) is the Science and Stewardship Department Manager for the Nantucket Conservation Foundation. Her research and management work focus on property conservation management planning, fire ecology, restoration and maintenance of early successional habitats, and rare wildlife ecology.

LITERATURE CITED

Barbour, H., T. Simmons, P. Swain, and H. Woolsey. 1999. Our Irreplaceable Heritage: Protecting Biodiversity in Massachusetts. Natural Heritage and Endangered Species Program, Massachusetts Division of Fish and Wildlife, and the Massachusetts Chapter of The Nature Conservancy, Boston, MA.

Baskin, J.M., and C.C. Baskin. 1989. Physi-ology of dormancy and germination. Pp. 53–66 in M.A. Leck, V.T. Parker, and R.L. Simpson, eds., Ecology of Soil Seed Banks. Academic Press, San Diego, CA.

Beattie, K.C., E.M. Steinauer, and R.S. Free-man. 2008. The effectiveness of mechanical brushcutting as a sandplain grassland and heathland restoration tool in a scrub oak/huckleberry shrubland on Nantucket Island. Unpublished document, Nantucket Conser-vation Foundation, Nantucket, MA.

Bernd-Cohen, T., and M. Gordon. 1999. State coastal program effectiveness in protecting natural beaches, dunes, bluffs, and rocky shores. Coastal Management 27:187-217.

[BioMap2] Natural Heritage & Endangered Species Program. 2012. Town Report for Nantucket. BioMap2: Conserving the Bio-diversity of Massachusetts in a Changing World. Natural Heritage & Endangered Species Program, Massachusetts Division of Fisheries and Wildlife, Westborough, MA. Accessed 20 January 2017 from <http://maps.massgis.state.ma.us/dfg/biomap/pdf/town_core/Nantucket.pdf>.

Bolker, B.M., M.E. Brooks, C.J. Clark, S.W. Geange, J.R. Poulsen, M.H.H. Stevens, and J.S. White. 2008. Generalized linear mixed models: A practical guide for ecology and evolution. Trends in Ecology and Evolution 24:127-135.

Briggs, J.M., A.K. Knapp, J.M. Blair, J.L. Heisler, G.A. Hoch, M.S. Lett, and J.K. McCarron. 2005. An ecosystem in transition:

Causes and consequences of the conversion of mesic grassland to shrubland. BioScience 55:243-254.

Carlson, L., M. Babione, P.J. Godfrey, and A. Fowler. 1991. Ecological survey of heathlands in Cape Cod National Seashore, MA. Technical Report NPS/NAROSS/NRTR-92/04, Prepared for the National Park Service. Botany Department, University of Massachusetts, Amherst.

Clark, C.J., J.R. Poulsen, D.J. Levey, and C.W. Osenberg. 2007. Are plant populations seed limited? A critique and meta-analysis of seed addition experiments. American Naturalist 170:128-142.

Clark, F.H. 2004. Scleria triglomerata Michx. (Tall Nutrush, Whip Nutrush) Conservation and Research Plan for New England. New England Wild Flower Society, Framingham, MA.

Clarke, K.R., and R.N. Gorley. 2015. PRIM-ER v7: User Manual/Tutorial. PRIMER-E, Plymouth, MA.

Daubenmire, R.F. 1959. A canopy-cover method of vegetational analysis. Northwest Science 33:43-46.

Dickson, T.L., and W.H. Busby. 2009. Forb spe-cies establishment increased with decreasing grass seedling density and with increased forb seeding density in a northeast Kansas, U.S.A. experimental prairie restoration. Restoration Ecology 17:597-605.

Dunwiddie, P.W. 1989. Forest and heath: The shaping of the vegetation of Nantucket Is-land. Journal of Forest History 33:126-133.

Dunwiddie, P.W. 1998. Ecological management of sandplain grasslands and coastal heath-lands in southeastern Massachusetts. Pp. 83–93 in T.L. Pruden and L.A. Brennan, eds., Fire in Ecosystem Management: Shifting the Paradigm from Suppression to Prescription. Proceedings, 20th Tall Timbers Fire Ecology Conference, Tall Timbers Research Station, Tallahassee, FL.

Dunwiddie, P.W., and C. Caljouw. 1990. Prescribed burning and mowing of coastal heathlands and grasslands in Massachu-setts. Pp. 271–275 in R.S. Mitchell, C.J. Sheviak, and D.J. Leopold, eds., Ecosystem Management: Rare Species and Significant Habitats. Bulletin 471, New York State Museum, Albany.

Dunwiddie, P.W., R.E. Zaremba, and K.A. Harper. 1996. A classification of coastal heathlands and sandplain grasslands in Massachusetts. Rhodora 98:117-145.

D’Odorico, P., G.S. Okin, and B.T. Bestelmey-er. 2012. A synthetic review of feedbacks and drivers of shrub encroachment in arid grasslands. Ecohydrology 5:520-530.

Dzwonko, Z., and S. Loster. 2008. Changes in plant species composition in abandoned and restored limestone grassland – The effects of tree and shrub cutting. Acta Societatis Botanicorum Poloniae 77:67-75.

Edwards, A.R., S.R. Mortimer, C.S. Lawson, D.B. Westbury, S.J. Harris, B.A. Woodcock, and V.A. Brown. 2007. Hay strewing, brush harvesting of seed and soil disturbance as tools for the enhancement of botanical diver-sity in grasslands. Biological Conservation 134:372-382.

Farnsworth, E.J. 2006. Plant life history traits of rare versus frequent plant taxa of sandplains: Implications for research and management trials. Biological Conservation 136:44-52.

Foster, D.R., and G. Motzkin. 2003. Interpret-ing and conserving the openland habitats of coastal New England: Insights from landscape history. Forest Ecology and Management 185:127-150.

Godfrey, P.J., and P. Alpert. 1985. Racing to save the coastal heaths. The Nature Conservancy News 35:11-13.

Goldstein, P.Z. 1997. Lepidopteran assemblages and the management of sandplain commu-nities on Martha’s Vineyard. Pp. 217–236 in P.D. Vickery and P.W. Dunwiddie, eds., Grasslands of Northeastern North America: Ecology and Conservation of Native and Agricultural Landscapes. Massachusetts Audubon Society, Lincoln.

Griffiths, M.E. 2006. Salt spray and edaphic fac-tors maintain dwarf stature and community composition in coastal sandplain heathlands. Plant Ecology 186:69-86.

Haines, A. 2011. Flora Novae Angliae. Yale University Press, New Haven, CT.

Hobbs, R.J., and L.F. Huenneke. 1992. Distur-bance, diversity, and invasion: Implications for conservation. Conservation Biology 6:324-337.

Hofmann, M., and J. Isselstein. 2004. Seedling recruitment on agriculturally improved mesic grassland: The influence of distur-bance and management schemes. Applied Vegetation Science 7:193-200.

IBM Corp. 2012. IBM SPSS Statistics for Windows, Version 21.0. IBM Corp., Ar-monk, NY.

Karberg, J.M. 2014. Prescribed fire management in sandplain grasslands and heathlands: Impacts of burn seasonality and intensity on vegetation composition, Head of the Plains, Nantucket MA. Unpublished docu-ment, Nantucket Conservation Foundation, Nantucket, MA.

Lang, N.L., and C.B. Halpern. 2007. The soil seed bank of a montane meadow: Conse-quences of conifer encroachment and im-plications for restoration. Canadian Journal of Botany 85:557-569.


Langlois, K.H., Jr. 1979. Soil Survey of Nan-tucket County, Massachusetts. USDA Soil Conservation Service in Cooperation with Massachusetts Agricultural Experiment Station, Washington, DC.

Laughlin, D.C. 2003. Lack of native propagules in a Pennsylvania, USA, limestone prairie seed bank: Futile hopes for a role in eco-logical restoration. Natural Areas Journal 23:158-164.

Leahy, C.W. 1993. Eden’s End: The Case for Ecological Protection in Massachusetts. Massachusetts Audubon Society, Lincoln, MA.

Lezberg, A.L., K.C. Buresch, C. Neill, and T. Chase. 2006. Mechanical land clearing to promote establishment of coastal sandplain grassland and shrubland communities. Res-toration Ecology 14:220-232.

Lunt, I.D. 1997. Germinable soil seed banks of anthropogenic native grasslands and grassy forest remnants in temperate south-eastern Australia. Plant Ecology 130:21-34.

Luzuriaga, A.L., A. Escudero, J.M. Olano, and J. Loidi. 2005. Regenerative role of seed banks following an intense soil disturbance. Acta Oecologica 27:57-65.

Manning, C. 2007. Generalized linear mixed models (Illustrated with R on Brennan et al.’s datives data). Accessed 22 January 2018 from <http://nlp.stanford.edu/manning/courses/ling289/GLMM.pdf>.

[MA NH&ESP] Massachusetts Natural Her-itage and Endangered Species Program. n.d. Massachusetts Division of Fisheries and Wildlife, 1 Rabbit Hill Road, Westbor-ough, MA, 01581. Accessed 26 June 2017 from <http://www.mass.gov/eea/agencies/dfg/dfw/natural-heritage/species-informa-tion-and-conservation/rare-plants/>.

[MA Office of Coastal Zone Planning]. 2013. Sea Level Rise: Understanding and Applying Trends and Future Scenarios for Analysis and Planning. Accessed 23 January 2017 from <http://www.mass.gov/eea/docs/czm/stormsmart/slr-guidance-2013.pdf>.

McCulloch, C.E., and S.R. Searle. 2001. Gen-eralized, Linear, and Mixed Models. Wiley, New York, NY.

Motzkin, G., and D.R. Foster. 2002. Grasslands, heathlands and shrublands in coastal New England: Historical interpretations and approaches to conservation. Journal of Biogeography 29:1-22.

Neill, C., W.A. Patterson, and D.W. Crary. 2007. Responses of soil carbon, nitrogen and cations to the frequency and seasonality of prescribed burning in a Cape Cod oak-pine forest. Forest Ecology and Management

250:234-243.

Neill, C., M.W. Wheeler, E. Loucks, A. Weiler, B. Von Holle, M. Pelikan, and T. Chase. 2015. Influence of soil properties on coastal sandplain grassland establishment on former agricultural fields. Restoration Ecology 23:531-538.

Noss, R.F., E.T. LaRoe, and J.M. Scott. 1995. Endangered ecosystems of the United States: A preliminary assessment of loss and degra-dation. Biological Report 28, USDI National Biological Service, Washington, DC.

O’Dell, D.I. 2014. Middle and Eastern Moors Harrow Management Treatment Project Report. Unpublished document, Nantucket Conservation Foundation, Nantucket, MA.

Oldale, R.N. 1985. Geologic Map of Nantucket and Nearby Islands, Massachusetts. Depart-ment of the Interior, US Geological Survey, Arlington, VA.

Omand, K.A., J.M. Karberg, K.C. Beattie, D.I. O’Dell, and R.S. Freeman. 2014. Soil seed bank in Nantucket’s early successional communities: Implications for management. Natural Areas Journal. 34:188-199.

Parris, A., P. Bromirski, V. Burkett, D. Cayan, M. Culver, J. Hall, R. Horton, K. Knuuti, R. Moss, J. Obeysekera, et al. 2012. Global Sea Level Rise Scenarios for the US National Climate Assessment. NOAA Tech Memo OAR CPO-1, Climate Program Office, Silver Spring, MD.

Ratajczak, Z., J.B. Nippert, and S.L. Collins. 2012. Woody encroachment decreases di-versity across North American grasslands and savannas. Ecology 93:697-703.

R Core Team. 2015. R: A language and environ-ment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. <http://www.R-project.org/>.

Rowe, H.I. 2010. Tricks of the trade: Techniques and opinions from 38 experts in tallgrass prairie restoration. Restoration Ecology 18(s2):253-262.

Sorrie, B.A., and P.W. Dunwiddie. 1996. The vascular and non-vascular flora of Nan-tucket, Tuckernuck, and Muskeget Islands. Massachusetts Audubon Society, Massa-chusetts Natural Heritage & Endangered Species Program, Nantucket Maria Mitchell Association, and The Nature Conservancy, Nantucket, MA.

Swain, P.C., and J.B. Kearsley. 2001. Classifi-cation of the Natural Communities of Mas-sachusetts. Natural Heritage & Endangered Species Program, Massachusetts Division of Fisheries and Wildlife, Westborough.

Tiffney, W.N., and D.E. Eveleigh. 1985. Nantucket’s endangered maritime heaths.

Pp.1093–1109 in O.T Magoon, H. Con-verse, D. Miner, D. Clark, and L.T. Tobin, eds., Coastal Zone ’85: Proceedings of the Fourth Symposium on Coastal and Ocean Management, July 30–August 2, 1985, Baltimore, MD.

[TNC] The Nature Conservancy. 1998. Veg-etation Community Map of Nantucket, GIS Data Layer. The Nature Conservancy, Arlington, VA.

[USDA] United States Department of Agri-culture. 2012. Plant Hardiness Zone Map. <http://planthardiness.ars.usda.gov/PHZM-Web/Default.aspx>.

[US Climate Data]. n.d. United States Climate Data for Nantucket, MA, 1961–1990 Nor-mals. <http://www.usclimatedata.com/cli-mate/nantucket/massachusetts/united-states/usma0271>.

Valko, O., P. Torok, B. Tothmeresz, and G. Matus. 2011. Restoration potential in seed banks of acidic fen and dry mesophilous meadows: Can restoration be based on local seed banks? Restoration Ecology 19:9-15.

von Blanckenhagen, B., and P. Poschlod. 2005. Restoration of calcareous grasslands: The role of the soil seed bank and seed dispersal for recolonisation processes. Biotechnol-ogy Agronomy Society and Environment 9:143-149.

Von Holle, B., and G. Motzkin. 2007. Historical land use and environmental determinants of nonnative plant distribution in coastal northeastern United States. Biological Conservation 136:33-43.

Wagner, D.L., M.W. Nelson, and D.F. Sch-weitzer. 2003. Shrubland Lepidoptera of southern New England and southeastern New York: Ecology, conservation, and man-agement. Forest Ecology and Management 185:95-112.

Wheeler, M.M., C. Neill, E. Loucks, A. Weiler, B. Von Holle, M. Pelikan, and T. Chase. 2015. Vegetation removal and seed addition contribute to coastal sandplain grassland establishment on former agricultural fields. Restoration Ecology 23:539-547.

Wilson, M.V., C.A. Ingersoll, and M.G. Wilson. 2004. Why pest plant control and native plant establishment failed: A restoration autopsy. Natural Areas Journal 24:23-31.

Zaremba, R.E. 2004. Scleria pauciflora Muhlen-berg ex Willdenow var. caroliniana Alph. Wood and Scleria pauciflora Muhlenberg ex Willdenow var. pauciflora (Fewflowered nut-rush) Conservation and Research Plan for New England. New England Wild Flower Society, Framingham, MA.

Documents

Harrowing and Seed Addition for Sandplain Grassland and