USING AN INTEGRATED MODELLING APPROACH FOR RISK ASSESSMENT OF THE ‘KILLER SHRIMP’ Dikerogammarus villosus

• Published in: River research and applications 2014-05, Vol.30 (4), p.403-412
• Main Authors: Boets, P, Pauwels, I. S, Lock, K, Goethals, P. L. M
• Format: Article
• Language: English
• Subjects: Animal and plant ecology; Animal, plant and microbial ecology; Applied ecology; aquatic invasive species; Biological and medical sciences; chemical oxygen demand; Conservation, protection and management of environment and wildlife; Crustacea; Dikerogammarus villosus; environmental factors; Flanders; Fresh water ecosystems; Freshwater; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; habitat suitability; habitats; High-speed machining; Invertebrates; macroinvertebrates; Mathematical models; Methods and techniques (sampling, tagging, trapping, modelling...); migration; Modelling; monitoring; network analyst tool; orthophosphates; Parks, reserves, wildlife conservation. Endangered species: population survey and restocking; prediction; regression analysis; regression trees; risk assessment; rivers; simulation models; Spreads; Synecology; Water quality; water quality model; Watercourses
• ISSN: 1535-1459; 1535-1467
• DOI: 10.1002/rra.2658

The aim of this study was to predict the future distribution of an invasive macroinvertebrate species, Dikerogammarus villosus, under changing environmental conditions. For this, we used an integrated modelling approach. First, a habitat suitability model (HSM) was constructed on the basis of a regression tree model, to determine the preferred chemical water quality conditions. Subsequently, this HSM was combined with a chemical water quality model that makes predictions on future water quality scenarios for the year 2015 and 2027. It was expected that the area of suitable habitat of D. villosus would increase with improving water quality conditions in the future. Finally, migration speed was incorporated to model the spatial–temporal spread of D. villosus based on a network analysis. On the basis of monitoring data of Flanders (Belgium), it was calculated that D. villosus is able to spread with an average speed of 5 km year⁻¹.The model simulations indicate that the species is primarily present in large rivers and canals with a good chemical water quality. With improving water quality, the species will be able to colonize additional watercourses, mainly because of a decrease in chemical oxygen demand and orthophosphate concentration. A validation based on the observed occurrence shows that the model accurately predicts areas with a high suitability that are most likely to be invaded by D. villosus. Our integrated modelling approach is useful as a practical method to perform risk assessment for watercourses that are vulnerable to invasions not only in Flanders but also in the whole world. Copyright © 2013 John Wiley & Sons, Ltd.