Reply to comment

I think I am seeing a common pattern emerging in many of the articles that we are reading. This pattern relates not to the primary focus of most of the papers, but to the non-linear relations that produce “threshold-type” relations related to loss of biodiversity.

1) Species loss functional diversity loss

2) Diversity function relations

3) Diversity stability

4) Coextinction and specificity of affiliate and host.

In each case the relation appears to involve the issue of redundancy in some way. Perhaps too it can be considered in the context of specialization and generalization. Take the paper by Petchy and Gaston (2002) exploring the relation between function diversity and species loss. Here the more redundancy in the system the less linear the relation between loss of functional diversity and loss of species. The relationship is affected by how specifically one defines the functional traits. If it is based on “crude” functional groups, then there is high redundancy.

In the Schwartz et al paper (2000), we as a group raised the interesting point that if the relation between diversity and function is non-linear (their type–b), then what becomes most important for conservation scientists to determine is where on this curve diversity lies. If it along the asymptote, then things look favorable, but if diversity lies on (or near) the accelerating part of the curve, then the system could be close to experiencing catastrophic losses of certain function. The degree of threat will certainly also depend on the shape of the curve (is it shallow or is it strongly bowed)

See my comment from Week 7 for a figure of the relationship.

The paper on stability also showed non-linearity –related to redundancy in the system. Higher diversity promoted stability, but the relation was a saturating one.

In the (Koh, et al. paper the non-linearity is in proportion of affiliate species extinctions and seems to depend on the degree of redundancy in hosts for a given affiliate (specificity of the affiliate), where the more specialized affiliate shows lower redundancy and a more linear relation between P(Affiliate extinction) and P(host extinction) (Koh, et al. 2004).

Neal Williams