The Ardeid Newsletter, 2004 ~ Audubon Canyon Ranch

8/9/2019 The Ardeid Newsletter, 2004 ~ Audubon Canyon Ranch

1/16

behavior in

heronries

Common Raven

spatial dimensions

Raven home

ranges

power of rainfall

Livermore

Marsh

a test case

Eliminating

Ehrharta

unseen soil

dwellers

Ehrharta

underground

2 0 0 4

the

Ardeid

Research and

Resource M anagement

at Audubon Canyon Ranch


2/16

2004 the ARDEID page 1

Its as if the ravens are not even there.

Although Common Ravens fly frequent-ly through some heronries in search of

opportunities to prey on eggs and chicks,nesting Great Egrets exhibit little or nodefiance. In fact, they seem to show a com-plete lack of interest or concern. In addi-tion, Great Egrets leave their nestsunguarded as nestlings approach three

weeks of age, several weeks before youngegrets can flyeven though many of thesebroods are taken by patrolling ravens.

New cohorts of herons and egrets con-tinue to disperse from heronries eachsummer and to contribute beautifully tothe life in our marshes and estuaries. Butraven numbers are booming throughoutmost of the San Francisco Bay area (seeThe Ardeid 2001). Whether heron andegret populations will continue to repro-duce adequately under the growing pres-ence of ravens or, alternatively, suffermajor declines in productivity has

become a common concern amongobservers of heronries. The particularthreat posed by ravens is difficult tounderstand fully, because most instancesof nest predation are phantom eventsrevealed only by sudden nest vacancies.

The wisdom of egret behaviors in thepresence of ravens is also mysterious.Their apparent carelessness is consistentamong individuals and structured by along evolution of adaptive responses tocomplex ecological forces, including nestpredation. Scientific attempts to under-stand how birds respond to the risk of

nest predation have failed to discover reli-able ways to predict nest success. There-fore, to learn more about the threat ofgrowing raven populations on heron andegret nesting colonies (see The Ardeid2002), I conducted an investigation ofpredatorrather than preybehaviors(Kelly et al. in review). Collaboratorsincluded ACRs Katie Etienne, JenniferRoth of PRBO, several ACR field biologists,and numerous volunteer observers.Specifically, we asked: under what condi-tions and to what extent do ravens exploitheronries for food?

Measuring every move

From 1999 through 2003, AudubonCanyon Ranch field observers meas-ured nest mortality and raven

occurrence at all of the known heronries

in the northern San Francisco Bay area(Figure 1). At ten of these sites, we con-ducted two all-day watches for ravenactivity. We then selected three heronries

with resident ravens for more intensivestudy: (1) ACRs Picher Canyon, near theshoreline of Bolinas Lagoon, where GreatBlue Herons, Great Egrets, and SnowyEgrets settle among the branches of coastredwoods in the bottom of a narrowcanyon; (2) Drakes Estero in the PointReyes National Seashore, where nestingGreat Egrets and Great Blue Heronscrowd into a small patch of bishop pines

surrounded by grazed prairie grasses andvegetated dunes; and (3) West MarinIsland, a 2.7-acre rock in the MarinIslands National Wildlife Refuge just offthe eastern Marin County shoreline,

where a blend of California buckeyes andwoody shrubs often supports the SanFrancisco Bay areas largest concentrationof nesting Great Egrets, Snowy Egrets,Black-crowned Night-Herons, and GreatBlue Herons (see The Ardeid 2003).

At each of the three heronries selectedfor close study, we recorded detailedinformation on raven occupancy, land-ings, patrol flights, and interactions withother species. At five-minute intervalsduring hundreds of two-hour observationperiods, we recorded the presence of

Common Raven behaviors in heronries

Vague Consequences of Omnipresence

by John P. Kelly

continued on page 2

Figure 1. Study area and heronries (circles) in the northern San Francisco Bay area, including colony sites

for extended watches (*) and intensive observations of raven activity (PICA: Picher Canyon; WMAR: West

Marin Island; DRES: Drakes Estero).


3/16

Audubon Canyon RanchResearch and ResourceManagement

Executive Staf fSkip Schwartz, Executive DirectorJohn Petersen,Associate DirectorYvonne Pierce,Administrative Director

Staff BiologistsJohn Kelly, PhD, Director, Research & Resource

ManagementRebecca Anderson-Jones, Bouverie PreserveKatie Etienne, Research CoordinatorDaniel Gluesenkamp, PhD, Habitat Protection and

Restoration SpecialistGwen Heistand, Bolinas Lagoon Preserve

Land StewardsBill Arthur, Bolinas Lagoon PreserveDavid Greene, Tomales Bay properties

John Martin, Bouverie PreserveField BiologistsEllen BlusteinLauren HammackMark McCaustlandMichael Parkes

Data ManagementMelissa Davis

Research A ssociatesJules EvensFlora MacliseHelen PrattRich Stallcup

Resource M anagement AssociatesLen Blumin

Roberta DowneyScientific A dvisory PanelSarah Allen, PhDPeter Connors, PhDJules EvensKent Julin, PhDGreg KammanRachel KammanChris Kjeldsen, PhDDoug Oman, PhDHelen PrattW. David ShufordRich StallcupWesley W. Weathers, PhD

the ARDEID 2004

Nancy Abreu (H); Ken Ackerman(C,G,S,W); Drew Alden (R); Judy Allen(C); Leslie Allen (R); Sarah Allen (S); Bob

Anderson (H); Janica Anderson (H);Autodesk employees (R); Bob Baez(S,W); Norah Bain (H); Bruce Bajema(H,S); Katy Baty (W); Tom Baty (W);Gordon Bennett (S,W); Shelly Benson(R,W); Sherman Bielfelt (G); Gay Bishop(S); Mike Blick (G); Julie Blumenthal(W); Len Blumin (R,W); Patti Blumin(H,R); Ellen Blustein (H,S); Noelle Bon(H); Janet Bosshard (H); Tom Bradner(E); Anna-Marie Bratton (E); EmilyBrockman (H,S,W); Ken Burton (S,W);Phil Burton (H); Denise Cadman (H);

Ann Cassidy (H); Bonnie Clark (S);Michelle Coppoletta (R); Rig Currie (S);Melissa Davis (H); Nick Degenhardt (H);Jack Dineen (H); Carolyn Dixon (H);Giselle Downard (H); Roberta Downey(C,R); Joe Drennan (W); Judy Dugan (E);

Bob Dyer (H); Cathy Evangelista (H);Jules Evens (R,S,W); Katie Fehring(S,W); David Ferrera (W); John Finger(R); Binny Fischer (H); Grant & GinnyFletcher (P,S); Carol Fraker (H); DanielGeorge (S,W); Tony Gilbert (H,S,W);Beryl & Dohn Glitz (H); Mike Green(W); Philip Greene (H); Bill Grummer(H); Karlene Hall (R); Madelon Halpern(H); Lauren Hammack (M); FredHanson (S,W); Roger Harshaw (S);Michael Henkes (G); Diane Hichwa(H,S); Judy Hiltner (C); Peter Horn (S);Roger Hothem (H); Steve Howell (W);Lisa Hug (S,W); Jeri Jacobson (H,S);Rosemary Jepson (G); Rick Johnson(H,S,W); Calvin Jones (G); Gail Kabat(W); Lynnette Kahn (H,S); BobKaufman (H); Guy Kay (H); Thomas

Kehrlein (G); Richard Kirschman (S);

Ellen Krebs (H); Carol Kuelper (S); AndyLaCasse (H); Joan Lamphier (H); Jean-Michel Lapeyrade (G); Laura Leek (W);

Bill Lenarz (H); Stephanie Lennox (H);Robin Leong (H); Liz Lewis (M); EileenLibby (H); Patricia Lonacker (R); FloraMaclise (H); Phil Madden (R); Art &Lynn Magill (R); Jean Mann (R); AlanMargolis (R); Marin Conservation Corps(R); Roger Marlowe (W); Janaea Martin(G); Mark Mc Caustland (H,M,S); DavidMcConnell (H); Jean Miller (H); AnnMintie (E); Len Nelson (H); WallyNeville (H); Linda Nicoletto (C); TerryNordbye (S,W); Richard Panzer (H);Jennie Pardi (H); Bob Parker (G); MikeParkes (H,R,S,W); Lorraine Parsons (M);Tony Paz (E); Joelle Peebles (G);Precious Peoples (H); Kate Peterlein(S,W); Jeff and Kathy Peterson (R);Myrlee Potosnak (H); Linda & JeffReichel (H); Rudi Richardson (W);

DeAnn Rushall (T); Ellen Sabine (H);Marilyn Sanders (H); Fran Scarlett (H);Cindy Schafer (S); Phyllis Schmitt (H,R);Gordon & Laurie Schremp (H); AliceShultz (H); Asha Setty (C); Joe Smith(W); Anne Spencer (S); Bob Spofford(H); Rich Stallcup (M,S); JeanStarkweather (H); Jim & JeanneSternbergh (H); Michael Stevenson(S,W); Ron Storey (H); John Sutherland(C); Lowell Sykes (H,S,W); Judy Temko(H,S); Janet Thiessen (H); Gil & DodieThomson (H); Peggy Thorpe (H); LisaToet (E,R); Travelsmith employeeservice day (R); Sue Tredick (H); SusanTremblay (G); Tanis Walters (S); Ryan

Watanabe (R); Penny Watson (W);Rosilyn & Tom White (W); Phyllis

Williams (H); Jessica Wilson (C); Ken

Wilson (W); Katy Zaremba (R).

In this issue

Vague Consequences of Omnipresence: Common Raven behavior in heronries

by John P. Kelly ......................................................................................................................page 1

The Spatial Dimensions of Raven Life: Home-range dynamics in western Marin County

by Jennifer Roth ....................................................................................................................page 4

Eliminating Ehrharta:Bolinas Lagoon Preserve as a test area for regional conservation

by Daniel Gluesenkamp ........................................................................................................page 6

EhrhartaUnderground:

by Gwen Heistand..................................................................................................................page 6

The Power of Rainfall: Influences on the hydro-geomorphology of Livermore Marsh

by Katie Etienne ....................................................................................................................page 10

Cover photo: Common Raven byPeter LaTourrette Ardeid masthead

Great Blue Heron ink wash painting byClaudia Chapline

Volunteers for ACR research or habitat restoration projects since The

Ardeid 2003. Please call (415) 663-8203 if your name should have beenincluded in this list.

PROJECT CLASSIFICATIONS:

C = Coastal Habitat Restoration at Toms Point x E = Ehrharta erecta Removal atBolinas Lagoon Preserve x G = Grassland Management at Bouverie Preserve xH = Heron/Egret Project x M = Livermore Marsh and Olema Marsh Surveys xx P = Photo Points x R = Habitat Restoration x S = Tomales Bay ShorebirdProject x T = Turkey Research at Bouverie Preserve x W= Tomales Bay

Waterbird Census

The Watch


4/16

page 2 the ARDEID 2004

ravens in and near colony sites, identifiedperching substrates, and noted their posi-tions above or below the tree canopy oron open ground. We measured how longravens remained at each landing site andestimated their distances to heron oregret nests. Whenever a raven landed inor near a nest, we recorded the nest con-tents, nesting stage, age of chicks, andnumber of adult herons or egrets flushedfrom the area. For each patrol flight, we

recorded the time of day, flight duration,height above nests, and alert responses ofadult or nestling herons or egrets. Whenravens interacted with other bird species,

we determined which individuals weredominant and subordinate, the durationof chases, and other details. And ofcourse, we recorded any evidence of nestpredation. From these data, we looked forevidence that the nest predatory behav-iors of ravens might be influenced bytheir foraging experience, energy needs,time of day or season, or the availabilityof prey.

Evaluating occurrence

Our surveys revealed substantialnest predation by ravens at someheronries in the region but high

variability among colony siteseven

though ravens are common throughoutthe San Francisco Bay area. We estimateda regional average of only about oneraven occurrence in each heronry everyfive hours and detected ravens in onlyabout one-fourth of the heronries each

year. At Picher Canyon, resident ravenswere present less than 5% of the time, butravens occupied colonies at West MarinIsland and Drakes Estero about 30% ofthe time. The resident ravens occupied allthree sites of intensive study more oftenafter mid-June, often in pairs or accom-panied by fledgling ravens. Keep in mind,

however, that ravens occurred only rarelyat most other heronries.

Resident ravens preyed on about 27Great Egret nests per year at West MarinIsland, 20 nests per year at PicherCanyon, and only two to three nests per

year at Drakes Estero. Great Blue Heronnest mortality was less than two nests per

year at each site and, for most of these,we did not determine the cause of failure.Of 11 instances of complete colony siteabandonment in the region, only oneata site with nesting Great Blue Herons andGreat Egretswas associated with repeat-

ed heavy disturbance by ravens, and asingle Great Blue Heron nest was estab-lished there the following year.

The low regional rates of raven occur-rence and nest predation in heron andegret colonies were associated with recentinfluxes of ravens. If ravens move into anarea without prior experience in heron-ries, they may exhibit neophobia for thefirst few to several nesting seasons. Neo-phobia is a characteristic cautiousnesstoward novel food items that subsides asravens learn through experience not tofear what may later become important

food. Because nesting ravens occupy thesame home ranges across years (see arti-cle on page 3), nest predation in waterbirdcolonies by new resident pairs of ravensmight increase as neophobia subsides.

We did not find significant annualincreases in nest predation, but we did

find significant annual increases in ravenpredatorybehaviors: longer patrol flights,more landings, more frequent movementamong landing sites, and more flushingand harassment of nesting Great Egrets. Inaddition, recovered (fresh) prey remains atthe Marin Islands suggested increasingpredation ofadult Snowy Egrets: ravenskilled at least four adult Snowies in 2000,seven in 2001, 15 in 2002, and eight in2003. On one occasion, ACR biologistMark McCaustland watched in awe as araven chased an adult Snowy Egret to theground and killed it. Such trends in behav-

ior are consistent with the attenuation ofneophobia in adult ravens and suggestthat nest predation might increase annu-ally for some years after ravens begin tooccupy heronries.

The notion that ravens might becomemore daring predators across years isconsistent with previously known behav-iors. For example, ravens may have diffi-culty overcoming neophobia by their typ-ical means of approaching slowly(Heinrich 1988, Condor 90: 950-952),because egrets build their nests in isolat-ed locations in the nest-tree canopy.

Ravens also require unusually long peri-ods of time, under experimental condi-tions, to accept carcasses of large birds(Heinrich et al. 1995,Auk112: 499-503).Perhaps most importantly, however, therisk of major injury if egrets becomedefensive may be enough to prolong theirneophobic behavior.

Assessing predation risk

The lack of annual increases in nestpredation at the three study sites, inspite of annual increases in associ-

ated behaviors, suggests that resident

ravens may interfere with the activities ofother nest predators, such as raccoons,owls, or raptors. Ravens were dominant in93% of interactions with visiting birdspecies such as Osprey, Golden Eagle,Red-tailed Hawk, Peregrine Falcon, and

American Crow. In one instance, a Red-tailed Hawk that was flushing egrets fromtheir nests at the Marin Islands washarassed repeatedly by the residentravens until it left the area. Althoughravens often chase visitors away, they alsorespond opportunistically to nest distur-

Figure 2. Raven predation on Great Egret nests,

length of raven patrol flights, and number of raven

interactions with other species increased signifi-

cantly (P< 0.05) with increased raven productivity.

Figure 3. Great Egret nest failures known to have

resulted from raven predation peaked as nestlings

approached three weeks of age, at ACRs Picher

Canyon, 2000-2001.

see Omnipresence, page 5


5/16


Early one morning, a pair of ravensflies over the heron and egretcolony at ACRs Picher Canyon in

search of unattended eggs or youngchicks. At Point Reyes National Seashore,ravens soar routinely over dune beachesin search of Snowy Plover nests. Mean-

while, at Tonys Seafood Restaurant onTomales Bay, a pair of ravens lingers eachday in the vicinity of a nearby dumpster.

What impact is raven predation having onheronries and rare bird species? What fac-tors are influencing their movements and

use of the landscape? These questionsprompted us to begin radio-trackingbreeding ravens in western Marin Countyin order to learn more about their impacton sensitive bird species and the factorsaffecting their distribution (Roth et al.2004).

Radio-tracking

We captured and attached back-pack-mounted radio-transmit-ters to 16 adult ravens from 15

resident pairs, including both members ofthe pair residing near Picher Canyon.

Capturing ravens was one of the most dif-ficult parts of the study, and we went togreat lengths to outsmart these intelligentbirds. We started by identifying birds that

we were interested in studying and thenwatched them for several days to learnmore about their habits and decide on asuitable trapping location. Once wedecided to attempt a capture, we roseseveral hours before dawn to place thebait and hide our traps and ourselvesbefore first light; the birds would avoidthe area for days at the first hint that

something was amiss. We hid in bushesand ditches, covering ourselves withbranches and dried grass. We often spentseveral hours in our makeshift blindsbefore the ravens came near the bait.Ravens are surprisingly cautious aroundunknown food sources, often approach-ing the area and backing away severaltimes before eating anything. Mean-

while, we waited breathlessly as theyapproached the traps or moved into aposition where we could shoot a net overthem. Our disappointment over nearmisses was tangible.

On one occasion, a bird that discov-ered our traps and escaped without beingcaptured flew into the air, circling andcalling angrily. Ravens came from alldirections. We soon had a large group ofravens circling the area and calling. The

warning had gone out! Conversely, imag-ine our excitement each time we handledone of these large, intelligent birds. Eachbird responded differently to being cap-tured; some birds struggled and othersquietly observed us. Each release was fol-lowed by a few tense moments as we

watched to see whether there would beany problems with the harnesses we used

to attach the radio-transmitters to thebirds. Fortunately, all radio-tagging activi-ties were successfully completed withoutmishap, but the challenges of followingtheir movements were just beginning.

We radio-tracked each bird for one tothree years. Each time we detected a bird,

we mapped its location on a topographicmap and recorded information on behav-ior and habitat use. For example, werecorded birds flying over grazed grass-lands, foraging along beaches, attendingnests in cypress trees, and roosting inpine groves. We also mapped nest sites

and concentrated food sources in thearea. We used the data we collected toestimate the home-range size of each birdand to evaluate some of the factors likelyto affect home-range size and space use

within home ranges.

Home-range size

Ahome range is the area used by ananimal during normal activitiessuch as food gathering, mating,

and caring for young. We measured eachhome range by calculating the area that

Home-range dynamics in western Marin County

The Spatial Dimensions of Raven Life

by Jennifer E. Roth

Figure 1. Locations, centers of activity (50%

occurrence), and home ranges (95% occurrence)

of the female (dashed lines) and male (solid lines)

ravens nesting near ACRs Picher Canyon heronry.

Unlabeled outlying areas are isolated portions of

home ranges.

continued on page 4


6/16


encompassed a 95% probability of anindividual ravens occurrence, assumingthat this area would include all normalactivities and exclude rare excursions tomore distant areas (Figure 1).

Home-range size was highly variablefor the ravens in our study (Figure 2).Sizes ranged from 0.24.9 km2 for femalesand 0.34.9 km2 for males. The homeranges of the Picher Canyon pair were

large compared to other pairs: 4.2 km2 forthe female and 4.9 km2 for the male.

Many factors may contribute to varia-tion in home-range size among individu-als. For example, food supply, competi-tion for food, territoriality, vegetationtype, and body size might influence thedistances birds must travel. We tested thepossibility that differences in sex and dis-tance to important sources of food couldexplain the large differences in home-range size among ravens in westernMarin County.

Sex differences

Sex may influence home-range size inbirds when there are differences inbody size or behavior between

females and males. When males andfemales differ greatly in size, associateddifferences in agility and power mightallow them to compete less for particularprey or perhaps increase their collectivemenu of available prey. However, maleravens are only slightly larger thanfemales. Nonetheless, we expected to finda difference between female and malehome-range sizes during the breeding

season due to differences inbehavior between the sexes:females are largely responsi-ble for incubation and malesleave the vicinity of the nestto forage while females areattending the nest. Contrary

to our expectations, home-range size did not differ sig-nificantly between femaleand male ravens in our study.

Also, we did not find a signifi-cant difference in the distri-bution of locations for thefemale and male of the PicherCanyon pair (Figure 1). It maybe that sex roles do not differenough during the breedingseason to affect overall home-range size in ravens. Alterna-tively, the similar home-range

sizes between male andfemale ravens may reflecttheir tendency to interact fre-quently with their mates

through courtship, pair-bond reinforce-ment, increased vigilance by the maleduring egg laying and incubation, orother social behaviors associated withnesting (Boarman and Heinrich 1999).

Distance to food sources

In general, prey abundance and dis-tance to food and water influencehome-range size and movements in

birds. Besides preying on egret eggs andchicks, ravens in our study fed on smallmammals and reptiles; grain, calf carcass-es, and afterbirths found at many localdairies; human refuse at beaches, parkinglots, and dumpsters; and eggs and youngof other birds in the area. On severaloccasions, we even saw ravens harassTurkey Vultures until they regurgitatedtheir food. The ravens were able to catchthe regurgitations in mid-air! However,despite the wide variety in raven diets andsome unusual foraging behaviors, home-range sizes are most likely to be influ-

enced by the locations of stable, concen-trated sources of food.

We determined the distance betweeneach raven nest and the nearest ranch,human food source, or waterbird colonyin order to evaluate the influence of con-centrated food sources on home-rangesize. We found a general trend indicatingthat ravens nesting farther away fromconcentrated food sources had largerhome ranges (Figure 2). For instance, thePicher Canyon ravens nested about 1 kmaway from the heron and egret colony,

which was the nearest concentrated food

source to their nest, and they occupiedrelatively large home ranges compared toother ravens in western Marin County(Figure 2).

Space use within home ranges

Birds are likely to use some areas

more intensively than others due tothe patchy distribution of food and

habitat areas suitable for feeding, cachingfood, and nesting. We evaluated space use

within home ranges to determinewhether ravens locations exhibited ran-dom, uniform, or clumped distributions.

We also estimated the size of core areas orcenters of activity, defined as areas with a50% probability of an individual ravensoccurrence (Figure 1). We found that loca-tions of females were significantlyclumped within 83% and 100% of homeranges in 2000 and 2001, respectively.

Males were more variable in their use ofspace, with locations clumped in only38% and 44% of home ranges in 2000 and2001, respectively.

We used these data to determine therelationship between centers of activity,nest sites, and concentrated food sources.

All 16 of the radio-tagged birds centeredtheir activities around their nest sites. Insome cases, there was a concentratedfood source within that center of activity.Other birds had two centers of activity,

with one centered around the nest siteand one centered around a concentrated

food source. The Picher Canyon ravenswere unique among the individuals westudied in having three distinct centers ofactivity that centered around (1) theirnest site, (2) the Picher Canyon heron andegret colony, and (3) a nearby horse ranch

where they could rely on a daily supply ofgrain at feeding time (Figure 1). Theseactivity centers were separated by habitatareas that the ravens only rarely occupied.

Annual variation

Like most things in nature, food sup-ply, habitat quality, and other factors

that affect home-range use, mayvary from year to year, causing correspon-ding changes in home-range size. Wecompared home-range sizes between

years and found no significant differ-ences-large home ranges remained largeand small ones remained small. However,there were significant small-scale shifts inhome-range placement for most females(67%) and males (63%). These shifts

were associated with changes in nestlocation and the distribution of desirablefood resources between years. Given the

Figure 2. Raven home-range size increases with the distance

between the nest and the nearest concentrated food source (P 0.91). So the timing of nestpredation reflects a general pattern ofvulnerability rather than the timing ofraven predation per se.

Most observers have noticed thatravens are now present throughout ourregion and that their overall numbersare increasing. But our closer look atraven behavior suggests that weshould be cautious about assumingincreases in nest predation in heron-ries. On the other hand, where ravens

have been present for only a few to sever-al years, nest predation might increase

with continuing declines in neophobia.To further complicate things, whetherravens add to, displace, or buffer nest pre-dation by other species remains a mys-tery. As a result, the extent to whichravens occupy heronries may not closelyreflect predation risk.

For those who watch the lives ofravens with particular interest, it shouldbe no surprise that their behavior chal-lenges as well as inspires our understand-ing of nature. In assessing the risk ofraven predation, herons and egrets as wellas human observers should look beyondthe mere presence of these prominentpredators.

Omnipresence, from page 2

Color-banded Common Raven at Abbotts Lagoon, Point

Reyes National Seashore.

PHILNOTT


8/16


In 1930, near the University of Califor-nia Berkeley campus, an alien invaderescaped from a government-sponsored

genetics experiment. Although theescapee was a genetically natural (notgenetically modified) species, created byevolution, its escape into a new land hashad serious consequences to the natural

inhabitants of our region. In the decadessince, the green monster has spread qui-etly, wiping out resident flora and faunaas it invades, and is now on the verge oftaking over much of western MarinCounty. The invader is a South Africangrass named Ehrharta erecta, and thisarticle tells the story of Audubon Canyon

Ranchs efforts to understand and defendagainst its spread.

Ehrharta erecta is a highly invasiveperennial grass native to South Africa.The species was part of an experiment atthe U.C. Berkeley botanical garden inves-tigating whether increasing chromosomenumbers increases invasive ability; as itturns out, Ehrharta erectawas tremen-dously invasive without any alteration

whatsoever. By 1950 the species wasabundant at the U.C. Berkeley campus,and in 1996 the tremendous abundanceofEhrharta erecta in San Francisco natu-ral areas compelled recognition of thespecies as a major conservation concern(Sigg 1996). Ehrharta erecta is currently

invading western Marin, with large popu-lations in Olema Valley, along thePanoramic Highway on Mt. Tamalpais,around the towns of Inverness andBolinas, and in Audubon Canyon RanchsBolinas Lagoon Preserve.

Ehrharta erecta is a prolific seed pro-ducer, and the small seeds likely are car-ried to new sites via soil on shoes, on deerhooves, and as contamination in pottedplants. Ehrharta can thrive in an extreme-ly wide range of habitats, from coastaldunes to closed-canopy forest, and formsrobust monospecific stands under full sunor in as little as 2.5% of daylight (Haubensakand Smyth 2000). Once established, popu-lations increase rapidly and Ehrharta

ACRs Bolinas Lagoon Preserve as a test area for regional conservation

EliminatingEhrharta

by Daniel Gluesenkamp

Ehrharta erecta. Ehrharta covering native habitat. Pike County Gulch.

Cryptosphereeven thename given to the soiland leaf litter environ-

ment is laced with the promiseof discovery. What creaturesinhabit this hidden world?Strange primitive insects with-out wings (some withouteyes), springtails (named fortheir anal appendages calledfurcula that propel themthrough soil pore spaces),nematodes, thousands ofspecies of mites, feather-

winged beetles, and larvalinsects of all sorts, to name afew. The diversity and com-plexity of the cryptosphere hasbeen compared to coral reefsand tropical rainforests. And

yet, we know very little aboutit.

As mentioned in theaccompanying article,Ehrharta erecta creates adense thatch of vegetationboth above and belowground, thereby altering the

environments physicalattributes. The dense swardof green grass sweeping up

ACRs Pike County Gulch isan obvious result ofEhrhartainvasion. Impacts of theinvader on soil chemistry andbiotic communities areunknown. While it is not fea-sible at present to perform adetailed study, one piece ofthe Ehrharta experimentinvolves collecting data onleaf litter invertebrates in an

attempt to understand theimpact of the invasive grasson soil fauna.

Hidden ecologies

One square meter of fer-tile soil can containmore individual organ-

isms than all the humans thathave ever lived. Various stud-ies have reported up to 1012

bacteria (1 trillion), 1012 proto-zoa, 107 nematodes, spring-tails and mites, 107 insects,

Ehrharta Underground: Theres more to an invasion than meets the eye

by Gw en Heistand


9/16


immediate action is required to preventEhrharta from expanding, merging, andsmothering the natural diversity of thissingular preserve.

Three lines of defense

In 2003 ACRs Habitat Protection andRestoration (HPR) Program initiatedan ambitious project to address the

threat posed by this aggressive invader.The project includes (1) scientificresearch targeted to identify techniquesfor controllingEhrharta erecta; (2) eradi-cation ofEhrharta on ACR lands; and (3)collaborating with other groups and landmanagers to develop regional solutionsto the Ehrharta invasion.

In June 2003 we began an experimen-tal assessment ofEhrharta removal tech-niques. The experiment compares three

promising techniques: manual removalby pulling, dilute herbicide treatment,and solarization (covering with blackplastic). Each treatment is applied to 10Ehrharta-dominated plots. Data collect-ed from this experiment will allow us todetermine which Ehrharta removal treat-ment is most effective and to evaluatehow each treatment influences recoveryof the natural communities that we aretrying to restore.

In addition to the 30 experimentalplots designed to evaluate removal tech-niques, the Ehrharta experiment

includes two different sets of unmanipu-lated controls with 10 plots each,Ehrharta-dominated plots and un-invaded plots. In each of the 50 plots, weare quantifying the abundance ofEhrharta, vegetation composition, theabundance of native plants and of other

becomes the dominant herbaceous plant,excluding native and non-native vegeta-tion (McIntyre and Ladiges 1985). Thereare currently no established techniquesfor controlling the plant or for restoringinvaded habitat (Pickart 2000).

We have recently discovered severalpatches ofEhrharta erecta in each of thethree canyons that comprise ACRsBolinas Lagoon Preserve. While most ofthe patches are still small, the Ehrhartainfestation in Pike County Gulch provides

a glimpse at onepossible future for

ACRs flagship pre-serve: nearly 2000m2 of the canyonfloor is covered

with a densemonospecificsward, and thebright greeninvader is climb-ing the sides of thecanyon into thebrown leaf litter of

the natural herba-ceous understory.Given the plantsamazing rate ofspread and abilityto thrive in arange of habitats,

Ehrhartaremoval plot. Uninvaded versus invaded habitat.

Figure 1. After one year, all three treatments significantly reduced the percent

cover of Ehrharta erecta, relative to the invaded control plots, while manual

removal also stimulated germination from the seed bank (see text). Solid bars

indicate the percent of plots covered by Ehrharta;error bars indicate one stan-

dard error (P< 0.001).

continued on page 8

1,000 earthworms,and 20,000 km of

fungal mycelia.(Neher 1999,Pennisi 2004).Bacteria, fungi,and nematodesare three of thetaxonomic groupsthat have the lowestpercentage of describedspecies, making classificationof soil communities difficult atbest. In addition, within eachtaxonomic group, somespecies feed on bacteria, some

on fungus, someon roots, and

some on otherinvertebrates.Where stud-ied, differentecosystem

types havebeen found to

support differentassemblages of soil

biota. Grasslands have showna marked abundance ofomnivorous soil invertebrates.

Agricultural soils can havegreater numbers of bacterial

feeders and root-feedingnematodes. Forests display a

relative abundance of fungalfeeders (Neher 1999).

Belowground food websand ecological relationshipsare intricate and intimatelylinked to the abovegroundenvironment (Wardel et al.2004, Wardel 2002). Science is

just beginning to delve intothese connections. Few stud-ies have attempted to assesssoil biota in relation to eitherhabitat restoration or invasivespecies. Invasion of non-

native plants may affect aboveand belowground feedbacks,

potentially changing soilchemistry and altering soilbiota to facilitate invasion (vander Putten 2002).

Hidden consequences?

During Februarythrough April of 2004,Dan and I worked with

several volunteers to collect350 1-cm2 litter samples fromthe center of each of the 50experimental plots used in theErharta study. Each sample


10/16


non-native invaders, important habitatcharacteristics, and leaf-litter inverte-brate communities.

Preliminary results indicate that aweak herbicide solution is most effectiveat reducingEhrharta abundance, while

manual removal actually stimulatesEhrharta seed germination and increasesabundance of the invader (Figure 1).

While Year One results show solarizationto be moderately effective at reducingEhrharta cover, some adult Ehrharta

plants are able to sur-vive beneath the plastictarp; the utility of thistechnique will dependon how well these sur-vivors recover in thenext year.

Ehrharta controlmay further depend onlimiting the ability ofEhrharta populationsto recover from buriedseeds, so we are experi-mentally assessing seed

longevity by buryingsmall bags of seeds attwo depths, exhumingseedbags at regularintervals, and germi-nating the exhumedseeds to determineseed survival. Twelve

bags were buried at each of 10 locations,and the first set of bags will be exhumedand tested for viability in November 2004.Early observations of buried seed bagsprovide reason for hope. Each of themarker stakes is surrounded by a constel-lation of small clusters ofEhrharta

seedlings, one cluster for each seed bag!This shows that the germination rate ofthese buried seeds is very high and thatrelatively few seeds remain dormant, andsuggests that control efforts may not haveto contend with a large and persistent soilseedbank.

Although the increasing ground coverbyEhrharta clearly reduced the cover ofnative plants (Figure 2), recovery of nativeplants one year after treatment did notdiffer significantly among treatments.However, differences in native coveramong treatments may become signifi-

cant as the small native plants mature.Probably the most interesting variable

that we will be monitoring is the responseof leaf-litter invertebrate critters (seeaccompanying article). Very few scientificstudies have quantified the impacts ofplant invasion on animal communities,and only a handful have assessed howanimal communities are affected by inva-sive plant control efforts. Terrestrial inver-tebrates are extremely important, both aselements of decomposition and nutrientcycles and as the basis of many terrestrialfood webs, and ACRs Ehrharta research

will be one of only a few studies thatassess how the terrestrial invertebratecommunity is affected by non-nativeplant invasion. It will be extremely inter-esting to observe how invertebratesrecover followingEhrharta removal!

While the scale of Bolinas LagoonPreserves Ehrharta invasion is daunting,

Figure 2. Native plant cover declines with increasing cover by Ehrharta

erecta. Data are from unmanipulated plots (r2 = 0.59, P< 0.001).

was run through a BerleseFunnel, which uses light to

drive soil fauna downwardinto a container, producingapproximately1/8 to 1/4 tea-spoon of invertebrates.Characterization, on a broadscale, of invertebrate assem-blages in each of these sam-ples has begun. Just to give aflavor of the numbers found inless than a teaspoon, one sam-ple currently being analyzedcontains 187 acari (mites), 214collembolans, 3 dipterans(flies), 12 coleopterans (bee-

tles), 13 diplurans (primitivewingless insects), and 11

unidentified larvaewithapproximately two-thirds ofthe sample still to be sorted!

Once all samples have allbeen processed, we will ana-lyze the results to determine

whether there is any differ-ence in soil fauna between thethree treatments (herbicide,solarization, and manualremoval) as well as betweenthe invaded control sites anduninvaded plots. We mightexpect to see more omnivores

in the grass dominated plotsversus more fungal feeders in

the native forest understory.Plots covered in black plasticmight be expected to have adifferent suite of creaturesduring and immediately aftertreatment. One aspect of what

we are hoping to ascertain ishow resilient the ecologicallyimportant leaf litter commu-nity is in response to restora-tion efforts. Because so little isknown about soil ecology on

ACR lands, whatever we dis-cover will be exciting. These

data, combined with data onthe abundance Ehrharta and

other non-native and nativeplant species, as well as datacharacterizing plant speciescomposition for each plot,should advance our under-standing of howEhrhartaerecta control, eradication,and dispersal might affectnative biota.

On a more personal note,it is difficult to describe thethrill of seeing a microscopicfly with beaded antennae andperfectly formed halteres or a

Ehrhartaplot. Buried seed bag sprouting (Ehrhartain circles).


11/16

our commitment to protecting the pre-serves natural diversity mandates that weact to restrain this harmful invader. In thelast year we have conducted surveys tolocate and map all ACR Ehrharta patchesand have begun work to treat the high

priority sites. ACR has followed a stitch intime policy with regard to control work,focusing on small easily- eradicatedpatches ofEhrharta that have the poten-tial to become intractable infestations.Results of this first year have been prom-ising, and we expect that several years ofconcerted effort by staff and volunteers

will be required to control this advancedinvasion.

Invasive species such as Ehrharta erec-ta do not recognize property boundaries,and so we are working with our neigh-bors and with other restorationists to

improve Ehrharta management acrossthe region. In the last 18 months, ACRstaff have helped make the Marin-Sonoma Weed Management Area (WMA)aware of the threat posed byEhrhartainvasion, and as a result the WMA con-tributed significant support to imple-menting ACRs Ehrharta experiment and

assigned 2 interns to map all Ehrhartaoccurrences in Marin County. As ACRsHPR Specialist, I communicate frequent-ly with biologists from California StateParks, the National Park Service, andother agencies to share information andinsights regardingEhrharta manage-

ment. When ACRs Ehrharta experimentis completed we will use the data todevelop a prescription for restoringEhrharta-invaded habitat, and we expectthat presentation of results will furtherincrease regional Ehrharta controlefforts.

Conservation commitment

The genus Ehrharta is largelyrestricted to the Cape region ofSouth Africa, which, like California,

is one of 5 regions on the planet with aMediterranean climate. Ehrharta erecta

has the greatest geographic range of anyplants in the genus, and it is interesting tothink that Ehrharta erecta may, by itsnature, be a colonizer of new sites. In amanner reminiscent of many ofCalifornias most marvelous natives, thetaxon named Ehrharta erecta emerged inan accelerated burst of speciation stimu-lated by the appearance of summer-aridclimate in southern Africa. Ehrharta erec-ta is not a bad plant but, rather, is anamazing and tremendously vital organ-ism that is the singular result of a uniqueevolutionary history. Unfortunately, its

aggressive spread in coastal Californiathreatens to eliminate some survivingrepresentatives of Californias own uniqueevolutionary history.

Harmful invaders like Ehrharta posevery significant challenges, but they alsoprovide important opportunities.Research on invasive species is helping us

understand how natural systems workand is creating a fruitful nexus of collabo-ration between pure scientists andapplied restorationists. The urgent needfor research and restoration volunteershas fueled a joyful reconnection of citi-zens with nature that is similar to the

amateur scientist revolution of 19th cen-tury England. Most importantly, the ubiq-uity and impact of biological invasions ishelping humans realize that they must beactive and conscientious stewards of thisplanet: where Rachel Carsons SilentSpringdemonstrated that human actionscan have global impacts, the replacementof a diverse natural community by a sin-gle species of grass is teaching us thatinaction is not without consequences.

Over decades and centuries, Ehrhartaerecta may or may not accumulate newpathogens and herbivores and become a

well-regulated and harmless part of ourcoastal plant communities. In the mean-time, conservation organizations such as

ACR remain committed to defendingCalifornias ancient natural diversity fromthe negative impacts ofEhrharta erectaand other runaway invaders.

References cited

Haubensak, K. and A. Smyth. 1999. Erharta erecta.Species assessment prepared for ChannelIslands National Park.

McIntyre, S. and P. Y. Ladiges. 1985. Aspects of thebiology ofErharta erecta. Lam. Weed Research

25: 21-32.Pickart, A. J. 2000. Ehrharta calycina, Erharta erec-

ta, and Ehrharta longifora. In: Bossard, C. C., J.M. Randall, and M. C. Hoshovsky (eds): Invasiveplants of California wildlands. University ofCalifornia Press, Berkeley, CA.

Sigg, J. 1996. Erharta erecta: sneak attack in themaking? California Exotic Pest Plant CouncilNews 4(3):8-9.

beetle larva with a wildlyintricate exoskeleton that

looks like it is made from tor-toise shell or the field of viewunder a dissecting scopecompletelyfilled with mites ofdifferent sizes and shapes.Knowing that these creatureslive their lives navigating thepore spaces in the soil or theunderside of decomposingleaves adds to my amaze-ment at the intricacy of lifeon this planet. And, along

with the sheer aestheticpleasure of viewing these

spectacular, minute beingsand a feeling of increased

respect for all that livesunderfoot, I cant help but

wonder what we may bedoing to alter this world as

we move invasive speciesaround the globe and prac-tice less than sustainablemethods of agriculture.

ACR is extremely fortunate tohave a crew of committed, andpossibly slightly crazy, volun-teers who share this sense ofrespect and adventure and

who enjoy spending Saturdaysand Sundays bent over micro-

scopes keying out and countingwee beasties found in the cryp-tosphere. Many thanks go toTom Bradner, Anna-MarieBratton, Judy Dugan, AnnMintie, and Tony Paz.

References cited

Neher, D. A. 1999. Soil communitycomposition and ecosystemprocesses: comparing agriculturalecosystems with natural ecosys-tems. Agroforestry Systems 45:159-185.

Pennisi, E. 2004. The secret life offungi. Science 304: 1620-1622.

van der Putten, W. H. 2002. How to

be invasive. Nature 417: 32-33.Wardle, D. A., R.D. Bardgett, J.N.

Klironomos, H. Setala, W.H. vander Putten, and D.H. Wall 2004.Ecological linkages betweenaboveground and belowgroundbiota. Science 304: 1629-1633.

Wardle, D. A. 2002. Communitiesand Ecosystems: Linking theAboveground and BelowgroundComponents. PrincetonUniversity Press, Monographs inPopulation Biology.

Volunteers set up the experiment.



12/16


The North Pacific Coast Railroadconstructed a levee along theTomales Bay shoreline in the 1870s

that modified the hydrology of east-shoremarshes. Although wooden trestles werebuilt across major tide channels, watercirculation was altered and sedimentbegan to accumulate behind the levees.Livermore Marsh at the Cypress Grove

Research Center is one of these wetlandsthat exhibit some interesting physical andbiotic characteristics. Before AudubonCanyon Ranch (ACR) purchased the prop-erty in 1971, previous owners installedfloodgates in the levee so they could usethe 26-acre wetland for grazing and fresh-

water storage. In 1982, the levee waseroded by El Nio flood waters. ACRdecided to repair the levee and constructa spillway at an elevation that would pre-vent tidewater from entering LivermoreMarsh. ACR excavated four ponds in thelower marsh to maintain coveted fresh-

water habitat during dry seasons.However, when the levee breached byflood waters again in 1998, ACR decidedto allow the wetland to be restored natu-rally to a tidal marsh system.

In 1999, grants from the Frank A.Campini Foundation and the MarinCommunity Foundation allowed ACR torestore trail access across the levee breachand to conduct a five-year study to docu-ment physical and biotic changes associ-ated with the reintroduction of tidal cir-culation. In recent issues ofThe Ardeid,

we presented reports summarizing our

research objectives (1999), watershedeffects (2001) and changes in bird use(2003). This article examines principalfactors influencing the shape (topogra-phy) of the lower marsh (Figure 1).

The physical characteristics of tidalmarshes have been studied extensivelyaround the world, and some of the mostimportant research on the morphology ofdiked salt marshes has been conducted inthe San Francisco Bay area (Coats etal.1989; Coats and Williams 1990, and

Williams and Orr 2002). These engineershave identified correlations between key

marsh features,such as channeldimensions andtidal exchange,that are oftenused to designmarsh restora-tion projects. Wemeasured the

same topograph-ic features todocumentchanges duringthe transforma-tion of LivermoreMarsh from afreshwater to atidal system.

This studyincluded month-ly measurementsof the tidal inlet,annual surveys

of developingtide channels,and two detailedtopographic sur-veys of the lowermarsh in 1999 and 2003 (see photo onback cover). Hydrologist LaurenHammack used these data to estimatefour characteristics of tidal marsh sys-tems: tidal inlet area, active tide channellength, active tide channel volume, andtidal prism volume (Table 1, page 12). Thetidal prism estimate represents the vol-ume of tide water that flows through the

levee below the elevation of the meanhigher-high water (3 ft NGVD29, or 5.4 ftNOS based on NOAA predictions).

Tidal versus fluvial ef fects

Tidal processes and terrestrial riverflow are important factors thatshould be evaluated before apply-

ing scientific models to particular sites.This is a complex and demanding chal-lenge, but at Livermore Marsh, we hadthe advantage of being able to routinelyand accurately monitor changes in theinlet area. We used a laser level mounted

at either end of the bridge to provide a

stable reference for depth measurementsat 47 locations across the 96-ft bridge.

One of the interesting questions we areexploring is whether different types oftidal patterns have predictable effects onthe cross-sectional area of the tidal inlet.Neap tides exhibit the minimum rangebetween high and low tides, whilespring tides exhibit the largest range

between highs and lows. Median tidesrange between mean-higher-high water(MHHW) and mean-lower-low-water(MLLW) and occur for several daysbetween the neap and spring tide cycles.

I measured the inlet each month dur-ing the last day of a median tide cycle andscheduled neap and spring measure-ments at least three times a year. We weresurprised to find that tidal action did notsignificantly influence either the cross-sectional area of the inlet or the rate ofchange in the inlet area.

From our analysis, it appears that

cumulative rainfall was the primary factor

Figure 1. Elevations associated with tidal inundation at Livermore Marsh, in 2003.

Influences on the hydro-geomorphology of Livermore Marsh

The Power of Rainfall

by Kat ie Etienne


13/16

influencing changes in the inlet area(Figure 2). In addition, the effects of rain-fall were not significantly influenced bytide conditions. Figure 3 illustrates thehigh variability in the size of the tidal inletduring the first five years, although theinlet area appears to be stabilizing aroundthe values predicted from mature marshsystems (see The Ardeid 2002).

The role of sediment

W

e calculated the volume of pri-mary and secondary tide chan-nels, as well as incipient chan-

nels that may eventually become tertiarychannels or blind sloughs. During the first

year after the levee breach, the primarychannel increased in length by 128 ft. Theprimary tide channel did not lengthenbetween 1999 and 2000, but there was asmall increase in channel volume as thechannel became deeper and wider. Thispattern of channel lengthening followedby widening is consistent with otherdeveloping marshes where changes in

width are also more rapid than changes indepth (Coats et al. 1995).

Although there was an increase in

channel length and volume in 2001 thatcontributed to the increase in tidal prism,the tidal inlet continued to fill in (Table1). This inverse relationship betweenchanges in the tidal prism and the inlet isdifferent from the direct relationship inmature marsh systems, where increasedinlet area corresponds to increased tidalprism volume (Coats et al. 1995).

Subsequent channel surveys indicatedthat although most channels were becom-ing wider and longer, the total channelvolume and tidal prism decreased in 2002.However in 2003, we measured a largeincrease in channel volume and length as

well as an increase in tidal inlet area(Table 1). These episodic changes inchannel development are probably asso-ciated with rainfall events and the cohe-sive nature of the sediment, which tendsto erode in blocks.

In 1999, Gian-Marco Pizzo identifiedthe significance of the dense substrate,

which is not easily eroded by runoff ortidal action. His calculation of tidal veloc-ity in the primary channel (0.2 m/sec)

was far less than the 1 m/sec normallyrequired to cause erosion. Pizzo conclud-

ed that channel development in Liver-more Marsh was primarily influenced bythe propagation of vertical head cuts thatproduce the stair-step shape in develop-ing channels. During the next five years,

we continued to document the lateralmovement of numerous head cuts andslumping banks. Over time, head cuts andchannel banks became taller and themusic of falling water became louder asturbulent flow scoured material awayfrom the base of new waterfalls.

Sporadic changes in tidal prism

Prior to analysis, we partitioned the

topographic survey data to distin-guish between active tide channels

and isolated or perched ponds that didnot drain during diurnal tidal cycles(Figure1). The small circular pond closestto the levee is one of four ponds con-structed in 1983, when ACR was manag-ing the system as a freshwater marsh. Thelarger, crescent-shaped pond is a remnantof a pre-existing channel that wasmapped by the U.S. Coast Survey inthe1862 (see The Ardeid, 1999). Because

water level in these and other smallponds in the marsh did not rise and fall

with the tides, we excluded their volumes

from tidal prism estimates in 1998through 2001.

The head cut of the primary tide chan-nel progressed very slowly through themiddle of the marsh, compared to chan-nel banks that developed in the perimeterof the marsh. The rate of erosion in themiddle marsh was reduced by a remnantstand of tules, because their dense stemsdecreased water velocity and their rhi-zomes bound the sediment together.

Another factor was probably the percola-tion of water from the perched pond,

which maintained soil saturation in themiddle of the marsh. Saturation withwater prevented cracking of the substratethat allows blocks of sediment to tip andfall downstream.

Meanwhile, channel banks around theperimeter of the marsh continued to col-lapse into the channel, and sediment wasslowly transported through the system.Finally, after several weeks of heavy rainand the spring tide cycle of 6-8 Jan 2004,the sill between the crescent pond andthe head cut eroded, and water previouslystored in the large pond was released

(Figure 4). As rain continued to fall, hightides began to circulate through the cres-cent pond, which increased the tidalprism by at least 0.36 acre-ft. This esti-mated increase does not include the addi-tional tidal volume resulting from the ero-sion of unconsolidated material in thepond by winter runoff and high tides.

This process of slow erosion and tidechannel development indicates theimportance of substrate composition.

Another major factor is the elevation ofthe marsh plain. Survey data show thatmost tide water never flows above the

deepest portions of the developing chan-


Figure 2. The tidal inlet area of Livermore Marsh

increased significantly with cumulative rainfall (P

Documents

The Ardeid Newsletter, 2004 ~ Audubon Canyon Ranch