rarest species is deleted first from the system, followed by the next species in the rank, in what we call an ascending fashion (ascending along a rank-abundance curve) (Figure 6.3). Our contention is that the biodiversity-ecosystem function model has a different shape if species are deleted in a descending fashion along rank abundance diagrams (Figure 6.4A). Deleting the most abundant species first, i.e. the one that channels the largest fraction of primary production, could result in an abrupt change in ecosystem function (Figure 6.4B). The biodiversity-ecosystem function mode! has the opposite pattern when species are deleted in a descending fashion, with large changes in ecosystem function as a result of few changes trj species richness, followed by a plateau at lower richness levels where further deletions do not result in further alteration of ecosystem processes.
The large impact on ecosystem function of deleting the dominant species is the result of deleting the species which is best adapted to modal environmental conditions, and is not the result of deleting a large fraction of biomass. A prediction of the mode! is that deleting the amount of biomass of the dominant species but from all species in proportion to their abundance will have a small ecosystem effect in comparison with removing the same amount of biomass but from only one species, the dominant one. For example, the model predicts that the removal of the dominant species in a hypothetical ecosystem which accounts for 40% of the biomass will have a larger effect on ecosystem function than removing 40% of the biomass from each individual species. In both cases the amount of biomass remo%'ed is the same, but in one case the removal is spread over all the community and in
Figure 6.4 (A) Rank-abundance diagram showing two alternative patterns for species deletions: an ascending pattern where the rarest species is deleted first followed by the next species in the rank, and a descending pattern where the most abundant species is deleted first followed by the next species in the rank. (B) The effect on ecosystem processes of deleting species in an ascending or descending fashion
Figure 6.4 (A) Rank-abundance diagram showing two alternative patterns for species deletions: an ascending pattern where the rarest species is deleted first followed by the next species in the rank, and a descending pattern where the most abundant species is deleted first followed by the next species in the rank. (B) The effect on ecosystem processes of deleting species in an ascending or descending fashion the other it is concentrated on the dominant species. We suggest that the latter has a larger ecosystem effect than the former.
There is an infinite number of models of biodiversity-ecosystem function, defined by the order in which species are deleted. The ascending and descending cases are the boundary cases. From this information the relationship between biodiversity and ecosystem function can be specified for any particular case simply by knowing the rank order of the species to be deleted.
Time is an important consideration in our conceptual model. The size of the response of ecosystem function to the deletion of one or more species will depend upon the time at which the response is measured. As the time between the deletion and the measurement increases, the size of the response should decrease. The explanation for this decrease lies in the compensatory response of the remaining species. The rapidity and magnitude of the compensatory response will be process- and ecosystem-specific. For example, deletion of the dominant plant species in the short-grass steppe of North America will have a large effect on net primary production during the year of the deletion and perhaps for several subsequent years. In less than 10 years rhe remaining plant species will probably completely compénsale, and net primary production will be back to pre-disturbance levels. In this case compensation is complete. Other processes or ecosystems may respond differently to the deletion of the dominant species. We can speculate that deletion of the dominant microbial species that accounts for nitrogen mineralization may produce a very different response depending upon the presence of other species that can perform the same function. If alternative species are not present, nitrogen mineralization will be decreased and over time the compensatory response will be small or absent.
Time is also related to environmental variability: the longer the time-scale of observation the greater the range of environmental conditions experienced by an ecosystem. The effect of removing species on ecosystem function depends on the prevailing environmental conditions. For example, removing drought-resistant species during a wet year will have small effects on ecosystem processes. However, removing them in a dry year may have major ecosystem effects. Therefore, the greater the time-period over which ecosystem responses are observed, the higher the probability of observing an effect of changes in biodiversity. This greater probability will be attenuated by the compensatory potential, which will also increase with time.
So far, our discussion has assumed that all species have similar roles and their impact on ecosystem function is solely related to their abundance. However, ecologists have long recognized the existence of similarities among species and the convenience of defining functional groups (Humboldt, von 1806). Species within functional groups share morphological, physiological and/or phenological characteristics which result in a common ecological role (Sala et al. 1989). Therefore, the deletion of an entire functional group could have a larger impact on ecosystem function than deleting the same number of species but drawing from a variety of functional groups. A species may belong to more than one functional group, and consequently the impact of deleting one species may be related to the number of species already existing in the functional group(s) and on the number of functional groups to which the species belongs. Again, the effect on ecosystem function is not simply related to the number of species, but to which species are added or deleted.
Functional groups within a community account for different fractions of total ecosystem processes. For example, perennial shrubs account for a large percentage of total above-ground net primary production in the Chihuahuan desert of North America (MacMahon 1988). We could rank functional groups according to their abundance and their contribution to individual ecosystem processes and construct a rank-abundance diagram for each. Functional groups can be deleted from the least to the most important in an ascending fashion along the rank abundance curve, or alternatively from the most important to the rarest. Deleting entire functional groups should result in abrupt changes in ecosystem function (Figure 6.5). The decrease in ecosystem function should be largest when deleting first the most abundant functional group.
So far we have considered the effects of changes in species richness which occur as a result of deleting species. This exercise assumed an initial condition of a system in the richest stage, and evaluated the effect of deleting species in different fashions. This follows the most common experimental approach to this question (Ewel el at. 1991; Tilman and Downing 1994). Equally important is the effect of species additions on ecosystem function. In most cases, the models developed for the species deletion case should be applicable for the species addition problem. There are three possible outcomes of species additions: increase, decrease, or no change in ecosystem processes. Increases in ecosystem processes should occur in those systems which have previously lost some species. The effect on eccosystem functioning of species additions
Was this article helpful?