Temporal Niche

Download Report

Transcript Temporal Niche

Temporal Niche
• Time as a resource axis
• Consider activity times of vertebrates
• Owls vs. hawks
Temporal Niche
• Time as a resource axis
• Seasonal differences in flowering times
(lowering competition for pollinators)
Temporal Niche
Temporal Partitioning
• Animals can move
– But animals must constantly forage
• Plants must cope with constant
competition from neighbors
– Most plants can persist through resource
droughts
Temporal Niche
A traditional model of competition
(exploitive) does not really view time as an
independent axis
Temporal Niche
• There are two ways in which time may be
partitioned:
• 1) seasonal differences may allow
predators to coexist by exploiting different
prey assemblages
• 2) diel differences may allow predators to
partition a rapidly renewing resources
Temporal Niche
• Schoener (1974) reviewed the early
literature and determined animals often
segregate food and habitat dimensions but
rarely segregate along temporal niche axis
• Predators utilized diel partitioning more
than other trophic groups
• So, what has gone on…
Temporal Niche
• You could treat time as any other niche
axis (and reshuffle observed activity data)
• For example, Pianka looked at seasonal
activity (or daily activity)
• Results, time was not important (food
overlap study showed significant
differences)
Temporal Niche
• Although RA1-RA4 are all valid null
models for looking many overlap studies
(e.g. diet or microhabitat) they may not be
appropriate for temporal niche studies
• Most species show strong modalities with
peak activity at certain times (again daily
or seasonal) or multiple times
• RA1-RA4 are going to ‘destroy’ the curves
Temporal Niche
• A more appropriate null model may be to
retain the shape of the activity curve and
randomize the placement of its peak
Temporal Niche
One important assumption of temporal
activity patterns is that animals that feed at
different times must have different diets
• For example, nocturnal and diurnal
predators should differ in their diets and
therefore face less competition from each
other than with species with an identical
activity level
Temporal Niche
• One appropriate null model would be the
dietary overlap between species with
similar activity patterns (matched) is no
different than for a random ‘mixed’ pair
(one diurnal and one nocturnal species)
• Would your alternative hypothesis be
temporal aggregation or segregation?
• Why and what is the mechanism?
Temporal Niche
• Jaksic (1982) tested these hypotheses for
dietary overlaps of diurnal falconiform and
nocturanal strigiform raptors.
• Results (using standard nonparametric
statistics) indicated there were no
significant differences in dietary overlap of
mixed versus matched species pair
• Thus no exploitative competition
demonstrated (what about interference?)
Temporal Niche
• However, standard statistics may not be
suitable for testing the mixed-matched
dietary overlap hypothesis (think
independence of Sp A,B and C)
A & B, B & C, C& A
• Also, nearest-neighbor statistic is biased;
consider large sample vs. small
Temporal Niche
• Calculating a new overlap score for each
random pair would yield a null model
(Pimm 1983) (e.g. consider a group
noctural and diurnal predators)
• We would expect matched pairs (lower)
and mixed (higher) should differ from
random. Furthermore, nearest neighbor
should be a smaller distance than random
Temporal Niche
• Huey and Pianka (1983) examined dietary
differences between nocturnal and diurnal
predators in assemblages of lizards,
raptors and water snakes
• For both African and Australian lizards,
significantly more first- and secondnearest neighbors in dietary similarity were
synchronous than were asynchronous in
the activity times
Temporal Niche
• Similarly, more synchronous pairs of
species tended to overlap in diet more
often than predicted, and synchronous
pairs less often than predicted (opposite of
the predictions of the mixed-matched
dietary hypothesis, suggesting that lizard
food resources are not being partitioned
by species that are active at different times
of the day)
Temporal Niche
• Fig 5.2
Temporal Niche
• For raptors, were not different than their
null model
• For water snakes, dietary overlap of
synchronous species were higher than
expected
• Consequently, the degree of synchrony in
activity periods is not a reliable indicator of
dietary overlap
Phenological Overlap
• There are alternative hypotheses for the
evolution of nocturnal and diurnal
segragation …
– Predator avoidance
– Alleviation of interference competition
– Physiological constraints
Temporal Niche
• Seasonal overlap may be more plausible
than daily
• Hypothesized that within functional feeding
guilds of stream insects, competition for
food will result in temporal segregation
(apparent sequential peak production of
six species in an Appalachian stream)
Phenological Overlap
• Robertson (1895) first proposed that
pollen transfer by animals was a
potentially limiting resource that could lead
to staggered phenologies of flowering
plants
• However, other hypotheses also exist…
Phenological Overlap
• 1) pollinator preference in which one
species is a winner (and others lose)
• 2) interspecific pollen transfer (resulting in
lower reproductive success of some sort)
• 3) formation of maladaptive hybrids (again
lowering fitness)
• Mechanism (1) suggests competition for
competitors, (2) does not; all 3 suggest
increases in abundance of one plant
species and its’ pollinators can be negative
Phenological Overlap
• However, there is a specific case when
some plants can be benefited from the
presence of others plants… when?
Phenological Overlap
• Segregation of flowering times will reduce
overlap in shared pollinators. However, is
the converse true?
• High overlap in flowering times need not
imply strong competition for pollinators
• Thompson (1982) found overlaps in
flowering times of subalpine meadow
plants were unrelated to relative visitation
rates by pollinators
Phenological Overlap
• Despite the
attention
phenological
studies have
received,
there had
been little
statistical
investigation
Fig 5.3
Phenological Overlap
• Poole and Rathcke (1979) summarized
overlap as a single index (the sample
variance of the distance between adjacent
flowering peaks; similar to body size
spacing).
• The distance from first to last flowering
represents the range (0 to 1) if competition
has lead to regular spacing of the other
species
Phenological Overlap
• If competition has led to regular spacing of
peak flowering times, the observed
(population) variance in the position of
lowering peaks should be significantly
smaller than expected by chance
Phenological Overlap
• The ratio of observed to expected variance
(P/E(P) is an idex that correspnds to
flowering peaks that aggregated (if (P/E(P)
> 1), random (P/E(P) =1, or staggered
(P/E(P) <1) with the growing season.
• **This metric is the same for determining
the spatial structure of distributions
Phenological Overlap
• Stiles claimed lowering peaks
exhibited an orderly, staggered
sequence
Phenological Overlap
• Later found to be
slightly staggered
Phenological Overlap
• Remember, this metric does not measure
overlap in flowering times, but regularity in
the spacing of peak flowering times
• Cole (1981) proposed a different metric,
which represents the flowering period as a
line segment rather than representing the
flowering peak as a single point, then
calculating overlap of the line segment
Phenological Overlap
• The null hypothesis is simply that the
flowering period of each species occurs
randomly and independently within the
growing season
• Statistical analysis is problematic (nonindependence of overlaps)
• There have been many other iterations of
various metrics
Phenological Overlap
• Ashton et al. (1988) pointed out that since
the boundaries of flowering times are
bound by observed data and
randomizations all occur within those
boundaries, they are biased towards ‘lowoverlap’ patterns
• Note: most simulations have detected
random or aggregated patterns
Phenological Overlap
• Their solution was to also use peaks, but
expand the boundaries of the ‘season’ to
possibly include all dates
Nonequilibrium
Analyses
• The preceding analyses all make the
implicit assumption that flowering times or
activity periods of species have reached
an ecological or evolutionary equilibrium
• There are a number of different
mechanisms that may structure a
community
Nonequilibrium
Analyses
• If communities are in equilibrium, the
phenological pattern of temporal overlap
from one year to the next would be
consistent (as well as the opposite pattern)
• If resources fluctuate, then competitive
effects may be only seen during ‘resource
crunches’
Nonequilibrium
Analyses
• What would constitute a valid null model for
assessing equilibrium?
• One approach would be to assess if cooccurring species exhibit compensatory
fluctuations in abundance, activity, or phenology
• If a community is in equilibrium, then relative
changes in abundance should be mirrored
(inversely) by other species
• Conversely, positive fluctuations may occur if
species are tracking shared resources
Nonequilibrium
Analyses
• What would the data look like?
•
Jan March May July Sept
• Sp A
• Sp B
• Sp C
5
6
5
10
12
10
15
14
15
10
12
10
4
6
4
Nov
4
6
4
• Patterns of covariance in this matrix can
be quantified by comparing the sum of the
individual species variances to the
variance of their sums
Nonequilibrium
Analyses
• Specifically, the ratio of variance of their
sums to the variance of the individual
species (V)
• This ratio describes whether species are
fluctuating independently (V=1),
concordantly (V>1), or compensatorily
(V<1)
• One important variant, the product of the
variance ratio and the number of censuses
(VT) can be used as a significance test
Summary
• There are good tests to compare diets of
synchronous and asynchronous predators
(Pimm’s Monte Carlo simulation)
• Flowering peaks (Poole and Rathcke,
although maybe better (chap 6))
• Flowering intervals (Ashton et al)
• Compensatory fluctuations of abundance
or activity (Schulter’s variance ratio)