Disturbance and refugia in the ecology of stream benthic communities
Natural habitats are spatially and temporally heterogeneous and exert powerful influences over the distribution, interactions and adaptations of organisms. Southwood's (1977, 1988) Habitat Templet was tested in twenty-two streams, using refugium availability and disturbance frequency as the two axes. Disturbance frequency was estimated as the number of days a year 50% of the stream bed was in movement, and refugium availability was estimated from shear stress frequency distributions using FST hemispheres A long-term sampling regime allowed persistence to be estimated in all streams. A comparison of sites revealed that all sites were approximately equally persistent, and no pattern was found with disturbance frequency or refugium availability. This suggested that some degree of adaptation to the habitat had occurred, and differences in species traits were consequently investigated. Extensive literature reviews were assessed to obtain traits for as many of the species as possible in the Ashdown Forest. Species traits were compared to environmental variables using multivariate statistics. Only weak patterns were shown, although one of the traits that was significantly different between sites was mobility. It is believed that the concept of 'trade-offs' between traits and the constant problem faced by ecologists of scale are the causes of this poor correlation. It is generally believed that frequently disturbed streams will have a higher proportion of mobile species to enable rapid recolonisation. Extensive field experiments revealed that there was indeed a difference between streams, but that communities in less disturbed streams were actually more mobile than those in disturbed streams. Disturbance frequency and refugium availability did not significantly contribute to explaining species variance at a large scale, but differences in mobility were revealed at a smaller scale. These results suggest that Southwood's Habitat Templet Theory may be important, but care must be taken when choosing the scale at which it is tested.