Niche picking: the importance of being temporal


Environmental fluctuation, temporal dynamics and ecological processes symposium, Ecological Society of America (ESA) Annual Meeting, Milwaukee, WI, USA August, 2008

It is increasingly evident that adaptation to the frequencies, not simply to the extremes, of environmental fluctuation can regulate the abundances of populations and the structure and thus function of ecological communities. A recent symposium at the 2008 Ecological Society of American annual meeting in Milwaukee, entitled Environmental fluctuation, temporal dynamics and ecological processes, presented a range of studies demonstrating practical ways of approaching such temporal dynamics. The goal of this symposium was to confront the general theory of temporal dynamics with data and with more specific theoretical applications, highlighting a means of investigating this fundamental ecological process.

‘The appearance of temporal dynamics in a diversity of communities suggests a ubiquitous presence, at least in the plant world’

Traditionally, a central aspect of species diversity is the differentiation of niches: species may co-exist over the long term when they possess sufficiently different requirements. Commonly invoked mechanisms of niche differentiation pinpoint species differences in a spatial or seasonal context, so that species co-exist because they complete essential parts of their life cycles in different parts of the habitat or at different times. Spatio-temporal dynamics allow the co-existence of species segregated into patches created by disturbance or edaphic variation.

By contrast, in temporal dynamics the population of established (reproductive) individuals is long-term stable over time; the competitive dynamic is played out by seedlings and juveniles, producing recruitment alternating between species over years, decades or even to the extent of a century or more in long-lived taxa. Unless explicitly taken into account, temporal niche dynamics can easily look like an absence of species interaction.

Temporal dynamics supported by noncatastrophic environmental fluctuation have been invoked as a regulatory agent in desert annual communities (Pake & Venable, 1996), Great Plains grasslands (Adler et al., 2006), Mediterranean-type scrubland (Facelli et al., 2006) and tropical forests (Kelly & Bowler, 2002), and involve a large proportion of the plant species within these communities (Kelly et al., 2008). The appearance of temporal dynamics in such a diversity of communities suggests a ubiquitous presence, at least in the plant world. Whether or not temporal dynamics also apply directly to nonplant taxa, the role of plants as mediators between above-ground and below-ground ecological processes provides the potential for carry-on effects between, as well as within, trophic levels.

The theory of temporal niche dynamics was set out long ago, in a lottery model for fish species’ co-existence (Chesson & Warner, 1981). The model specified environmental fluctuations affecting recruitment differently for different species as the starting point for dynamic long-term co-existence in which the populations continually fluctuate about a long-term mean. Given this background, the first necessity is ‘storage’ of reproductive capacity to allow recovery from unsuccessful seasons: long-term reproductive viability or persistent propagules. The second necessity is that species recruit at different times (or at different rates which also fluctuate) as a result of environmental differences; for example, species 1 recruits strongly when times are good for species 1, and species 2 recruits strongly when times are good for species 2.

To promote or permit co-existence, further criteria need to be satisfied. During good times a species at high density (considerably higher than the long-term mean) must not benefit strongly because the tendency must be to fall back towards the long-term mean. The per capita recruitment must therefore be suppressed at high density and this is accomplished through competition; mostly intra-specific. At low density (much lower than the long-term average) a species must be able to recover strongly and the population density grow towards the mean. The suppressing effect of competition needs to be minimal in good times at low density. Competition should thus be strongest under good conditions and weakest under poor conditions – both stabilizing influences. This is called covariance between competition and the environment [cov(EC)], which, to be effective, increases with increasing population density.

As a determinant of the co-existence of similar species, temporal dynamics are likely to be a major factor in community diversity and function. Nonetheless, temporal dynamics have yet to be well integrated into the larger ecological canon. Among recent ecology texts, only one treats temporal niche dynamics in any detail (Gurevitch et al., 2006); out of a number of recent reviews on co-existence mechanisms, only one directly addresses temporal niche processes (Chesson, 2000). Anecdotal evidence from the general ecological community suggests that this is because temporal dynamics are seen to be difficult to understand and require long-term investment – longer term than is available within the structure of most careers in ecology. As shown at this symposium, this need not be the case.

Demonstrating the effects of temporal dynamics

It is self-evident that in many cases, direct observation of temporal dynamics is a long-term endeavour. However, community ecology has a long and ongoing history of tackling large-scale and long-term processes through demonstrating instead the outcomes or effects of those processes, and several presentations focused on determining such effects for temporal dynamics. After a brief review of general theory behind the temporal storage effect, Peter Chesson (University of Arizona, USA) described a manipulative experiment geared to show the effects on herbaceous perennials of the storage effect over a relatively short term. Chesson and associates had reasoned that recruitment in a monoculture where competition is necessarily wholly intraspecific, and thereby stonger, should be more closely regulated and less variable in response to the same environmental conditions than recruitment in a polyculture (Chesson, 2008). In an Australian herbaceous community, Chesson and associates created monocultures of a native perennial Lagenifera species to compare Lagenifera recruitment over time under this treatment with simultaneous recruitment in natural polycultures of Lagenifera and Poa, finding support for their expectation in two experiments over separate time sequences of 14 months and 26 months.

In the same vein, Colleen Kelly (University of Oxford, UK) and Michael Bowler (University of Oxford, UK) used static data to show that the distribution of relative abundance within closely related species in a Mexican dry forest is best explained by pairwise focused competition within temporal dynamics. With this result, they were thereby able to tie together recent work on fractional abundance and an independent signal for storage dynamics within the same community (Kelly & Bowler, 2005; Kelly et al., 2008). That distribution is not explained by community-wide neutrality, or by two-species interchangeability or competitive exclusion. Yet, the species abundance distributions for the whole community and for species that are members of congeneric pairs are both lognormal, an unplanned result with implications for the current on debate on niche vs neutral processes regarding community organization (Leibold, 2008).

Investigating the necessary conditions for temporal storage dynamics

Another approach to investigate temporal dynamics is to establish whether the necessary conditions for their action are present, and whether those conditions are strong enough to have an effect on community dynamics. Peter Adler (Utah State University, USA) and colleagues used a heritage data set – post dustbowl grassland censuses across the US heartland – to search for conditions necessary for co-existence through the action of temporal dynamics. For Kansas prairie grasses they found the presence of the key conditions (Adler et al., 2006), but little evidence for environment–competition covariance, in an Idaho sagebrush community. They investigated whether the storage effect in the Kansas grasslands is sufficiently strong to permit co-existence, but the results were not conclusive.

Temporal dynamics may also be induced by biotic factors, which in turn are likely to be regulated by environmental conditions. Earlier work by Michael Hanley (University of Plymouth, UK) showed slug herbivory by seedlings on the UK's Salisbury Plain to have significant long-lasting effects on the cover of mature herbaceous plants (Hanley et al., 1996 and subsequent data), with great variation from year to year. More recently, Hanley and Rebecca Sykes (University of Plymouth, UK) found experimental support for the role of differential sensitivity (DS) temporal dynamics by demonstrating that slug grazing on seedlings is sufficient to reverse the relative cover, at maturity, of congeneric pairs, with the more palatable, faster-growing species achieving more cover when grazing is absent or low and the less-palatable species gaining the upper hand when grazing pressure is higher.

Direct observation

Norma Fowler (University of Texas, USA) and Craig Pease (Vermont Law School, USA) did not set out to document temporal dynamics, but nonetheless did so through an ingenious approach that was as interesting for its novelty as its results. They used a simple model of population dynamics in studying 16 yr of data on eight species in a Texas grass-dominated herbaceous community. Year by year they solved the model equations with a separate carrying capacity, K, for each species and each year. They found that the population density lagged behind the carrying capacity K, growing after an increase in K and being too high to sustain after a drop. The lags and overshoots were not synchronous, and these results indicate an important role for temporal niche separation in the community structure of these arid grasslands.

Focused questions

Akiko Satake (Hokkaido University, Japan) and colleagues used theory to investigate the conditions under which pollination needs might drive the temporally varying patterns of mast flowering and fruiting. They showed that the probability of plant species which share pollinators enhancing synchronized flowering and reproduction increased when the cost of producing a single flower increases simultaneously with a decrease of pollination rate with lower flower intensity (Satake & Iwasa, 2000, 2002; Iwasa & Satake, 2004). The model offers a promising approach to the ubiquitous phenomenon of intermasting intervals. It also offers another instance where interactions between tropic levels – and biodiversity – may mediate or be mediated by temporal fluctuations in the environment.


The essential message of this symposium, summed up by Gordon Fox (University of South Florida, USA), was that variability over time, as much as variability in space, allows species to co-exist, and the effect of temporal fluctuation appears to be ubiquitous. The studies presented and discussed above exemplify a growing body of work demonstrating this observation.

Although evidence is accumulating, temporal niche dynamics are still little studied, and this should not and need not be. Revelations of temporal dynamics are not restricted only to those already initiated into its mysteries. Some of the strong evidence here for temporal niche dynamics was ‘found’, rather than the result of efforts directed at testing the theory. While the signals of temporal dynamics may entail studies carried out over a substantial number of years (N. Fowler and C. Pease; detailed in the section entitled ‘Direct observation’) or long-term data sets (Adler et al., 2006), they may also be found in static data (Kelly & Bowler, 2002) or in studies of only a few years (Chesson, 2008; M. Hanley and R. Sykes, detailed in the section entitled ‘Investigating the necessary conditions for temporal storage dynamics) or in the targeting of specific theoretical questions (Satake & Iwasa, 2000, 2002).


I gratefully acknowledge the assistance of the speakers in preparing this report (all errors are, of course, my responsibility). I thank Chamela Biological Station, Jalisco, México, for its hospitality.