Site description and experimental design
This study was carried out in Carbon in Permafrost Experimental Heating Research (CiPEHR) site, established in September 2008. This site is a typical moist acidic tundra located in the northern foothills of Interior Alaska (63°52’59”N, 149°13’32”W), near the Denali National Park and Preserve of Eight Mile Lake. The annual mean temperature of the year 1976-2009 was -1.0 °C, with monthly mean temperature ranging from -16 °C (December) to 15 °C (July). The annual mean precipitation of 1976-2009 was 378 mm. The soil type is gelisol, with a 45-65 cm thick organic horizon covered by a mixture of glacial till and loess. The active layer thickness is 50-60 cm. Dominant plant species included the tussock-forming sedge Eriophorum vaginatum (Cyperaceae) and the small deciduous shrub Vaccinium uliginosum (Ericaceae). Other plants included Carex bigelowii(Cyperaceae), bryophytes Pleurozium schreberi (Hylocomiaceae),Sphagnum spp. (Sphagnaceae), and lichens Cladonia spp. (Cladoniaceae).
The experimental soil warming study was initiated in October 2008. A total of 6 snow fences (1.5 m tall × 8 m long) were established and arranged into 3 blocks, which was ~ 100 m apart to mitigate small-scale spatial variability. For each snow fence, the warmed plot was located at the leeward side with 5 m distance from the fence, and the paired control plot was located at the windward side with 8-14 m distance from the fence. Snow fences increased soil temperature via trapping insulating layers of snow in the leeward side. For example, snow fences led to a threefold increase in snow depth of warmed plots in 2009 (1.35 m in warmed plots versus 0.38 m in control plots) (Natali et al., 2011). Different from traditional snow fence experiments, both snow and snow fences were removed before snowmelt to ensure that soil hydrology, light conditions, and plant phenology were not affected by extra snow. On March 8-15 of 2010, extra snow in warmed plots was shoveled to ensure comparable snow cover depth between warmed and control plots. Snow fences were removed at the time of snow removal to avoid shading of plots during the growing season (Natali, Eag, & Rubin, 2012; Natali et al., 2011). Snow removal allowed this experiment to focus on the effects of soil warming.