Biodiversity Conservation in the Face of Dramatic Forest Disease: An Integrated Conservation Strategy for Tanoak ( Notholithocarpus densiflorus ) Threatened by Sudden Oak Death

2013·13.5 MB·English

by  Richard C Cobb

1
Madrono, Yol. 60, No. 2, pp. 151 164, 2013
BIODIVERSITY CONSERVATION IN THE FACE OF DRAMATIC FOREST
DISEASE: AN INTEGRATED CONSERVATION STRATEGY FOR TANOAK
NOTHOLITHOCARPUS DENSIFLORUS) THREATENED BY SUDDEN
(. OAK DEATH
Richard C. Cobb and David M. Rizzo
Department of Plant Pathology, One Shields Ave, University of California Davis,
CA
Davis, 95616
[email protected]
Katherine Hayden and Matteo Garbelotto
J.
Department of Environmental Science, Policy, and Management, 137 Mulford Hall,
CA
University of California, Berkeley, 94720
JoAo A. N. Filipe and Christopher A. Gilligan
Department of Plant Sciences, University of Cambridge, Cambridge CB2 3EA, U.K.
Whalen W. Dillon and Ross K. Meentemeyer
Department of Geography and Earth Sciences, University of North Carolina Charlotte,
NC
9201 University City Boulevard, Charlotte, 28223
Yana Valachovic
S.
CA
University ofCalifornia Cooperative Extension, 5630 South Broadway, Eureka, 95503
Ellen Goheen
USDA OR
Forest Service Forest Health Protection, Central Point, 97529
Tedmund Swiecki
J.
CA
Phytosphere Research, 1027 Davis Street, Vacaville, 95687
Everett M. Hansen
Department of Botany and Plant Pathology, Oregon State University, Corvallis,
OR
97331
Susan Frankel
J.
USDA, Forest Service Pacific Southwest Research Station, 800 Buchanan Street,
West Annex Building Albany, CA 94710-001
Abstract
Non-native diseases ofdominant tree species have diminished North American forest biodiversity,
structure, and ecosystem function over the last 150 years. Since the mid-1990s, coastal California
forests have suffered extensive decline of the endemic overstory tree tanoak, Notholithocarpus
densiflorus (Hook. & Arn.) Manos, Cannon & S. H. Oh (Fagaceae), following the emergence ofthe
exotic pathogen Phythophthora ramorum and the resulting disease sudden oak death. There are two
centralchallengesto protectingtanoak: 1) thepathogenP. ramorum hasmultiple pathways ofspread
andisthusverydifficulttoeradicate,and2)theloweconomicvaluationoftanoakobscuresthecultural
and ecological importance ofthis species. However, both modelingand field studies have shown that
pathogen-centricmanagementandhost-centricpreventativetreatmentsareeffectivemethodstoreduce
rates of spread, local pathogen prevalence, and to increase protection of individual trees. These
managementstrategiesarenotmutuallyexclusive,butwelackpreciseunderstandingofthetimingand
extent to apply each strategy in orderto minimize disease and the subsequent accumulation offuels,
lossofobligatefloraandfauna, ordestructionofculturallyimportantstands. Recentworkidentifying
heritable disease resistance traits, ameliorative treatments that reduce pathogen populations, and
silvicultural treatments that shift stand composition hold promise for increasing the resiliency of
tanoak populations. We suggest distinct strategies for pathogen invaded and uninvaded areas, place
these in the context oflocal management goals, and suggest a management strategy and associated
research priorities to retain thebiodiversity and cultural values associated with tanoak.
Key Words: California Floristic Provence, disease ecology, genetic diversity, pathogen-caused
extinction, pathogen management, Phytophthora ramorum restoration, tanoak population decline.
,
MADRONO
152 [Vol. 60
To the detriment of cultural, economic, and occurs on tanoak twig lesions at levels sufficient
silvicultural interests, North America has experi- to spread the pathogen as well as cause bole-
enced multiple declines of dominant tree species lesionswhich are the principal infections thatlead
following the introduction and spread of exotic to mortality in tanoak and oak. In California
pathogens, many of which are innocuous mi- forests, sporulation can be an order ofmagnitude
crobes or weak pathogens in their native ranges greater on bay laurel and peak amounts for both
(Loo 2009). The quarantine ofknown pathogens species occur during warm spring rain events
and other techniques to avoid introduction of (Davidson et al. 2005, 2008; Mascheretti et al.
pathogens into novel environments are important 2008). In California, both tanoak and bay laurel
actions to reduce pathogen spread on interna- are central to pathogen spread within stands and
tional and inter-continental scales (Brasier 2008). across landscapes although risk of pathogen
However, the host range, invasion history, and establishment and spread rates are more strongly
persistence in the environment of many forest affected by baylaurel density (Meentemeyeretal.
pathogens suggest these invasive organisms have, 2008a; Davidson et al. 2011). Host impacts are
or will become, naturalized in their new ranges. strikingly different across species; P. ramorum
This challenging natural resource problem de- infection often leads to tanoak mortality, but has
mands well-crafted management efforts to avoid no known deleterious impacts on bay laurel at
the most extensive or severe impacts of forest either the individual or the population level
pathogens. (DiLeo et al. 2009; Cobb et al. 2010). In
The pathogen Phytophthora ramorum S. California, the stand-level densities of tanoak
Werres, A.W.A.M. de Cock (Werres et al. 2001) and bay laurel greatly influence tanoak mortality
exemplifies the destructiveness, epidemiological rate and population decline (Cobb et al. 2012b).
complexity, and difficulty in managing the most Pathogen dynamics are somewhat different in
problematic invasive microorganisms (de Castro Oregon forests where sporulation occurs year-
and Bolker2005; Smithet al. 2006). Phytophthora round and bay laurel does not play a significant
ramorum was first introduced into California and roleinspreadatthelandscapescale(Hansenetal.
spread into wildlands via infected nursery stock 2008). However, in both Oregon and California
in Santa Cruz and Marin Counties from a yet forests, tanoak population decline is expected
unknown native range (Garbelotto and Hayden across much of its endemic range (Fig. 1;
2012; Griinwald et al. 2012). The pathogen Meentemeyer et al. 2004; Vaclavik et al. 2010;
subsequently spread to Big Sur (Monterey Lamsal et al. 2011; Meentemeyer et al. 2011;
County) and Humboldt County, probably on Dillon et al. this volume).
infected plants, and has since become established Independent studies using repeated measure-
in central-California coastal forests. Establish- ments of individual tanoak show that mortality
ment of P. ramorum has been rapid during the rate ofinfected trees increases with size (McPher-
~20years sinceits arrival; spread ofthepathogen son et al. 2010; Cobb et al. 2012b). Although the
is aerial with extensive local dispersal (20-50 m) mechanism for this increase is not known, the
and less frequent but consistent long-distance pathogen’s selective removal of large tanoak
events of 1-3 km (Hansen et al. 2008; Meente- leads to rapid, stand-level loss oftanoak biomass
meyer et al. 2008a; Garbelotto and Hayden in P. ramorum invaded stands (Ramage et al.
2012). Phytophthora ramorum causes the disease 2011; Cobb et al. 2012b; Metz et al. 2012). Basic
sudden oak death (Rizzo et al. 2002) which has biodiversity information such as acorn produc-
been responsible for a regional-scale population tion rates, knowledge of species obligate to
decline of tanoak, Notholithocarpus densiflorus tanoak, and other ecological functions provi-
(Hook. & Arn.) Manos, Cannon & S. H. Oh sioned by mature tanoak trees have not been
(Fagaceae); formerly Lithocarpus densiflorus reported in the peer-reviewed literature. Howev-
(Manos et al. 2008), as well as the mortality of er, ecological functions of individual trees often
hundreds of thousands of oak trees Quercus increase with tree biomass, which scales expo-
(
spp.) from the red oak clade, primarily coast live nentially with tree size (Lamsal et al. 2011; Cobb
oak (Q. agrifolia Nee) throughout the invaded et al. 2012a). Selective removal of large trees is
areas of California (Meentemeyer et al 2008b; likely to accelerate loss ofecosystem characteris-
Davis et al. 2010; Metz et al. 2012). tics directly associated with tanoak, such as
Phytophthora ramorum has a broad host range, ectomycorrhizal associations or provisioning of
with over 140 known native and non-native habitat for other native flora and fauna which
plants including many trees, shrubs, forbs, rely on or are obligate to tanoak (Rizzo et al.
grasses, and ferns common to the California 2005; Bergemann and Garbelotto 2006; Cobb
floristic province (Griinwald et al. 2012). In et al. 2012b; Wright and Dodd this volume).
California, sporulation occurs on multiple hosts, The distribution of tanoak varies across the
the most important of which is the common landscapes of the California floristic province
forest tree California bay laurel (Umbellularia (Fig. 1), increasing in prevalence with latitude up
californica. Hook, and Arn.). Sporulation also to the northern range limit of the species in
;
2013] COBB ET AL.: TANOAK CONSERVATION STRATEGIES 153
match the scale of the outbreak increase the
likelihood ofsuccessful eradication ortheefficacy
ofslowingpathogen spread (Gilliganand vanden
Bosch 2008; Filipe et al. 2012). This requires
rigorous study to identify effective management
actions and minimizeunintended impactsthat are
counter to the overall goal. Our intent here is not
to review individual management techniques and
their efficacy, rather we seek to bring together
different management actions and place them
in the context of local disease conditions and
management goals. We begin by highlighting
several historical and contemporary mistakes in
forest disease management and emphasize the
importance of incorporating historical lessons
into sudden oak death management. We suggest
how judicious use of established techniques can
alter disease conditions and increase tanoak
population resiliency across different forest types
and community assemblages. We use the term
‘resiliency’ to mean retention of tanoak and the
ecological functions provisioned by this tree and
recognize that specific management objectives
will vary among ownerships, agencies and
situations. Part of our goal is to aid in the
selection of when, where, and how to increase
tanoak resiliency with existing management
actions. We stress that management resources,
land-ownercooperation, andchemical treatments
Fig. 1. The California floristic province with approx- will often be insufficient to control pathogen
imate extent of P. ramorum invasion and tanoak spread and disease emergence. Rather, when
dfiisrtmreidbutiinofne.ctPedhyttorepehst;hotrhae rmaamporwuamspogienntesrataerde cwoint-h management actions aim to retain tanoak as a
publically available data accessed in January 25, 2013 component of forest structure and composition,
from oakmapper.org and suddenoakdeath.org. we advocate for integration of host-centric
actions into management strategies. These strat-
egies rely on increasing tanoak resiliency through
southwestern Oregon. Several notable disjunct stand manipulation to reduce sporulation, in-
populations are located in the northern Sierra creasing distance among tanoak to reduce the
Nevada foothills and in small, isolated popula- probability oftransmission, and identification of
tions in Santa Barbara County, CA (Fig. 1 resistancein tanoak populations. Lastly, we stress
Lamsal et al. 2011; Meentemeyer et al. 2011; that continuous feedback among sudden oak
Dillon et al. this volume). While P. ramorum death research, management, and pathogen
establishment and disease emergence have been monitoring is needed to maintain tanoak as a
extensive in central coastal California, P. ra- component ofbiodiversity in many forests and to
morum has not yet invaded the majority of the reduce the spread ofP. ramorum.
tanoak range (Meentemeyer et al. 2011). How-
ever, the pathogenisexpected to invadeculturally Learning from Failure
important, highly susceptible North Coast forests
over the next several decades reflecting the The consequences ofinvasion ofCryphonectria
opportunity and importance of planning and parasitica(Murrill) Barrand theresultingdisease,
employing a proactive strategy to reduce disease chestnut blight, is a notorious example ofdisease
impacts. causing North American overstory tree loss
Phytophthora ramorum populations in newly (Anagnostakis 1987). Experience with chestnut
invaded areas are often below thresholds where blight can inform present-day challenges posed
the pathogen can be detected using landscape- by exotic pathogens, especially P. ramorum.
scale surveys. These cryptic invasions play an Cryphonectriaparasitica is a generalist pathogen,
important role in landscape-level spread, and selectively removes American chestnut Castanea
(
represent a critical challenge to eradication dentata Marsh.) from forest stands, and readily
attempts and other pathogen-centric disease disperses across local and regional scales (Loo
management strategies (Filipe et al. 2012). For 2009). Within 30 years of introduction by the
many invasive pathogens, rapid responses that pathogen, thesecharacteristicscausedawidespread
MADRONO
154 [Vol. 60
decline of American chestnut and loss of the Natural resource management relies on adap-
considerable ecological, cultural, and economic tive management to address imprecise knowledge
value that was associated with this species (Ana- and to avoid the field’s own version of “the twin
gnostakis 1987; Freinkel2007). Therateandextent traps of overtreatment and therapeutic nihilism”
of chestnut decline was unprecedented when the (e.g., the Hippocratic Oath, Edelstein 1943). The
disease emerged circa 1900, and remains as one of backbone of this approach is feedback among
the most infamous diseases in North American researchers, managers, and policy makers who
forest history because of the significant economic together establish experimental treatments, quan-
and cultural impacts caused by the epidemic tify their efficacy, and direct resources. An
(Freinkel 2007). analogous framework was not in place to address
Chestnut blight is also notable as a case of either chestnut blight or much of the manage-
failed management that stemmed from decisions ment response to the decline of eastern hemlock
with incomplete knowledge of pathogen biology populations. In the case ofsudden oak death, an
and host relationships. For example, the state adaptive management approach is more likely to
of Pennsylvania initiated an ambitious plan to identify treatments which judiciously use limited
remove infected chestnut trees and create a management resources (Filipe et al. 2012) and
barrier zone with the intention of restricting avoid loss of genetic diversity (Hayden et al.
pathogen spread into the western portion of the 2011; Kjaer et al. 2012) in comparison to reactive
state (Freinkel 2007). However, understanding of actions such as the attempt to construct a barrier
C. parasitica transmission pathways, estimates of to arrest the spread of chestnut blight (Freinkel
pathogen spread rate, and the capacity to detect 2007). The benefits of this approach are already
cryptic infection were all inadequate and patho- very clear; one of the most important initial
gen spread was unabated. A consequence of breakthroughs regarding sudden oak death was
dedicated and extensivechestnut cuttingwas, at a the discovery that P. ramorum was the cause of
minimum, an acceleration of chestnut removal widespread oak and tanoak mortality (Rizzo
from the overstory ofthese forests, an action that et al. 2002) which laid to rest spurious hypotheses
may have even expedited pathogen spread that the mortality was caused by insects or native
(Freinkel 2007). Today, American chestnut re- pathogens. Adaptive management further relies
mains common in eastern forests as an understo- on clear articulation of management objectives
ry shrub, but overstory trees and their associated while also recognizing the need to evaluate
ecological functions such as provisioningoflarge, treatments to achieve these objectives. The extent
nutritious, low-tannin nuts, have been lost from ofthe P. ramorum invasion suggestsmanagement
most ofthe species’ formerrange (Paillet2002). A goals must be comprehensive enough to match
full accounting of changes to native flora and the magnitude of the disease and recognize
fauna is hampered by the lack ofnatural history technical and practical limitations in order to
investigations of species associated with the tree identify where, when, and in what combination
prior to the epidemic. Although other species available techniques will be most effective in
replaced American chestnut in the overstory, loss slowing pathogen spread and maintaining the
of chestnut probably impacted small mammal cultural and ecological value oftanoak.
and invertebrate populations and may have
resulted in the extinction of several insects Tanoak Mortality: Patterns and Impacts
obligate to this overstory tree (Orwig 2002;
Ellison et al. 2005). Tanoak has lowtimbervalue and theeconomic
Regrettably, chestnut blight is not the sole value is further lowered because of competition
example of misdirected or ineffective manage- with commercial timber species, i.e., Douglas fir,
ment actions. Eastern hemlock Tsuga canadensis Pseudotsuga menziesii (Mirb.) Franco (Harring-
(
L.) forests are in decline across its native range ton and Tappeinier 2009). Before the 1950s,
from Georgia to southern Vermont and New tanoak-bark extracted tannins were important
Hampshire due to the invasion and subsequent to California’s leather industry, but the advent
regional-scale outbreak of hemlock woolly adel- of synthetic tanning compounds has shifted the
gid Adelges tsugaeAnnand), anexoticinsect pest economic perspective on this tree (Alexander and
(
(Fitzpatrick et al. 2012; Orwig et al. 2012). Lee 2010; Bowcutt 2011). Prior to the emergence
Eastern hemlock decline has been accelerated in of sudden oak death, most recent research
southern New England by pre-emptive hemlock involving tanoak focused on quantifying and
harvesting, regardless of stand invasion status reducing competition with coniferous species
(Orwig et al. 2002). From a conservation central to commercial timber production often
perspective, indiscriminant host tree removal will through the use of herbicides (Harrington and
accelerate population declines oftrees threatened Tappeiner 2009; Bowcutt 2011). Little basic
by insect or disease outbreak as well as degrade research has been conducted on tanoak and the
the genetic diversity of remaining populations measurement of growth characteristics, their
(Foster and Orwig 2006; Broadhurst et al. 2008). relationship with disease resistance, and descrip-
2013 COBB ET AL.: TANOAK CONSERVATION STRATEGIES 155
]
tion offlora and fauna obligate to or associated County, CA and Curry County, OR are notable
with tanoak have only recently been undertaken for their lower abundance ofbay laurel especially
(Bergemann and Garbelotto 2006; Cobb et al. compared to forests in the central coast, espe-
2012b; Wright and Dodd this volume). Although cially Big Sur (Monterey County) and Sonoma
the risk of pathogen-driven, regional-scale re- County (Lamsal et al. 2011). Phytophthora
moval of tanoak from the forest overstory is ramorum is likely to invade and cause disease
becoming increasingly clear, we face a critical even when tanoak is the sole sporulation-
lack ofunderstanding as to what and how many supporting species (Rizzo et al. 2005; Goheen
species may be threatened by the widespread et al. 2009; Meentemeyer et al. 2011) although
mortality ofthis common tree. mortality rates are likely to eventually slow
Even though P. ramorum has not yet invaded because of a negative feedback on pathogen
most of the tanoak range (Meentemeyer et al. populations when hosts die (Cobb et al. 2012b).
2011), the pathogen has significantly impacted Additionally, pathogen-killed trees or cut tanoak
tanoak from the Big Sur region (Monterey often develop prolific basal sprouts (Harrington
County, CA) north to Sonoma County, CA. In and Tappeiner 2009; Cobb et al. 2010; Ramage et
these regions, sudden oak death has caused al. 2011), which may be sufficient to perpetuate
widespread loss of overstory tanoak trees, pathogen populations and maintain spread in
significant shifts in overall forest composition, stands that do not include bay laurel (Cobb et al.
and outright extirpation of tanoak from some 2012b).
stands (Meentemeyer et al. 2008b; Ramage et al. Sudden oak death also has far-reaching eco-
2011; Cobb et al. 2012b; Metz et al. 2012). These nomic, cultural, and ecological impacts. Spread
forests occur across topographically, environ- ofP. ramorum via ornamental nursery plants was
mentally, and biologically diverse landscapes responsible for the initial introduction to Cali-
where several different conifer and broad-leafed fornia and continues to play a role in spread
trees are likely to experience increases in density, within the United States and internationally
biomass, and dominance as a result ofthe decline (Garbelotto and Hayden 2012; Griinwald et al.
of tanoak populations. Metz et al. (2012) and 2012). Regulation of nursery stock, a costly but
Cobb et al. (2010) examined sudden oak death- necessary management action, has been initiated
impacted California stands ranging from rela- to address this pathway ofpathogen spread (see
tively dry, mixed-evergreen forests to relatively also Alexander and Lee 2010). For individual
cool and moist redwood-dominated forests; in property owners, the loss of oak and tanoak
both studies, tanoak mortality was associated directly impacts property values (Kovacs et al.
with increased dominance of bay laurel. Given 2011), and increased disease-caused fuel loadings
that P. ramorum establishment and disease at wildland-urban interfaces could augment the
emergence is associated with bay laurel abun- risk of property loss during wildfire (Metz et al.
dance (Maloney et al. 2005; Meentemeyer et al. 2011; Valachovic et al. 2011). Similarly to
2008a; Cobb et al. 2010), bay laurel is often American chestnut, tanoak has historically been
situated to benefit from tanoak population an important component ofhuman nutrition and
decline. This shift in species composition has local culture; tanoak resources retain significant
the additional impact of improving habitat for cultural importance to many Native American
P. ramorum suggesting that the pathogen will communities in California (Bowcutt 2011). From
persist even, if tanoak becomes locally extinct a functional ecology perspective, tanoak is the
(Cobb et al. 2012b). sole ectomycorrhizal host in many forests (Rizzo
Additional shifts in species composition due to et al. 2005) and supports ectomycorrhizal diver-
sudden oak death are also likely in North Coast sity on par with Quercus species (Bergemann and
forests from Mendocino County to Curry Coun- Garbelotto 2006). Tanoak-wildlife associations
ty, OR. While the current distribution and are poorly described, but the large nutritious
prevalence of P. ramorum in the North Coast is acorns are likely used by many species and a
lower compared to the central coast where the recently described insect pollination pathway
poagtehnogienvnaswiaosnianintdialalyssionctiratoedducetdan(oFaikg.m1o)r,tpaaltiht-y saungdgeesctoslogyeitcalunindtoecruacmteinotnsedthaitnsmecatycboemmiumnpiatciteesd
have mobilized management actions by state, by the loss of tanoak (Wright and Dodd this
private, and federal land managers (Rizzo et al. volume).
2005; Valachovicet al. 2008; Goheen et al. 2009).
InNorth Coast forests, Douglas-fir and redwood Disease Treatments: Models
(Sequoia sempervirens [Lamb, ex D. Don] Endl.) and Experiments
have greater importance in terms of density,
biomass, and basis in local economies; these A central problem of managing emergent
characteristics are likely to favor an increase in pathogens is that the available biological or
dominance by these species in areas following epidemiological understanding is usually insuffi-
tanoak mortality. Notably, forests in Del Norte cient to control disease at the earliest stages of
MADRONO
156 [Vol. 60
outbreak. When epidemiological understanding in the surrounding area in an attempt to account
is flawed, seemingly appropriate management for undetected infections (treatment buffers) and
actions may actually increase pathogen spread have been implemented in southern Humboldt
and disease intensity (Ndeffo Mabah and Gilli- County, Redwood Valley (northern Humboldt
gan 2010), or simply be insufficient to control County), and Curry County, OR (Hansen et al.
pathogen spread despite large investments of 2008; Valachovic et al. 2008). These treatments,
management resources (Filipe et al. 2012). when done at an early disease stage, are clearly
Research is essential to addressing these ineffi- effective in reducing local pathogen prevalence
ciencies and maximizing the benefits of disease and the likelihood ofspread within treated stands
management. Epidemiological models can help (Hansen et al. 2008; Goheen et al. 2009),
guide management by estimating the efficacy of suggestingoverall rates ofspread in thelandscape
treatments and forecasting the spread and are slowed by aggressive and rapid interventions
impacts of pathogens based on current data. Of (Filipe et al. 2012). However, in each case the
equal importance are field experiments designed pathogen has not been eradicated by the treat-
to test treatment efficacy and model structure. ments and spread throughout the landscape has
Field experiments are also essential in establish- continued. Undetected, or cryptic, infections are
ing a feedback among treatment application, an important source ofthis continued spread into
monitoring, and research efforts, including mod- uninfected stands (Filipe et al. 2012). The broad
eling, which comprise adaptive management. For P. ramorum host range, asymmetric impacts on
example, field studies improve the understanding hosts (some hosts support sporulation but are
of pathogen biology and epidemiology, which not killed by infections), cryptic infections, and
can be applied in models to identify the most survival in habitats such as soil and watercourses
appropriate treatments thus generating hypothe- make eradication from a region exceedingly
sesthat are testablewith further field experiments difficult (Hansen et al. 2008; Goheen et al.
(Sniezko 2006; Gilligan and van den Bosch 2008). 2009). However, disease management need not
Models ofdisease risk based on host distribu- attain 100% pathogen reduction to be effective.
tion, environmental conditions, and pathogen Slowing pathogen spread within the landscape
spread are useful tools for understanding where is valuable, as it provides time to plan and
and when tanoak mortality will occur. The initial proactively manage for ecosystem impacts, such
risk models for sudden oak death in California as increased fuel loads, stress to biodiversity
andOregon(Meentemeyeretal. 2004; Vaclaviket associated with tanoak, and the accompanying
al. 2010) identified locations that are optimal for loss ofacorn production.
pathogen establishment. These models have been As new isolated outbreaks emerge in northern
reliable tools because the underlying epidemio- California and southern Oregon it will become
logical assumptions have subsequently been necessary to focus limited management resources
demonstrated to be biologically accurate (Meen- to achieve specific cultural and biological conser-
temeyer et al. 2012). More recent models incor- vation goals. Inoculum-reduction treatments
porate spread following establishment in a given such as attempts to locallyeradicate the pathogen
setoflocations(Meentemeyeretal. 2011; Filipeet are expensive to apply and demand significant
al. 2012). These dynamical models enable estima- investments from land managers, researchers,
tion of a probable time of invasion for specific and field staff (e.g., Valachovic et al. 2008)
tanoak populations and suggest that the majority suggesting a shift from pathogen-centric to
ofthetanoakrangeisatriskofpathogeninvasion host-centric management will be more effective
and disease emergence over the next several to reaching conservation goals when an individ-
decades. In conjunction with tanoak distribution ual outbreak reaches some yet undefined thresh-
models (Lamsal et al. 2011), these P. ramorum old. However, limiting spread into specific unin-
spread models hold promise to predict the timing vaded landscapes will be aided by attempting
of disease emergence, risk of tanoak population eradication of isolated outbreaks as these treat-
decline (Dillon et al. this volume), and changes in ments are likely to reduce inoculum (Filipe et al.
ecosystemfunctions suchascarboncycling(Cobb 2012). For example, the Redwood Valley out-
et al. 2012a). Further development ofmodels and break is strategically important given its proxim-
datasets to identify levels ofbiodiversity, cultural, ity to culturally and ecologically valuable tanoak
or ecological value would be useful as an overlay stands on tribal lands, within Redwood National
with risk models (e.g., Meentemeyer et al. 2004; Park, and an other nearby public and private
Vaclavik et al. 2010) to aid in identification of lands with large tanoak populations. Pathogen-
specific stands where resources for tanoak con- centric treatments in Redwood Valley have likely
servation should be focused. slowed P. ramorum spread into these valuable
Management of isolated outbreaks of P. nearby areas by reducing inoculum loads. How-
ramorum has largely relied on removal ofinfected ever, further removal of isolated infections
hosts to reduce inoculum. These treatments are outside of the treated area may be needed to
often combined with removal ofsusceptible hosts attain maximum inoculumreduction oras part of
2013] COBB ET AL.: TANOAK CONSERVATION STRATEGIES 157
host-centric management that increases the resil- nate, in general protection ofindividual trees has
iency of these tanoak populations and protects been shown to be effective in slowing spread
their associated biodiversity, ecological function, within some populations. Genetic variation
and cultural value. When to shift from pathogen- within host species may also reduce susceptibility
centric to host-centric management oridentifying and subsequent pathogen spread. Bay laurel
the optimal balance of each approach is an populationsin Oregon (knownlocallyas “myrtle-
outstanding question for land managers working wood”) are not as abundant, susceptible, and
to address consequently not as important in spreading P.
P. ramorum and other forest disease outbreaks. ramorum as they are in much of California
Epidemiological models can inform both path- (Hansen et al. 2005; Hiiberli et al. 2011),
ogen-centric and host-centric management ac- suggesting genetically based differences among
tions (Fig. 2). Often, treatments for each man- host populationsmayinfluencedisease severityin
agement type will use the same techniques (stand different regions. Similarly, Hayden et al. (2010)
thinning, infected host removal), but the timing found evidence of resistance to P. ramorum
and location of treatments may differ. Some infection within tanoak populations, indicating
insights into the design ofthese treatments can be that patterns ofresistance could influence tanoak
gleaned from stand-level epidemiological models susceptibility or mortality within stands. Similar-
which often describe the amount of inoculum ly to chemical protection treatments, further
thatcanreach a given individualhostthroughthe work is needed to understand how resistance
concept offorce ofinfection (FOI) (forexamples, influences pathogen spread rates.
see Meentemeyer et al. 2011; Cobb et al 2012a; Recent modeling results suggest the degree to
Filipe et al. 2012). FOI often has two compo- which management actions reduce P. ramorum
nents: the number ofinfected hosts weighted by a infection rates can directly influence the number
dispersal kernel (a function of relative distance of tanoak that maintain an overstory canopy
between hosts), and the rate at which each position in P. ramorum invaded forests (Fig. 3).
infected individual transmits the pathogen to Following a line oftested epidemiological models
susceptible individuals. Sporulation rates vary (Filipe and Gibson 2001; Gilligan and van den
across host species and are dependent on Bosch 2008; Ndeffo Mabah and Gilligan 2010;
environmental conditions (rainfall, temperature), Filipe et al. 2012), Cobb et al. (2012a) developed
therefore formulations of FOI may include a stand level model of disease outbreak with
species specific pathogen spread factors as well dynamic vegetation composition and tanoak
as environmental variation. In addition, the size structure. These models demonstrate that
length of time an individual remains infectious, a desired outcome of host-centric management
known as the infectious period, also determines may be feasible: retention of tanoak, including
the number of individuals which spread a large tanoak, even though P. ramorum has
pathogen. Infectious period is poorly quantified invaded the stand. The model shows a critical
for bay laurel and tanoak, but models demon- level oftanoak density (tanoak stems ha-1) below
strate that this parameter can influence rates of whichthepathogenisnot ableto maintain spread
pathogen spread at stand-to-landscape scales as from host to host and consequently, does not
well as the persistence of the pathogen within diminishtanoakpopulations orremove overstory
invaded stands (Cobb et al. 2012b; Filipe et al. tanoak. These results assume bay laurel is not
2012). The rate of spread and tanoak mortality present in, or has been removed from the stand
within a population or landscape increases with and that remaining tanoak trees are evenly
FOI and management actions can be directed at distributed across the stand. The models also
one or more of its components. Therefore, the suggest that slowed host-to-host infection rates
achievement ofspecificmanagement goals will be result in better retention of the habitat and
improved byunderstandingwhichcomponents of ecosystem function provided by tanoak (Fig. 2).
FOI are being acted upon by a given treatment. Slower infection rates could be achieved by
In general, eradication treatments reduce the increasing tanoak resistance or stand-level appli-
number of hosts spreading the pathogen but do cation of effective chemical protection. How-
not change the rate ofspread from each infected ever, identification of tanoak resistance prior
host. This is important because infected individ- to treatments is currently limited by the lack of
uals that remain untreated support continued molecular-resistance markers; this technical lim-
local and regional pathogen spread (Hansen et al. itation must be overcome to use resistance
2008; Filipe et al. 2012). Prophylactic chemical measurements in broader-scale preemptive treat-
treatments, such as phosphonate injections, have ments. Regardless, the epidemic threshold pre-
been shown to reduce the susceptibility of dictions ofthe model (e.g.. Fig. 3) should still be
individual tanoak against P. ramorum (Garbe- tested with field experiments, including current
lotto and Schmidt 2009); while further research is phosphonate applications (Garbelotto and
needed on the efficacy, longevity, phytotoxicity Schmidt 2009) and tanoak thinning treatments
and environmental interactions with phospho- to understand under which conditions these
MADRONO
158 [Vol. 60
Fig. 2. Landscapes and stands that have not been invaded by Phytophthora ramorum (A) offer different
opportunities and require different management approaches compared to invaded forests (B) when the goal is to
retain tanoak and its associated flora and fauna. Variation in the likelihood ofpathogen invasion into pathogen-
freelandscapes and standsislargelydrivenbyproximitytopathogen invadedstandsandcommunitycomposition.
Community compositioncan be manipulated to slow invasion and disease impacts. Disease impacted stands (e.g.,
B) require a host-centric management approach that avoids removal of surviving tanoak, protects potential
genetically based resistance, and retains ecological functions and species obligate to tanoak. In both cases,
successfulconservation oftanoakwill requirecontinuous feedback betweenresearch and managementto establish
baselines, test hypotheses, and determine the efficacy ofmanagement actions.
density thresholds are overcome through long- be measured at least in part by the success in
distancedispersalevents (Filipeand Gibson2001; retaining biodiversity and function associated
Gilligan and van den Bosch 2008), sporulation with tanoak; retention of tanoak is likely to be
from species not thought to be epidemiologically easier in community or environmental conditions
significant, or human-mediated transport of where pathogen spread is lower. Currently, many
infected plants (Cobb et al. 2012b; Filipe et al. culturally and ecologically significant tanoak
2012). populationshaveyettobeinvadedbyP. ramorum
(Lamsaletal. 2011; Meentemeyeretal. 2011), and
Management Goals Defined: Conserve timely actions have the potential to reduce future
Biodiversity Associated with Tanoak impacts ofthis disease.
Phytophthora ramorum has, and will continue Invaded Stands and Landscapes
to alter the distribution, prevalence, and biomass
of tanoak across California and Oregon. A Invasion status is a useful first distinction for
difficult reality of this disease is that much of management efforts and techniques as P. ra-
this tanoak loss has and will occur regardless of morum invasion of a stand eliminates the
the amount and efficacy ofmanagement actions. usefulness of many treatments (Fig. 2). For
However, the available management tools and example, prophylactic treatments are ineffective
current understanding of the disease may be on trees that are already infected (Garbelotto and
sufficient to avoid the worst possible outcomes Schmidt 2009). Similarly, limiting inoculum
such as the removal oflarge trees and consequent influx into stands where the pathogen is already
loss of ecological function (cf., Orwig 2002; established may have a minimal effect in reducing
Ellison et al. 2005). Management efficacy should disease because most inoculum is produced
2013] COBB ET AL.: TANOAK CONSERVATION STRATEGIES 159
ments (such as eradication attempts) at the initial
invasion orpreemptive host removal ahead ofthe
infection, and construction ofa host-free barrier
at the landscape-scale. Among these treatments,
preemptive removal of hosts ahead of the initial
invasion and removal of infected hosts at the
initial invasion, similar to
the strategy in Oregon, were found to be most
effective in slowing P. ramorum spread as long as
the treatments were applied early during the
epidemic and the scale oftreatments matched the
scale of the invaded area. The landscape-barrier
infection rate yr'1 treatments were not effective in slowing the
spread or protecting particular areas because
continued spread from undetected infections, or
long-distance dispersal across the barrier over-
came these treatments.
In regions with the most extensive P. ramorum
invasion, eradication attempts are a poor choice
of management. In highly invaded stands from
Big Sur to Sonoma County, P. ramorum popu-
lations are so large and widely established that
any benefit from local removal of infected trees
would be overcome by reinvasion from adjacent
stands. Eradication attempts always represent a
tradeoff between causing tree mortality through
cutting and herbicides vs. allowing the pathogen
to cause tree mortality. In highly invaded
Fig. 3. The epidemiological characteristics of Phy- landscapes, management should be aimed at
tophthora ramorum affect the threshold density oflarge reducing disease impacts and protecting high-
tanoaktrees(alltrees > 30cmdiameterat 1.3m)below value individual stands or trees. For example, in
which P. ramorum outbreak cannot be sustained in the fire-prone Big Sur landscape, high-density
2a01p2abr)a.me(tAe)rTizheedepeipdiedemmiicotlhorgeischaolldmsofdoerlta(nCooabkbdeentsiatly. tanoak stands are problematic because sudden
decreases with increasing infection rate. (B) Similarly, oak death generates significant amounts ofdead
the density of large tanoak decreases with increasing fuels (Metzet al. 2011; Cobb et al. 2012a). Stands
infectious period (log scale; infection rate was kept with accumulated dead fuels also suffered greater
constant at 0.3). amounts of soil damage during the 2008 Basin
Fire (Metz et al. 2011; Big Sur). These patterns
suggest treatments which increase resiliency of
locally (Davidson et al. 2005; Mascheretti et al. tanoak populations to P. ramorum while also
2008; Cobb et al. 2012b). Phytophthora ramorum- reducing dead fuels may have substantial benefits
invaded stands are potential pathways of spread in terms of reducing ecological costs caused by
and the quarantine of infected, sporulation- wildfire in disease impacted forests. Disease-
supporting plants from these stands is ajudicious generated fuels are dependent on the rate of
action to reduce, though not eliminate, this mortality and the amount of tanoak biomass
pathway oflong-distance pathogen spread. present before invasion. This suggests preemptive
Managers increasingly face isolated outbreaks reduction of tanoak density combined with bay
that threaten to spread into the wider landscape. laurel removal is likely to slow disease-driven
In Oregon, emphasis has been placed on moni- dead fuel accumulation and reduce maximum
toring and rapid implementation of eradication- amounts (Valachovic et al. 2011; Cobb et al.
type treatments within a designated quarantine 2012a). These treatments could easily be struc-
zone (Hansen et al. 2008). In northern Humboldt tured to retain specific trees such as larger or
County, theisolated outbreak in Redwood Valley more resistant individuals and increase spacing
was also addressed with a rapid eradication between individual trees in order to reduce
attempt response, but discovery of the pathogen within-stand spread. However, the effects of
outside of the treatment area raises the question increasing stand openness are unresolved; path-
of what follow-up treatments are needed to ogen dispersal may be less impeded in more open
maximize the benefit of the initial intervention. stands or, conversely, microclimate conditions
The model designed by Filipe et al. (2012) was could be less suitable to sporulation and infection
used to examine a number of landscape-level (Rizzo et al. 2005). This uncertainty reflects the
treatments including inoculum-reduction treat- need for further experimental treatments to
MADRONO
160 [Vol. 60
understand the interactions of management tanoak populations following challenge by P.
actions with the epidemiology of P. ramorum. ramorum (Fig. 3). In a study quantifying patterns
The Big Sur region is a strong candidate for of tanoak resistance, Hayden et al. (2010; 2011)
experimental fuel and disease reduction treat- found variation in susceptibility within tanoak
ments given the frequency of fire and the extent populations, but fairly equivalent susceptibility
of sudden oak death in this region (Rizzo et al. among populations. If this pattern is broadly
2005; Meentemeyer et al. 2008b; Metz et al. representative of tanoak genetic patterns, indi-
2011). vidual stands are unlikely to resist invasion by
P. ramorum on the basis of genetic composition
Uninvaded Stands and Landscapes alone. However, resources from within stands
could be used to maximize tanoak population
The most effective strategy to prevent impacts resiliency in conjunction with other treatments,
of sudden oak death is to prevent P. ramorum or to develop less susceptible tanoak growing
invasions into new, at risk environments alto- stock suitable for restoration in P. ramorum-
gether. However, the history of P. ramorum is invaded areas and preemptive planting treat-
notable for numerous and often surprising ments in uninvaded landscapes. Further work is
invasions (Werres et al. 2001; Rizzo et al. 2005; needed to increase the ease and rate of identify-
Hansen et al. 2008; Brasier and Webber 2010) ing tanoak resistance, to determine if rates of
and current models suggest many stands and sporulation differ in less susceptible trees, and to
landscapes will be invaded in the coming decades develop tanoak suitable for restoration.
(Meentemeyer et al. 2011). Proactive manage- Where the management goal is to retain
ment ofuninvaded stands may be much easier to biodiversity and function associated with tanoak,
implement where manipulation of stand and treatments need not conserveeverylivingtanoak.
community structure are already planned to Rather, these treatments must retain tanoak
achieve management goals apart from disease. populations large enough to maintain both the
Where ever possible, host-centric management species and its valued ecological functions.
should aim to achieve multiple goals (Fig. 2). Moreover, sampling and conservation efforts
The design, implementation, and evaluation of should be targeted to retain genetic diversity
preemptive disease treatments are difficult from a acrossregions ofknowntanoak differentiation. A
research perspective because invasion dynamics geneticanalysis ofneutralmarkersbyNettelet al.
are idiosyncratic. Without pathogen invasion ofa (2009) has shown deep divisions in chloroplast
stand-whichis neverguaranteed the efficacy ofa markers between central coastal California ta-
preemptivetreatmentcannot be evaluated. At the noak populations and northern coastal-Califor-
same time, intentional pathogen introduction to nia/Klamath/Sierra tanoak populations, along
uninvaded areas must be ruled out on ethical with evidence ofconsiderable pollen flow among
grounds. We suggest two actions that help populations. The variance structure in neutral
overcome this practical problem. First, disease nuclear genetic markers was similar to the
management should be put in thecontext oflong- variance in resistance measured on leaves cut
term management goals, such as the reduction from wild trees, with more variance within
of fuel loads, growth of timber species, desired populations than there was among populations
community composition or canopy structure, (Hayden et al. 2011). In pathogen-invaded
and/or enhancement of biodiversity. Second, landscapes, disease may be a useful force to select
further effort should be allocated to developing for more resistant tanoak. Hayden et al. (2011)
models for management evaluation. Models suggested that greater prevalence ofresistance at
provide insight into the consequences of land- some sites could reflect a selection effect of the
scape-level management when experiments at pathogen, and McPherson et al. (2010) reported
this spatial and/or temporal scale are not thesurvivalofseveralmaturetanoaksafternearly
possible (Meentemeyer et al. 2012). Lastly, to a decade ofmonitoring in a P. ramorwm-invaded
implement treatments for emerging outbreaks, forest. These results argue for retention oftanoak
we caution that management goals must be where P. ramorum has already killed a large
consistent with the priorities and level of portion of the tanoak population. In highly
cooperation within the local community (Alex- diseaseimpactedstands, treatmentswhichremove
ander and Lee 2010). survivingtreeswouldclearlyexacerbatethelossof
tanoak and may reduce genetic resources impor-
tant for developing molecular markers of resis-
Genetically-Based Resistance
tance and restoration planting stock.
Genetically-based resistance to P. ramorum in
tanoak has the potential to be of great practical Protect High Conservation Value Stands
value. Stands with greater innate resistance are
likely to experience lower rates of pathogen Targeted conservation acquisitions, such as
spread, fewer disease impacts, and have larger land purchases or conservation easements, are