Virtually all of earth's biotic systems have been impacted by human activities (Vitousek et al. 1997). The unprecedented extent and magnitude of these impacts has led ecologists to develop new research approaches, conceptual frameworks and theory (Peters, Bestelmeyer & Knapp 2011) in response to the realization that many of the drivers of ecological dynamics today differ from those studied in the past (Smith, Knapp & Collins 2009). One particularly important consequence of global scale changes in biogeochemical and hydrological cycles is that key resources (water, CO2, N, etc.) are being altered chronically, in most cases directionally and for some resources such as N, they may even accumulate over time (Smith, Knapp & Collins 2009). There are abundant lines of evidence that entire ecosystems, or components of the biota they contain, have already responded to these environmental alterations, and the consensus expectation is that they will continue to do so in myriad ways (IPCC 2007). It is also clear that the nature and pace of responses to global change drivers can vary dramatically among ecosystems. Indeed, an important unresolved issue in ecology involves identifying the determinants of differences in ecosystem sensitivity to global change (National Research Council 2001; Smith, Knapp & Collins 2009). In other words, what are the underlying mechanisms and attributes of ecosystems that lead to some appearing relatively resistant to global changes, while others respond rapidly and to a much greater extent?
Recent theoretical and analytical analyses have identified a number of biogeochemical feedback mechanisms that can constrain ecosystem responses to altered resources (Hungate et al. 2003; Luo et al. 2004; Luo & Weng 2011; Wu et al. 2012). Broader frameworks have also been proposed that incorporate a hierarchy of mechanisms to account for both ecosystem resistance to changing resources, as well as nonlinear and potentially large responses in ecosystem function to global change drivers that alter resources [Hierarchical Response Framework (HRF), Smith, Knapp & Collins 2009]. One approach for increasing our understanding of how ecosystems may respond to chronic resource alterations is to conduct long-term manipulative experiments. Collins et al. (2012) report results from one such experiment in which water has been added to a tallgrass prairie ecosystem to alleviate growing season water stress for 19 consecutive years. They found, surprisingly, that community structure changed very little during this time. Any changes in species diversity and community structure that were detected varied from year to year and were inconsistent with the treatment effects over time. Thus, despite complete removal of growing season water limitation for almost two decades, the tall, perennial C4 grasses maintained dominance in this system. This resistance to chronic water addition is quite remarkable given that ecosystem structure and function across central US grasslands are widely accepted to be driven by precipitation amount (Sala et al. 1988; Burke et al. 1991). In addition, at the Konza Prairie where this long-term experiment is ongoing, above-ground productivity has been shown to be water limited 75% of the time (Knapp, Briggs & Koelliker 2001).
Despite the overall lack of change in plant community structure, Collins et al. (2012) did highlight the dynamics of one C4 tallgrass species (Panicum virgatum) in the lowland topographic position of the experiment. This grass species was present but was less abundant than the dominant C4 grass, Andropogon gerardii, in the community for the first 10 years of the experiment. However, after a decade of water additions, P. virgatum became dominant and remained so for the next 9 years. This response, they argued, is consistent with one HRF prediction – namely that with chronic resource alterations, extended lag periods might be expected before a period of community reordering occurs. Collins et al. (2012) state ‘Our results provide strong empirical support for the HRF. After years of water addition species reordering occurred after a 10-year lag period and P. virgatum now dominates the lowland community once dominated by A. gerardii’.
Collins et al. (2012) were not able to test another prediction of the HRF; that responses in ecosystem function to resource addition will be relatively modest initially (driven by ecophysiological processes), but much greater alterations in function will be evident when community reordering occurs. Such reordering is expected to occur as species that are better able to exploit changes in resources increase in abundance. To test this prediction, we compared responses of above-ground net primary production (ANPP, data available at http://www.konza.ksu.edu/knz/) to irrigation for two key time periods identified by Collins et al. (2012) – the initial 10-year period of dominance by A. gerardii and the subsequent 9 years of dominance by P. virgatum (Fig. 1). The response of this key ecosystem function was consistent with HRF predictions: water addition increased ANPP by 37% in the first decade of the experiment and by 64% after reordering of the dominant C4 grasses. This is a dramatic increase in ecosystem function given such a modest shift in community structure (i.e., only the identity of the most abundant grass changed with no overall change in richness or composition). Indeed, at the functional group level (the taxonomic resolution of the best earth systems models), there was no change in community structure because one C4 perennial tall grass replaced another as the dominant. The increase in ANPP observed in the second half of the experiment is particularly noteworthy when placed within the broader context of regional patterns of ANPP; levels of ANPP in the last 9 years of this study far exceeded those expected for the amount of water received (Fig. 1). Such dramatic and rapid shifts in ANPP have been documented previously, but only with significant community change such as occurs with shrub or forest encroachment into grasslands (Briggs et al. 2005; Knapp et al. 2008) or with substantial nutrient additions (Seastedt, Briggs & Gibson 1991). However, analyses of plant and soil samples from this experiment (J. Blair, personal communication, Kansas State University) have not provided any evidence that increased soil N availability might explain this response, suggesting that a simple change in identity of the dominant grass species has resulted in the large ANPP response.
While Collins et al. (2012) provide an intriguing example of plant community structure being surprisingly resistant to chronic alterations of a limiting resource, additional insight can be drawn when these results are combined with the data in Fig. 1. Ecologists have long recognized that changes in limiting resources can lead to major alterations in community composition and structure, often with significant consequences for ecosystem function. But Collins et al. (2012) and the analyses in Fig. 1 show that such community shifts are not required for chronic resource alterations to lead to dramatic changes in ecosystem function. In other words, overall stability in community richness, diversity or growth form dominance does not preclude a large functional response to chronic resource alterations – such as those expected to occur with global change (Smith, Knapp & Collins 2009). The results in Collins et al. (2012) also demonstrate that relatively simple experiments can be valuable for testing global change theory, particularly if they are long-term (Knapp et al. 2012). There have been recent calls for complex, multisite and multifactor global change experiments (e.g., Luo et al. 2011), and there is little doubt that these types of experiments would be valuable. However, as Collins et al. (2012) show, significant advances in our understanding of ecosystem responses to global change can occur with single factor experiments as well. But regardless of whether future experiments manipulate single or multiple factors, if they are deployed to span multiple ecosystem types and they are conducted long-term, ecologists will be better able to identify those mechanisms underlying differences in ecosystem sensitivity to global change (Smith, Knapp & Collins 2009).