Distribution and abundance of tropical freshwater shrimp along a stream corridor: Response to disturbance

Different intensities and frequencies of disturbances can alter spatial and temporal patterns of abundances for coexisting species. We observed the effects of a high-flow event generated by Hurricane Hugo (19 September 1989) and below-normal stream flow during 1934 on freshwater shrimp populations in a tropical, montane scream. To determine if these different hydrologic regimes altered distributions of populations, we compared shrimp densities during three periods: pre-Hurricane Hugo (20 mo), post-Hurricane Hugo (50 mo), and low-flow (12 mo). There were significant differences in the relationships between locations of scream pools along an elevational gradient (300 to 470 m) and the abundances of two species of shrimp (Atya lanipes and Xiphocaris elongata) during these three periods. Atya increased in density with increasing elevation in a consistent fashion during all three periods. Densities also increased during the post-Hugo and low-flow periods relative to the pre-Hugo baseline. Xiphocaris increased in density with increasing elevation during the pre- and post-Hugo periods, but density decreased with elevation during the low-flow period. Palaemonid species of predatory shrimp, Macrobrachium carcinus and Macrobrachium crenulatum, consistently decreased in density with elevation during all three periods of observation. Multiple regression analyses demonstrated high predictability of Atya density distributions based on volume, pool depth, and coefficient of variation of pool depth during the pre-Hugo baseline observations. Xiphocaris densities were positively associated with pool width and negatively associated with the coefficient of variation of pool width in the pre-Hugo period. However, physical variables did not predict Atya or Xiphocaris densities during the post-Hugo period. During the dry period, there was a negative association between Atya densities and pool depth and width, and a positive association with the coefficient of variation of pool depth. During the dry period, Xiphocaris densities were best predicted as a function of maximum depth; there was a negative association with the coefficient of variation of pool width. Pre-Hugo Macrobrachium densities were negatively associated with the coefficient of variation of pool width, pool width-to-depth ratio, and elevation; in the post-Hugo and dry periods, Macrobrachium densities were best predicted by elevation. Mechanisms that likely cause these patterns of distribution include avoidance of predators coupled with active preference by prey species for pool habitats with low frequency of washout by storm flows, and with sufficient storage of food resources (microbially conditioned leaf detritus).