Genetic Structure of Notholithocarpus densiflorus (Fagaceae) from the Species to the Local Scale: A Review of Our Knowledge for Conservation and Replanting

2013·8.3 MB·English

by  Richard S Dodd

Madrono, Vol. 60, No. 2, pp. 130-138, 2013
GENETIC STRUCTURE OF NOTHOLITHOCARPUS DENSIFLORUS
(FAGACEAE) FROM THE SPECIES TO THE LOCAL SCALE: A REVIEW OF OUR
KNOWLEDGE FOR CONSERVATION AND REPLANTING
Richard Dodd
S.
Department of Environmental Science Policy and Management, University of California,
CA
Berkeley, 94720
[email protected]
Alejandro Nettel
Campus del Mar, Universidad de Ciencias y Artes de Chiapas. Prolongacion Juan Jose
Calzada s/n., Tonala, Chiapas, Mexico
Jessica W. Wright
USDA-Forest Service, Pacific Southwest Research Station, 1731 Research Park Drive,
CA
Davis, 95618
Zara Afzal-Rafii
Department of Environmental Science Policy and Management, University of California,
CA
Berkeley, 94720
Abstract
Tanoak, Notholithocarpusdensiflorus(Hook. &Arn.) Manos, Cannon&S. H. Oh(Fagaceae), isan
important component ofmixed-evergreen forests and woodlands in coastal California and Oregon,
withincursionsintotheSierraNevadaandtheKlamathRanges. SuddenOakDeath(SOD)iscausing
severe dieback and mortality in tanoak and could transform these ecosystems in areas where the
pathogenPhytophthoraramorum S. Werres, A.W.A.M. de Cockcan becomeestablished. Knowledge
ofgenetic diversitywithin the species is important forboth disease resistance screening, conservation
and replanting in sites with high mortality. Here we review what has been learned about the genetic
structure within tanoak since SOD has caused disease epidemics in the species. We review published
work on genetic structure at the species level and provide some re-analyses ofthese data that show
divergenceacrossthegeographicrange.Wealsoreviewrecentlypublisheddataongeneticstructureat
a fine spatial scalethat provides some guidelines for the selection oftrees as seed sources. Finally, we
interpret a range of seed provenancing strategies in the light of our knowledge of tanoak genetic
diversity.
Key Words: Conservation, fine-scale genetic structure, genetic divergence, Notholithocarpus
densiflorus replanting, tanoak.
,
Human-induced perturbations as a result of keystone species have been threatened or reduced
resource use, habitat fragmentation and climate from overstory trees to understory shrubs.
change are leading ecosystems to functional Chestnut blight is a notable example on the
tipping points with potentially far-reaching con- American continent (Garnas et al. 2011) and
sequences (Barnosky et al. 2012). In forested pandemics of Dutch elm disease have been
ecosystems, the fine balance between a function- catastrophic across the northern hemisphere
ing system in which hosts and pathogens co-exist (Brasier and Buck 2002).
and the rapid decline that ensues when environ- The current epidemic of Sudden Oak Death
mental conditions push the system out of (SOD) caused by the exotic pathogen Phy-
equilibrium is an example of such a tipping tophthora ramorum S. Werres, A.W.A.M. de
point. Recently, this has been exacerbated by the Cock has a very wide host range and is shaping
movement of organisms globally, resulting in up to be a potentially devastating disease on
exotic diseases with catastrophic consequences. tanoak, Notholithocarpus densiflorus (Hook. &
Foresttreesareparticularlyvulnerable because of Arn.) Manos, Cannon & S. H. Oh (Fagaceae), in
long generation times and sedentary life histories, central and northern California (Rizzo et al.
but they are also critically important as keystone 2005). Epidemiological risk analyses suggest that,
species of the broader ecosystem. Over the last like other systems, tanoak forests are likely to be
century, several forest tree diseases have caused transformed through the loss of overstory trees
such severe host mortality that, what were once and may only avoid extinction through successful
2013] DODD ET AL.: GENETIC STRUCTURE OF NOTHOLITHOCARPUSDENSIFLORUS 131
resprouting (Cobb et al. 2012). If the disease organization ofpedigrees and the effective limits
takes the course predicted by Cobb et al. (2012), to recent seed and pollen dispersal. The lattercan
replanting will be necessary to augment genetic be important in resistance screening, where
diversity that may be lost stochastically as stands sampling of relatives may be desirable and in
regenerate vegetatively and to re-introduce diver- conservation and restoration where maximizing
sity in stands that fail to resprout. Selection of for genetic diversity in seed selection is the goal.
suitable seed sources will be critical to the We have been studying genetic diversity in
conservation of genetic resources as well as for tanoak, with aview to understandingpartitioning
successful replanting. There is considerable de- of genetic diversity at the broad (Nettel et al.
bate as to the optimum strategy for selecting seed 2009; Dodd et al. 2010) and fine-scale landscape
sources that will maximize the likely success of levels (Dodd et al. 2013). Here, we review these
local populations, while promotinga diverse gene earlier reports, re-analyze some of our earlier
pool on which selection can act ifenvironmental data with new tools and discuss the consequences
conditions change (Breed et al. 2012). Seed for conservation and replanting strategies in the
collection from local sources to maximize likely face of mortality from disease outbreaks. We
locally-adapted genotypesmay lead to inbreeding refer to replanting rather than restoration be-
depression ifpopulation sizes have fallen below a cause we are not proposing strategies to restore
threshold level (Eckert et al. 2010; Breed et al. the ecosystem to its former state, but we are
2012), whereas mixing seeds from different looking specifically at replanting of tanoak in
geographic sources introduces the risk of out- areas where it has been lost through perturba-
breeding depression ifthe source populations are tions such as SOD.
from different locally adapted gene pools (Kra-
mer and Havens 2009). Reciprocal transplant The Species
studies are helpful to gain some insights among
these competing scenarios. However, for many Tanoak Notholithocarpus densiflorus was
( )
tree species, such approaches are too time- recently attributed to its monospecific genus by
consuming, so molecular marker data need to Manos et al. (2008) in recognition of its closer
be used. Molecular methods allow the rapid affinities to Quercus Castanea and Castanopsis
,
cdaosceusmesnetqautenicoensooffgeDnNetAi,c doirverasniatly,ysebsutoifnDmNosAt TthaannoatoktihseaArseisatrnicstteodneenodaekmsic(LtiothtohcearCpaulsifsoprpn.i).a
fragment sizes are confined to regions of the Floristic Province; the tree form, N. densiflorus
genome that are not transcribed and are assumed var. densiflorus ranges from disjunct stands in
DNA
to be selectively neutral. Although such Ventura County, California to more continuous
markers do not necessarily indicate genetic stands as far north as Coos County, Oregon
adaptations, they do provide us with information (Tappeiner et al. 1990). Although it is best
on evolutionary divergence that would indicate represented in the Coastal Ranges, it extends
zones for conservation management (Moritz inland to the foothills ofthe Sierra Nevada. The
1999) and seed sources that should not be mixed dwarfform, N. densiflorusvar. echinoides is more
for planting if outbreeding depression is a common at higher elevations in the north-eastern
concern. range of the species, particularly on serpentine.
Spatial organization of genetic diversity is a Occurrence reports of the dwarf form should be
property of species that derives from contempo- treated with caution as there is likely to be
rary demographic processes and past environ- confusion over ecological dwarfism and geneti-
mental events. At a broad spatial scale, encom- cally distinct forms (Dodd unpublished data).
passing a range of environmental conditions, Tanoak is predominantly insect-pollinated (see
evolutionary forces operate on the different gene Wright and Dodd this volume) and can produce
pools and can lead to more or less divergent heavy crops ofacorns that have a relatively high
lineages (Hampe and Petit 2005). Detecting viability; in addition it is a prolific basal sprouter.
genetically divergent groups of individuals or
populations is an area ofinterest in evolutionary Genetic Structureatthe Species Fevel:
biology (Pritchard et al. 2000; Manel et al. 2003; Conservation Units
Waples and Gaggiotti 2006) and has consequenc-
es for the evolutionary trajectory ofhost respons- In defining management units for conservation
es to disease (Parker and Gilbert 2004) and for under the pressures of climate change, Moritz
designing appropriate conservation programs (1999) argued for the greater importance of
(Grivet et al. 2008). At a local scale, genetic preserving distinct lineages over phenotypic or
structure varies as a result primarily of the adaptive traits: while adaptations can be recreat-
balance between dispersal and stochastic process- ed given an adequate gene pool, historical
es of chance representation of alleles (genetic processes cannot be repeated. Neutral genetic
drift). Knowledge of fine-scale genetic structure markers have been used extensively to detect
(FSGS) provides information on the spatial lineages that have resulted from an overlay of
MADRONO
132 [Vol. 60
past events on biogeographic processes. Today, northern coastal chlorotype) in response to
one ofthe greatest problems facing conservation inoculation (Hayden et al. 2011).
geneticists is how to determine what constitutes
distinct population lineages. In plants, chloro- Do Nuclear DNA Markers Support Divergence
plast and nuclear genomes are the most com- Revealed By The Chloroplast Genome?
monly studied, but they may show distinctly
different underlying DpaNttAerns. Maternal inheri- As for othermembers ofthe Fagaceae (Sork et
tance of chloroplast in most angiosperms al. 2010),onlyweakgeneticstructurewasdetected
commonly results in very strong geographic amongpopulations oftanoak (Nettel et al. 2009).
structure, particularly for heavy-seeded genera This is generally interpreted to be a result of
with short dispersal distances such as Quercus widespread pollen dispersal in many trees species
(Dodd et al. 2008; Grivet et al. 2008) and (Ashley 2010). Field sampling for phylogeo-
Notholithocarpus (Nettel et al. 2009). On the graphic studies is commonly designed to cover
other hand, bi-parental inheritance of nuclear the range of distribution of the species and to
DNA results in much weaker geographic struc- capture any potentially, naturally isolated groups
ture, as a result of large amounts of pollen with ofindividuals. Unfortunately, although sampling
much greater dispersal distances than seed. locations are commonly referred to as popula-
In tanoak, Nettel et al. (2009) detected four tions, there is no a priori knowledge of what
major and two rare chlorotypes based on four constitutes ageneticpopulation, ofteninterpreted
chloroplast microsatellite repeats (cpSSR) and as apanmicticgroup(but seePalsbolletal. 2007),
one sequence region. No additional chlorotypes which then becomes a problem of post hoc
were detected in more extensive sampling of the inference. To infer panmictic groups, Nettel et
species, particularly to the northern limit of its al. (2009) usedthe Bayesianapproachimplement-
STRUCTURE
range in southern Oregon (Dodd et al. 2010). The ed in (Pritchard et al. 2000).
four major chlorotypes included: A coastal However, it is well accepted that when genetic
California type from Santa Barbara t1o. Humboldt structure is weak, the program STRUCTURE
County, 2. A northern type from Humboldt does not perform well in detecting the number of
County to the northern limit of the range in groups. An alternative Bayesian analysis BAPS
southern Oregon, 3. A Klamath Mountains performed on a larger data set, including more
chlorotype and, 4. A Sierra Nevada chlorotype extensive sampling in southern Oregon, detected
t(Faingo.ak1).diTshpelraeyfsorae,weltlh-edecfhilnoerdopgleaosgtrapgheincomsteruci-n nailntehoucglhusttheersBA(PDSodadssiegtnmaeln.ts20i1n0di)c.ateHoswpaetviealrl,y
ture, but low variation compared to California consistent grouping of populations, they do not
show the relative importance ofthe partitions.
Floristic Province endemic oaks, such as coast
live oak Quercus agrifolia Nee), for which we Here, we have extended our analyses using the
have dete(cted 31 chlorotypes (Dodd et al. 2008) software Barrier vs. 2.2 (Manni et al. 2004). The
and valley oak Q lobata Nee), for which 22 advantage of this analysis is that a hierarchy of
( . importance of discontinuities and their direction
cHholwoervoetry,pescowmepraerisdoetnesctaemdonGrgivseptecieetsasl.ho(u2l0d08b)e. on the landscape can be inferred. Barrier uses
centralcoordinates fora sampled population first
treated with caution because of potential varia-
to calculate a Voronoi tessellation and then a
tions in diversity at the selected gene loci. Delaunay triangulation that draws a network
Although the geographic partitioning is informa- connecting all the sample localities. The Mon-
tive on barriersto seeddispersal, it does notmean monier’s (1973) maximum-difference algorithm
that seed dispersal can occur throughout the
then takes a distance matrix (genetic distance) to
geographic range of a chlorotype. Analysis of identify boundaries, where differences between
more of the chloroplast genome could reveal pairs ofpopulations are greatest. The significance
further variation within any of the four major of these barriers can be tested by bootstrapping
types detected. Perhaps the most interesting the distance matrix after re-sampling genotypes
finding from these results is the marked break in each population. We used the population
nbeeatrweAerneatthae,noCrtAhe(rnHuamnbdolsdotuthCeorunntcoya)staalndtyptehse caonmdpLairsihseorn2s0o1p0t)itoonoibntAarilneaqumiantr3i.x5.o1f.3p(oEpxucloaftfiioenr
divergence between coastal and interior popula- pairwise genetic distances (Slatkin’s linearized
itnitoenrsesotffotranSoOakD.oTuhtebrefaokrsmearndspsluitggeisstosfngorretahteersnt FsaSmt)p-lTedogbeonoottsytpreaspwdiitshtarnecpelmaactermiecnets,1w00etfiirmsetsrien-
and southern coastal lineages that have been each of the populations using the Excel add-in
separated and only recently come into secondary provided by Resampling Stats, Inc., Arlington,
contact. Whether the two lineages differ in their Virginia, USA. We then re-calculated the matri-
tolerance to P. ramorum remains to be seen, but ces ofpopulationpairwise Slatkin’s linearizedFst
preliminary data do suggest low, but significant for each ofthe 100 re-sampled data sets. We ran
variation among populations (not including the Barrier to detect the locations of the first five
.
2013] DODD ET AL.: GENETIC STRUCTURE OF NOTHOLITHOCARPUSDENSIFLORUS 133
0.62
0.54
0.46
0.38
36 40
Latitude
Fig. 1. Map ofchloroplast andnuclearDNAbreaks in the distribution oftanoak(Notholithocarpusdensiflorus)
Filledcircles showsamplinglocationswithcolorsdepictingdifferentchlorotypes (adapted fromDoddet al. 2010).
Linesshowthemajorbreaks innuclearDNAfrom BARRIER analysis ofpopulationpairwise Slatkin’s linearized
Fst conducted in this report. Thickness ofline segments are proportional to the ratio ofFsr ofthat segment to
maximumEVt. Numbers nextto thickest line segments showproportional bootstrap support afterresampling 100
times within populations. Inset shows relationship between expected heterozygosity (He) and latitude oforigin of
population (adapted from Nettel et al. 2009).
MADRONO
134 [Vol. 60
barriers in the genetic distance matrix and used influenced by changes to a drier climate in the
the resampled data to determine bootstrap late Holocene (Kirby et al. 2007). Indeed, Nettel
support. et al. (2009) inferred decreases in effective
The breaks detected by Barrier are shown in population size since the mid-Holocene. Today
Figure 1 as lines of variable thickness corre- these populations are fragmented and relatively
sponding to the ratio of the FST for each line small and represent the trailing edge of the
segment divided by the maximum FSt detected, species. Genetic diversity, measured as expected
as recommended by Manni et al. (2004). For heterozygosity, is least in these southern small
clarity, bootstrap support values are shown only and fragmented populations and increases north-
for the thicker line segments for each of the ward (Fig. 1). Divergence across the Great
breaks. Interestingly, the first two barriers Central Valley of California is not unexpected.
detected were in the southern range of tanoak Many taxa display deeper divergence between the
between Santa Barbara and Lompoc and be- Sierra Nevada and Coastal Mountain Ranges
tween Lompoc and Nacimiento respectively. than within the mountain ranges, although this
Both of these breaks received strong bootstrap tends to be less clear for tree species (Calsbeek et
support. The third barrier separated interior al. 2003). However, it was surprising that a
populations from the Klamath Mountains and barrier was detected within the Sierra Nevada
the Sierra Nevada from Butte County from Range, between the Butte County and El Dorado
coastal populations. The thickness of these County populations. More extensive sampling is
segments tended to decrease northwards indicat- needed to confirm this. Finally, the divergence
ing a weaker divergence northwards, consistent between north coastal and south coastal popula-
with our earlier STRUCTURE analyses that tions shown most clearly with chloroplast DNA,
grouped thesepopulations with the northernmost and supported by nuclear DNA, shows a
coastal California group of populations and remarkable delineation over a very short dis-
those from Oregon (Nettel et al. 2009). Interest- tance. Indeed Dodd et al. (2010) reported only
ingly, the Barrier analysis detected divergence one population in which they found a mixture of
between Sierra Nevada populations from Butte the two chlorotypes. The most likely explanation
and El Dorado Counties. The fourth barrier is that climatic changes in the past resulted in
continued the separation of interior and coastal restriction of tanoak north and south of this
populations southwards, with the greatest diver- region that has only very recently become a zone
gence between the El Dorado County population of secondary contact. Taken together, the chlo-
and those from the San Francisco Bay Area. The roplast and nuclear data suggest that the range of
thickness of the segments decreased northwards tanoak can be divided into at least four groups
as this barrier joined the third barrier and also that should be considered in conservation efforts:
decreased southwards. The fifth break that was 1. A northern group from Areata to Oregon, 2. A
detected separated north coastal and south northern and central California coastal group
coastal populations east of Areata, California from Areata south to Monterey County, 3. One
(Humboldt County). Notably, this barrier passed or more interior groups (additional population
between two nearby populations from Korbel sampling is required to determine the limits of
and Hoopa (Humboldt County), consistent with these lineages) and 4. Two or more trailing edge
the partition between the northern and southern populations in the southernmost range of the
coastal chlorotypes shown in color in Figure 1. species.
These new analyses ofnuclearDNA support the
divergencebetweencoastalandinteriorpopulations Within Population Genetic Structure:
and between north and south coastal populations Seed Sources for Replanting
detected from chloroplast DNA. In addition, they
reveal probable isolation and fragmentation in the Conventional thinking dictates that local seed
southernmost range of the species that was not sources for replanting should provide gene pools
evident from the chloroplast data. that are well-adapted to the local environment
(Broadhurst et al. 2008; Kramer and Havens
What Can We Infer From Divergence? 2009). However, severe mortality from disease
may lead to drastic reductions in effective
Following Moritz (1999), we are looking for population sizes, with consequent risks of in-
divergent lineages that have arisen because of breeding (including mating among close rela-
past events; the most important for many plant tives). In outbreeding species that typically have
taxa were the glacial cycles that have occurred high genetic load, inbreeding depression is likely
over the last approximately 2 million years and to be expressed. The risks of inbreeding depres-
the more recent Holocene warming. The major sion are even greater in species that spread
divergences in population lineages in tanoak do through vegetative reproduction because of the
suggest correlations with these past events. The increased pollen supply from spatially clustered
southern range of tanoak has likely been most genetically identical ramets.
2013] DODD ET AL.: GENETIC STRUCTURE OF NOTHOLITHOCARPUSDENSIFLORUS 135
Tanoaks can produce large acorn crops and
also reproduce prolifically by basal resprouts.
Thelarge heavy seeds are dispersed by gravity, or
by birds and small mammals, so that seed
dispersal distances are likely to be short (Scofield
et al. 2010; Moran and Clark 2012). This,
combined with the potential local accumulation
of stems of resprout origin should result in fine-
scale genetic structure composed of demes of
related individuals. The spatial scale of clonality
and ofrelatedness become important parameters
in determining the minimum distances that
should be observed in selecting trees for seed
collection to minimize risks ofincreased inbreed-
ing in re-vegetated sites.
Clonal Spread
Detection of members of a clone (ramets)
requires molecular markers with sufficient power
that a pair ofindividuals with identical genotypes
can be confidently assigned to a clone and are
unlikely to be ofseed origin. Using a set ofeight
nuclearmicrosatellitemarkers, Dodd et al. (2013)
reported an average of 1.6 ramets per tanoak
A
genet in mature upland and lowland stands.
frequency histogram of numbers of ramets per
genet shows large numbers ofsingle-ramet genets
and a maximum of five ramets per genet on
lowland sites and eight ramets per genet on
upland sites (Fig. 2a). The maximum spatial
distance over which the ramets were distributed,
estimated as the zero intercept of a logistic
functionfittedto a plot offrequencydistributions
m
of different distance classes (Fig. 2b) was 6.5
on upland sites and 5.5 m on lowland sites. One Fig. 2. (a) Frequency distribution of numbers of
problem with rapidly mutating markers such as ramets per genet (solid bars: lowland sites and unfilled
nuclear microsatellites is ramets may not have bars: upland sites); (b andc) frequencydistributionsfor
identical genotypes because ofsomatic mutations clone mates within distance classes for lowland and
(Ally et al. 2008). Therefore, if resprouting has upland sites (three replicate sites shown as grey, solid
been occurring over generations, different muta- and open bars. Fitted logarithmic function y = —13.8
tions could arise among the ramets and the ln(x)+22.8 for lowland site andy = —39.9 ln(x)+74.5
estimates oframet number and distance based on for upland site. Adapted from Dodd et al. (2013).
identical genotypes could be underestimated.
Although Dodd et al. (2013) allowed for a single ofmost ofthe individuals reside, can be estimated
step mutation at their microsatellite loci for using classical population genetic theory for
assignment of non-identical ramets to a genet models ofcontinuous populations (Wright 1943).
lineage (multilocus lineages), this did not increase A two-dimensional neighborhood size (Nb =
significantly the estimates ofnumbers and spread 47icfd) is normally distributed with radius 2c>,
of ramets. Based on the single location for their where a (standard deviation of dispersal) defines
study, clonal spread appears not to haveexceeded thedisplacement ofprogenyfromparents, and d is
about 6-7 m. the population density. Neighborhood size (Nb)
can be inferred from spatialpatterns ofrelatedness
Spatial Genetic Structure estimated from molecular data that allows infer-
ence of a2 (Vekemans and Hardy 2004) and
Fine-scale genetic structure for neutral genetic effective plant density. Wright’s model can be
markers arises mainly as a result of the balance extended to include the composite effects of seed
between gene dispersal and stochastic processes dispersal, pollen dispersal and vegetative spread
of genetic drift. Assuming a population to be in (Gliddon et al. 1987). Dodd et al. (2013) estimated
drift-dispersal equilibrium, genetic neighbor- neighborhood sizes in tanoak as a function of
hoods that define a circle within which parents clonal spread and of sexual dispersion. Clonal
MADRONO
136 [Vol. 60
m m
spread averaged 1.7 on lowland sites and 2.1 al limit. There are two major disadvantages of
on upland sites, whereas sexual dispersion aver- this approach. First, if populations are small,
aged 69.3 m on lowland sites and 46 m on upland they may not be optimally adapted to their local
sites. These dispersal distances are evolutionary environment because ofgenetic drift. This will be
measures that make a number of assumptions, exacerbated by the selection ofa relatively small
most importantly that populations are in equilib- seed pool for replanting. Ifadopted, seed should
rium and that dispersal follows a normal distribu- be selected from trees that are least likely to be
tion. Neither of these assumptions is realistic; related. Ideally, research is needed to determine
populations need long periods of environmental dispersal distances of seed and pollen to deter-
stasis to reach equilibrium and dispersal distances mine minimum distances among seed trees for
aremorelikelytobeleptokurticwithmostprogeny each population in need of replanting. If this is
being close to parental individuals. Nevertheless, not possible, the minimum distances described
the estimates of dispersal distances from genetic above from the study at Jackson Demonstration
neighborhoods providethebestindications ofhow State Forest can be applied. However, we know
far apart trees for seed collection should be to very little about pollen dispersal in this insect-
minimize relatedness, in the absence of direct pollinated species and factors affecting pollina-
estimates from progeny arrays (multiple seeds tors will play an important role in the size of
from a single parent) to infer parentage. genetic neighborhoods. The second disadvantage
oflocal provenancing is that it does not account
Implications for Replanting for environmental change. This is likely to be
very important in the southern range of the
Present indications are that the SOD pathogen species, including areas ofhigh mortality levels in
will continue to cause epidemics resulting in Monterey County.
severe mortality of tanoak trees and that the
ecosystems in which tanoak is dominant are Predictive Provenancing
likely to be transformed, with tanoak being
represented only in the understory as resprouts This approach requires information on plant
(Cobb et al. 2012). So far, ecological monitoring performance under a range of experimental
is insufficient to determine the long-term survival conditions. This can be provided by reciprocal
of resprouts, or whether some genotypes are transplant studies, but these have not been
likely to be more persistent than others as clones. performed for tanoak. The great advantage of
Assuming continued mortality of adult trees, this techniqueis thatphenotypescan beidentified
stands throughwhichdiseasehas spreadwill have that should perform well at the replanting site.
smaller effective population sizes, as fewer Furthermore, climate modelingcan provide some
individuals will reach seed-producing age. To inferences on future climate at the site so that
augment the loss oftanoak, vegetation managers phenotypes can be selected that are expected to
need guidelines for the selection of seed for match future climates. Substantialwork isneeded
replanting. Breed et al. (2011) describe a range of before this approach can be implemented for
seed sourcing strategies aimed at minimizing the tanoak. However, reciprocal transplant plots are
effects ofinbreeding and outbreeding depression, a worthy investment to obtain valuable informa-
avoiding or mitigating introduction of mal- tion on translocation success.
adapted genotypes and providing a diverse
genetic base for adaptation to environmental Composite Provenancing
change. Weexaminethese optionsfortanoak and
recommend under what circumstances they may The objective ofthis approach is to attempt to
be appropriate and what further research is mimic landscape patterns of gene flow and to
needed to optimize success ofreplanting. incorporate a mix of planting stock that reflects
the probable pool of genotypes that would
Local Provenancing naturally contribute to the evolutionary potential
ofthe site (Broadhurst et al. 2008). This requires
Although the use of local seed sources is the knowledge ofpatterns ofgene flow and differen-
most commonly adopted strategy, it is best tiation at scales from 100s to 1000s of meters.
reserved for a system in which only local This scale is intermediate between the species-
populations exist (Breed et al. 2012). Under these scale and the fine-scale described above for
conditions risks of outbreeding depression and tanoak. Advantages of this approach are that
introducing mal-adapted genotypes are mini- seed provenances can be detected that are tied
mized. For tanoak this may be important in together by gene flow and quantities ofseed from
peripheral populations such as the southern the different sources can be selected so that the
range where populations are more isolated (see risk of genetic swamping and introduction of
discussion above) and in the interior coastal mal-adapted genotypes is low, and provenances
range where populations reach their distribution- can be selected that account for climate change.
2013] DODD ET AL.: GENETIC STRUCTURE OF NOTHOLITHOCARPUS DENSIFLORUS 137
Genetic data at this scale are not available for Literature Cited
tanoak, but they could be obtained at relatively Ally, D., K. Ritland, and S. P. Otto. 2008. Can
low cost. clone size serve as a proxy for clone age? An
exploration using microsatellite divergence in
Admixture Provenancing Populus tremuloides. Molecular Ecology 17:4897-
4911.
selAedctmsixsteuerdefrproomvedinfafnecrienntge(nBvriereodnmeetnatla.l20c1o2n)- Ashldteihysep,elraMsna.dl:sVc.ahp2oe0w.10C.DriNPtliAacnatlmpiRcaerrvoesinaettwaegslel,iitnpeosPlllhaianntvateiSocanil,etneacrneedsd
ditions within large populations. Unlike com- 29:148-161.
posite provenancing, no account is taken of Barnosky, A. D., E. A. Hadly, J. Bascompte, E. L.
dispersal distances, but sampling can be skewed Berlow, J. H. Brown, M. Fortelius, W. M.
to favor seeds from sites whose climate is likely Getz, J. Harte, A. Hastings, P. A. Marquet,
to match future climate at the replanting site. N. D. Martinez, A. Mooers, P. Roopnarine,
Although, this approach is likely to introduce G. Vermeij, J. W. Williams, R. Gillespie, J.
mclaaliamdathpattedthegseensohtoyuplesd,rapBirdeleydbeetpuarlg.ed(2f0r1o2m) eKlilt,zesE,.C.ReMvairlslhaa,lla,nNd. MA.atzBk.e,SmDi.thP.. M2i0n1d2-.
Approaching a state shift in earth’s biosphere.
the population, but this would require large Nature 486:52-58.
numbers of seedlings and would introduce the Brasier, C. M. and K. W. Buck. 2002. Rapid
risk of outbreeding depression in the next evolutionary changes in a globally invading path-
generation. Therefore, provenances taken at ogen, the causal agent of Dutch elm disease.
great distances from the replanting site should Biological Invasions 3:223-233.
first be tested in crossing studies and reciprocal Breed, M. F., K. M. Ottewell, M. G. Gardner,
transplants. For tanoak, we highly recommend AND A. J. Lowe. 2011. Clarifying climate change
that this type of approach should not mix adaptationresponsesforscatteredtreesinmodified
landscapes. Journal of Applied Ecology 48:637-
provenances from the five major groups (Santa
6—41.
Barbara, Lompoc, central and northern coastal M. G. Stead, K. M. Ottewell, M. G.
California, extreme northern California and Gar, dner, and A. J. Lowe. 2012. Which prove-
Oregon, interior populations from the Sierra nancewhere? Seed sourcingstrategies forrevegata-
Nevada) described under Genetic structure at tion in a changing environment. Conservation
the species level above. Genetics 14:1-10.
Broadhurst, L. M., A. Lowe, D. J. Coates, S. A.
Cunningham, M. McDonald, P. A. Vesk, and
Conclusions C. Yates. 2008. Seed supply for broadscale
restoration: maximizing evolutionary potential.
Ourknowledge ofgenetics oftanoak is still at Evolutionary Applications 1:587-597.
an early stage, but it does provide some Calsbeek, R„ J. N. Thompson, and J. E. Richard-
guidelines that will be important for future son. 2003. Patterns of molecular evolution and
management oftanoak stands ifSOD continues diversification in a biodiversity hotspot: The
to cause high levels of mortality. Flere we have California Floristic Province. Molecular Ecology
reviewed some of the published work that 12:1021-1029.
addresses population processes and the impli- CobbC,.R.A.C.,GiJl.lAi.gaNn.,Fialinpde, DR.. KM..MeReinzztoe.mey20e1r2,.
cations for choice ofseed sources for replanting Ecosystem transformation by emerging infectious
after mortality from SOD. In this issue, Wright disease: loss of large tanoak from California
and Dodd provide evidence that tanoak is forests. Journal ofEcology 100:712-722.
insect pollinated that contrasts with the polli- Dodd, R. S., Z. Afzal-Rafii, and W. Mayer. 2008.
nation syndrome in the true oaks. This is likely Molecular markers show how pollen and seed
to have implications for patterns of pollen dispersal affect population genetic structure in
dispersal that in turn will be influenced by coast live oak (Quercus agrifolia Nee). Pp. 485-
environmental changes associated with the 495 in A. Merenlender, D. McCreary, and K. L.
Purcell (eds.). Proceedings ofthe sixth symposium
wdiistehasec,limwaitteh chpaonpguel.atiWoen sftrilalgmheanvteatmiounchantdo oopnpooratkunwitoioedsl.andGse:netroadlay’TsecchhnailclaelngesR,eptoormtorrPoSwW’-s
learn about the ecological processes that affect GTR-217. USDAForestService, PacificSouthwest
genetic diversity in this species and the conse- R—esearch Station, Albany, CA.
quences for conservation and revegetation , ,and . 2010. Ancestralseedzones
under the combined challenges of disease, andgeneticmixture oftanoak. Pp. 171-182 in S. J.
development and climate change. Frankel, J. T. Kliejunas, and K. M. Palmieri (tech,
cords). Proceedingsofthesuddenoakdeathfourth
Acknowledgments science symposium. General Technical Report
PSW-GTR-229. USDA Forest Service, Pacific
This work was supported by the USDA Forest S—outhwest Research Station, Albany, CA.
Service, Pacific Southwest Research Station (06-JV- W. Mayer, A. Nettel,andZ. Afzal-Rafii.
,
11272138-047 and 06-JV-l1272138-069 to R.S.D.). 2013. Clonal growth and fine-scale genetic struc-
)
MADRONO
138 [Vol. 60
ture in tanoak (Notholithocarpus densiflorus: Faga- Moran, E. V. and J. S. Clark. 2012. Between-site
ceae). Journal ofHeredity 104:105-114. differencesinthescaleofdispersalanddeneflowin
Eckert, C. G., S. Kalisz, M. A. Geber, R. Sargent, red oak. PLoS ONE7:e36492. doi:10.1371/journal,
E. Elle, P.-O. Cheptou, C. Goodwillie, M. O. pone.0036492.
Johnston, J. K. Kelly, D. A. Moeller, E. Moritz, C. 1999. Conservation units and transloca-
Porcher, R. H. Ree, M. Vallejo-Marin, and tions: strategies for conserving evolutionary pro-
A. A. Winn. 2010. Plant mating systems in a cesses. Hereditas 130:217-228.
changing world. Trends in Ecology and Evolution Nettel, A., R. S. Dodd, and Z. Afzal-Rafii. 2009.
25:35^43. Genetic diversity, structure, and demographic
Excoffier, L. and H. E. L. Lischer. 2010. Arlequin change in tanoak, Lithocarpus densiflorus (Faga-
suite ver 3.5: A new series ofprograms to perform ceae), the most susceptible host to the sudden oak
population genetics analyses under Linux and death disease in California. American Journal of
Windows. Molecular Ecology Resources. 10: Botany 96:2224-2233.
564-567. Palsboll, P., M. Berube, and F. W. Allendorf.
Garnas,J. R., M. P. Ayres,A. M. Liebhold,andC. 2007. Identification of management units using
Evans. 2011. Subcontinental impacts of an inva- population genetic data. Trends in Ecology and
sive tree disease on forest structure and dynamics. Evolution 22:11 16.
J. Ecol. 99:532-541. Parker, I. M. and G. S. Gilbert. 2004. The
Gliddon, C., E. Belhassen, and P. H. Gouyon. evolutionary ecology of novel plant - pathogen
1987. Geneticneighborhoods inplantswithdiverse interactions. Annual Review of Ecology, Evolu-
systemsofmatinganddifferentpatternsofgrowth. tion, and Systematics 35:675-700.
Heredity 59:29-32. Pritchard, J. K., M. Stephens, and P. J. Don-
Grivet, D., V. L. Sork, R. D. Westfall, and F. W. nelly. 2000. Inference of population structure
Davis. 2008. Conservingtheevolutionarypotential using multilocus genotype data. Genetics 155:
945-959.
of California valley oak Quercus lobata Nee): a
multivariate genetic app(roach to conservation Rizzo, D. M., M. Garbelotto, and E. Hansen.
planning. Molecular Ecology 17:139-156. 2005. Phytophthora ramorum. integrative research
Hampbieo,divAer.sitayndundRe.r cJl.imPaetteitc.han2g0e0:5.theCornesaerrveidngge aCanldifomranniaagaenmdenOtregoofnafnoreestmse.rgAinnnguaplatRheovgieenw oinf
Phytopathology 43:309-335.
Haydmaetnt,ersK..MoJ.l,ecAu.laNretEctoello,gyR.8:4S.61-D4o6d7.d, and M. ScofIineflldu,enDc.eGo.f,aVc.orLn.wSooordkp,eacknedrPs.ocEi.alSmboehuasvei.ou2r01o0n.
hGreiasgrihsbltyealnscouetsctetopo.tiab2ln0e11ih.onstWtri.loldFuoaclreledstthefEotrcreoeseltsogfdayilslae?anVsdaerMiiaanbnl-ea tarcaonrsnpsortin oaf csooausttherlniveCaolaikfor(nQiuaerocausk asgarviafnolniaa.
Journal ofEcology 98:561-571.
Kirbayg,emMe.ntE.2,61S:.17P8.1-L1u7n91d., M. A. Anderson, and SorkI,keV.gaLm.i,,F.HW..WDaanvgi,s,aRn.dWeDs.tGfrailvle,t.A.20F1l0i.ntG,eMn.e
B. W. Bird. 2007. Insolation forcing ofHolocene movement and genetic association with regional
csltiumdaytefrcohmanLgaekeinElSsoiuntohree.rnJoCuarlinfaolrnoifa:Paalesoeldiimmneonlt- clloibmaattae gNreaed)ientisn inthCealiffaocreniaofvalclleiymoataek (cQhuaenrgceu.s
ogy 38:395^117. Molecular Ecology 19:3806-3823.
Kramer, A. T. and K. Havens. 2009. Plant Tappeiner, J. C., P. M. McDonald, and D. F. Roy.
conservation genetics in a changing world. Trends 1990.Lithocarpusdensiflorus(Hook. &Arn.)Rehd.
in Plant Science 14:599-607. Tanoak. Pp. 417^425 in R. M. Burns and B. H.
Manel, S., M. K. Schwartz, G. Luikart, and P. Honkala (tech, coords.), Silvics ofNorthAmerica,
Taberlet. 2003. Landscape genetics: combining vol. 2, Hardwoods. Agriculture Handbook 654,
landscape ecology and population genetics. Trends USDA, Forest Service, Washington, DC, USA.
in Ecology and Evolution 18:189-197. Vekemans, X. and O. J. Hardy. 2004. New insights
Manni, F., E. Guerard, and E. Heyer. 2004. from fine-scale spatial genetic structure analyses in
Geographic patterns of (genetic, morphologic, plant populations. Molecular Ecology 13:921-935.
linguistic) variation: how barriers can be detected Waples, R. S. and O. Gaggiotti. 2006. What is a
by using Monmoniers algorithm. Human Biology population? An empirical evaluation of some
76:173-190. geneticmethodsforidentifyingthe numberofgene
Manos, P. S., C. H. Cannon, and S.-H. Oh. 2008. pools and their degree of connectivity. Molecular
Phylogenetic relationships and taxonomic status of Ecology 15:1419-1439.
the paleoendemic Fagaceae of Western North Wright, S. 1943. Isolation by distance. Genetics
America: recognition of a new genus, Notholitho- 28:114-138.
carpus. Madrono 55:181-190. Wright, .1. W. and R. S. Dodd. 2013. Could tanoak
Monmonier, M. 1973. Maximum-difference barriers: mortality affect insect biodiversity? Evidence for
an alternative numerical regionalization method. insect pollination in tanoaks. Madrono (this
Geographical Analysis 3:245-61. volume).