Introduction
Evidence is mounting that rapid human-caused environmental changes, such as climate (e.g., temperature warming and shifting precipitation patterns) and land use change, cause substantial species redistributions in mountain areas (Elmendorf et al. 2012; Hedenås et al. 2016; Kowarik, 2003; Pearson et al. 2013; Thuiller et al. 2008). Climate change is four times the global rate in Arctic regions (Rantanen et al. 2022) and is especially pronounced in cold-climate mountain areas (Callaghan et al. 2013). Signs of the impact of warming on communities are already evident in many of these mountain regions, where migration of species from lower to higher elevations have been well-documented (Dainese et al. 2017; Frei et al. 2010; Rixen & Wipf, 2017). The effects of land use changes are usually less profound at higher elevations compared to lower elevations, the rate of change in communities often lags that of climate (Bertrand et al. 2011). This interchange between predominantly disturbance-driven populations at lower elevations and the largely climate-driven populations along elevational ranges, makes mountain areas especially suitable for the study of the synergy between both.
Rising temperatures as a consequence of global change have favored introductions of warm-adapted, non-native species to become established in mountain areas, especially where disturbance was high (Elmendorf et al. 2012; Heijmans et al. 2022; Pearson et al. 2013; Taylor et al. 2017; Thuiller et al. 2005). The vegetation productivity and the length of the growing season have also increased, as a result of both higher temperatures in summer and on average a decrease in snow cover in winter (Elmendorf et al. 2012; Pearson et al. 2013). For some species this leads to increased growth rates or an extended distribution, while other species may have adverse effects on fitness. The impact of climate change on vegetation communities per se is thus quite difficult to predict, as responses to the changing environment can vary widely in speed and magnitude across species and functional groups (Klanderud & Totland, 2005; Parmesan & Hanley, 2015).
Mountain regions are becoming increasingly accessible through improved infrastructure. Roads and hiking trails are major conduits for human-mediated dispersal in these regions (Dainese et al. 2017; Lembrechts et al. 2014 and 2016a, Liedtke et al. 2020; Wedegärtner et al. 2022), allowing for rapid uphill migration (Hulme, 2014). These disturbed sites are often characterized by changes in soil conditions, such as compaction and chemistry, which affect species diversity and composition, by creating an environment that often promotes ruderal species (Frenkel, 1977; Guo et al. 2018; Rendeková et al. 2019). Roadside dispersion is related to traffic intensity and the size of the road network (Chiuffo et al. 2018; Pauchard et al. 2009), while hiking trails often facilitate ruderal plant dispersal from roads or settlements further into the mountains (Liedtke et al. 2020).
Human-mediated dispersal facilitates non-native plant species influxes from all over the world as tourists are often bringing in hitch-hiking seeds that stick to their clothing, boots, or the tires of cars (Frenkel, 1977). Most of these non-native species have a ruderal growth strategy (Alexander et al. 2016; Chiuffo et al. 2018; Kowarik, 2003). Consequently, non-native ruderals mostly appear first near train stations (Brandes, 2002), parking lots (Frenkel, 1977), roadsides (Lembrechts et al. 2014), and other places where human displacement is most abundant (Guo et al. 2018; Liedtke et al. 2020). The degree of invasion in a community is thus related to the intensity of human activity (Kowarik, 2003).
Due to their long, harsh winters, and short, relatively cold summers, subarctic mountain ecosystems were previously believed to be relatively resistant to the influx of non-native species (Pauchard et al. 2009), but climate change and increased anthropogenic disturbance are gradually changing this view (Pauchard et al. 2009; Walther et al. 2009). Many non-native ruderal species are known to be good dispersers that can reach high elevations twice as fast as native species (Dainese et al. 2017), although their climatic tolerance may constrain their survival to the next growing season (Rendeková et al. 2019). Nevertheless, a widespread uphill migration of non-native species has been observed along elevational gradients in response to climate change (Alexander et al. 2016; Dainese et al. 2017; Kueffer et al. 2013; Pauchard et al. 2009). Indeed, introductions tend to take place in the lowlands (Alexander et al. 2010; Guo et al. 2018; Liedtke et al. 2020; Pauchard et al. 2009), and from these sites species either move through human-mediated dispersal or spread out on their own.
The Directional Ecological Filtering (DEF) process describes the unidirectional uphill expansion of non-native species (Alexander et al. 2010). Non-native species richness gradually declines with increasing elevation. With their lower elevational limit consistently in the lowlands, non-native species spread over an elevational range until they reach their upper elevational limit. As a result of this directional movement starting in the lowlands, only climatic generalists are likely to reach high elevations. Non-native species are thus gradually filtered out along the elevational gradient, probably due to increasing climatic harshness (Alexander et al. 2010), although evidence shows that a longer residence time also inevitably results in higher elevational limits (Pyšek et al. 2011).
Testing the interactive effects of climate change and anthropogenic disturbances on native and non-native ruderal species expansion requires detailed knowledge on the history of disturbance events, as well as long-term data on ruderal species distributions. Such data is available for a mountain region in the north of Sweden, around Abisko – a small village known for its hiking trails and the Abisko Scientific Research Station (Andersson et al. 1996). The local climate is defined as subarctic with cool summers and relatively mild winters with extensive snow cover. The Scandes mountain range to the west, creates a rain shadow effect directly over Abisko, making it the sunniest area in northern Sweden (Callaghan et al. 2010 and 2013). However, similar to other high latitude regions, Abisko has been subject to increasingly severe climate warming in combination with substantial anthropogenic disturbance since the early 1900s (Callaghan et al. 2013). This makes it an ideal study area to test the interaction of these global change drivers, specifically on the introduction and changes in ruderal species compositions over time.
In 1903, a railroad was completed from Kiruna to Narvik, soon followed by the first tourist hotel in Abisko (Callaghan et al. 2013). The Rallarvägen trail - the focus of the underlying study - runs parallel to the railroad and served as a transport road during construction. The accessibility of the Abisko region was further improved with the opening of the first paved road (the E10 highway) from Kiruna to Riksgränsen in 1982, which followed the existing Rallarvägen and the railroad line. The effects of the E10 on roadside vegetation were studied in 1989 (Bäck & Jonasson, 1998). Yet, in contrast to other studies that have examined the role of roads on the influx of non-native species (e.g, Lembrechts et al. 2014 and 2016a), the effects of the E10 construction were very limited (Bäck & Jonasson, 1998). Since its opening, the upgraded infrastructure has contributed to an increase in tourism in the region, so it is possible that changes in vegetation composition resulting from the road may have become noticeable only now. Additionally, over the past decades, the average annual air temperature in the region has gradually increased from 0 to 1°C (ANS, 2019; see also Callaghan et al. 2010). These rising temperatures already resulted in significant upward shifts in the treeline and the distribution of a range of plant species, as well as substantial changes in their phenology (MacDougall et al. 2021). The effects of the railroad in combination with the E10, tourism, and climate change may have caused a steady increase in ruderal species in the vegetation and dynamic changes in the ruderal composition over the past 120 years.
Importantly, a unique historical time series is now available for the region: we know the exact timing of major disturbance events (railroad and road building), we have a clear view of the changes in climate in the region (weather data has been continuously measured by the Abisko Scientific Research Station) since 1913 (ANS, 2019), and we have vegetation surveys along the Rallarvägen dating back till 1903 (Lewejohann & Lorenzen, 1983; Sylvén, 1904 and 1913-15). With these datasets in hand, augmented with a recent resurvey of the Rallarvägen trail and additional vegetation monitoring along trails leading from the Rallarvägen into the mountains, we set out to answer three key research questions about the history of ruderal species along the Rallarvägen and in the broader Abisko region:
We hypothesized that railroad building at the beginning of the 20th century would have facilitated the establishment of significant amounts of ruderal species – both native and non-native – along the Rallarvägen trail. Additionally, we expected later disturbance events, such as the building of the E10 in the early 1980s, to have created a new influx of largely non-native ruderals, with in the last decades an influx of mostly warm-adapted ruderal species as a result of climate change.
We expected these ruderal species to be concentrated around points of introduction with continuous disturbance, such as the main train stations, with a progressive decline in richness with increasing distance to these introductory points. Additionally, more recent introductions and warm-adapted ruderals were expected to be restricted to low-elevation and/or warmer environments.