In the face of increasing temperature at high latitudes, ecosystem respiration (RE) is a key to determining sink and source dynamics of a boreal forest. In this paper, we analyzed four-year RE data obtained in an open black spruce forest-a typical boreal forest ecosystem with permafrost in Interior Alaska. RE measured as nighttime CO2 effluxes for both the ecosystem and the forest floor were clearly situated along the exponential temperature dependent curve, except for the data obtained in extremely dry conditions in mid-summer. More than 93 % of RE data measured at nighttime with high vapor pressure deficit (VPD > 400 Pa) were lower than the values predicted from the temperature-dependent curve. Consequently, the year 2013 (with an unusually dry summer) had a 15 % lower amount of annual RE than that could be expected from the temperature-dependent curve without considering the effect of high VPD. The suppression of RE under dry conditions was also related to decreases in soil moisture and net ecosystem productivity. Finally, assuming daytime RE could be extrapolated from the temperature-dependent curves, annual daytime RE estimated with the effect of high VPD was decreased by up to 62% from RE estimated without the VPD effect. The down-regulation of RE presented in this study postulates a possible negative feedback for the carbon budget of boreal forests in response to climate warming.

In the future, climate models predict an increase in global surface temperature and during winter a changing of precipitation from less snowfall to more raining. Without protective snow cover, freezing can be more intensive and can enter noticeably deeper into the soil with effects on C cycling and soil organic matter (SOM) dynamics. We removed the natural snow cover in a Norway spruce forest in the Fichtelgebirge Mts. during winter from late December 2005 until middle of February 2006 on three replicate plots. Hence, we induced soil frost to 15cm depth (at a depth of 5 cm below surface up to -5 degrees C) from January to April 2006, while the snow-covered control plots never reached temperatures < 0 degrees C. Quantity and quality of SOM was followed by total organic C and biomarker analysis. While soil frost did not influence total organic-C and lignin concentrations, the decomposition of vanillyl monomers (Ac/Ad)(V) and the microbial-sugar concentrations decreased at the end of the frost period, these results confirm reduced SOM mineralization under frost. Soil microbial biomass was not affected by the frost event or recovered more quickly than the accumulation of microbial residues such as microbial sugars directly after the experiment. However, in the subsequent autumn, soil microbial biomass was significantly higher at the snow-removal (SR) treatments compared to the control despite lower CO2 respiration. In addition, the water-stress indicator (PLFA [cy17:0 + cy19:0] / [16:1 omega 7c + 18:1 omega 7c]) increased. These results suggest that soil microbial respiration and therefore the activity was not closely related to soil microbial biomass but more strongly controlled by substrate availability and quality. The PLFA pattern indicates that fungi are more susceptible to soil frost than bacteria.

Arctic climate change is expected to lead to a greater frequency of extreme winter warming events. During these events, temperatures rapidly increase to well above 0 degrees C for a number of days, which can lead to snow melt at the landscape scale, loss of insulating snow cover and warming of soils. However, upon return of cold ambient temperatures, soils can freeze deeper and may experience more freeze-thaw cycles due to the absence of a buffering snow layer. Such loss of snow cover and changes in soil temperatures may be critical for litter decomposition since a stable soil microclimate during winter (facilitated by snow cover) allows activity of soil organisms. Indeed, a substantial part of fresh litter decomposition may occur in winter. However, the impacts of extreme winter warming events on soil processes such as decomposition have never before been investigated. With this study we quantify the impacts of winter warming events on fresh litter decomposition using field simulations and lab studies. Winter warming events were simulated in sub-Arctic heathland using infrared heating lamps and soil warming cables during March (typically the period of maximum snow depth) in three consecutive years of 2007, 2008, and 2009. During the winters of 2008 and 2009, simulations were also run in January (typically a period of shallow snow cover) on separate plots. The lab study included soil cores with and without fresh litter subjected to winter-warming simulations in climate chambers. Litter decomposition of common plant species was unaffected by winter warming events simulated either in the lab (litter of Betula pubescens ssp. czerepanovii), or field (litter of Vaccinium vitis-idaea, and B. pubescens ssp. czerepanovii) with the exception of Vaccinium myrtillus (a common deciduous dwarf shrub) that showed less mass loss in response to winter warming events. Soil CO2 efflux measured in the lab study was (as expected) highly responsive to winter warming events but surprisingly fresh litter decomposition was not. Most fresh litter mass loss in the lab occurred during the first 3-4 weeks (simulating the period after litter fall). In contrast to past understanding, this suggests that winter decomposition of fresh litter is almost nonexistent and observations of substantial mass loss across the cold season seen here and in other studies may result from leaching in autumn, prior to the onset of true winter. Further, our findings surprisingly suggest that extreme winter warming events do not affect fresh litter decomposition. Crown Copyright (c) 2009 Published by Elsevier Ltd. All rights reserved.