Scale, time, space, and predictability: species distributions of preimaginal black flies (Diptera: Simuliidae).

Abstract

This study examines species-specific distributions of preimaginal black flies (Diptera: Simuliidae) over two large spatial scales (within and across ecoregions) and two seasons (spring and summer). The study area in the western two-thirds of South Carolina, United States, was divided into three ecoregions (Mountains, Piedmont, Sandhills). The mean correct classification of species distributions among stream sites was 81.3%. Predictability of species distributions varied with spatial scale, location, and time. Species occurrence was not independent of ecoregion; distributions of 12 of the 13 species examined at this scale were predictable on the basis of changes in water column (temperature, percentage dissolved oxygen) and channel characters (velocity, streambed-particle size). However, with the exception of the Mountains during the summer, predicting species distributions within ecoregions was far less successful than predicting distributions across these regions; predictability was particularly poor in the Sandhills. Seasonal effects on predictability were most pronounced in the Mountains. As stream sites became more homogeneous, predictive capability declined. Species-level identifications, aided by cytogenetic techniques, were of paramount importance in recognizing the association between species distributions and stream conditions, emphasizing that considerable ecological information is lost when species are not considered. This study emphasizes the critical role of taxonomic resolution, linked with spatial and temporal scale of observation, in elucidating patterns of species assemblages. It also demonstrates that predictability of species distributions at a fixed scale can vary with geographic location and time.