Search

Understanding and predicting the geographic distribution of taxa in hierarchical stream landscapes is a cornerstone of river ecology. A central issue is to tease apart the unique and shared effects of local and catchment predictors over species distributions. Here, we tested Hynes's influential hypothesis (1975, Baldi Memorial Lecture) that ‘In every respect, the valley rules the stream'. We predicted that if catchment features exert a major control on in-stream local conditions, the shared effect of local and catchment predictors should largely surpass their unique effects. To test this prediction, we used logistic regression models and variation partitioning to quantify the unique and shared effects of local and catchment predictors on the distribution of two hyporheic crustacean taxa (Bogidiellidae, Amphipoda and Anthuridae, Isopoda) in streams of New Caledonia. We sampled the two taxa at 228 sites. At each site, we quantified nine local predictors related to habitat area and stability, sediment metabolism and water origin, and eight catchment predictors related to geology, area, primary productivity, land use and specific discharge. When analyzed separately, the two predictor types explained the same amount of model variation in occurrence in both taxa. When analyzed jointly, the shared effects of the two predictor types explained twice as much model variation as the unique effect of each. The overriding contribution of shared effects was notably due to controls exerted by catchment area and geology on local habitat size and sediment metabolism, respectively. For both taxa, a model with only these two catchment predictors provided occurrence distribution as reliable as models containing only local predictors or both predictor types. Our findings pave the way for predicting reliably from catchment predictors alone the geographic distribution in local occurrence of taxa in difficult-to-access habitats and landscapes, such as here the hyporheic zone of tropical str