Functional groups are a useful tool for interpreting and predicting shifts in plant community composition arising from linked changes in climate and stream hydrology. While methods for classifying species into functional groups are nearly as diverse as the studies that implement them, there have been some attempts at standardization. We are exploring two published methods—Hodgson et al’s (1999) multi-trait method and Westoby’s (1998) leaf-height-seed method—that are based on the classic model of plant life history strategies developed by J.P. Grime. Both methods utilize easily measured traits to approximate resource allocation and classify each species as ruderal, competitor, or stress tolerator. The methods are intended to be transferable to other ecosystems, but neither has been tested in riparian ecosystems in the American Southwest. Our goal was to determine if the methods are adequate for classifying dryland riparian plant species according to Grime’s strategies.

Plant trait, abundance, and environmental data were sampled for 158 plant species along gradients of moisture and disturbance levels in Southwestern riparian zones. Species were classified into functional groups according to Westoby’s and Hodgson’s methods. Correlation analysis was used to determine if the results reflect patterns in resource availability and disturbance levels in the plants’ environment.  

Results/Conclusions

Hodgson’s method was more successful than Westoby’s at classifying the species according to Grime’s life history strategies. Water is a limiting resource in dryland regions, and we expected soil moisture to strongly correlate with the competitor and stress tolerator scores. While Hodgson’s C and S scores significantly correlated with soil moisture levels, Westoby’s did not. Flood disturbance also structures riparian plant communities. Ruderal plants are adapted to high disturbance, and while Hodgson’s R scores correlated with two flood-related variables, Westoby’s did not significantly correlate with either.

Although Hodgson’s method performed well when correlated with environmental variables, neither method was ideal for use in our area, indicating that such methodologies will require regional tailoring. For example, several succulent species which we consider stress tolerators were classified as ruderals due to low leaf dry matter. Additionally, we have low confidence that flowering season has the same relationship to a plant’s “ruderality” in an arid, monsoon-dominated climate as it does in humid northern Europe where the model was developed. We will directly test the relationships between plant traits and environmental variables as a next step towards our goal of refining the classification scheme for use in riparian ecosystems of the arid Southwest.