Here we present the data used in the manuscript "Removal of grazers alters the response of tundra soil carbon to warming and enhanced nitrogen availability", Ecological Monograps, Early view in October 2019 by H. Ylänne, E. Kaarlejärvi, M. Väisänen, M. K. Männistö, S. H. K. Ahonen, J. Olofsson & S. Stark. In this paper we studied, how five years of experimental warming and increased soil nitrogen availability interact with both long- and short-term differences in grazing intensity in shaping ecosystem carbon stocks and the processes underlying the changes. We used an over 50-year-old reindeer fence that separates a lightly grazed shrub-dominated tundra from a heavily grazed graminoid-dominated tundra, where the different grazing histories on the two sides of the fences have created different ecosystem states. In addition to the long-term grazing difference, we also established short-term grazer exclosures on the heavily grazed side of the fence to account for the effect of a sudden grazing cessation.

This file includes data of ecosystem carbon stocks, soil properties, and fungal and bacterial copy numbers. It also provides data on development of the vegetation through the course of the experiment (2010-2014) and presents the activities of six extracellular enzymes measured on three occasions in 2013.

Experimental set-up and site characteristics

The data is from an experimental set-up, created in 2010 to a tundra heath on the northern slope of  a hill Raisduoddar in northern Norway (69°31’29 N, 21°19’16 E; altitude 430–570 m a.s.l.). This is in a non-permafrost area with glacial tills as the dominant mineral soil type. Only weak podzolic profiles are developed and a continuous soil organic layer of approximately 4.25 cm is found above the mineral soil layer. The area is bisected by a reindeer pasture rotation fence, that has been in place at least since 1966, therby creating a legal border between the reindeer summer ranges and their migration range. The summer range side of the fence experiences heavy grazing (HG) intensity for a few weeks during annual reindeer migrations in August, whereas the migration-range is subjected to grazing for only a short period of time (referred to as light grazing intensity, LG).

The set-up consists of eight blocks bisecting the fence (5 × 10 m large, spaced 20 m apart). On each block to both sides of the fence (HG and LG), treatments of control (Ctrl), warming (W), fertilization (F) and combined warming and fertilization (WF) were performed on plots sized betweem 0.935 and 1 m². Additionally, plots under HG were divided into grazed (HG) and ungrazed (HGexc) subplots with short-term exclosures (height 0.9 m, mesh 40 × 40 mm) that were built each year (2010–2014) before reindeer migration and removed after the grazing event to avoid winter-time snow accumulation.

Warming was implemented with open-top chambers (OTCs, diam. 120 cm, height 40 cm) set to the plots after snow melt (early June) and kept in place until the arrival of reindeer (early August) to avoid OTCs affecting the feeding behaviour of reindeer. By this way, the warming treatment’s direct effect was limited to approximately two months, encompassing the early and peak growing season.

Fertilization was applied yearly, early in  the growing season, by dissolving ammonium nitrate (NH₄NO₃) equivalent to 10 g N m^-2 in one L of water and applying it to the plots. Due to logistical reasons and the high mean annual precipitation in the study area, the unfertilized plots were not watered. 

Ecosystem carbon stocks in 2014

Ecosystem carbon stocks were sampled after five years of the experiment, in 2014, before the annual grazing period at the study site. Therefore, these do not account for the immediate grazer-induced biomass removal. Vascular aboveground biomass was collected from an area of 250 cm² and sorted into growth forms. Bryophyte, lichen and litter biomass was hand-picked from cored ground layer samples (diam. 119 mm). Litter included both cryptogam and vascular litter. Soil and root biomass was collected with 3–4 soil cores (diam. 29 mm) underneath the litter layer until the corer hit large stones and separated later into organic and mineral soil layers. The soil layers were later homogenized separately (2 mm sieve), and the root biomass was obtained from the sieving residue after washing. All vegetative, litter and soil biomass was dried (60 °C, 70 h), weighed, milled (RETSCH 2000) and analysed for carbon and nitrogen contents (C-H-N Element Analyser EA1110, CE Instruments, SFS-EN 15104:2011).

Soil properties in 2014

From the sieved soil samples, we determined soil moisture (drying at 100 °C, 12 h), organic matter content (loss on ignition at 475 °C, 4 h) and bulk density (dw volume^-1). We also conducted soil extractions with 0.5 M K₂SO₄ for 2 h (within 48–72 h after sampling), and analyzed the inorganic soil ammonium (NH₄⁺-N) and nitrate (NO₃^–-N) concentrations with colorimetrical analyses (SFS 3032, Shimadzu UV-1700 spectrophotometer for NH₄⁺-N; SFS-EN ISO 133395CFA, Seal Analytical AA3 for NO3^–-N).

Fungal and bacterial copy numbers in 2013

Total genomic DNA for bacterial and fungal quantitative PCR (qPCR) was determined from plots under LG and HG in July 2013 from the same composite samples from which N concentrations were analyzed. The composite soil samples were sieved in the field and subsamples of  0.1–0.5 g were immediately frozen in liquid nitrogen for later analysis. The DNA was extracted using a modified phenol-chloroform-isoamylalcohol (PCI) protocol (Griffiths and Whiteley 2000) with 600 µL of CTAB buffer in a mixture of beads (0.7 g ceramic beads (1.0 mm), 0.3 g glass beads (0.1 mm) and two large glass beads (3.5 mm); BioSpec). The fungal ITS2 and bacterial 16S rRNA gene copies in the soil samples were quantified using a Bio-Rad CFX96 Real-time thermal cycler (Bio-Rad, Hercules, CA, USA). All qPCRs were run in technical triplicates of 20 µL containing 10 µL SSoFast EvaGreen qPCR Supermix (Bio-Rad) for bacterial and SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) for fungal reactions, 0.5 µL of forward and reverse primers (10 mM) (Eub341F and Eub534R)for bacteria and fITS7 and ITS4 for fungi), 7 µL ddH₂O and 2 µl template in a 100-fold dilution. The conditions were 98 °C for 2 min, followed by 40 cycles of 98 °C for 5 s, 56 °C for 20 s for bacterial qPCR and 98 °C for 3 min, followed by 40 cycles of 98 °C for 15 s, 61 °C for 60 s for fungal qPCR with a melt curve analysis as the final step. The standard curves were generated using genomic DNA from bacterial isolate Granulicella mallensis and fungal isolate Laccaria laccata.

Vegetation abundance (2010-2014)

Vegetation abundance was analyzed throughout the course of the experiment (2010–2014), with a modified point intercept method, where ten vertical pins are positioned 10 cm away from each other on 50 cm wide rows (eight rows on LG, four rows on HGexc and HG in 2010–2013, and two rows in LG, HGexc and HG in 2014) and where all hits per species are counted. The recordings were done in the first half of August every year. We normalized the number of hits to hits per 100 pins and report species abundances pooled into growth forms (evergreen shrubs, deciduous shrubs, graminoids, forbs, bryophytes and lichens).

Extracellular enzyme activity in June, July and August 2013

To determine soil  microbial activity, we collected composite soil samples (3–7 soil cores, diam. 25 mm) from the organic layer three times during the growing season in 2013 (June 19^th, July 16^th and August 6^th). We recorded depth of the organic layer, sieved the samples (mesh 2 mm) and stored these at +4 °C for further analyses. We analyzed the potential activities of β-glucosidase (BG), β-N-acetylglucosaminidase (NAG), leucine aminopeptidase (LAP), acid phosphatase (AP), amidohydrolase (commonly known as urease; U) and phenol oxidase (POX) within 5 days after sampling. We used the following chromogenic substrates: 5 mM paranitrophenyl(pNP)-β-glucopyranoside for BG, 2 mM pNP-β-N-acetylglucosaminide for NAG, 5 mM leucine p-nitroanilide for LAP, 5 mM pNP-phosphate for AP, 30 mM urea for U, and 50 mM pyrogallol for POX and conducted the assays in sodium acetate buffer (50 mM, pH 5.0 that corresponds to the study site’s mean soil pH, 5.13). The soil-substrate aliquots were incubated at room temperature, five microliters of 1.0 M NaOH was added to AP, BG and NAG before the supernatants were measured spectrophotometrically for their potential activities (405 nm for BG, NAG, LAP and AP, 450 for POX, Multiscan FC microplate reader, Thermo Scientific). U activity was verified by measuring the formation of NH₄⁺ after 5 h. The 100 µL aliquots of supernatant were incubated with 10 µL of sodium citrate, phenol nitroprusside and hypochlorite (19 °C, 1 h), and absorbance was measured at 620 nm (Multiscan FC microplate reader, Thermo Scientific). The absorbances of homogenate and substrate controls were subtracted from the assay absorbance. Extinction coefficients for calculating potential activities were obtained based on standard curves for paranitrophenol (BG, NAG, AP), paranitroaniline (LAP) and NH₄Cl (U), and oxidation of pyrogallol by mushroom tyrosinase (POX). All potential activities were counted as μmol h^–1 g SOM^–1.