Experimental design
In 2004 five temporary fences were constructed along the permanent fence, extending approximately 6 m into the summer (heavily) grazed side of the fence, before reindeer entered the area in the autumn. Each of the temporary fences is between 85 and 185 meters in length and built in areas that allowed for the capturing of terrain and elevational differences. The temporary fences were taken down every other year to allow for grazing in the area closest to the permanent fence every second year (Figure 1a). The area on the inside of the temporary fence was thus exposed to acute (i.e. pulse grazing) after more than forty years of chronic (i.e. press grazing), while the areas outside of the temporary fence continued to experience press grazing. We placed permanent plots in the sites along the permanent fence where the temporary fences were constructed (n = 5 sites). Each site contained two sets of triplets (n = 6 plots) except for the fourth fence, which contained four sets of triplets (n = 12 plots), for a total of 36 permanent plots (Figure 1b). At each site, one set of triplets was placed in a drier area (18 plots in total) and the other in a moist/wet area (18 plots in total); we refer to this covariate as wet/dry. Within each triplet, each of the three plots was exposed to a different grazing intensity: the plot on the side of the permanent fence in the spring and autumn ranges was lightly grazed (12 plots in total), the plot inside the temporary fence was grazed every other year (pulse grazing; 12 plots in total), and the plot on the side of the permanent fence in the summer grazed area was grazed every year (press grazing; 12 plots in total). All plots were placed no more than 12m from the permanent fence, but plots on the lightly grazed side of the fence were placed 3m away from the fence due to the man-made disturbance in this area from herders walking next to the fence. Each plot was 3m x 3m (9m2) and was evenly split into nine 1m x 1m (1m2) sub-plots.