Invasibility of Mediterranean-climate rivers by non-native fish: the importance of environmental drivers and human pressures

Abstract

Invasive species are regarded as a biological pressure to natural aquatic communities. Understanding the factors promoting successful invasions is of great conceptual and practical importance. From a practical point of view, it should help to prevent future invasions and to mitigate the effects of recent invaders through early detection and prioritization of management measures. This study aims to identify the environmental determinants of fish invasions in Mediterranean-climate rivers and evaluate the relative importance of natural and human drivers. Fish communities were sampled in 182 undisturbed and 198 disturbed sites by human activities, belonging to 12 river types defined for continental Portugal within the implementation of the European Union's Water Framework Directive. Pumpkinseed sunfish, Lepomis gibbosus (L.), and mosquitofish, Gambusia holbrooki (Girard), were the most abundant non-native species (NNS) in the southern river types whereas the Iberian gudgeon, Gobio lozanoi Doadrio and Madeira, was the dominant NNS in the north/centre. Small northern mountain streams showed null or low frequency of occurrence and abundance of NNS, while southern lowland river types with medium and large drainage areas presented the highest values. The occurrence of NNS was significantly lower in undisturbed sites and the highest density of NNS was associated with high human pressure. Results from variance partitioning showed that natural environmental factors determine the distribution of the most abundant NNS while the increase in their abundance and success is explained mainly by human-induced disturbance factors. This study stresses the high vulnerability of the warm water lowland river types to non-native fish invasions, which is amplified by human-induced degradation.

Figure 1. Map of the river types defined to continental Portugal showing undisturbed (grey dots) and disturbed (black dots) sampling sites.

Figure 2. Response pattern of the most abundant NNS to human-induced pressure gradient, established according to 5 environmental quality classes: a) Mean density of L. gibbosus; b) Mean density of G. lozanoi; c) Mean density of G. holbrooki.

(▪): Mean; box:+SE; whisker:+SD. Significance of ANCOVA results with total human pressure as a covariable is shown.

Figure 3. Veen diagrams showing the partition of the variance in the occurrence (a) and mean density (b) of the most abundant NNS explained by environmental variables, human-induced pressure and spatial trends.