Individual species and ecosystem function

In "The co-distribution of species in relation to the neutral theory of community ecology," published in Ecology (the most classical academic journal of ecology) Graham Bell summarizes a current discussion among ecologists:

Most species are unevenly distributed over the landscape at any given spatial scale, and communities differ in composition from site to site. These familiar facts can be given two rival interpretations, the two poles of community ecology theory. The first is that each species occupies only those sites to which it is well adapted, and from which it is able to exclude competitors. The community appears to be well engineered, so to speak, in the sense that the loss of any of its specialized components is likely to reduce the productivity or the stability of the system as a whole. […] The second interpretation is that most species are able to grow at most sites, so that community composition is determined largely by the accidents of dispersal, and local diversity is strongly influenced by the composition of the regional species pool. [...][According to this interpretation] the loss of species will have no effect on higher-level community properties.

Bell compares these two models with data sets from real world plant communities. He observes differences in species composition among habitats that he attributes to species-specific differences (which may be, for example, physiological differences). However, he finds that the characteristics of individual species have no effect on bulk properties of ecological communities such as the distribution of abundance, the overall species diversity, and the average strength of species co-distribution in space. Rather, these properties seem to depend on drift (changes in the relative abundance of competing species that are not due to differences in their characteristics) and limited dispersal.

A similar dichotomy exists in other sciences. The nature of interactions between molecules in a gas will depend on the chemical nature of the molecules involved, whereas the kinetic law of gases predicts bulk properties without requiring a knowledge of chemical composition. Somewhat closer to home, describing the mass balance of phosphorus in an estuarine ecosystem would normally require no more than the crudest categorization of the organisms involved, paying little if any heed to variation in the phosphorus economy among the innumerable species in the community. In some cases, this heuristic principle of equivalence will break down: The presence of zebra mussels rather than unionid clams, for example, may make a very large difference to the dynamics of phosphorus in a lake. In most places and at most times, however, the bulk properties of communities may be insensitive to the details of interactions among species, and are then adequately predicted, and correctly explained, in terms of neutral processes.

If Bell is right the loss of species will often have no effect on ecosystem services, although it will diminish our (or at least my) enjoyment of landscape diversity and the distinctiveness of places. Other sorts of data that support the same conclusion also come to my mind. Many animal species that help pollinate the flowers or disperse the seeds of plants provide largely equivalent services to plants. Up to some point if a given species disappears another one will keep playing the same role. The details of that "some point" are obviously very important.