C t l f b b d t (A il t i i li ) i ti l dControl of barbed goatgrass (Aegilops triuncialis) in serpentine grasslandsControl of barbed goatgrass (Aegilops triuncialis) in serpentine grasslandsg g ( g p ) p gPaul A Aigner and Rhett J Woerly University of California Davis Donald and Sylvia McLaughlin Reserve Lower Lake CAPaul A. Aigner and Rhett J. Woerly, University of California Davis, Donald and Sylvia McLaughlin Reserve, Lower Lake, CA

ff i f bSerpentine grasslands are known for being relatively free of the invasive species that typify other  Effects on native forbsCalifornia grasslands.   Recently the special status of serpentine grasslands as strongholds of native plant diversity has been threatened by the spread of barbed goatgrass (Aegilops triuncialis), a Eurasian annual  Native forbs were especially diverse in the study area, with 51 species recorded across all plots, and ay y p g g ( g p ),that is unique in its tolerance of serpentine soils (Plate 1). Carefully timed prescribed fire, especially

Native forbs were especially diverse in the study area, with 51 species recorded across all plots, and a median of 13 species per plot As a group native forbs increased in frequency dramatically on hand‐that is unique in its tolerance of serpentine soils (Plate 1).  Carefully timed prescribed fire, especially 

applied over two consecutive years has been shown to reduce goatgrass while benefitting many nativemedian of 13 species per plot.  As a group, native forbs increased in frequency dramatically on hand‐pulled early season Fusilade and early season Envoy plots (Fig 3A) All other treatments showed aapplied over two consecutive years, has  been shown to reduce goatgrass while benefitting many native 

species (DiTomaso et al 2001 ) At the McLaughlin Reserve in the inner North Coast Range of Californiapulled, early‐season Fusilade, and early‐season Envoy plots (Fig. 3A).  All other treatments showed a trend toward increasing native forb frequency but none differed statistically from controls Of thespecies (DiTomaso et al. 2001 ).  At the McLaughlin Reserve in the inner North Coast Range of California, 

we found prescribed fire cannot be applied for two consecutive years because the low productivity oftrend toward increasing native forb frequency, but none differed statistically from controls.    Of the i ht f b i ith ll l t f 1% t h d i ith lwe found prescribed fire cannot be applied for two consecutive years, because the low productivity of 

ti l d t d t f l b ild i th d S ti l deight forb species with an average cover across all plots of >1%, two showed increases with early‐

serpentine grasslands prevents adequate fuel build up in the second year.  Some serpentine grasslands d i h fi ill ll

season application of grass specific herbicides (Fig. 3B&C).  Navarretia jepsonii, a serpentine endemic are so unproductive that fire will not carry at all.   annual, had higher cover on early‐season Fusilade plots compared to  controls, mowing, and all of the 

M ti t d b th d f f ll / lt ti t t t t fi l t d i t t t flate‐season  herbicide plots (Fusilade, Envoy, and Roundup).  N. jepsonniimay have been negatively 

Motivated by the need for a follow up/alternative treatment to fire, we evaluated nine treatments for p ( , y, p) j p y g y

affected by Roundup, as it was still actively growing at the time of application. Holocarpha virgata, a their effectiveness controlling barbed goatgrass and for restoring native species cover and diversity:  

y p, y g g pp p g ,summer annual, had higher cover on early‐season Fusilade and Envoy plots relative to controls.

two grass‐specific herbicides —fluazifop [Fusilade II ] and clethodim [Envoy +]—applied before summer annual, had higher cover on early season Fusilade and Envoy plots relative to controls.   

goatgrass flowering (early‐season), at flower initiation (mid‐season), and at early seed development  80.5 6g g g ( y ), ( ), y p(late‐season), a broad‐spectrum herbicide (glyphosate) applied late‐season, mowing , and hand pulling.    ve

r a2009Plate 3 Control plot summer 2009 Plate 4 Early‐season herbicide application March erA B C( ), p (g yp ) pp , g , p g

cov

0.4

cy

20092010

Plate 3.  Control plot , summer 2009. Plate 4.  Early‐season herbicide application, March 2009 co

v aA B C

h d6en

t a

0 3 enc 20102009.

ent c

MethodsEff t er

ce

ab0.3 0.3

eque 4

erce

Effects on  a pe

ab

b fre ii pe ab

In fall of 2007 we established a grid of 100, 2m x 2m plots  with 2‐m wide buffers between plots in a  goatgrass4

rgat

a

0.2cy

0.2forb

son

g , p pserpentine grassland with high cover of barbed goatgrass but few other non‐native species (Plates 2, 3). goatgrass

a vi

r

ab

0.2

nc 2009 e in

f

jeps abcserpentine grassland with high cover of barbed goatgrass but few other non native species  (Plates 2, 3).   

We counted goatgrass tillers within a 0 25‐m2 frame in the center of each plot and stratified the rpha ab

ab

0 1en

20092010

0.1

nge 2

etia

j

abcWe counted goatgrass tillers within a 0.25 m frame in the center of each plot, and stratified the experiment into 10 blocks based on tiller density The nine treatments and a non‐treatment control After two years of treatment ,  2

ocar0.1

que 2010

chan

arre abc abc

experiment into 10 blocks based on tiller density.  The nine treatments and a non‐treatment control were assigned randomly within each block Treatments were applied in spring 2008 and 2009 In early all treatments except for 

Hol

o ab ab ab ab beq

0.0c

Nav

a

bc bc bc bcwere assigned randomly within each block. Treatments were applied in spring 2008 and 2009.  In early M h 2008 2009 d 2010 ( i t th li ti f t t t i 2008 d 2009) d th

pRoundup and early‐season  0

H b0.0fre 0 1 0

N bc bc c

March 2008, 2009, and 2010 (prior to the application of treatments in 2008 and 2009) we measured the f f b b d i f b i b h ( l l i l ll l h

p yEnvoy resulted in reduced Env

EFus

EHan

dEnv

MEnv

LFus

MFus

LRou

ndMowon

trol

0s f

Hand

FusE

EnvE

usM

EnvL

nvM

Mow ound

FusL

ntrol

-0.1

FusE

FusM

Hand

EnvM

EnvE

FusL

EnvL

ontro

lMow

Round

0

frequency of barbed goatgrass, native forbs, native bunchgrasses (almost exclusively Nassella pulchra Envoy resulted in reduced goatgrass frequency relative to

En Fu Ha En E Fu F Rou MCon

treatment0 1ass Ha Fu En Fus En

Env MoRou FuCon

t

t t t

Fu Fu Ha En En F E

Con M Rou

treatmentand Poa secunda), and other annual grasses  (Bromus hordeaceus and Vulpia microstachys) in 25, 27‐mm  goatgrass frequency relative to controls (Fig 1) Mid season

treatment-0.1ra treatment treatmentFigure 3 (A) Change in native forb frequency after one and two years of treatment Error bars represent SEM

diameter circular subplots within each plot.  In early May 2010, we made visual estimates of cover of all  controls (Fig. 1). Mid‐season Fusilade hand pulling and tg Figure 3.  (A) Change in native forb frequency after one and two years of treatment.  Error bars represent SEM.  

Solid lines below the bars group treatments that are not significantly different in 2010 by Tukey’s HSD test atspecies within a 1‐m2 frame at the center of each plot.   Fusilade, hand pulling, and 

l F il d h d-0.2oa

Solid lines below the bars group treatments that are not significantly different in 2010 by Tukey s HSD test at α=0 05 (B & C) Percent cover of two native forbs after two years of treatment Error bars represents SEMp p

early‐season Fusilade showed  go

α=0.05.  (B & C) Percent cover of two native forbs after two years of treatment.  Error bars represents SEM.  Letters above the bars group treatments that are not significantly different by Tukey’s HSD test at α=0 05the best performance, reducing 

0 3n g Letters above the bars group treatments that are not significantly different by Tukey s HSD test at α=0.05. 

Abbreviations: EnvE Envoy early season; EnvM Envoy mid season; EnvL Envoy late season; FusE Fusilade earlygoatgrass frequency by 66%,  -0.3 in Abbreviations: EnvE, Envoy early season; EnvM, Envoy mid season; EnvL, Envoy late season; FusE, Fusilade early 

season; FusM Fusilade mid season; FusL Fusilade late season; Hand hand pull; Round Roundup62% , and 52% respectively  geBarbed goatgrass

season; FusM, Fusilade mid season; FusL, Fusilade late season; Hand, hand pull; Round, Roundup., p y

from the pre‐treatment  -0.4ng Effects on other non nativesg g

pfrequency of about 0.7. Early‐ an Effects on other non‐nativesfrequency of about 0.7.  Earlyseason Envoy and Roundup -0 5chseason Envoy and Roundup produced no significant

0.5c Besides goatgrass, the only non‐native species recorded in the plots were the annual grass, Bromusproduced no significant reduction relative to controls 0 6

hordeaceus, and the annual forb, Lactuca serriola, and both species were a minor component of the reduction relative to controls.  E l li ti f M d E w M L L E d l

-0.6 community with an average of 0.4% and 0.06% cover, respectively.  B. hordeaceus increased on hand‐Early‐season application of 

FusM

Hand

FusE

MowEnv

MFus

LEnv

LEnv

Eou

nd ntrol

y g , p ypulled plots , late‐season  Envoy plots, and Roundup plots relative  to plots that were treated early with 

Envoy resulted in damage to  Fus Ha Fu M Env Fu En EnRouCon

t p p , y p , p p p ygrass‐specific herbicide (data not shown). B. hordeaceus was susceptible to all three herbicides, but

goatgrass, followed by partial 

C

treatmentgrass specific herbicide (data not shown).   B. hordeaceus was susceptible to all three herbicides, but sets seed in April and May so it was unaffected by the late season applications

recovery.  Late‐season  treatment sets seed  in April and May, so it was unaffected by the late season applications.y

application of all herbicides  Figure 1.  Change in barbed goatgrass frequency after one and two ppfailed to prevent at least some

g g g g q yyears of treatment.  Error bars represent SEM.  Solid lines below the failed to prevent at least some 

viable seed productiony pbars group treatments that are not significantly different in 2010 by viable seed production. g p g y yTukey’s HSD test at α=0.05.  Abbreviations: EnvE, Envoy early season; 

Plate 1.  Barbed goatgrass invasion front in a  Plate 2. Experimental plots in summer 2009, after the y , y y ;

EnvM, Envoy mid season; EnvL, Envoy late season; FusE, Fusiladeserpentine grassland at the McLaughlin Reserve  first season of treatment

, y ; , y ; ,early season; FusM, Fusilade mid season; FusL, Fusilade late season; y ; , ; , ;Hand, hand pull; Round, Roundup.

Eff t ti, p ; , p

Early‐season herbicide treatments were applied in mid‐March, three to four weeks before flower  Effects on native grassesinitiation. Mid‐season herbicide treatments were applied in early May as inflorescences were beginning 

Both grass specific herbicides caused damage to native perennial grasses but virtually all of thesepp y y g g

to extend from the sheath.  Late‐season herbicide treatments were applied at the end of May, when all Both grass‐specific herbicides caused damage to native perennial grasses, but virtually all of these 

d ft th d f li ti L t E lt d i ipp y,

goatgrass had flowered, and most were showing signs of early seed development (a small, soft, greengrasses recovered, even after the second year of application.  Late‐season  Envoy resulted in an increase 

goatgrass had flowered, and most were showing signs of early seed development (a small, soft, green embryo was apparent) Mowing occurred at the peak of flowering in late May before any seed

in total bunchgrass frequency relative to controls (data not shown), whereas all other treatments embryo was apparent). Mowing occurred at the peak of flowering in late May before any seed development was evident and was done with gas powered string trimmers at a height of approximately

produced changes that were statistically indistinguishable from controls.  There was no suggestion that development was evident, and was done with gas‐powered string trimmers at a height of approximately 2 cm Hand pulling occurred in mid May

bunchgrass frequency was reduced by any treatment, with the possible exception of early‐season 2 cm . Hand pulling occurred in mid‐May.

g q y y y , p p yEnvoy.  Individually, neither N. pulchra nor P. secunda was negatively affected by treatments (Fig. 2,  Plate 6 Early‐season Fusilade plot in early summerPlate 5 Late season Envoy plot after two years ofy y, p g y y ( g ,Plate 5), except that N. pulchra was completely eliminated by Roundup (P. secunda escaped this effect

Plate 6.  Early‐season Fusilade plot in early summer 2009 showing growth of N jepsonii and H virgata

Plate 5.  Late season Envoy plot after two years of treatment showing persistence of N pulchra

Herbicides were applied using a spray to wet technique with a backpack sprayer using the Plate 5), except that N. pulchra was completely eliminated by Roundup (P. secunda escaped this effect because it was dormant during the late‐season herbicide applications) The native annual grass V

2009 showing growth of N. jepsonii and H. virgata.treatment showing persistence of N. pulchra.pp g p y q p p y g

recommended spot‐spray rate for Fusilade II, the near‐maximum spot‐spray rate for Envoy +, and abecause it was dormant during the late‐season herbicide applications).  The native annual grass, V. microstachys also was not negatively impacted by any of the treatments (Fig 2) For V microstachys Conclusionsrecommended spot spray rate for Fusilade II, the near maximum spot spray rate for Envoy  , and a 

1 6% solution for Roundup (Table 1 Plate 4) Envoy and Fusilade were mixed with 0 33 oz/gal of amicrostachys, also was not negatively impacted by any of the treatments (Fig. 2).  For V. microstachys, hand pulling produced a benefit relative to early application of Envoy but neither treatment differed

Conclusions1.6% solution for Roundup (Table 1, Plate 4).   Envoy and Fusilade were mixed with 0.33 oz/gal of a modified vegetable oil/silicone surfactant (Syl‐tac)

hand pulling produced a benefit relative to early‐application of Envoy, but neither treatment differed f t lmodified vegetable oil/silicone surfactant (Syl‐tac).

4from controls.

1015 • Fusilade applied early or mid‐season and Envoy applied mid‐season controls barbed goatgrass, 4

a

10

ver a

15

r

pp y y pp g g ,while sparing native perennial grasses and forbs. Early‐season Envoy failed to control goatgrass.

er

a

8cov

over

while sparing native perennial grasses and forbs. Early season Envoy failed to control goatgrass.Table 1. Application rates used for herbicide treatments.

3cov

a8

ent c

nt c

o • Application of grass‐specific herbicides early in the season provides the most benefit to native Table 1.   Application rates used for herbicide treatments.

Herbicide Mix concentration Approximate per‐acre rate ent c a

aa er

ce10cen forbs (probably because of their early release from competition), but may set back native grasses, 

oz concentrate/gallonpp p

oz concentrate/acre erce a

ab ab 6s pe

abperc

which benefit more from applications later in the season when they have set seed or are nearer to oz concentrate/gallon oz concentrate/acre

E 0 8 64 802a

pe

chys ab

hra

pp ydormancy. Early‐season application can also provide control of other non‐native annual grasses,Envoy + 0.8 64‐80

unda

4stac

abulch

dormancy.  Early season application can also provide control of other non native annual grasses, such as B hordeaceus that mature earlier than goatgrass

F il d II 0 75 60 75

ecu

cros ab ab ab5a pu such as B. hordeaceus, that mature earlier than goatgrass.

Fusilade II 0.75  60‐75 1

oa s ab ab

ab 2a m

i

ab ab ab bsell • Mowing provides acceptable goatgrass control, but less benefit to native forbs.

d

Po ab 2

ulpi

a ab ab

bNas

s

L t li ti f R d f il d t t l t li i t d N l h d d dRoundup Pro 2.1 168‐210  b Vu bN • Late‐season application of Roundup failed to control goatgrass, eliminated N. pulchra, and reduced 

h f i f b h ill i l i h i f li i

und

usL

Mow nvL sM nvM and trol

nvE

usE

0

and

und

usL

nvL sM Mow usE tro

lvM nv

E

0

usL sM usE nvL

and

Mow trol

vM nvE

und

0 the cover of some native forbs that were still actively growing at the time of application.

Roun

Fus Mo

Env Fus Env HanCon

trEnv Fus

t t t

HanRou

nFus Env Fus MoFus

Contr

Env EnvFus Fus Fus Env Han MoCon

trEnv EnvRou

n

• For all treatments significant goatgrass cover remains after two years of treatment Outside oftreatment treatmenttreatment • For all treatments, significant goatgrass cover remains after two years of treatment.  Outside of this study our experience suggests that five to six years of hand pulling is necessary to deplete theReferences Fi 2 P t f th ti ft t f t t t E b t SEM L tt this study, our experience suggests that five to six years of hand pulling is necessary to deplete the 

d b k T id l ti f h bi id i t lti t t t h ldFigure 2.  Percent cover of three native grasses after two years of treatment.  Error bars represent SEM.  Letters b th b t t t th t t i ifi tl diff t b T k ’ HSD t t t 0 05 Abb i ti seed bank.  To avoid selection for herbicide resistance, a multi‐year treatment program should 

b i i h ib d fi d l h bi id i h iDiTomaso J M K L Heise G B Kyser A M Merenlender and R J Keiffer 2001 Carefully timed burningabove the bars group treatments that are not significantly different by Tukey’s HSD test at α=0.05.  Abbreviations: E E E l E M E id E L E l t F E F il d l F M begin with prescribed fire and alternate herbicide use with mowing.DiTomaso, J.M., K.L. Heise, G.B. Kyser, A.M. Merenlender, and R.J. Keiffer. 2001.  Carefully timed burning 

can control barb goatgrass California Agriculture 55(6): 47 53EnvE, Envoy early season; EnvM, Envoy mid season; EnvL, Envoy late season; FusE, Fusilade early season; FusM, F il d id F L F il d l t H d h d ll R d R dcan control barb goatgrass.  California Agriculture 55(6): 47‐53. Fusilade mid season; FusL, Fusilade late season; Hand, hand pull; Round, Roundup.