Supplementary material 3 from: Rato J, Brandão P, Gama M, Banha F, Anastácio P (2025) Effectiveness of legislative tools to stop biological invasions: freshwater turtles' invasion in Europe as a study case. In: Anastácio P, Ribeiro F, Chainho P (Eds) Invasions in Aquatic Systems. NeoBiota 102: 313-329. https://doi.org/10.3897/neobiota.102.143330

Our data on FTs imports was retrieved from the CITES database, which includes all imports for the species on its appendices for all ratifying countries. However, instead of the number of shipments like Mair et al. (2019), we opted to use an approach similar to the one of Slábová et al. (2021) and used the number of traded individuals. By doing it, not only do we increase the representation of the most threatened species, but we also increase the representation of small traded species that can be imported in large shipments (Slábová et al., 2021). Nevertheless, the use of CITES database still has several constrains. First, it does not include all traded species. One example is Pseudemys concinna, a non-listed species in CITES, but a commonly sold pet (Banha et al., 2019; Kopecký et al., 2013). Similarly, one species may be added or deleted to the CITES appendices, with no alteration to this species market (e.g. Pelodiscus sinensis) but its imports stop being reported. The constrains above described do not affect our conclusions and might even mean that our import numbers are underestimated, with a higher number of traded FTs. If so, the introduction of FTs may be much more problematic in the future. Although citizen science platforms, such as GBIF, give useful information on wild populations (Fraisl et al., 2022; Ivanova & Shashkov, 2021), there are several constraints to that data type. These platforms rely on photo identification to distinguish the several species, that often are misidentified by users (Dickinson et al., 2010; Durso et al., 2021). In some cases, citizen science platform users may intentionally upload the wrong location or even upload photos of turtles in captivity (Rato et al., 2024). Also, when turtles bask in groups, multiple individuals, and even different species, may be captured by a photo (Rato et al., 2024). The opposite may also occur, with several photos of the same individuals, resulting in multiple records of the same individual at a given location (Durso et al., 2021). These constraints are very hard to deal with when working with large datasets, because they imply verification of each photo, the manual exclusion of duplicate records, and count of all wild individuals in all photos. Given our dataset dimension, here we considered each GBIF record as one individual of the reported species, and we did not evaluate each photo individually. Apart from the constraints of citizen science platforms themselves, records of wild species tend to be biased to highly populated areas, to daytime and weekends and to “prettier” and rare species (Callaghan et al., 2021; Dickinson et al., 2010; Durso et al., 2021; Fraisl et al., 2022). Additionally, species characteristics that increase detectability also may affect user records on citizen science platforms (Callaghan et al., 2021).