My current work looks at rapid changes in forests that can cause them to shift to non-forest states. To get into the literature on resilience I have been reading studies on rapid shifts both in forests and non-forest systems. A large number of papers on ecosystem resilience are studies of lake or coral systems where there are apparent clear shifts from one state to another. In the case of lakes rapid shifts from clear to murky systems can occur as a result of a gradual addition of nutrients. Similarly, coral systems have been shown to shift from reefs with high coral cover to algal dominated systems as a result of combinations of disturbance by cyclones or humans and climatic change. However, there seem to be relatively few studies investigating regime shifts in forest systems.
The examples that I have read and know most about are transitions between tropical forest and savannah and from boreal forest to tundra. In tropical forest-savannah transitions the theory goes that disturbance can lead to drying out of forest floors, leading to increased risk of fires and inflated tree mortality (see papers here, here and here for details on this). If fires are severe enough they can cause death of large trees, following which there tends to be an increase in seedling and subsequently sapling numbers. These small trees are often very densely packed together which can further encourage the spread of fire resulting in a positive feedback. This then results in the further loss of trees which eventually impedes seedling propagation. The establishment of these feedbacks is dependant upon the climate being dry and hot enough to cause fires to occur as well as to impede seedling recruitment.
Similarly the mechanisms for transitions from boreal forest to tundra may be the result of fire which encourages loss of trees and increased flammability of remaining forest areas (see Scheffer et al for details of this). At the Northern limits of boreal forest temperature is a limiting factor. Both low temperatures and soils that are frozen in winter reduce growth of any potential colonising trees. However, once established trees may increase heat input to soils creating a positive feedback that results in increased likelihood of recruitment of seedlings.
One thing that occurred to me from reading about these two cases of transitions was that they occur at the extreme opposite ends of a temperature gradient, either in polar or tropical regions. This raises interesting questions like: Are forests found at extreme temperatures likely to show relatively low resilience? And, does this mean that transitions are also more likely in mountain tree-lines? In some ways this is a fairly obvious point. Just as it is assumed that most species have lowest population densities at the edge of their geographic range, it is likely that forest cover is sparser close to the biological limits of tree species. Given that extreme climates are less favourable to forest development this also suggests that feedbacks between local and climate processes are more important in certain locations.
Now, I have no data to back any of this hypothesis up, just a hunch. However, I feel like the point I have made is a bit of an obvious obvious so am convinced that there must be something else out there on this topic. If you know of anything, let me know.
*Edit – I was right in fearing that I may have been re-inventing the wheel. Didham et al. (2005) came up with the same kind of ideas to those I have had here, only stating them more eloquently. They posit that systems where abiotic controls are particularly important may be more likely to show regime shifts. However, I’m not sure this has been empirically tested yet.