Spatial patterns along the secondary trails
In support of the original DEF hypothesis along elevational gradients (Alexander et al. 2010), we observed declining ruderal species richness with increasing elevation on the trails branching off the Rallarvägen (Fig. 6) (Alexander et al. 2010; Dainese et al. 2017; Lembrechts et al. 2014; Liedtke et al. 2020; Wedegärtner et al. 2022). Interestingly, only few non-native ruderals were observed along these trails, especially above 500 m a.s.l., rendering the decline in non-native richness less obvious.
In the traditional view of the DEF-hypothesis, high anthropogenically disturbed sites in the lowlands ensure quick and repeated introductions of non-native ruderals, from where they move uphill until they reach their (current and species-specific) elevational limits. This results in decreasing richness with elevation. However, our results showed that neither the non-native nor all ruderal species’ elevational maxima in our system were correlated with their climatic affinity (EIV-T). Instead, elevational limits unexpectedly related positively with their first year of observation in the whole region (Fig. 7): newest introductions were residing on average farther uphill. These patterns evoke two possible interpretations. First, there has not been incremental, linear uphill migration over time. This incremental migration might here be overruled by human-mediated long-range dispersal, which usually results in faster uphill migration than by climate change (Hulme, 2014). This is in accordance with what we said earlier and suggests that the DEF is often as much disturbance-driven as it is climate-driven. These higher elevation populations might however not be stable as climatic conditions might not allow interannual survival, and long-term monitoring is thus needed to verify these patterns (Lembrechts et al. 2016a). Second, our analysis could be missing key variables responsible for the elevational distribution. The goodness-of-fit for these regression analyses was very low (Fig. 7), substantiating the fact that the first year of observation to the whole region does not fit the elevational maximum very well.