Rationale. Shrub growth and expansion is occurring in the Arctic, with implications for global climate. Multiple strands of evidence demonstrate an important role of recent increases in tundra air temperatures in enhanced growth of deciduous shrubs at the expense of graminoids and mosses, commonly termed ‘shrubification’. However, shrub growth responses are heterogeneous through both space and time, as evidenced by: (a) meta-analyses of the climate-sensitivity of shrub wood production through time; and (b) progressive declines in the strength of the relationship between remotely sensed vegetation productivity and warming air temperatures. It is plausible that – over decades – soil nitrogen (N) availability may contribute to the observed heterogeneity of shrub growth sensitivity to climate. My key aim was to quantify and identify the presence of N-limitation to shrub growth and its variability in space and through time in Western Siberia, Russia. Method. First, to determine the status of the evidence base for alternative controls on shrub growth, including soil N, I created and followed an evidence mapping protocol to generate a spatial- temporal evidence map. Second, I analysed stable nitrogen isotope ratios (δ15N) in Salix lanata L. shrub wood rings, which provided annually resolved time-series to indicate local-scale N availability. I created annual resolution δ15N time series for ten individuals at Yuribei (1980-2013), Yamal Peninsula, and three-yearly resolution time-series for ten shrub individuals at each of three sites within Western Siberia (1940-2013). I interrogated the δ15N and ring width time-series for shrub individuals by creating mechanistic representations of shrub-nitrogen dynamics; I then identified the most appropriate representations of N-dependency and other environmental constraints on individual shrub growth. Third, I assessed the role of snow dynamics in controlling the relations between nitrogen and shrub growth by extending the modelling approach to incorporate seasonality. Here, snow variability was proxied at a regional scale by earth observation and a local scale by the investigation of δ18O and δ13C isotope time series from wood rings. A software library, Bristlecone, was developed and used to conduct model-fitting and model-selection of mechanistic models. Results. First, the evidence map identified that there were spatial-temporal limitations within the recent evidence for controls on Arctic shrubification processes, including constraints on the temporal extent and resolution of N-limitation effects on shrub growth. Second, reconstructed δ15N time-series indicated that declines in N availability to the Salix lanata shrubs have occurred over decades. Although declines in soil N were observed in Arctic bioclimatic subzones D and E, different trajectories occurred at fine spatial resolution. Interrogation of the δ15N and ring width time-series indicated that the growth of all shrub individuals assessed has been N-limited; and for a majority of these, plant-soil litter feedbacks were a significant determinant of shrub-nitrogen dynamics. Model-inferred dynamics also suggested that the ratio of shrub N uptake to soil N supply has increased over decades. I identified summer air temperature as an important control on individual shrub growth rates in Western Siberia but that this has declined over recent decades. In addition, the growth of shrub individuals at Yuribei was sensitive to the length and timing of the snow-free growing season, but analysis of nitrogen-shrub dynamics did not identify snow depth effects to soil N replenishment rates nor shrub biomass protection. Implications. The findings challenge a key assumption that rising air temperatures have led to universal increases in N availability within tall shrub-dominated Arctic tundra environments. Although all Salix lanata individuals studied were N-limited over decades, the strength and form of limitation varied, with implications for modelling of these dynamics in Earth System Models. The presence of strong litter feedbacks and relative insensitivity of shrub-nitrogen relations to snow-induced microclimatic effects suggest that shrub establishment and maturation provides a greater long-term control on N availability than climatic controls.