6 wr f n...santa cruz island from small nursery stock matthew l. james1,3, david m. hubbard1, and...

Experimental Planting of Native Shrubs on Santa CruzIsland from Small Nursery Stock

Authors: James, Matthew L., Hubbard, David M., and Cory, Coleen

Source: Monographs of the Western North American Naturalist, 7(1) :477-488

Published By: Monte L. Bean Life Science Museum, Brigham YoungUniversity

URL: https://doi.org/10.3398/042.007.0137

BioOne Complete (complete.BioOne.org) is a full-text database of 200 subscribed and open-access titlesin the biological, ecological, and environmental sciences published by nonprofit societies, associations,museums, institutions, and presses.

Your use of this PDF, the BioOne Complete website, and all posted and associated content indicates youracceptance of BioOne’s Terms of Use, available at www.bioone.org/terms-of-use.

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

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofitpublishers, academic institutions, research libraries, and research funders in the common goal of maximizing access tocritical research.

Downloaded From: https://bioone.org/journals/Monographs-of-the-Western-North-American-Naturalist on 12 Aug 2021Terms of Use: https://bioone.org/terms-of-use

Monographs of the Western North American Naturalist 7, © 2014, pp. 477–488

EXPERIMENTAL PLANTING OF NATIVE SHRUBS ON SANTA CRUZ ISLAND FROM SMALL NURSERY STOCK

Matthew L. James1,3, David M. Hubbard1, and Coleen Cory2

ABSTRACT.—The natural vegetation of Santa Cruz Island was severely disturbed by nonnative herbivores for wellover a century. As the livestock and feral ungulates (primarily sheep, cattle, and pigs) were removed from the island overthe last 30 years, many of the native plant communities began to recover naturally. Recovery has been extremely slow inother areas, especially where nonnative annual grasses and fennel (Foeniculum vulgare) dominate several hundredhectares that were under intense agricultural use (pastures and farmed lands). We experimentally tested the feasibility ofspeeding up the recovery process in a postagricultural area of the island’s Central Valley using active restoration tech-niques. We assessed how weed control via herbicide application and planting of small nursery stock without irrigationmight contribute to restoring natural plant assemblages in 3 different areas of the Central Valley (valley bottom, uppersouth-facing slope, and midsouth-facing slope). In February 2009, we planted the same 21 species of native plants inexperimental plots with and without weed control at each location. In December 2009, we planted 28 species in adja-cent plots after 2 seasons of weed control. We assessed natural recruitment in weeded and unweeded plots that were notplanted. At all 3 locations, a single early season herbicide treatment prior to planting had strong positive effects on thesurvival, cover, and reproduction of planted natives compared to no herbicide treatment; the effects persisted and grewstronger in the second and third years. Repeated herbicide treatments over 2 years before planting did not result in anyadditional significant positive effects on native survival or growth compared to only one herbicide treatment. We sawvirtually no natural recruitment of native shrubs in unplanted plots. We found that typical coastal sage scrub speciesperformed best at the upper slope site and poorly at the valley bottom site. Grasses and shrubs tolerant of poorlydrained soil did better in the valley bottom. We found that planting native species from small nursery stock without irri-gation is effective for a wide range of grassland and coastal sage scrub species. All of the restoration techniques we usedare cost effective and can be scaled-up to restore large areas of postagricultural lands.

RESUMEN.—La vegetación natural de la Isla Santa Cruz fue gravemente alterada por herbívoros no nativos durantemás de un siglo. Muchas de las comunidades de plantas endémicas empezaron a recuperarse naturalmente durante losúltimos 30 años, cuando el ganado y los ungulados ferales (principalmente ovejas, vacas y cerdos) fueron eliminados dela isla. La recuperación ha sido extremadamente lenta en otras áreas, especialmente donde hierbas anuales no nativas yel hinojo (Foeniculum vulgare) dominan varios cientos de hectáreas que estaban bajo uso agrícola intenso (pastos y tie-rras labradas). Comprobamos experimentalmente la viabilidad de acelerar el proceso de recuperación en un área post-agrícola del Valle Central de la isla, usando técnicas de restauración activa. Evaluamos cómo el control de la maleza, apartir del uso de herbicidas y la siembra de pequeñas plantas de vivero sin irrigación, podrían contribuir a restaurar gru-pos de plantas naturales en 3 áreas diferentes del Valle Central (en el fondo del valle, ladera superior mirando al sur yladera media mirando al sur). En febrero del 2009, plantamos las mismas 21 especies de plantas nativas en terrenosexperimentales con y sin control de hierbas. En diciembre de 2009, plantamos 28 especies en terrenos adyacentes trasdos temporadas de control de hierbas. Evaluamos el reclutamiento natural de los terrenos con maleza y los que notenían hierbas que no fueron plantados. Una sola temporada temprana de tratamiento con herbicida previa a la planta-ción tuvo fuertes efectos positivos en la supervivencia, cobertura y reproducción de las plantas nativas plantadas, encomparación con el tratamiento sin herbicida en las tres localidades. Los efectos persistieron y se hicieron más fuertesen el segundo y tercer año. Reiterados tratamientos con herbicida durante dos años antes de la plantación no tuvieronningún efecto positivo adicional significativo en la supervivencia nativa o en el crecimiento, en comparación con elúnico tratamiento con herbicida. Vimos que no había ningún desarrollo natural de los arbustos endémicos en los terre-nos sin plantar. Encontramos que las especies típicas costeras de breña de salvia obtenían un mejor resultado en la de laladera superior y un menor resultado en el fondo del valle. Las hierbas y arbustos que toleraban las tierras pobrementedrenadas tenían mejores resultados en el fondo del valle. Descubrimos que plantar especies nativas de pequeños viverossin irrigación es efectivo para un amplio rango de praderas y especies de breña de salvia costera. Todas las técnicas derestauración que usamos se pueden aplicar para restaurar grandes áreas de tierras post-agrícolas.

1Coastal Restoration Consultants, Inc., Monte Vista Ave., Ventura, CA 930032The Nature Conservancy, Santa Cruz Island Project, 532 E. Main St. Suite #200, Ventura, CA 93001.3E-mail: [email protected]

477


Successful large-scale restoration of nativeshrub and grassland communities in postagri-cultural lands in coastal southern Californiarequires the restoration of natural ecosystemprocesses. Every project area has a different his -tory of disturbance and, therefore, different pro -cesses in need of restoration. Because manyrestoration projects occur on postagriculturalland in southern California, the majority ofprojects require actions in 4 main areas: (1)allowing disturbed soil to regenerate a naturalsoil profile by ending soil disturbance (tilling,rooting, etc.); (2) controlling or eradicatinginvasive nonnative plants that outcompetenative plants and alter natural nutrient andwater cycles; (3) removing overgrazing bynonnative herbivores such as sheep, cattle, orgoats; and (4) reintroducing native plantswhere there are no longer native propagulespresent (D’Antonio and Vitousek 1992, Stylin-ski and Allen 1999, Corbin and D’Antonio2004, Cox and Allen 2008, Eliason and Allen2008, Yelenik 2008, Yelenik and Levine 2010).Addressing each of these issues (and often oth-ers) may be crucial to successful restoration. Avariety of restoration approaches have beenused on the California mainland with widelyvarying success (Cox and Allen 2008). Whatworks on one site may fail at another. Adaptiveapproaches (collecting experimental, monitoringdata and adjusting techniques) are importantin restoration (Erskine Ogden and Rejmánek2005). Small- and medium-scale experimentalpilot studies are valuable for fine tuning tech-niques that can then be applied at largerscales. The object of this study is to evaluatetechniques that are most likely to lead to suc-cessful, cost-effective restoration across largeareas of postagricultural landscape on SantaCruz Island (SCI).

The vegetation on California’s ChannelIslands is similar to that of the nearby main-land. Forest, woodland, scrub, grassland,riparian, and coastal vegetation communitiescan all be found on one or more of the Chan-nel Islands (Junak et al. 1995). However, theislands’ long isolation from the mainland hasresulted in numerous endemic species andsubspecies occurring on the islands, and onlya subset of the mainland flora is present. Untilquite recently, SCI had large populations ofsheep, cattle, and feral pigs that decimated thenative plant communities through overgrazingand rooting. Over the years, many nonnative

invasive plants were accidentally or intention-ally introduced to the island as well (Junak etal. 1995). A few areas of the island, includingmost of the Central Valley, were used for agri-culture, including vineyards in the late 1800sand early 1900s. There has been significantnatural recovery of native plant populationssince nonnative herbivore removal on SCI andon nearby San Miguel and Santa Barbaraislands (Corry and McEachern 2009); how-ever, introduced nonnative weeds continue todominate the postagricultural areas.

Nonnative annual grasses and fennel(Foeniculum vulgare) now dominate the Cen-tral Valley and other large areas including thelower Christy and Sauces drainages, parts ofthe isthmus, and the east end of the island. Inthe 1990s, a fennel control experiment wasconducted in the Central Valley and suc-ceeded in eliminating some dense patches offennel ( J. Randall personal communication).However, nonnative grasses and forbs quicklymoved in and replaced the fennel. There hasbeen very little colonization of these annualgrasslands by native shrubs despite abundantintact habitat (and presumably abundant propa -gules) on the hills to the north (coastal sagescrub) and the south (chaparral). The ability ofthe annual grasses to modify the habitat inways that favor their own persistence andexclude other species (e.g., by generating deepthatch and altering soil moisture and nutrientcycling) does not require continued distur-bance (D’Antonio and Vitousek 1992). Thesepostagricultural areas will likely remain domi-nated by annual grasses for decades or longerif left alone (Eliason and Allen 2008).

We set out to determine what types of actionswould be needed to effectively restore nativeshrub and grassland communities in the Cen-tral Valley of SCI. Some form of weed controlis necessary for restoring native communitiesin annual grasslands (Cox and Allen 2008).Without weed control, annual grasses inhibitgermination and establishment of na tive seed -lings by outcompeting them for light (Eliasonand Allen 1997) and water (Davis and Mooney1985, Eliason and Allen 1997). This is trueboth for naturally dispersed propagules andplanted or seeded restoration sites. The amountof weed control required (one or multiplerounds of treatment) and the best technique(herbicide, tilling, mowing) vary by site anddepend on the goals of the restoration project.

478 [Volume 7MONOGRAPHS OF THE WESTERN NORTH AMERICAN NATURALIST


For this experiment, we chose to reintro-duce native plants from nursery stock in smallplugs (76 cm3) as opposed to the much larger1-gallon pots (3000 cm3) commonly used inrestoration. Smaller stock is preferred forlarger projects because the plugs (1) are eas-ier to plant, (2) require significantly less nurs-ery space and fewer supplies, (3) require lesssoil disturbance when planting, (4) are less ex -pensive per unit, and (5) are easier to trans-port to remote areas. We also conducted acompanion experiment testing the efficacy ofdirect seeding, but the results are not in -cluded here.

We developed our planting palette basedon observations in nearby intact native vegeta-tion. We limited ourselves to perennial nativespecies occurring in and around the CentralValley on south-facing slopes and in valley-bottom locations, and we included SCI andnorthern Channel Islands endemics. We simul -taneously repeated the experiment in 3 areas:high on a south-facing slope, in the valley bot-tom, and midway up the south-facing slope.By introducing a wide variety of species atmultiple locations, we expected to developrecommended planting palettes for the differ-ent areas that would each be a subset of the 28species we used.

The goal of the experiment described herewas to determine an effective strategy forrestoring nearly 200 ha of disturbed habitatin the Central Valley and similar areas of SCI.Specifically, we asked (1) whether nativeshrubs are reinvading the dense annual grass-lands in the Central Valley, (2) whether nativeshrubs will reinvade after fennel and annualgrasses are suppressed, (3) how much weedcontrol is needed to successfully reintroducenative shrubs and grasses from small nurserystock, (4) whether different species will growbetter in different areas of the Central Valley,and (5) whether large-scale restoration ofshrub and grassland habitats is feasible inthe Central Valley. To answer these questions,we implemented a medium-scale experimentusing a range of weeding and planting treat-ments at 3 locations in the Central Valley. Weused only planting and weed-control tech -niques that can be implemented on a largeproject. Our findings are applicable to coastalsage scrub and grassland habitats on otherdisturbed areas of the island and on the main-land in California.

METHODS

Study Site

The study site is on Santa Cruz Island,located approximately 40 km off the coast ofSanta Barbara, California. At 250 km2, it is thelargest of the 8 islands that make up the Chan-nel Islands. The study site is located in theCentral Valley of SCI at an elevation of ap -proximately 80 m.

SCI has a Mediterranean climate with cool,wet winters and warm, dry summers. The rela -tively tall ridges surrounding the Central Valleylimit marine influence on the climate, leadingto much warmer daytime temperatures andcolder nighttime temperatures than areas alongthe coastline. With the exception of occa-sional hail and infrequent snow on the tallestpeak, all precipitation on SCI falls as rain pri-marily between October and April (averaging502 mm per water year). The prevailing winddirection is WNW, with only occasional epi -sodes of sustained winds over 4.5 m ⋅ s–1

(~10 mph). The Central Valley is occasionallyaffected by “Santa Ana” wind events, whenvery dry and hot easterly winds cause tempera -tures to spike and humidity to plummet.

Annual rainfall amounts and timing variedwidely during the first 4 years of the study.The 2008–2009 season (the first year weplanted) was drier than average (208 mm). Sig-nificant rain fell the week before planting inearly February (53.8 mm), but very little rainfell after planting (25 mm in total, with thelargest event dropping 7 mm). Rainfall in2009–2010 (the second year of planting) wasabout average (478 mm). Most of the rain fellin December, January, and February afterplanting. The summer of 2010 was very cool,with many overcast days and a general lack ofhot, dry Santa Ana events. The 2010–2011rain season was above average (651 mm), withstorms spaced throughout the growing season.The 2011–2012 season was drier than average(320 mm), with soaking early season rains, anespecially harsh midwinter drought, and somesoaking spring rains. The midwinter drought,which occurred from late January throughmid-March, featured only one rainfall eventover 2 mm (6 mm in mid-February) and had8 days with high temperatures over 26.7 °C.

The postagricultural areas of the CentralValley are currently dominated by invasive, non -native annual grasses and fennel (Foeniculum

2014] PLANTING NATIVE SHRUBS FROM NURSERY STOCK 479


vulgare). There are several other non nativeplant species and a few native grass and shrubspecies that are locally dominant. We chose3 experimental locations. Each was form erlyplowed, had relatively consistent fenneldensity with few native plants, and was domi-nated by either slender wild oat (Avena bar-bata) or ripgut brome (Bromus diandrus).

Experimental Design

The same experimental design was used at3 locations. Happy is located in the valley bot-tom adjacent to an incised ephemeral drainage.The topography is generally flat with a slight(~1%) west-facing aspect. The soil is a clayloam and includes some microtopography,probably the remnants of pig rooting. Sneezyis located on a moderately steep (~10%),south-facing slope on the north side of theCentral Valley. The soil is a sandy clay loamwith some rocks and pebbles. Clumpy is closeto the valley bottom and has a slight (~1%)south-facing slope and clay loam soil. It is far-ther from the stream channel than Happy andbelow Sneezy.

We used a randomized complete blockdesign with 6 treatments and 6 blocks perlocation for a total of 36 plots per location and108 plots total. Each plot was 7 × 7 m, with a2-m buffer between plots and between blocks.Each block at Sneezy and Clumpy contained6 plots in an east–west row on contour withthe slope. Blocks were arranged 2 across and3 down. At Happy, each block was 2 plots wideand 3 plots long, with the long axis in the east–west direction. Blocks were arranged end toend, east to west. The corners of each plot weremarked with steel rebar and color-coded flags.

The 6 treatments were No Spray/No Plant(no weed treatment and no outplanting ofnative plants; control); No Spray/Plant Yr 1(weeds not treated with herbicide and nativeseedlings outplanted in year 1); Spray Yr 1/NoPlant (all weeds sprayed once, and fennel cutand then sprayed multiple times, but no nativeseedlings outplanted); Spray Yr 1/Plant Yr 1(all weeds sprayed once, fennel cut and thensprayed multiple times, and native seedlingsoutplanted in year 1); Spray 2 Yrs/No Plant (allweeds sprayed multiple times the first seasonand one time in the second season, and nonative seedlings outplanted); and Spray 2 Yrs/Plant Year 2 (all fennel cut down and all weedssprayed multiple times the first season and

one time in the second season, and nativeseedlings outplanted in year 2). These final2 treatments are referred to as the “year 2plots.” Buffers between plots were not sprayedor otherwise treated.

In the 4 sprayed treatments (total of 72plots), fennel and all other plants, both nativeand nonnative, were sprayed with the herbi-cide glyphosate (1% Roundup) in January2009. All plants, including fennel in the year2 plots, were again sprayed with herbicide (1%Roundup) in April 2009 to kill a second cropof annuals, and a third time in November 2009to kill newly sprouted annuals. Fennel in all 4sprayed treatments was sprayed with triclopyr(1% Garlon 4 Ultra) in May 2009 and again inAugust 2009 (2% Garlon 4 Ultra). Fennel wascut down, and the stems were removed fromthe plots before the first herbicide treatment(resprouts were sprayed).

Each No Spray/Plant Yr 1 and Spray Yr 1/Plant Yr 1 plot was planted with 140 nativeplants of 21 species between 10 and 12 Feb-ruary 2009. The mix of plants (Table 1) wasdetermined by a combination of seed avail-ability, success in propagation, and a desire toemphasize species we thought would do best.The number of plants per plot was calculatedbased on our desired spacing (~50 cm). Thisspacing is denser than most coastal sage scrubrestoration sites that are planted. Our reasonfor the extra-dense planting was to balanceexpected higher-than-normal mortality due tosome plants being too small at the time ofplanting. The planting mix for Clumpy wasslightly modified (Table 1) compared to theother 2 sites due to the number of plants avail-able when it was planted (last). A total of 5040plants were installed in year 1.

Each Spray 2 Yrs/Plant Yr 2 plot was plantedwith 130 plants of 28 species on 16 and 17December 2009 (2340 plants). When moreplants were ready in early January 2010, weadded 10 additional plants to each plot tobring the total up to 140 plants per plot (Table3) and 2520 total plants in year 2. Addition-ally, on 7 December 2009, we planted 5 acornsof Quercus pacifica and 5 acorns of Q. agrifoliainto each Spray 2 Yrs/Plant Yr 2 plot.

All plants were grown from seed collectedon Santa Cruz Island. Seed was collected andimmediately sown or stored in a freezer for2–5 days to kill any invertebrates collectedwith the seed. After sprouting, seedlings were



transplanted into small plugs (50-plug trays)and allowed to establish for 4–18 weeks beforeplanting. Plants ranged from 3 to 15 cm tallwhen planted. We used Sunshine #5 soillessmix (peat moss and fine perlite) in seed flatsand plugs. Three slow-growing species plantedin year 2 (Heteromeles arbutifolia, Rhus inte-grifolia, and Rhamnus pirifolia) were sown infall and winter 2008 and transplanted intoliners (~30-cm deep and 8-cm diameter) inearly 2009. No seed, plant material, or soil wereimported to the island. All nursery suppliesand tools were new and free of soil.

Planting holes were made in moist soil viaa steel breaker bar with a diameter similar tothe plugs. In year one, a small area (~10-cmdiameter) of annual grass thatch and any livinggrass was removed around each planting hole,mainly to allow the planters to see the holes.In the second year, we minimized disturbanceof existing thatch during planting. We did notplant within about 30 cm of fennel plants orexisting native grasses but otherwise planted

on a grid with approximately 50 cm betweenplants. In plots with dense fennel or nativegrass, plants were planted much closer togetheron average. Plantings were never irrigated afterplanting in either season, and we did no re -placement plantings after the initial planting.

Monitoring Design

We estimated the percent cover by speciesin all plots in October 2008 before we carriedout any experimental manipulations. We esti-mated the percent cover to the nearest per-cent (any species present with <0.5% coverwas given a cover score of 0.1%) by speciesfor all green or senesced plants found withineach 7 × 7-m plot. We used a similar method-ology in February 2013 to assess relativecover. Instead of identifying plants to species,we classified them as native shrubs, grasses,or forbs and nonnative annual grasses, forbs, orperennials.

We carried out posttreatment monitoringin each plot on 6–8 May 2009, 6–8 October


TABLE 1. Planting list (same for No Spray/Plant Yr 1 and Spray Yr 1/Plant Yr 1 treatments).

Plant Yr 1 plots Plant Yr 2 plotsa_____________________ _____________Happy/

Family Species Sneezyb Clumpyb All locationsb

Fabaceae Acmispon dendroideus var. dendroideus 2 2 0Asteraceae Artemisia californica 3 3 13Asteraceae Artemisia douglasiana 10 8 15Asteraceae Baccharis pilularis 10 10 5Asteraceae Baccharis salicifolia 10 10 11Alliaceae Bloomeria crocea 1 1 0Poaceae Bromus carinatus 12 12 5Convolvulaceae Calystegia macrostegia ssp. macrostegia 0 0 4Ranunculaceae Clematis ligusticifolia 1 1 3Poaceae Elymus condensatus 13 13 4Poaceae Elymus glaucus 5 5 8Onagraceae Epilobium canum 6 6 5Polygonaceae Eriogonum grande var. grande 14 15 16Polygonaceae Eriogonum arborescens 12 15 8Asteraceae Grindelia camporum var. bracteosum 9 9 11Asteraceae Hazardia detonsa 5 8 3Asteraceae Hazardia squarrosa var. grindelioides 6 4 0Rosaceae Heteromeles arbutifolia 0 0 2Asteraceae Isocoma menziesii var. vernonioides 10 7 4Fabaceae Lupinus albifrons var. douglasii 3 3 5Asteraceae Malacothrix saxatilis var. implicata 3 3 1Cucurbitaceae Marah macrocarpus 0 0 1Phrymaceae Mimulus aurantiacus var. pubescens 0 0 2Asteraceae Pseudognaphalium californicum 0 0 1Asteraceae Pseudognaphalium microcephalum 2 2 0Rhamnaceae Rhamnus pirifolia 0 0 3Anacardiaceae Rhus integrifolia 0 0 5Poaceae Stipa lepida 3 3 5aAll Plant Yr 2 plots also had 10 acorns planted (5 each of Quercus pacifica and Q. agrifolia).bPlants per plot.


2009, 12–16 May 2010, 5–8 October 2010, 5–8November 2011, and 23–27 September 2012.Four sample quadrats, each 1 × 2.5 m, wererandomly located in the 10 contiguous loca-tions within the central 5 × 5-m core of eachplot. Within each sample quadrat, we esti-mated the percent cover for every species tothe nearest percent (any species present with<0.5% cover was given a cover score of 0.1%).The cover estimates from the 4 quadrats wereaveraged for each plot. To determine initialsurvival rates of planted plants, we countedthe number of planted plants alive in theplanted plots in the spring and fall followingplanting.

For the preproject monitoring, we consid-ered dead or senesced plants as part of thevegetated cover. For all other monitoring, wecounted only living plants and considereddead or senesced plants as part of the “un -vegetated” cover. All plants were identified tospecies except needlegrasses (Stipa pulchraand S. lepida), which were present in some ofthe plots but, in some seasons, were very diffi-cult to tell apart. Therefore, we clumped bothspecies together into the “Stipa spp.” category.All taxonomy follows Baldwin et al. (2012).

Statistical Analyses

The individual 7 × 7-m plots were the ex -perimental unit for all analyses. The 6 treat-ments were the 6 spray/no spray and plant/noplant combinations. Data were analyzed witha 2-way ANOVA in Excel to test for differ-ences in total native cover between the 6 treat-

ments and differences in planted native coverbetween the 3 treatments that included plant-ing. We found no strong block effects in anyof the analyses. Each of the 3 locations wasanalyzed separately. We used a linear regres-sion (in Excel) to explore the relationship be -tween native and nonnative plant cover withinindividual plots.

RESULTS

Prior to experimental manipulations inOctober 2008, all of the locations were domi-nated by nonnative plant species (Fig. 1).Ripgut brome, Bromus diandrus, was thedominant plant species at 2 of the locationsand was a close second to F. vulgare at theother (Fig. 1). Fennel cover was highest inSneezy (44.4%), lowest in Happy (11.0%), andintermediate in Clumpy (29.5%). There wasvery little unvegetated area at any of the loca-tions (Fig. 1). Total native cover was highest inSneezy and very low in Happy and Clumpy(Fig. 1). There were no native shrubs in anyplot. Only 9 native plant species were encoun-tered and only needlegrass, Stipa spp., had anaverage cover >0.5% at any of the sites (7.0%in Sneezy and 0.8% in Clumpy).

One round of herbicide treatment had astrong negative effect on nonnative cover forone year at Sneezy (Fig. 2). Multiple rounds ofherbicide over 2 seasons had a strong negativeeffect on nonnative cover for 2 years in Sneezy(Fig. 2). By 2013, nonnative cover in the year 2treatments had increased to near preproject


Fig. 1. Vegetation cover in experimental plots prior to manipulations. Error bars represent –+2 standard errors.


levels (Fig. 2). Nonnative cover followed simi-lar patterns at Happy and Clumpy.

Initial survival of planted natives was low-est in the No Spray/Plant Yr 1 treatment at alllocations (Table 2) after 3 months (May) and 8months (October). Initial survival in the Spray2 Yrs/Plant Yr 2 treatment was similar to theSpray Yr 1/Plant Yr 1 treatments in Clumpyand Sneezy and lower in Happy (Table 2). Ini-tial survival of year 2 plantings in Happy waslower than at the other 2 locations.

The estimated total native cover (plantedplus nonplanted natives) varied greatly between locations and treatments in fall 2012(Fig. 3). Nonplanted natives include perennialspecies that were in the plots at the beginningof the study and were not sprayed or survivedspraying and native seedlings that sproutedduring the experiment (from the seed bank orfrom seed produced in the plots in the firstyear). All 3 planted treatments had highertotal native cover than the 3 nonplanted treat-ments at all 3 locations (Fig. 3). We found a

total of 2 native shrub individuals establishedin unplanted plots (both were Eriogonumarborescens in Spray 2 Yr/No Plant plots atSneezy). Differences between treatments werestatistically significant at all 3 locations (2-wayANOVA df = 5, 25; Sneezy P < 0.001; ClumpyP < 0.001; Happy P < 0.001).

The average cover of planted natives in allplanted treatments increased at all locationsbetween fall 2009 and 2012 (Fig. 4). Thelargest year-on-year increases were between2010 and 2011. The No Spray/Plant Yr 1 treat-ment continues to have the lowest plantednative cover at each location. The Spray Yr 1/Plant Yr 1 treatment always had the highestaverage cover, though the difference wasslight at Sneezy and moderate at Clumpy. Dif-ferences between treatments were statisticallysignificant at all 3 locations in September 2012(2-way ANOVA df = 2, 10; Sneezy P < 0.001;Clumpy P = 0.001; Happy P = 0.05).

The average cover by species varied widelybetween the 3 locations in the Spray Yr 1/


TABLE 2. First-year survival of planted natives in spring and fall.

No Spray/Plant Yr 1 Spray Yr1/Plant Yr 1 Spray 2 Yrs/Plant Yr 2____________________ _____________________ _____________________May October May October May October

Sneezy 42% 23% 63% 59% 62% 55%Clumpy 39% 18% 44% 43% 51% 45%Happy 38% 14% 52% 39% 39% 28%Overall 40% 18% 53% 46% 51% 43%

Fig. 2. Total nonnative cover in Sneezy. Error bars represent –+2 standard errors.

0

25

50

75

100

125

No Spray/No Plant

No Spray/Plant Yr 1

Spray Yr 1/No Plant

Spray Yr 1/Plant Yr 1

Spray 2 Yrs/No Plant

Spray 2 Yrs/Plant Yr 2

Perc

ent

Cov

er

Pre-project 2008 Spring 2009 Spring 2010 Winter 2013

125

75

100

revoC

t

2ct ejopr-ePr9002g niprS0102g niprS3102rteniW

8002

25

50

t n

cereP

0 prSNoaPl

No/yaprt na

/yaprSNo1rYt naPl

No/1rYyaprSt naPl

prSPl

/1rYyapr1rYt naPl

2yaprSaPlNo

s/rYt n

s/rY2yaprS2rYt naPl

Pe

rce

nt

Co

ve

r


Plant Yr 1 treatment in fall 2012 (Fig. 5). Forexample, Artemisia douglasiana and Baccharispilularis are both doing best in Happy,whereas all other species are doing poorlycompared to the other locations (Fig. 5).Species characteristic of coastal sage scrub(Artemisia californica, Eriogonum arborescens,Hazardia detonsa, Isocoma menziesii, Elymuscondensatus, and Lupinus albifrons) are doingbest in Sneezy (Fig. 5). Sampling in the fallbiases these estimates against species that aredormant at that time of year, such as Bromuscarinatus, Elymus glaucus, and Stipa lepida.Cover of other species that are partially senes-

cent, such as Eriogonum grande, Grindeliacamporum, and Epilobium canum, is also underestimated compared to peaks in the spring orsummer.

We found that as the total native cover increased in plots, the total nonnative covertended to decrease (Fig. 6). The effect wasstrongest in plots that had large shrubs withdense canopies (Eriogonum arborescens andLupinus albifrons) and where native plantrecruitment from seed produced by the out-planted plants led to dense patches of nativegrasses (Bromus carinatus and Elymus glau-cus) and small shrubs (Eriogonum grande).


Fig. 3. Native vegetation cover in experimental plots in September 2012. Error bars represent –+2 standard errors.

Pe

rce

nt

Co

ve

r

0

10

20

30

40

No Spray/Plant Yr 1



No Spray/Plant Yr 1



No Spray/Plant Yr 1



Sneezy Clumpy Happy

% C

over

Oct-09 Oct-10 Nov-11 Sep-12 30

40

90-ctO01-ctO11-vNo21p-eS

20

30

re

vo

C%

21p-eS

0

10

%

0

y/aNo Sprrar 1Plant YYr

Spr

r 1/y YYraSprrar 1Plant YYr

aSprraPlant Y

y Sneez

rs/y 2 YYrar 2ant YYr

No SprPlant Y

y/aprrar 1 YYr

r y YYraSprrar Plant YYr

y pClum

r 1/r 1

rs/y 2 YYraSprrar 2 Plant YYr

y

rs/ y/aNo Sprrar 1Plant YYr

r 1/y YYraSprrar 1 Plant YYr

SprPlant Y

yppHa

rs/y 2 YYrarrar 2 lant YYr

Perc

ent

Cover

Fig. 4. Percent cover of planted natives at all 3 sites in the fall. Error bars represent –+2 standard errors.


DISCUSSION

The first of our study questions concernswhether native shrubs are reinvading theCentral Valley on their own. We found no evi-dence that this is occurring. We found no newnative shrubs in any of the control plots overthe 4 years we monitored. We did not test themechanism limiting establishment of nativesin the Central Valley, though we suspect thatin the areas where we conducted this experi-ment, any native seeds in the plots wouldeither be inhibited from germinating by thethick annual grass thatch or die after germi-nating from competition with annual non -natives (Yelenik and Levine 2010).

We also asked whether suppressing weedswould lead to natural recruitment of nativeshrubs and grasses. We did find 2 shrubs(both Eriogonum arborescens) in weeded andunplanted plots: one each in year 3 and 4 ofthe monitoring. This recolonization rate, 2plants in a total of 1764 m2, is much too slowfor us to expect that suppressing weeds willbe a sufficient strategy for restoration ofnative habitats in the Central Valley. Again,we did not test the mechanism limiting

establishment of natives in the Central Valleywhere we suppressed weeds, though ourresults suggest that there is little or no seedbank of native shrubs and that there is verylittle seed rain (significant seed sources areapproximately 40 m from Sneezy, >100 mfrom Clumpy, and >200 m from Happy).However, it is possible that native speciesgerminated and died before they were largeenough for us to detect and that competitionis a more important mechanism. Further,given that there was significant seed pro-duction in adjacent planted plots by the sec-ond year, which production led to subse-quent establishment of additional nativeplants in those planted plots, we might expecthigher rates of seed rain in the adjacentunplanted plots than in most areas of theCentral Valley.

We are confident that native species willneed to be introduced as part of an activerestoration project in the Central Valleybecause native species will not reinvade ontheir own (Erskine Ogden and Rejmánek2005, Yelenik 2008, Yelenik and Levine 2010).The most common way to reintroduce nativespecies is from seed or nursery stock. We


0

2

4

6

8

10

Art

cal

Art

dou

Bac

pil

Epi ca

n

Eri a

rb

Eri g

ra

Haz

det

Iso m

en

Ely

con

Lup a

lb

Art

cal

Art

dou

Bac

pil

Epi ca

n

Eri a

rb

Eri g

ra

Haz

det

Iso m

en

Ely

con

Lup a

lb

Art

cal

Art

dou

Bac

pil

Epi ca

n

Eri a

rb

Eri g

ra

Haz

det

Iso m

en

Ely

con

Lup a

lb

Sneezy (31.9%) Clumpy (20.9%) Happy (20.5%)

% C

over

8

10

4

6 re

vo

C%

2

0 lca

t rA

do

A

udo

t rA

lpi

c aB

nca

pi

E

b

rair

E

agr

irE

t de

zaH

31.9%y (Sneez

t de

zaH

nem

soI

nco

ylE

b

lap

uL

lca

t rA

udo

t rA

)31.9%

lpi

c aB

nca

pi

E

b

rair

E

agr

irE

t de

zaH

nem

soI

%) .90(2y pClum

nem

soI

nco

ylE

b

lap

uL

lca

t rA

udo

t rA

lpi

c aB

%) H

pi n

capi

E

b

rair

E

agr

irE

t de

zaH

nem

soI

)%5.02(yppaH

nco

ylE

b

lap

uL

Pe

rce

nt

Co

ve

r

Fig. 5. Vegetation cover of 10 native species in Spray Yr 1/Plant Yr 1 plots in fall 2012. Numbers in parentheses aretotal native cover for each location.


have identified several challenges to directseeding in the Central Valley, including thelack of and prohibition of using commerciallyavailable mainland-sourced seed (and hencegreater difficulty and cost in collecting suffi-cient seed on-island), the need to removethatch (which then releases more weeds fromthe seed bank and creates bare soil surfaceswhich dry quickly between rains), and theneed for fortuitous wet years with well-timedrain. However, our results from a companionexperiment suggest that seeding some speciesmay be feasible (unpublished data). Given therisks though, we still question the utility ofrelying on seeding alone.

Our third question was how much weedcontrol is needed to successfully reintro-duce native shrubs and grasses from smallnursery stock. Planting directly into habitatdominated by fennel and annual grass with-out weed control is probably not a viableapproach. We saw very high mortality ofplanted plants, and cover estimates are con-sistently lower in plots without weed controlthan in plots where we controlled weedsbefore planting.

We saw very good survival, growth, andreproduction (we observed flowering in overhalf of the planted species) from severalspecies even in a very dry year, after treating

fennel and applying only one round of annualweed control. We saw a very similar responsefrom planted natives after 2 seasons of annualweed control. After 4 growing seasons, we areconvinced that the native cover in weededtreatments is still increasing, as is seed pro-duction. We found evidence that the nativeplants were suppressing nonnatives, indicatingthat the restoration process will be sustain-able in the long term.

We were surprised that nursery stock in -stalled after 2 years of weed suppression didnot perform any better than those installedafter one year of weed suppression. We ex -pected there to be significant benefits to mul-tiple rounds of herbicide treatment over 2years in terms of reducing nonnative competi-tion and increasing growth of native plants.Further, with over 25 cm of rain after planting(versus less than 3 cm the first year), we ex -pected much higher survival and growth. Itmay be that the vigorous growth of the re -maining weeds in the wetter year suppressedthe growth of planted natives. It is also possi-ble that the extra herbicide treatments hadsome detrimental effect on growing conditions,such as a reduction in beneficial microbes(Irvine et al. 2013). We conclude that the extratime and cost for 2 years of weed control is notworthwhile.


y = -0.7557x + 110.13 R = 0.62144

40

60

80

100

120

140

0 10 20 30 40 50 60

Perc

en

t C

over

of

No

n-n

ati

ves

Percent Cover of Natives

120

140

se

viat

-n

-0.7557x +y =126.0=R

3 110.13441

80

100

-nn

oN

fo

re

vo

Ct

n

40

60

0

necr

eP

10 20 30

evoCtnecreP

30 40 50

seviatNfore

50 60

Fig. 6. Linear regression of total native cover versus total nonnative cover in February 2013. All plots in Sneezy,Clumpy, and Happy with at least 1% native cover are included.

Pe

rce

nt

Co

ve

r o

f N

on

na

tive

s

Percent Cover of Natives


Our results indicate that different areaswithin the Central Valley will support differ-ent suites of native species. Our study designused only one site each in the 3 general areas(high on the slope, midslope, and valley bot-tom). Because we did not replicate the experi-ment at multiple locations within each generalarea, we cannot be sure that this pattern isnot just an artifact of the particular locationswe chose. Nevertheless, our findings are con-sistent with observations of vegetation pat-terns in less disturbed areas elsewhere in theCentral Valley, and they provide a useful start-ing point for fine-tuning planting palettes infuture projects. Coastal sage scrub species,especially Artemisia californica, Eriogonumarborescens, Hazardia spp., Lupinus albifrons,and Elymus condensatus, should be empha-sized on the south-facing slopes. Baccharispilularis, Artemisia douglasiana, and perennialbunchgrasses will probably be good species inthe valley bottom areas. Midslope areas willbe a transition between these 2 vegetationtypes where the above species will overlapand others, such as Eriogonum grande, willlikely do best. We did not test the mecha-nisms that led to differences in species’ per-formance at different locations, but we sus-pect soil texture and moisture patterns (betterdrained soils on the slopes) played an impor-tant role.

Our study suggests that restoring largeareas of shrubland habitat in the Central Val-ley of Santa Cruz Island can be done using

limited weed control and planting. Thoughwe did see survival and growth of nativeplants without weed control, we believe thatin the long term, a weeding strategy, alongwith planting, will be needed to restore self-sustaining habitats.

There are 2 general ways to predict long-term responses with short-term data. A com-mon approach is to fit monitoring data tocurves and extrapolate into the future. How-ever, attempting to predict such trajectoriesis fraught with ecological pitfalls (Zedler andCallaway 1999, Matthews and Spyreas 2010).Stochastic processes (e.g., weather, fire) candramatically change ecological trajectoriesin restoration projects. Successional processesare probably important on most restorationsites, though they are usually difficult orimpossible to predict and can be responsiblefor sites diverging dramatically from observedshort-term trajectories (Matthews and Spyreas2010). Once we consider the myriad waysthese nondeterministic and deterministic pro -cesses might interact with each other overseveral decades, it is clear why our predictivepowers are limited. Nevertheless, there isalmost always a practical need to use availablemonitoring data either to predict the long-term “success” of a restoration or mitigationproject or, as in this case, to determine whatrestoration approach will yield the best long-term results in future projects. Based on thedata we have collected, we strongly believenative plants will need to be reintroduced,


Fig. 7. Hypothetical long-term trends in native cover with and without weed treatment.

Pe

rce

nt

Co

ve

r

Year

No Spray/Plant Yr 1




and we hypothesize that an approach of short-term weed suppression and planting fromsmall nursery stock will be more effective thanplanting alone for restoration of the CentralValley (Fig. 7).

We wanted to test the feasibility of usingsmall nursery stock without irrigation torestore shrublands in annual grasslands giventhe unique challenges associated with restora-tion on Santa Cruz Island. Our results after4 years suggest that (1) several species intro-duced from small nursery stock do not needmuch rain after planting to establish success-fully; (2) planting small nursery stock whenthere are no annual weeds alive (i.e., after agrow-kill cycle) creates a sufficient window fornative plants to establish; (3) planted nativesmay begin to outcompete annual grasses after2 or 3 years; and (4) different areas of theCentral Valley will support different plantcommunities.

ACKNOWLEDGMENTS

We thank the state of California’sWildlife Conservation Board for generousfunding to initiate this project. Additionalfunding was provided by the California StateCoastal Conservancy. We appreciate the con-tinued support of The Nature Conservancy’sSanta Cruz Island staff, both past and pres -ent, and The Nature Conservancy’s statechapter.

LITERATURE CITED

BALDWIN, B.G., D.H. GOLDMAN, D.J. KEIL, R. PATTERSON,T.J. ROSATTI, AND D.H. WILKEN, EDITORS. 2012. TheJepson manual: vascular plants of California. 2ndedition. University of California Press, Berkeley, CA.

CORBIN, J.D., AND C.M. D’ANTONIO. 2004. Competitionbetween native perennial and exotic annual grasses:implications for an historical invasion. Ecology 85:1273–1283.

CORRY, P.M., AND K. MCEACHERN. 2009. Patterns in post-grazing vegetation changes among species and envi-ronments, San Miguel and Santa Barbara Islands.Pages 201–214 in C.C. Damiani and D.K. Garcelon,

editor, Proceedings of the Seventh California IslandsSymposium. Institute for Wildlife Studies, Arcata, CA.

COX, R.D., AND E.B. ALLEN. 2008. Stability of exoticannual grasses following restoration efforts in south-ern California coastal sage scrub. Journal of AppliedEcology 45:495–504.

D’ANTONIO, C.M., AND P.M. VITOUSEK. 1992. Biologicalinvasion by exotic grasses, the grass/fire cycle, andglobal change. Annual Review of Ecology and Sys-tematics 23:63–87.

DAVIS, S.D., AND H.A. MOONEY. 1985. Comparative waterrelations of adjacent California shrub and grasslandcommunities. Oecologia 66:522–529.

ELIASON, S.A., AND E.B. ALLEN. 1997. Competition as amechanism for exotic grass persistence followingconversion from native shrubland. Restoration Ecol-ogy 5:245–255

______. 2008. Exotic grass competition in suppressingnative shrubland re-establishment. Restoration Ecol-ogy 5:245–255.

ERSKINE OGDEN, J.A., AND M. REJMÁNEK. 2005. Recoveryof native plant communities after the control of adominant invasive plant species, Foeniculum vulgare:implications for management. Biological Conservation125:427–439.

IRVINE, I.C., M.S. WITTER, C.A. BRIGHAM, AND J.B.H.MARTINY. 2013. Relationships between methylobac-teria and glyphosate with native and invasive plantspecies: implications for restoration. Restoration Ecol-ogy 21:105–113.

JUNAK, S., T. AYERS, R. SCOTT, D. WILKEN, AND D. YOUNG.1995. A flora of Santa Cruz Island. Santa BarbaraBotanic Garden. Santa Barbara, CA.

MATTHEWS, J.W., AND G. SPYREAS. 2010. Convergence anddivergence in plant community trajectories as aframework for monitoring wetland restoration prog -ress. Journal of Applied Ecology 47:1128–1136.

STYLINSKI, C.D., AND E.B. ALLEN. 1999. Lack of nativespecies recovery following severe exotic disturbancein southern Californian shrublands. Journal of AppliedEcology 36:544–554.

YELENIK, S.G. 2008. Plant-soil feedbacks and native shrubrecovery of exotic grasslands, Santa Cruz Island, CA.Doctoral dissertation, University of California, SantaBarbara, CA.

YELENIK, S.G., AND J.M. LEVINE. 2010. Processes limitingnative shrub recovery in exotic grasslands after non-native herbivore removal. Restoration Ecology 18:418–425.

ZEDLER, J.B., AND J.C. CALLAWAY. 1999. Tracking wetlandrestoration: do mitigation sites follow desired trajec-tories? Restoration Ecology 7:69–73.

Received 11 April 2013Accepted 9 April 2014


6 wr f n...santa cruz island from small nursery stock matthew l. james1,3, david m. hubbard1, and...

Documents