Savanna dynamics in central Texas: just succession?

Question: What is the best way to model savanna dynamics? Specifically, under what conditions is a traditional succession model, i.e., a model of ordered, uni-directional change in the plant community, better than a state-and-transition model? Location: Central Texas savannas. Methods: We describe three examples of successional processes in central Texas savannas: (a) woody plant encroachment, (b) invasion by a non-native grass, and (c) establishment of different grass species in highly disturbed sites. Results and Conclusions: Savanna dynamics are now commonly conceptualized with state-and-transition models. However, in some situations a traditional succession model may be more appropriate or more useful. Succession models may better fit current ecological reality, as found in central Texas. Succession models emphasize transient dynamics rather than an (often unknown) endpoint, and direct us towards relevant literature from non-savanna ecosystems. Succession models may be particularly useful for land management and restoration, and where woody plant encroachment and/or invasions by non-native species control vegetation dynamics. [ABSTRACT FROM AUTHOR]