Decoupled long‐term effects of nutrient enrichment on aboveground and belowground properties in subalpine tundra

Fertilizer experiments aid understanding of the role of nutrient limitation in tundra ecosystems, and of the effects that global change may exert on them through enhancing nutrient availability. However, little is known about whether fertilizer additions influence tundra ecosystem properties directly, or indirectly through altering plant community composition. To separate these direct and indirect effects, we used an ongoing fertilization experiment in subalpine tundra in northern Sweden initiated in 1989. Here, the slow‐growing dwarf shrub Empetrum hermaphroditum (hereafter Empetrum ) dominates, but fertilization causes its replacement by the fast‐growing grass Deschampsia flexuosa (hereafter Deschampsia ). Therefore, in 1999, we set up three subplots in every plot of the fertilizer experiment subjected to different treatments, i.e., regular removal of Empetrum or of Deschampsia , or no removal. An interactive effect between fertilizer and removal treatments means that fertilization indirectly influences ecological properties through its effects on the dominant species whereas no interactive effect means that fertilization effects are more likely direct. We measured plant community properties four times between 1999 and 2010, and belowground properties in 2010. Fertilization exerted large effects on nearly all variables aboveground (i.e., vascular plant community properties, bryophyte and lichen cover) and belowground (i.e., measures of nutrient availability, mass of humus and major elements per area, microbial community properties). Meanwhile, loss of either Deschampsia or Empetrum reduced total vascular plant cover, with greatest effects in those fertilization treatments where they otherwise had the most mass. Deschampsia removal affected several other plant community properties, but mostly independently of fertilizer treatment; Empetrum removal had little effect on any other plant community variable. Belowground, both removal treatments reduced humus mass per area but had no effect on any other variable. Our finding of few interactive effects of fertilizer and removal treatments aboveground and none belowground means that fertilizer effects on most community and ecosystem properties operate independently of even large shifts in the plant community. These results also point to a decoupling between aboveground and belowground communities in tundra, and therefore offer insights about the mechanistic basis through which global change impacts on soil nutrient availability may transform tundra ecosystems.