1

THIS IS AN AUTHORS’ COPY OF THE ACCEPTED MANUSCRIPT.

This manuscript has been accepted for publication in Wetlands Ecology & Management. The final publication is available at

Springer via http://www.springer.com/-/5/AVBOthJw2brxj7RSZsuv. This document will be updated when final publication

information is available.

Controlling perennial pepperweed (Lepidium latifolium)

in a brackish tidal marsh

V.D. Tobias*1, G. Block2, E.A. Laca1

1 Department of Plant Sciences University of California, Davis.

One Shields Avenue, Plant Sciences Mailstop 1, Davis, California 95616 2 U.S. Fish and Wildlife Service National Wildlife Refuge System Pacific Southwest Region Inventory and Monitoring Program

735B Center Blvd. Fairfax, CA 93930

*corresponding author: V.D. Tobias email address: vtobias@ucdavis.edu phone: (530)302-5460

ABSTRACT

Perennial pepperweed (Lepidium latifolium) is an aggressively invasive species that spreads by

vegetative growth and seeds. Common methods for removal such as hand-pulling and mowing are

impractical in brackish marsh environments. We evaluated the effects of two herbicide treatments

(imazapyr and imazapyr + glyphosate) against a non-herbicide control (flower head removal) on invasive

pepperweed and native vegetation in three habitats (bay edge, channel edge, and levee) in brackish

marshes. Both herbicide treatments produced significantly better control of pepperweed than the

control, but imazapyr alone took two years of treatment to produce levels of control that were similar to

one year of the imazpyr + glyphosate treatment. Both herbicide treatments also reduced native cover,

but the effects were more severe in plots treated with imazapyr + glyphosate than in plots treated with

imazapyr alone. Effects on pepperweed were similar across the three habitats, but impacts on native

vegetation were less severe in bay edge environments. Managers should consider the tradeoffs when

choosing a treatment plan for pepperweed: the quick reduction of pepperweed achieved by the

combination of imazapyr and glyphosate may come at the expense of creating opportunities for

reinvasion by causing bare ground and/or patches of litter that are slowly recolonized by native species.

KEYWORDS: invasive species, wetland, management, brackish marsh, imazapyr, glyphosate

2

INTRODUCTION

Perennial pepperweed (Lepidium latifolium, LELA) is an aggressive invasive plant in the mustard family

(Brassicaceae). It is native to Europe, Asia (temperate and tropical), and Africa, but has been declared a

noxious weed in 16 U.S. states, including California (USDA 2014). L. latifolium is a key species of concern

for San Francisco Bay because of its wide distribution throughout the estuary and high potential for

spreading (Grossinger et al. 1998). It has been documented in a wide range of habitats, including

rangelands, riparian areas, and coastal wetlands. Since it was first reported in the Sacramento-San

Joaquin Delta in 1941 (Robbins et al. 1941) it has spread throughout the San Francisco Bay-Delta

complex. Large monocultures of L. latifolium are found in tidal marshes of Suisun Bay and San Pablo

Bay, as well as along waterways that feed into the estuary (Grossinger et al. 1998; Vanderhoof et al.

2009). Control of invasive plant species such as L. latifolium is considered one of several important

strategies to conserve tidal marsh ecosystems of the San Francisco Bay Estuary (Goals Project 1999;

Vasey et al. 2012; U.S. Fish and Wildlife Service 2013).

Pepperweed can create monocultures, displacing native plants and reducing habitat for native species.

For example, invasion of Rush Ranch (Suisun Bay) by L. latifolium displaced endangered Suisun thistle

(Cirsium hydrophilum var. hydrophilum) in tidal wetlands (Fiedler et al. 2007) and displaces Sarcocornia

pacifica (Reynolds and Boyer 2010). One of the main ways L. latifolium succeeds is by altering the

structure of marshes because it is taller, has a more diffuse structure, and senesces in the fall, unlike

native species which do not produce substantial seasonal changes in cover (Reynolds and Boyer 2010).

Of particular concern are effects on habitat for the endangered salt marsh harvest mouse

(Reithrodontomys raviventris) and the reduction of suitable nesting areas for waterfowl species (Trumbo

1994). There is also potential for L. latifolium to invade habitat for endangered California clapper rails

(Rallus longirostris obsoletus), although additional studies are needed to determine impacts (Spauz and

Nuur 2004). In addition to changes in habitat, invasion of marshes by L. latifolium can cause shifts in

food availability for species that eat insects. The presence of L. latifolium causes shifts in community

composition and dominant species of invertebrates to those that are less desirable food sources in

Sarcocornia-dominated marshes (Reynolds and Boyer 2010). In some cases, the presence of L. latifolium

may have no impact or even some favorable impacts on wildlife. For example, sites invaded by

pepperweed had similar probabilities of presence for songbirds as those that were uninvaded and

pepperweed might provide better habitat for song sparrows (Spauz and Nuur 2004). Alternatively,

wildlife species that have specific habitat requirements that include native salt marsh plant species may

be adversely affected by expansive L. latifolium monocultures.

L. latifolium thrives in a wide range of environmental conditions and can reproduce and disperse by

rhizomes and roots as well as seeds. Once established, L. latifolium roots alter soil chemistry, favoring

its own growth and survival (Blank 2002). For example, changes in soil chemistry may enable rhizomes

to spread into saline marsh that would otherwise be too stressful for vegetative parts to establish or

seeds to germinate. L. latifolium reduces soil salinity by increasing calcium and magnesium in the soil

relative to sodium (Renz and Blank 2004; Blank and Young 2004). Although much of the previous

research on the biogeochemical impacts of L. latifolium took place in seasonally-flooded dry lake bed

wetlands, patterns of expansion from well-drained channel edges into otherwise Sarcocornia-dominated

3

saline marsh plane have been documented in several marshes (Reynolds and Boyer 2010). In seasonal

wetlands, areas with wetter soils have higher potential for L. latifolium infestation than areas with drier

soils (Renz et al. 2012). This pattern supports the hypothesis that L. latifolium is able to spread into

more stressful areas farther from channels by clonal integration (Reynolds and Boyer 2010). Changes in

soil chemistry may also reduce the ability of native plants to compete with L. latifolium. L. latifolium

roots increases the availability of soil nitrogen by producing nitrogen-cleaving enzymes (Blank 2002).

Over time, this leads to reduced availability of phosphorus in soil, which can reduce growth of L.

latifolium as well as other species (Blank et al. 2002).

Several studies have recognized the need for and difficulty of establishing effective methods of

eliminating L. latifolium from tidal marshes (e.g. Boyer and Burdick 2010; Whitcraft and Grewell 2012)

and rangelands (e.g., Young et al. 1998). Land managers commonly employ hand removal if they have

access to many volunteers; however, hand removal is often impractical for several reasons. First, the

time involved is prohibitive for areas with small staffs and few volunteers. Second, L. latifolium’s

extensive root system makes removal problematic. Not only do the plants spread horizontally by

rhizomes that are easily fragmented during hand removal (Francis and Warwick 2007), but the tap roots

of pepperweed can also extend more than three meters into the soil (Blank et al. 2002). Mowing plus

herbicides has been effective in rangelands (Renz and DiTomaso 2004), but this is impractical for coastal

wetlands because saturated wetland soils do not support heavy mowing equipment. Soil disturbance

can facilitate the spread of L. latifolium. The rate of expansion of L. latifolium patches in areas that are

disked is 0.85-m per year, approximately three times greater than in areas that are not disked (Renz et

al. 2012). Herbicides are the most common and effective means of controlling pepperweed in wetlands.

Although several herbicides can reduce or eliminate L. latifolium in uplands, imazapyr has been shown

to be more effective for controlling L. latifolium in marshes than other herbicides such as 2,4-D

(Whitcraft and Grewell 2012) and glyphosate (Boyer and Burdick 2010). Imazapyr has also been shown

to be more effective than glyphosate for controlling other invasive plant species, such as smooth

cordgrass (Spartina alterniflora) in areas that are frequently inundated by tides (Patten 2002).

Even the most effective herbicides may not consistently remove every plant, and results may vary based

on elevation and habitat type (Boyer and Burdick 2010). Previous studies of the effectiveness of

herbicides on L. latifolium in wetlands have focused on tidal marsh. In a disturbed formerly tidal area,

imazapyr reduced L. latifolium cover by more than 90% after a single year of treatment (Whitcraft and

Grewell 2012). Glyphosate alone is more effective for controlling L. latifolium in tidal marsh

environments with higher salinities (Spenst 2006). Following hand removal of pepperweed plants from

brackish marshes in the San Francisco Bay, application of imazapyr was more effective than application

of glyphosate for controlling regrowth (Boyer and Burdick 2010). The negative impacts of imazapyr on

native cover are far more persistent (up to 2 – 3 years) in higher elevation habitats with less flooding

(Whitcraft and Grewell 2012). In contrast, in estuarine conditions imazapyr decays exponentially and

areas where it has been used to eradicate non-native plants can be colonized by Salicornia (Patten

2003). Glyphosate is much more persistent in estuarine sediments, with a half-life of 119 days as

compared to the 1.6 day half-life of imazapyr (Paveglio et al. 1996). Imazapyr only adsorbs to acidic soils

(pH< 5) and it is not retained by soils with higher pH because it becomes ionized (Pless 2005). Uptake of

4

glyphosate from soil is limited but some root uptake may occur depending upon soil characteristics;

plants mainly take up glyphosate through leaves (Pless 2005).

The purpose of this experiment was to test the efficacy of the herbicide imazapyr alone and in

combination with glyphosate in controlling L. latifolium in brackish tidal marsh habitats. We also

examine changes in non-target vegetation that result from herbicide application. Results of this study

can be used to inform control and eradication of invasive L. latifolium in brackish marshes.

METHODS

Site Description

San Pablo Bay National Wildlife Refuge (38.1297°N, 122.3931°W) is on the northern coast of San Pablo

Bay, which is part of the San Francisco Bay Estuary. Water in San Pablo Bay is brackish to saline, and

salinity varies depending on tides, freshwater flow from the Sacramento Delta, and weather. A set of 36

plots was established during May 2007 in the lower reaches of Tolay Creek, divided evenly across three

habitat types: bay edge, channel edge, and adjacent to levees. Plots were randomly placed within

known patches of L. latifolium. Each plot was 4 m x 4 m in size and a 1 m x 1 m subplot, within which

measurements were taken, was placed in the center of each plot. Plots were marked with pin flags and

wooden stakes (1.32 m). One meter subplots were marked with stakes and GPS coordinates were

recorded at the plot center (average of 30 readings, UTM NAD83).

Treatments

We applied one of three treatments to each 4 m x 4 m plots and any pepperweed plants within a buffer

of at least 1.5 m of the plots to reduce edge effects. The three treatments were: flower head removal

(control), imazapyr (trade name Habitat; I), or imazapyr mixed with glyphosate (trade name Rodeo; I+G).

Herbicide was applied with a hand sprayer in late May 2007 and repeated in late May 2008. Application

rates followed prescriptions in the San Pablo Bay National Wildlife Refuge Lepidium latifolium Control

Plan (p. 17; Hogl et al. 2007). For the I treatment, imazapyr was mixed at a concentration of 0.25% with

added surfactants. For the I+G treatment, a 1.5% solution of glyphosate was mixed with a 0.25%

solution of imazapyr with surfactants. Applications were conducted on ebb tides to allow sufficient

drying time and on days with calm wind to reduce the potential for spraying non-target species.

Inflorescences in control plots were removed on May 8, 2007, prior to seed set. While inflorescence

removal may not be seen as a control in the sense that there were no plots where no actions were

taken, it was necessary to prevent dispersal by seeds to areas outside the experiment plots, where they

could potentially start new infestations. Removing flower heads from L. latifolium plant prevented them

from dispersing seeds outside the plots, but because pepperweed primarily spreads by rhizomes in

wetlands (Pennings and Callaway 2000; Francis and Warwick 2007), removal of flower heads should not

impact the ability to spread within the plots. The removal of inflorescences is a control in the sense that

it represents a neutral, non-herbicide treatment for comparison that introduces minimal physical impact

on pepperweed plants and no impact on non-target native species. Although the intent was to have an

5

equal number of plots among treatment types, treatment crew error resulted in an unbalanced design

for the herbicide treatments (Table 1).

Measurements

Number of L. latifolium stems was counted and percent foliar cover of all other plant species present

was visually estimated for each 1 m x 1 m subplot in May 2007 prior to treatment application. L.

latifolium was not included in the percent cover measurements. The same measurements were

repeated in May 2008 and May 2009 on the same 1 m2 subplot as the pre-measurement.

Statistical Analysis

We used generalized linear mixed model (lmer in package lme4 in R; R Core Team 2014, Bates et al

2014) to model the effects of time, environment, and treatment type on the number of L. latifolium

stems in a plot (stems + rosettes) and separately on percent cover. We chose this package because it

can be used for repeated measures designs with count data (family = poisson). A poisson distribution is

more appropriate than the normal for count data because counts cannot be negative and tend to have a

long right tail. Incorporating random effects allowed us to account for the repeated measures design in

the model. The intercept as well as the linear and quadratic effects of time for each plot vary about the

estimates of the fixed effects. This allowed the model to take into account that each plot started with a

different number of stems and that stem counts over time, within a plot are likely to be correlated.

Years were standardized and centered on 0 to ensure orthogonality of linear and quadratic effects of

time. We used a backwards model selection with likelihood ratio tests and Wald tests of fixed effects

(Anova in package car, Fox and Weisberg 2011) to guide variable removal. The initial (full) model

contained treatment, year (linear and quadratic), environment, and the two-way interactions among

those factors as well as plot random effects for intercept and linear and quadratic coefficients for year.

RESULTS

L. latifolium Stem Counts

Treatment type significantly affected the number of L. latifolium stems over time, but the type of

environment did not (Type III Wald test of fixed effects: Treatment x Year – X2=27.95, df=2, p<0.001;

Treatment x Year2 – X2=10.27, df=2, p=0.006), after controlling for the effects of time on individual plots

(variances of random effects: Intercept – 1.07; Year – 0.60; Year2 – 0.50). While the number of stems

increased in the control plots, the number of stems decreased in plots where herbicides were applied

(Figure 1). By 2009, virtually no L. latifolium stems remained in the plots that were sprayed with either

herbicide treatment.

There was a steep decline in the number of stems between the first two years and the less dramatic

decline between the second two years for each treatment. The significant interaction between the

squared year and the treatments accounts for this change in slopes. The combination of I+G reduced

6

the number of stems faster, but by the second year of spraying, both the I and I+G treatments reduced

the number of stems to zero.

Percent Cover

Percent cover varied significantly by the interaction of treatment type and environment (Type III Wald

X2= 19.9, df=4, p<0.001), after accounting for the effects of time on individual plots (variances of random

effects: Intercept – 0.07; Year – 0.15; Year2 – 0.25).

The most dominant native species, pickleweed (Sarcocornia pacifica; SAPA), declined with the

combination I+G treatment (Figure 2). The second most common native species, Frankenia salinia

(FRSA), declined with both the I and I+G treatments. Dead organic matter also covered a larger

percentage of plots that were treated with herbicides over time, while it was rarely present in control

plots.

Within each environment, there was a higher percentage of bare ground in plots where both herbicides

were applied than other treatments in the third year of the experiment (Figure 3). There was essentially

no bare ground in the control plots in 2009 and no bare ground was reported for any plots in the first

two years of the study. There was less bare ground in bay plots treated with both herbicides than the

other environments that were also treated with both herbicides. There was a large amount of variation

in the amount of bare ground, in all environments and both herbicide treatments.

DISCUSSION

Overall, treatment of L. latifolium infestations with imazapyr alone was a better choice than imazapyr +

glyphosate in tidal marsh habitats. Over two years of treatment, imazapyr alone achieved similar levels

of control to the I+G treatment, but without the side effect of reduced cover of native species.

Managers should particularly consider the variable effect of herbicides on native vegetation among

various types of habitat within tidal marshes.

The nearly complete removal of L. latifolium stems where herbicides were applied indicates that

imazapyr alone or in combination with glyphosate is an effective means of removing L. latifolium from

tidal marsh habitats. The steep decline in the number of L. latifolium stems from the first to the second

year shows that the herbicide is generally effective. The shallower slope between years two and three

reflects the fact that there were relatively few stems remaining after the initial treatment. Increases in

the number of L. latifolium stems in control plots are consistent with previous studies where unsprayed

L. latifolium patches expanded (Renz et al. 2012). Removing flower heads from L. latifolium plant

prevents them from setting seeds, but the primary method of expansion in wetlands is vegetative

growth of rhizomes (Pennings and Callaway 2000; Francis and Warwick 2007). The lack of significant

effect of type of environment on the number of L. latifolium stems indicates that the treatments are

similarly effective in bay edge, channel edge, and levee-adjacent environments. Treatment with

imazapyr alone in two consecutive years reduced the number of L. latifolium stems as effectively as a

7

single year of application of the imazapyr + glyphosate treatment. This is important for management

because it shows that adding glyphosate is not necessary for controlling L. latifolium. Similarly, imazapyr

alone achieved better control than glyphosate alone and similar levels of control to the combination of

imazapyr and glyphosate for West Indian marsh grass (Hymenachne amplexicaulis) in Florida (Sellers et

al. 2008) and reed canary grass (Phalaris arundinacea) in South Dakota (Bahm et al. 2014). Although the

combination of imazapyr with glyphosate was not tested, imazapyr alone has been shown to be more

effective than glyphosate for controlling invasive smooth cordgrass (Spartina alterniflora; Roberts and

Pullin 2008). Considering these studies together, treatment with imazapyr alone appears to be a good

choice when developing plans to control multiple invasive species in wetland habitats.

Managers should weigh side effects of each herbicide treatment before choosing a plan: although

imazapyr alone required two years of treatment to achieve the same level of control for L. latifolium as a

single year of treatment with imazapyr + glyphosate, the combination of imazapyr and glyphosate had

larger and more immediate impacts on non-target native species. The quick reduction of L. latifolium

achieved by the combination of imazapyr and glyphosate may come at the expense of causing a

reduction in native plant cover and increase in bare ground. Increased bare ground is likely to be

temporary, but slow regrowth of native species may create an opportunity for invasion by other species

or reinvasion by L. latifolium. In a similar study, although imazapyr alone was more effective for

reducing pepperweed in brackish marshes than glyphosate alone, sites sprayed with imazapyr showed

no signs of native plant recovery after one year while sites treated with glyphosate showed increased

native plant cover (Boyer and Burdick 2010). The combination of our results with those of Boyer and

Burdick (2010) suggest that if resources exist to allow multiple applications, the imazapyr treatment

should be preferred over a single application of imazapyr + glyphosate or glyphosate alone. While any

treatment including imazapyr seems to produce negative impacts on native vegetation in the short

term, the nearly complete removal of pepperweed with two applications will prevent spread to other

sensitive areas. Managers should also consider the type of habitat when choosing a management plan

for L. latifolium. Although the I and I+G treatments control L. latifolium well, their effects on native

plants vary by environment. The negative impacts of the both treatments on native species were

ameliorated by the bay edge environment.

Although no previous study has tested the effects of the combination of imazapyr and glyphosate on L.

latifolium in tidal marshes, each herbicide alone has been shown to reduce native recovery following L.

latifolium removal. Imazapyr has previously been shown to reduce L. latifolium cover drastically in a

single year of application, but persistent impacts on native plants were observed for the two years of

the study (Whitcraft and Grewell 2012). Similar results were shown in a previous study where plots

sprayed with imazapyr alone showed no signs of native recovery after one year, and areas sprayed with

glyphosate alone allowed slow recovery by S. pacifica (Boyer and Burdick 2010). Another study of the

effectiveness of imazapyr and glyphosate applied independently to invasive cordgrass found no

indication of limited recolonization of natives after applying herbicide and found that recolonization of

Salicornia spp. was greater for imazapyr-treated areas than for glyphosate-treated areas (Patten 2003).

In our study plots, declines in S. pacifica were most severe in plots treated with I+G and S. pacifica was

replaced by litter and dead organic matter. In contrast, treatment with imazapyr alone showed no

8

persistent impacts on S. pacifica in channel and levee habitats in our study. In fact, percent cover of S.

pacifica doubled in levee habitats treated with imazapyr alone. It should be noted, however, that levee

plots in our study were placed adjacent to levees, rather than on top of levees, and thus received some

tidal flushing. In both the H. amplexicaulis (Sellers et al. 2008) and P. arundinacea (Bahm et al. 2014)

studies, lack of establishment of native vegetation following imazapyr treatments left openings for re-

establishment of non-natives.

Many studies, like ours, have investigated the impacts of control plans that include herbicides on the

most common native species, but additional work focusing on how herbicides affect the native plant

community are also needed. Additional studies are needed to determine long-term impacts of imazapyr

on native tidal marsh vegetation. Although bare ground that resulted from herbicide treatments was

persistent throughout this study, the relatively short study period did not allow for observations of

longer-term colonization. Studies of long-term impacts could include potential methods for reducing

impacts of herbicide persistence in drier environments such as tops of levees and roadsides.

ACKNOWLEDGEMENTS

This experiment was part of a set of studies on perennial pepperweed control at San Pablo Bay National

Wildlife Refuge. Funding for these studies was provided by U. S. Fish and Wildlife Service Inventory and

Monitoring Program, National Fish and Wildlife Foundation, California Department of Fish and Wildlife,

U.S. Fish and Wildlife Service Invasives Program, U.S. Fish and Wildlife Service Coastal Program, and the

Marin-Sonoma Mosquito and Vector Control District. We also wish to thank our partners, Point Blue

(formerly Point Reyes Bird Observatory)’s Students and Teachers Restoring a Watershed Program,

Sonoma Land Trust, Friends of San Pablo Bay NWR, Renee Spenst (Ducks Unlimited), Ingrid Hogle

(Invasive Spartina Project), and Shelterbelt Builders.

REFERENCES

Bahm MA, Barnes TG, Jensen KC (2014) Evaluation of Herbicides for Control of Reed Canarygrass

(Phalaris arundinacea). Natural Areas Journal 34(4):459-464.

Bates D, Maechler M, Bolker B, Walker S (2014) lme4: Linear mixed-effects models using Eigen and S4. R

package version 1.1-6. http://CRAN.R-project.org/package=lme4

Blank RR (2002) Amidohyrolase activity, soil N status, and the invasive cruicifer Lepidium latifolium. Plant

and Soil 239: 155-163.

Blank RR, Qualls RG, Young JA (2002) Lepidium latifolium: plant nutrient competition-soil interactions.

Biology and Fertility of Soils 35: 458-464.

Blank RR, Young JA (2004) Influence of three weed species on soil nutrient dynamics. Soil Science 169(5):

385-397.

9

Boyer KE, Burdick AP (2010) Control of Lepidium latifolium (perennial pepperweed) and recovery of

native plants in tidal marshes of the San Francisco Estuary. Wetlands Ecology and Management 18: 731-

743.

Fiedler PL, Keever M, Grewell BJ, Partridge DJ (2007) Rare plants in the Golden Gate Estuary (California):

the relationship between scale and understanding. Australian Journal of Botany 55:206–220

Fox J, Weisberg S (2011) An {R} Companion to Applied Regression, Second Edition. Thousand Oaks CA:

Sage. URL: http://socserv.socsci.mcmaster.ca/jfox/Books/Companion

Francis A, Warwick SI (2007) The biology of invasive alien plants in Canada 8. Lepidium latifolium L.

Canadian Journal of Plant Science 87: 639-658.

Goals Project (1999) Baylands Ecosystem Habitat Goals. A report of habitat recommendations prepared

by the San Francisco Bay Area Wetlands Ecosystem Goals Project. U.S. Environmental Protection Agency,

San Francisco, California. Bay Regional WaterQuality Control Board, Oakland, CA.

Grossinger R, Alexander J, Cohen AN, Collins JN (1998) Introduced Tidal Marsh Plants in the San

Francisco Estuary. Richmond, CA: San Francisco Estuary Institute.

Hogle I, Spenst R, Leininger S, Block G (2007) San Pablo Bay National Wildlife Refuge Lepidium latifolium

Control Plan. U. S. Fish and Wildlife Service, San Pablo Bay National Wildlife Refuge, Petaluma, Calif.

http://www.fws.gov/invasives/staffTrainingModule/pdfs/planning/SPBNWR_Control_Plan_061807.pdf

Patten K (2002) Smooth cordgrass (Spartina alterniflora) control with imazapyr. Weed Technology

16(4):826-832.

Patten K (2003) Persistence and non-target impacts of imazapyr associated with smooth cordgrass

control in an estuary. Journal of Aquatic Plant Management 41:1-6.

Paveglio FL, Kilbride KM, Grue CE, Simenstad CA, Fresh KL (1996) Use of Rodeo and X-77 spreader to

control smooth cordgrass (Spartina alterniflora) in a southwestern Washington estuary: 1.

environmental fate. Environmental Toxicology and Chemistry 15(6):961-968

Pennings SC, Callaway RM (2000) The advantages of clonal integration under different ecological

conditions: a community-wide test. Ecology 81(3): 709-716.

Pless P (2005) Use of imazapyr herbicide to control invasive cordgrass (Spartina spp.) in the San

Francisco Estuary: water quality, biological resources, and human health and safety. Berkely, CA: Leson

& Associates.

R Core Team (2014) R: A language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, Austria. http://www.R-project.org.

Renz MJ, Blank RR (2004) Influence of Perennial Pepperweed (Lepidium latifolium) Biology and Plant–

Soil Relationships on Management and Restoration. Weed Technology 18(sp 1): 1359-1363.

10

Renz MJ, Steinmaus SJ, Gilmer DS, DiTomaso JM (2012) Spread dynamics of Perennial Pepperweed

(Lepidium latifolium) in two seasonal wetland areas. Invasive Plant Science and Management 5:57-68.

Renz MJ, DiTomaso JM (2004) Mechanism for the Enhanced Effect of Mowing Followed by Glyphosate

Application toResprouts of Perennial Pepperweed (Lepidium latifolium). Weed Science 52(1): 14-23.

Reynolds LK, Boyer KE (2010) Perennial pepperweed (Lepidium latifolium): properties of invaded tidal

marshes. Invasive Plant Science and Management 3: 130-138.

Robbins WW, Bellue MK, Ball MS (1941) Weeds of California. California Department of Agriculture,

Sacramento, CA.

Roberts PD, Pullin AS (2008) The effectiveness of management interactions for the control of Spartina

species: a systematic review and meta-analysis. Aquatic Conservation: Marine and Freshwater

Ecosystems 18: 592-618.

Sellers BA, Diaz R, Overholt WA, Langeland KA, Gray CJ (2008) Control of West Indian marsh grass with

glyphate and imazapyr. Journal of Aquatic Plant management 46: 189-192.

Spautz H, Nur N (2004) Impacts of non-native perennial pepperweed (Lepidium latifolium) on

abundance, distribution, and reproductive success of San Francisco Bay tidal marsh birds. Pt Reyes Bird

Obs report to US FWS. Sausalito, CA

Spenst RO (2006) The Biology and Ecology of Lepidium latifolium L. in the San Francisco Estuary and their

Implications for Eradication of this Invasive Weed. Dissertation thesis, University of California, Davis

Davis, CA.

Trumbo, J (1994) Perennial pepperweed: a threat to wildland areas. California Exotic Pest Plant Council

Newsletter 2(3):4–5.

USDA, ARS, National Genetic Resources Program (2014) Germplasm Resources Information Network -

(GRIN) [Online Database]. National Germplasm Resources Laboratory, Beltsville, Maryland. URL:

http://www.ars-grin.gov/cgi-bin/npgs/html/tax_search.pl?Lepidium+latifolium (accessed 11 July 2014)

U.S. Fish and Wildlife Service (2013) Recovery Plan for Tidal Marsh Ecosystems of Northern and Central

California. Sacramento, California. xviii+605 pp. URL:

http://www.pacific.fws.gov/ecoservices/endangered/recovery/plans.html

Vanderhoof M, Holzman BA, Rogers C (2009) Predicting the distribution of perennial pepperweed

(Lepidium latifolium), San Francisco Bay, California. Invasive Plant Science and Management 2(3):260-

269. DOI: 10.1614/IPSM-09-005.1

Vasey MC, Parker VT, Callaway JC, Herbert ER, Schile LM (2012) Tidal wetland vegetation in the San

Francisco Bay-Delta Estuary. San Francisco Estuary and Watershed Science 10(2): 1-15.

11

Whitcraft CR, Grewell BJ (2012) Evaluation of perennial pepperweed (Lepidium latifolium) management

in a seasonal wetland in the San Francisco Estuary prior to restoration of tidal hydrology. Wetlands

Ecology and Management 20:35-45.

Young JA, Palmquist DE, Blank RR (1998) The ecology and control of perennial pepperweed (Lepidium

latifolium L.). Weed Technology 12: 402-405.

12

Table 1 Number of plots that received each herbicide treatment within the three types of environment.

Environment

Bay Channel Levee Total

Trea

tmen

t Control 6 5 4 15

I + G 3 5 3 11

I 3 2 5 10

Total 12 12 12 36

Figures:

Fig. 1 Mean number of total LELA stems (with 95% confidence intervals) by treatment and year. C=control, IG=Imazapyr+

Glyphosate, I=Imazapyr only.

13

Fig. 2 Mean percent cover (with 95% confidence intervals) of the two most dominant native species in experimental plots

(SAPA=Sarcocornia pacifica and FRSA=Frankenia salinia) and litter and dead organic matter by treatment and environment

over time. C=control, IG=Imazapyr+ Glyphosate, I=Imazapyr only.

14

Fig 3 Effects of environment and treatment type on percent bare ground (with 95% confidence intervals) in a plot for

2009 only. Graphs are not shown for 2007 or 2008 because bare ground was not present in field plots during these

years. C=control, IG=Imazapyr+ Glyphosate, I=Imazapyr only.