COS 166-4 - Dispersal as a deterministic or stochastic process: The influence of nutrient additions on dispersal traits

Thursday, August 9, 2012: 2:30 PM
Portland Blrm 258, Oregon Convention Center
Lauren Sullivan and W. Stanley Harpole, Ecology, Evolution and Organismal Biology, Iowa State University, Ames, IA
Background/Question/Methods

The debate between the relative roles of deterministic and stochastic processes in many ecological fields such as community assembly has led to assumptions that local interactions are governed by niche-based deterministic processes, while regional-scale interactions are regulated by neutral or stochastic processes.  With these assumptions, dispersal is often considered stochastic as it is seen to be a random process by which propagules arrive at new locations and thus contribute to ecological drift.  While assumptions of neutrality and stochasticity have important merit for null model development, we argue for the consideration of alternative, deterministic hypotheses for dispersal.  More simply, we assert that dispersal is a major component of fitness and thus under selection pressure and therefore an important aspect of a species’ niche.  To investigate the extent to which dispersal is a neutral process, we quantified how increased soil resources altered potential dispersal of an entire community.  We measured four common dispersal traits on grassland species receiving factorial additions of nitrogen, phosphorous and potassium.  We then used these trait differences to parameterize simulation models where all species interactions were equal except dispersal.

Results/Conclusions

We found nutrient additions reliably increased dispersal traits of grassland species.  Height at seed release, an important predictor of dispersal distance, increased with nitrogen (N) and phosphorous (P) addition in forbs (p<0.01 and p<0.05 respectively), while higher order interactions (NP, NK and NPK) increased this height in grasses (p<0.05 for all).  Full fertilizer addition (NPK) also increased the number of flowering heads produced in forbs (p<0.05), which increased the probability with which a seed may disperse. The niche-based simulation models that included fitness-related dispersal traits showed qualitatively different outcomes for coexistence when compared to alternative neutral and stochastic models. Our results show that potential dispersal distance can increase deterministically with resource inputs.  Results also indicate that dispersal itself can act as a stabilizing niche mechanism.