Modeling seed dispersal by wind in herbaceous species

Wind can be regarded as the most important vector for seed dispersal in ope n grassland vegetation. Experimental estimates of seed dispersal distances in this environment are complex because of low arrival probabilities at lar ge distances. Therefore, a proper mathematical generalization would be esse ntial to give insight in dispersal probability distributions. Hence a promi sing individual-based model for seed dispersal presented by Andersen was te sted for different wind velocities and seeds. Simulation results from the s eed dispersal model were compared with observations in a horizontal wind tu nnel. Considering the large variation in seed morphology and mass, the simu lation results fitted wind tunnel results reasonably well, indicating the g eneral applicability of the tested model for herbaceous species. Model sens itivity was evaluated with respect to wind speed and vegetation height. Dif ferences in wind speed had a larger impact on the tail of the seed shadow t han on median dispersal. However, vegetation height had little impact on th e tail of the seed shadow compared to the median. Terminal velocity (V-t) is the crucial species specific parameter in wind d ispersal models. There are two frequently used methods to determine V-t: a dropping method and a method to float seeds in an upward air stream. Howeve r, these methods have never been compared directly. This paper presents V-t values determined with both methods. In general results were in the same o rder. Only for high values of V-t the results of the floating method were f ound to be lower than the results of the dropping method. Simulation result s showed that the intraspecific differences in V-t values were an important factor in determining the seed shadow.