Accounting for flow intermittence in freshwater species distribution modelling

Abstract

To forecast species responses to environmental changes, it is crucial to understand drivers of species distribution acting at multiple spatial and temporal scales. While hydrology is acknowledged as a key driver for freshwater species, most studies have been limited to perennial river flow regimes when exploring species distribution in rivers. We developed a method to characterize stream flow to subsequently account for flow intermittence in species distribution models (SDMs). First, we used the hydrological Soil and Water Assessment Tool (SWAT) to simulate flow in the French Pyrenees at fine spatial grain and large spatial extent. Several metrics characterizing flow intermittence were subsequently included in a SDM along with topographic, hydrographic, climatic and land use variables. We applied this methodology to an extensive data set of the Pyrenean brook newt (Calotriton asper) presence in France.This amphibian is highly dependent on aquatic habitats, and its distribution could be mainly driven by hydrological and climatic variables. We found that 60% of the stream network of the study area is intermittent. Maximum air temperature and proportion of agricultural areas contribution in explaining species distribution were the most important (26.3% and 21.2%, respectively). Both variables negatively influenced the probability of presence of the Pyrenean brook newt. We also found a negative correlation with the frequency of zero-flow events (13%), emphasizing the role of flow intermittence in driving aquatic species distribution. Our study provides new insights into fine-grained hydrology in the Pyrenees and illustrates intermittence characterization and mapping methods that could be applied elsewhere.