Long-term effects of fertilization and herbivory on the boreal forest understory: Results of a 20-year experiment
Knowledge of the factors that control vegetation composition and productivity are becoming increasingly important as global climate change influences both environmental conditions and also herbivore distribution and abundance. Understory vegetation in the boreal forest grows in stressful conditions with low levels of light and fertility, prolonged winters with low temperatures, and brief growing seasons with cold soils. This vegetation also undergoes elevated levels of herbivory by snowshoe hares every 10 years when hares reach peak population densities during their population cycle. We investigated whether vegetation composition in the boreal forest understorey was controlled by resource level alone (bottom up control), herbivory alone (top down control) or by both (interactive control). We present results from a 20-year period of a 2 x 2 factorial experiment manipulating fertilizer addition and mammalian herbivore exclosures where we examined long-term responses of vegetation composition and ecosystem properties (including soil nutrients and microbial activity) over a 20-year period. We also examined the potential for recovery from these long-term manipulations; after 10-years of treatments, treatments were discontinued on half of each plot and recovery in vegetation and soil properties monitored for the following 10 years.
While long-term fertilization had significant effects on most variables, herbivore exclosures had few detectable effects, even after 20 years including two snow-shoe hare population peaks. With fertilization, plant species richness declined dramatically and the community became dominated by two species, Epilobium angustifolium and Mertensia paniculata. Some species initially showed positive responses to fertilizer (e.g. Festuca altaica) which later reversed. Predictably, fertilized plots showed higher levels of soil nutrients, although the enzymes involved in the acquisition of N and P were mostly unaffected by fertilizer. Total microbial biomass declined with fertilization, as did the ratio of fungi to bacteria, indicating a more bacterial dominated food web. Recovery of vegetation and soil properties after cessation of fertilization was slow, and in some cases, non-existent. In plots where treatments were ceased, recovery of different ecosystem and community properties appears to follow a ‘cascade of responses’. Soil chemistry recovered first, followed by plant tissue chemistry, followed more slowly by plant community composition. After 10 years, however, the soil microbial community shows no signs of recovery. With this experiment, we highlight the necessity of long-term research, as short-term effects were not necessarily representative of those after two decades of treatments.