2.3. Data analysis
The index of relative importance (IRI) was used to represent the dominance of different fish species. This index is a composite measure that reduces bias in descriptions of animal dietary data, but has been proposed as a standard method for studying fish species diversity and community ecology (Paul & Haefner, 1990; Hart et al., 2002; Hill et al., 2017). The IRI value for each species combines its frequency of occurrence, percentage of total biomass (PW = W1-I / WT), and percentage of total numbers of prey consumed (PN = N1-I / WT):
IRI = (PW + PN) FO1-I * 104.
A species for which the value of IRI is >300 can be recognized as a dominant species; in this study, an IRI value of >1000 was used to denote an absolute dominant fish species.
A Bray–Curtis similarity measure was also computed. This was composed of a similarity coefficient matrix based on the ratio of non-native fish to the total number of fish (RN = NN1-X/ NT), as well as the ratio of different non-native species in the total number of non-native fish species catch (RNT = NN1-X / NNT). The RN and RNT were square-root transformed in the analysis of the similarity matrix across the different rivers. After that, the similarity indices across the different rivers was used to perform a clustering analysis with Primer software (Clarke et al., 2006).
The numbers of non-native fish species (X) in the data from the Hainan Island rivers and the Continent coastal rivers was compared using one-way analysis of variance (ANOVA). The ratios of the most-common non-native native fishes (RNT = NN1-X / NNT) in the Hainan Island rivers and the Continent coastal rivers were compared using a nonparametric test (two independent samples). Linear regressions were applied to determine the relationship between mean minimum temperature in the coldest month (data for January 2018) and abundances of the most widely distributed non-native fish species. All statistical analyses were performed with SPSS version 16.0, and the results were considered significant at p ≤ 0.05.