Marcot, B.G.; Croft, L.K.; Lehmkuhl, J.F.; Naney, R.H.; Niwa, C.G.; Owen,
W.R.; Sandquist, R.E.   19xx (in press).  Macroecology, paleoecology, and
ecological integrity of terrestrial species and communities of the interior
Columbia River basin and portions of the Klamath and Great Basins.  Gen.
Tech. Rep. PNW-GTR-XXX.   Portland, OR: U.S. Department of Agriculture,
Forest Service, Pacific Northwest Research Station.  XX p.
 

Ecological Functions of Species Among Communities

We present species function profiles that relate key ecological functions (KEFs) --
those major roles that each species plays in their environment -- of vertebrates to
occurrence by vegetation community.  KEFs of species can affect ecosystem
productivity, diversity, and sustainability; and function profiles of species help
summarize the diversity of species functions in ecological communities of the basin
assessment area.  Function profiles of species display the degree of functional
redundancy (number of species sharing the same KEF) within each community.
Some functions, particularly cavity excavation in snags and primary borrow
excavation in soil, are highly variable among communities in terms of number of
species.  Late seral forests provide the greatest redundancy in carrion feeding,
general nutrient cycling, and primary cavity excavation functions of vertebrates.  We
discuss functions and functional redundancy of other communities as well.  We pose
several hypotheses relating species functional redundancy to ecosystem resiliency
and changes in species community structure, as one facet of ecosystem
management.  The greatest vertebrate functional diversity is found in early seral
montane forest, followed by upland woodlands and riparian woodlands, and upland
shrublands.  No one community, however, contains all ecological functions of
species.

...

The SER database was queried to produce ecological profiles of species by
taxonomic group, habitat condition, ecological function, ecological status and
geographic occurrence, legal status, and other attributes.  Because few studies have
been conducted on most of the basin species, the SER database is incomplete and
consists largely of categorical information based on the judgments and experiences
of experts (see Marcot and others, in prep., and Marcot, in prep. for caveats on use
and interpretation of the SER database information).

...

Key ecological functions of species were further explored by cross-tabulating
number of vertebrate species by ecological functions and by terrestrial vegetation
communities.  This resulted in "species-function profiles" that were used to
determine which vegetation communities had the greatest variation in KEFS of
species, and which ecological function categories varied the most over vegetation
communities.

Species function profiles were developed for all vertebrates combined by tallying
the number of species with selected specific KEFs, by terrestrial vegetation
community.  Greater numbers of species with the same KEF in a specific community
were interpreted as greater redundancy of that function.  Similarly, Huston (1994:3)
referred to species with the same KEFs as "functional types of organisms," and
redundancy of species with a particular KEF as "functionally analogous species."
Franklin and others (1989:93) noted that "forest ecosystems which have redundancy
in structure and function are more likely to be able to absorb stresses, including
species losses, without catastrophic damage...."  Silver and others (1996:17)
similarly concluded that "functional diversity, and not just species richness, is
important in maintaining the integrity of nutrient and energy fluxes," and that high
species richness affords alternative and redundant pathways for the flow of
resources in an ecosystem.  Our species function profiles helped (1) identify
communities with the greatest (and the least) redundancy in ecological functions of
species, and (2) determine variation in redundancy of each KEF among
communities.

Total functional diversity of all vertebrate species was calculated by multiplying the
number of KEFs of vertebrate species in each terrestrial vegetation community by
the average number of vertebrate species performing each function (after Huston
1994:4ff).  Species function profiles and total species functional diversity were not
calculated for rare or potentially rare plants and allies, nor for invertebrates, because
(a) scientific knowledge and our databases of these taxa are incomplete, and (b)
most of these species likely respond to microhabitats, microclimates, and substrates
at far too fine a resolution than afforded by the available broad-scale data in this
study.

...

Species function profiles--Numbers of species with selected KEFs were tallied by
terrestrial vegetation community.  We term these depictions "species function
profiles" (fig. 30), which depict variations in functional redundancy among
terrestrial vegetation communities (table 2).  For this analysis, we selected a range
of KEFs that seemed particularly pertinent to vertebrates and for which variations in
numbers of vertebrate species among communities might prove instructive (table 8).

We explored patterns of two trophic relations functions.  The first was presence of
heterotrophs (table 8), which can include primary consumers or herbivores,
secondary consumers or primary predators, tertiary consumers or secondary
predators, omnivores, carrion feeders, cannibals, and coprovores (feed on fecal
material).  As expected, the number of vertebrate heterotroph species did not differ
significantly among terrestrial vegetation communities (fig. 31a) because
heterotrophy is a generalized function of all vertebrate species (380).

We then narrowed the focus to carrion feeders, which include at least seven species:
Great Basin spadefoot (Spea intermontana), which also feeds on insects, small
invertebrates, and sometimes each other (cannibalism); coyote (Canis latrans), an
opportunist scavenger for which carrion is most important during winter; wolverine
(Gulo gulo); turkey vulture (Cathartes aura), whose feeding behavior significantly
contributes to decomposition of carcasses; bald eagle (Haliaeetus leucocephalus),
which can also take live prey; common raven (Corvus corax); and gray jay
(Perisoreus canadensis), an omnivore.  Other vertebrate species doubtless feed on
carrion as well.  Presence of the seven species listed here was more variable among
terrestrial vegetation communities (fig. 31b) than was the general function of
heterotrophy.  The fewest number of carrion-feeders--that is, the least functional
redundancy of the vertebrate community for this particular ecological function--
occurred in early seral ponderosa pine forest and early seral subalpine forest
vegetation communities, and reached its maximum in late seral forests and
woodlands.

Herbivory is another subset of heterotrophy, and an ecological function that
displayed a profile different from that of carrion feeding.  Herbivory reached its
greatest functional redundancy in early seral montane forest and in upland shrub
vegetation communities (fig. 31b).  As an ecological force, herbivory can alter the
vegetation structure and succession of habitats.  For example, moose (Alces alces)
heavily browse early seral species such as aspen and willow which can result in
increases of conifers in habitats otherwise dominated by deciduous shrubs and trees
(also see Pastor and others 1993).  In semi-palustrine habitats, plant communities
can be altered by herbivory of several species of waterfowl, including greater white-
fronted goose (Anser albifrons), brant (Branta bernicla), snow goose (Chen
caerulescens), and Ross' goose (C. rossii).  Montane voles (Microtus montanus)
and meadow voles (M.
 pennsylvanicus) can modify grassland structures through herbivory.  Mule deer
(Odocoileus hemionus), white-tailed deer (O. virginianus), Rocky Mountain elk
(Cervus elaphus nelsonii), and mountain goat (Oreamnos americanus) can have
major influences on composition and succession of vegetation through herbivory.
Mountain goats can alter alpine and some subalpine plant communities.  Herbivores
can also interact with pollinators and affect plant reproduction and the evolution of
plant traits (Brody 1996).

Another KEF is that of nutrient cycling relations, particularly species that
substantially aid in physical transfer of substances for nutrient cycling.  Essentially,
all organisms cycle nutrients.  At least some 32 vertebrate species, however, affect
nutrient cycling substantially in their environments (see Marcot and others, in prep.
for discussion).  The vertebrate species function profile for generalized nutrient
cycling relations suggests a fair amount of variability of functional redundancy
among communities (fig. 31c).  Communities with the greatest redundancy in
nutrient cycling functions among vertebrate species include late seral stages of most
forest types, upland and riparian woodlands, and early seral montane forest.
Communities with the least redundancy include agricultural, alpine, exotic, upland
herb and shrub, and riparian herb and shrub communities, possibly because of
overall lower vertebrate species richness in at least some of these communities.

We developed species function profiles for six categories of interspecies
relationships:  potential insect population control, vertebrate population control,
pollination vectors, transportation of plant or animal disseminules, primary cavity
excavation, and primary burrow excavation (fig. 31).  Some 22 species of
insectivorous amphibians, birds, and mammals have the potential for controlling
some nonirruptive insect populations; some woodpeckers can dampen the amplitude
of irruptive insect populations (Koplin 1969).  Functional redundancy of potential
insect control was greatest in upland and riparian woodlands and least in upland and
riparian herbland, exotic, and agricultural communities.

Predatory activities of at least 10 species can control vertebrate populations.  The
species include five mammalian carnivore predators and two raptors.  Examples
include American kestrel (Falco sparverius) which may aid in population control of
some insects and rodents, and American badger (Taxidea taxus) which may help
control ground squirrel populations.  Redundancy of vertebrate population control
was less variable among communities but reached its peak in upland and riparian
shrub, upland woodland, early seral montane forest, and--interestingly--alpine
communities; and was least in most mid and late seral forest communities.

At least six birds--five hummingbirds and an oriole--serve as pollination vectors for
flowering plants (see Marcot and others 1996 for discussion).  This species
assemblage did not show much variation in redundancy among vegetation
communities (fig. 31d).  Another, much larger set of 153 species serve to transport
plant or animal disseminules; this set showed a fair amount of variation in
redundancy among communities, with the greatest redundancy (most species) in
early seral montane forest, upland shrub, riparian woodland, and upland woodland
communities, and the least redundancy in early seral ponderosa pine forest and early
seral subalpine forest (fig. 31a).

Primary cavity excavators--mostly woodpeckers, nuthatches, and chickadees--
number 17 species (see Marcot and others, in prep. for discussion) and showed by
far the greatest variation in functional redundancy among vegetation communities of
all ecological functions explored here (fig. 31d).  Primary cavity excavation reached
its greatest redundancy by far in all late seral forests and upland and riparian
woodland communities; and its lowest redundancy in agricultural, alpine, exotics,
upland and riparian herbland, upland and riparian shrubland, and early seral forest
communities.

Primary burrow excavators, a set of 39 species (see Marcot and others, in prep. for
discussion), also showed significant variation among communities.  The greatest
functional redundancy was reached in early seral montane forest, upland and
riparian shrub, and alpine communities, and the least functional redundancy was
reached in late seral forest communities, a rather inverse pattern to that of primary
cavity excavators (fig. 31e).  As with invertebrate soil excavators, among
vertebrates the primary burrow excavators can be important to soil turnover and
mixing of organic matter.  Gophers, for example, can process an enormous quantity
of soil material.  In one study in coastal San Diego County, California, Cox (1990)
found that the total soil mined by pocket gophers (Thomomys bottae) in 10 Mima-
type mounds amounted to 8.23 Mg-1?ha-1?yr-1 and subsurface deposition was 20.31
Mg-1?ha-1?yr-1, so that total soil mining equalled 28.54 Mg-1?ha-1?yr-1.

Finally, soil relations and wood relations showed relatively little variation in
functional redundancy among communities (fig. 31e).  Soil relations functions
showed the greatest redundancy in early seral montane forest, and upland and
riparian shrub communities, and the least redundancy in early seral ponderosa pine
forest and early seral subalpine forest communities.

What general trends can be summarized from the patterns of these selected
ecological functions?  First, no one vegetation community harbors the greatest
redundancy of all ecological functions.  Most vegetation communities contribute to
greatest redundancy of some specific ecological function.  Thus, collectively, all
vegetation communities combined help provide for redundancy of all ecological
functions.  Exceptions seem to be agricultural and exotic (and urban) communities,
which typically provide poorly for most functions explored here.

As a corollary, late seral forest communities seem to provide the greatest
redundancy in carrion feeding, general nutrient cycling, and primary cavity
excavation functions.  Late seral forest communities provide poorly for vertebrate
population control and primary burrow excavation functions which reach their peaks
of redundancy in other communities, principally early seral forest communities.

Alpine communities provide for high redundancy in vertebrate population control
and in primary burrow excavation functions.  Out of six mammal species that
regularly function as secondary burrow users (that is, use burrows created by
primary burrow excavators), however, only two species (ermine (Mustela erminea)
and long-tailed weasel (M. frenata)) occur in alpine communities, and these species
occur widely throughout other vegetation communities as well.  Thus, although
alpine communities provide for high functional redundancy of primary burrow
excavation, it does not provide for high redundancy of secondary burrow use.

Upland and riparian herblands do not provide for maximum redundancy of any of
the functions explored here.  Upland and riparian shrublands provide for high levels
of redundancy of herbivory, vertebrate population control, pollination, primary
burrow excavation, and soil-relations functions.  Upland and riparian woodlands
provide for high levels of redundancy of carrion feeding, general nutrient cycling,
potential insect population control, vertebrate population control, pollination, and
primary cavity-excavation functions.  Thus, shrublands and woodlands provide for
complementary sets of ecological functions, just as do late and early seral forest
communities.

Early seral montane forest seems to provide for an anomalously wide array of
conditions and ecological functions of all forest types in the basin assessment area.
It provides for high redundancy of general nutrient cycling, vertebrate population
control, pollination, primary burrow excavation, and soil-relations functions,
whereas early seral ponderosa pine forest and early seral subalpine forest
communities do not.  This may have to do with the overall greater habitat and biotic
diversity found in early seral montane forests than in early seral stages of other
forest types, but this needs verification.

Which KEFs considered here are the most variable among terrestrial vegetation
communities in terms of number of vertebrate species (redundancy of species)?  The
answer is found by plotting the standard error (SE) of number of species among
vegetation communities for each KEF.  Results (fig. 32) show that, of the functions
considered here, primary cavity excavation is by far the most variable in functional
redundancy of species among all terrestrial vegetation communities.  This means
that only specific vegetation communities are likely to support most of the species
with this function.  The next two functions with high variation in redundancy are the
carrion feeder function and the primary burrow excavator function.  Thus, the
vegetation communities in which these most-variable functions occur are likely to be
important for maintaining these species' ecological functions within the basin
assessment area.  On the other hand, KEFs with the lowest SEs include transport of
plant or animal disseminules, heterotrophic consumer, and pollination vector
functions.  These are the least variable in redundancy and the most reliable among
communities; none of the vegetation communities is particularly critical for
maintenance of these functions within the basin assessment area (although specific
composition of species with these functions may differ among communities).

Which terrestrial vegetation communities are the most variable in terms of
redundancy among ecological functions?  The answer is found by plotting the SE of
species richness (number of species) among KEFs for each vegetation community.
Results (fig. 33) suggest that early seral montane forest has the greatest variation,
and three communities (mid seral montane forest, mid seral ponderosa pine forest,
and mid seral subalpine forest) have the least variation in redundancy among the
ecological functions discussed above.  This does not mean that low-variation
vegetation communities contain all KEFs, but only that the functions they do provide
for tend to vary the least in number of species per function; there is a greater
"evenness" (relative proportion) but not necessarily "richness" (number of different
kinds) of species' ecological functions.

Overall, it is unclear how the degree of functional redundancy of KEFs, either
among species or among vegetation communities, specifically affects long-term
resiliency of ecosystems to perturbations of the environment and to perturbations of
species community structures.  A working hypothesis might suggest that a greater
redundancy promotes greater resiliency in ecosystem ecological processes to
stochastic environmental events and to short-term or localized losses of some
species.  This, however, is largely unstudied in the basin assessment area.  Our
approach provides an analytic framework and a repeatable means of posing such
testable hypotheses on ecosystem processes and functional redundancy among
communities.

Total species functional diversity--Huston (1994) characterized total species
functional diversity as the number of different ecological functions performed by all
species in a community, times the mean number of species per function.  Figure 34
presents estimates of total species functional diversity of vertebrates for terrestrial
vegetation communities of the basin assessment area by using the ecological
functions discussed above.  Over all communities, the greatest vertebrate functional
diversity is found in early seral montane forest, followed by upland woodlands and
riparian woodlands, and upland shrublands.  Historically, most vegetation cover
types associated with early seral montane forest and upland woodlands have
increased in total area in the inland West since early historic times, whereas those of
riparian woodlands and upland shrublands have decreased.  Vertebrate species
within these cover types have shifted in relative abundance even if vertebrate
functional diversity has remained more constant.

The lowest vertebrate functional diversity is found in early seral ponderosa pine
forest, early seral subalpine forest, and agricultural lands (fig. 34).  The vegetation
cover types associated with early seral ponderosa pine forest communities have
decreased in area since early historic times; those associated with early seral
subalpine forest communities have both increased and decreased; and agricultural
lands have greatly increased.  Interestingly, exotic vegetation communities do not
rank particularly low in overall vertebrate functional diversity (fig. 34), although
there are many vertebrate species that do not find suitable conditions in such
habitats (Marcot and others, in prep.).  Thus, levels of vertebrate functional diversity
among vegetation comunities are complementary to--not coincident with--patterns of
vertebrate species diversity (see Marcot and others, in prep. for species diversity
descriptions).

To what extent can we expect functional redundancy among vertebrate species with
the same KEF?  This question was addressed by Marcot and others (in prep.) who
conclude that general functions, such as herbivory, cavity excavation, and
pollination, might be considered redundant if performed by different species, but
because each species has its unique combination of habitat associations and life
history patterns, no two species can be expected to be exactly interchangeable.
Thus, patterns of species functional diversity presented here must be interpreted
only as broad-scale geographic and macroecological trends and not interpreted as
evidence for allowing any particular species loss or replacement in communities.

...

The use of species function profiles and related analyses of species' KEFs can help
to (1) determine the degree of functional diversity within communities;  (2) identify
communities with high (or low) redundancy in specific ecological functions, and
therefore communities with high (or low) resiliency or buffering capacity in the face
of disturbance regimes and systematic environmental changes; and (3) identify
ecological functions with high (or low) redundancy among species, in particular
communities.  Walker (1995) suggests focusing conservation attention on
functionally important species groups with little redundancy; that is, with few
representative species.  We may extend his suggestion to include focusing
conservation attention on communities with low redundancy in (or high variation
among) ecological functions.  Risser (1995) notes that biodiversity can affect
ecosystem functions, and suggests focusing initial management attention on the key
structuring processes at intermediate scales of space and time.

Determining presence and redundancy of ecological functions also could be of help
in posing specific goals for ecosystem restoration, particularly for augmenting
natural ecosystems with restored ecosystems (Sinclair and others 1995) and by
matching functional diversity in ecosystems undergoing restoration to that in
undisturbed ecosystems.  Examples of this include the use of species and ecosystem
functions to monitor restoration progress in wetlands (Simenstad and Thom 1996)
and to monitor trophic levels and feeding functions (shredders, grazers, and
predators) of aquatic invertebrates in stream restoration projects (Murphy and
Meehan 1991, Reeves and others 1991).  Among terrestrial communities of the
basin assessment area, studying patterns of ecological functions (functional
redundancy and functional diversity) of forest seral stages targeted for restoration,
such as old single-stratum ponderosa pine forests, could prove useful as monitoring
benchmarks or to help define management targets.