Quantitative habitat models for the conservation of the endangered European crayfish Austropotamobius pallipes complex (Astacoidea: Astacidae)

Abstract

Crayfish are the largest mobile freshwater invertebrates and are keystone species in European aquatic ecosystems particularly in small streams and rivers. The white-clawed crayfish Austropotamobius pallipes (a species complex) is currently classified by the IUCN Red List as an endangered species (EN), because its populations have decreased significantly over the last decades in a number of European countries including Italy, due mainly to habitat modifications and the introduction and spread of alien species. Data on the ecological requirements of A. pallipes are needed to quantify the effects of habitat alteration, to simulate restoration scenarios, and to implement effective conservation measures for this species. We describe here a new methodology for modelling the habitat requirements for this endangered crayfish using the mesohabitat scale approach based on data from crayfish living in small streams draining the Italian foothills of the Alps (Lombardy region) and in streams in the mountainous areas of the Gran Sasso and Monti della Laga National Park (Abruzzo region). Data from seven morphologically different streams were used to train and validate the habitat models. The Random Forests algorithm was used to identify the best and most parsimonious habitat model, and to define the lowest number of variables to be surveyed in the future. The best habitat models were applied to each stream and used to classify each mesohabitat into suitability categories. Habitat flow-rating curves were developed to analyze spatio-temporal variation of habitat availability, and habitat time series analysis were used to define detailed management schemes for environmental river management. Flow releases and water temperature regimes were assessed for individual water diversions in order (1) to represent how physical habitat changes through time, and (2) to identify stress conditions for A. pallipes created by the persistent limitation of habitat availability. Results indicated that the kind of substrate in the stream bed (such as the proportion of fine-grained substrates), the water depth (whether shallow or deep), and the available cover (such as the presence of boulders, woody debris, and undercut banks) were all significant factors governing the occurrence of crayfish. The habitat models performed well in both calibration and validation phases (with accuracy ranging from 71 to 79 % in training and from 69 to 73 % in validation) and can be considered to be a valuable tool to predict the distribution of A. pallipes over a wide range of stream types. An example of how to establish environmental standards for small streams is presented. The proposed habitat model provides a useful tool that can be applied even when other commonly used methodologies are unsuitable. As such, this habitat model can be used to develop regional rules for the conservation of the endangered crayfish A. pallipes complex and for defining more site-specific management criteria.