Fig. 1 Geographical locations of the sampling sites. The color of the circle indicates the dominant species of exotic plants

We sampled a total of 2267 quadrats in the 236 sites. For each quadrat, first, plant species were recorded. Then, the coverage of each species (C ) was determined by visual estimation on 22-degree scales: 0.5%, 1%, 5%, 10%, 15%, 20% … 100%. Next, all plant materials, including shoots, leaves, and roots, in the quadrat were collected by careful manual digging. Finally, all plants were classified, washed, and dried at 70 °C for more than 48 h to determine the biomass of each species (B ). The total biomass of native/exotic plants in each quadrat (TBQ_N /TBQ_E ) = ∑B _n, total coverage of native/exotic plants in each quadrat (TCQ _N/TCQ _E) = ∑C _n, n was the species number of natives or exotics in a quadrat. The mean biomass of native/exotic plants at each site (MBS_NMBS_E )\(\ \frac{\sum\text{TBQ}_{i}}{i}\), the mean overage of native/exotic plants at each site (MCS_N /MCS_E ) =\(\frac{\sum\text{TCQ}_{i}}{i}\), and i is the number of quadrats at a site. When the abundance of exotic plants in a site was too low to appear in any quadrat, we determined the values ofMBS_E and MCS_E of the site were zeros. In addition, the defoliation (D_leaf , % of leaf area removed) of exotic plants by herbivores was assessed. 10~20 fresh shoots (all shoots when the total shoots < 10) of each exotic plant were randomly selected for defoliation assessments in each quadrat, andD_leaf was measured using our previously described method (Fan et al. 2016).