A detailed methods supplement is available online (link here).

Study area: The Anaktuvuk River Fire scar is located on the North Slope of the Brooks Range, Alaska, USA , approximately 23 km NW of Toolik Field Station (68.5833 oN, 149.7167 oW). It is bounded on the east by the Nanushuk River and on the west by the Itkillik River, extending approximately 65 km from North to South (NW corner, 69.4273 oN , 151.0619 oW; NE corner, 69.4274 oN, 150.6980 oW; SW corner, 68.8637 oN, 150.5114 oW; SE corner, 68.9122 oN, 150.1311 oW). This region is underlain by permafrost. Mean annual temperature is approximately -10° C and mean annual precipitation is 30 cm. Long-term average growing season temperature (1971-2000) is about 10° C and tends to be somewhat warmer (1-2° C) at lower elevations, which range from 130 m asl at the north end of the burn to 450 m asl at the south end.
 
In July of 2007, the Anaktuvuk River Fire was started by lightning and sustained by growing season temperatures that were the warmest recorded over the 19 year record and precipitation that was only 25% of the long term average33. Warm and dry conditions were maintained by anomalous high pressure over the North Slope that appeared to be related to summer sea ice recession34. By the time snowfall extinguished the fire in early October, it had burned 1,039 km2 of arctic tundra, doubling the cumulative area burned in this region over the past 50 years . This fire was an order of magnitude larger than the average fire size in the historic record for the North Slope and remotely sensed indices of severity were substantially higher than for other recorded tundra burns33.
 
Prior to the fire, 54% of the vegetated area within the burn perimeter was classified as upland moist acidic tundra (MAT; soil pH <5.5), 15% as moist non-acidic tundra (soil pH>5.5), and 30% as shrubland35. MAT is pan-Arctic in distribution and covers as much as 336 x 106 km2 of the tundra biome36. We focused our study on MAT because of its widespread distribution and because the surviving growing points of the dominant plant species, Eriophorum vaginatum L., provided a benchmark of pre-fire soil organic matter depth and plant biomass. From this, we developed a method (described below) for estimating soil and plant organic matter consumption during fire using techniques similar to those described for boreal spruce forests37. Upland tundra soils contain, on average, 7,500 g C m-2 in the combustible organic horizon, but this varies substantially across the landscape, ranging from 100 to 63,000 g C m-2 (ref. 38). Mineral soils contain approximately 18,000 g C m-2 above the permafrost and a similar amount from the surface of the permafrost to 1 m depth38. Relict glacial ice wedges and lenses are common across the North Slope, as is high ice volume in near-surface permafrost39.
 
Field sites: Twenty MAT sites within the burn were accessed via helicopter from either Umiat or Toolik Lake in July and August of 2008. Burn severity was mapped using the differenced Normalized burn Ratio (dNBR) method40 and sites 1-12 were randomly chosen to represent the range and frequency of dNBR values as described in Jones et al. (2009). The remaining sites, 13-20, were chosen randomly along hillslope transects from areas where collaborators had previously established eddy covariance towers and lake/stream monitoring.
 
To obtain empirical relationships between ecosystem structure and the element pools necessary for reconstruction of pre-fire soil C and N pools, 10 unburned MAT sites were also sampled Two unburned sites were adventitiously encountered within the burn perimeter, an additional site was located near the 2007 Kuparuk Fire (69.2974 °N, 150.3221 °W), and seven others were systematically selected along the Dalton Highway. The latter sites were randomly selected from a GIS database that included all MAT areas along the Dalton within the elevation and climate range of the burn scar and were allocated to span the same elevations as the burn scar. All sites were >300 m from the road to minimize the effects of dust and other disturbances41,42.
 
Tussock morphology: Our method for reconstructing pre-fire soil organic matter pools in burned sites was based on the relationship between the morphology of E. vaginatum tussocks and SOL depth. Tussocks are a columnar, caespitose growth form. At the top, many layers old leaf sheaths tightly clasp vertically growing short rhizomes (unthickened corms) that bear the tussock's living leaves and have apical and axillary growing points. These rhizomes are nestled in what we term the crown of the plant, a dense, compact mass of old leaf sheaths (upper arrow, Figure 4-A). The tussock's growing points are located 4.39 ± 0.26 cm (mean ± 1 SE, n=12 individual tussocks) below the crown's surface . The crown's dense, usually moist, mass of leaf sheaths resists burning and protects the growing points from heat damage during the fire, enabling rapid re-sprouting of leaves after fire . In unburned tundra, tussock crowns are located somewhat above the surface of the inter-tussock green moss (see below), and new leaf blades are further elevated above the moss layer . Beneath the crown, a deep, fibrous network of sheath-enclosed, proximal portions of corms tangles with a network of live and dead roots, to form a column that extends down through the organic soil layer into the mineral soil. Eriophorum vaginatum is one of the few MAT species whose roots grow down into the mineral soil43. The tussock columns, topped with re-sprouting leaves, were a striking feature of the post-fire landscape .
 
Field sampling—unburned sites: We quantified the relationship between E. vaginatum crowns and SOL depth or biomass, and characterized SOL bulk density and element concentration across the 10 unburned sites. Along a 50 m transect in each site, we measured the depth of thaw and SOL at 5 m intervals (random point) and directly adjacent to the tussock nearest to the random point (n=10 random and 10 tussock points). We measured at both random and tussock points to determine whether the relationship between soil organic matter depth and tussock crowns was related to distance to tussock. Thaw depth was measured by inserting a metal rod into the soil until it hit ice or rock (differentiated by the sound and texture of the hit), marking the surface of the green moss on the rod, removing it, and measuring the distance from the tip to the mark with a meter stick. Soil organic layer depth was measured by slicing a square pit with a serrated knife, removing a monolith of organic soil, exposing the surface of the mineral soil, and measuring the distance from the surface of the green moss to mineral soil on two sides of the pit. The two depth measurements per pit were averaged to yield one SOL depth measurement per point.
 
At each point where SOL depth was measured, we used two meter sticks attached at a sliding right angle and fitted with a tubular spirit level to measure the depth of the green moss below a plane parallel to the crown of the nearest E. vaginatum tussock . Use of the level ensured that the right angle was parallel to the crown and orthogonal to the ground. For the randomly located sample point, we also measured the distance to the nearest tussock using this apparatus. We measured the distance to, crown diameter of, and survivorship of the three next closest tussocks and used a nearest-neighbor method to estimate tussock density44. For each tussock, two crown diameter measurements were made (and averaged) at right angles by compressing the leaves and manually locating the sides of the crown.
 
To determine soil bulk density and element concentration, organic soil horizons were sampled volumetrically with a serrated knife. At 10 m intervals, a pit was dug and a 10 x 20 cm soil monolith was excised from the side of the pit, extending from the surface of the green moss to the surface of the mineral soil (roughly 5-30 cm depth depending on location). This monolith was wrapped in tinfoil to preserve structure, returned to the field station and frozen prior to shipping to the University of Florida (UF) for analyses of bulk density, moisture, C and N concentration, and C isotopes. All aboveground plant material attached to the surface of the soil monolith was included in the sampling. Tussocks were also harvested at 15 m intervals to develop allometric relationships between tussock diameter and combustible biomass (see below). Biomass was shaved from the live tussock with a serrated knife and returned to the field station, where it was dried at 70° C for 48 hours before weighing and shipping to UF for analyses of C and N concentration.
 
Field sampling—burned sites: Measurements in burned sites were similar to those in unburned sites except that measurements were made on the surface of the residual burned organic layer rather than the surface of the green moss , and tussock leaves were not sampled.
 
Laboratory analyses: Approximately 155 soil monoliths comprising ~1000 individual 5 cm increment soil samples were collected in total from the 20 burned and 10 unburned sites. In the lab, each monolith was bisected depth-wise with an electric carving knife. One half of the monolith was processed for radiocarbon measurements, as described below, and re-frozen for archival purposes. In the remaining half, green moss and dwarf shrubs were sliced off and the remainder of the core was sliced into 5 cm depth intervals with the last sample of variable depth depending on the location of the organic/mineral interface. Samples were homogenized by hand and coarse organic materials (>2.5 cm twigs and roots) and rocks were removed. Coarse and fine organic fractions were weighed wet, dried at 70° C for 48 hours to determine dry matter content, then ground on a Wiley mill with a 40 mm sieve. Carbon and N content was measured on a Costech Elemental Analyzer (Costech Analytical, Los Angeles, California, USA) calibrated with the NIST peach leaves standard (SRM 1547, National Institute of Standards and Technology, Gaithersburg, MD, USA). The volume of each monolith layer was calculated as depth times area minus the volume of rocks. Bulk density, C or N pools were calculated for both fine and coarse organic fractions.
 
Estimating pre- and post-fire organic matter pools and losses: Post-fire SOL C and N pools were calculated for each 5 cm depth increment of residual organic soil at each sampling point in the burned sites (Supplementary Table 2 ). Total residual soil profile depth of randomly located samples did not differ significantly from those measured adjacent to tussocks (paired-t=0.68, P=0.50, n=20 sites), so we used the randomly located sampling points for our calculations. Coarse and fine fraction pools were summed at each depth and depths were summed across the profile to estimate total SOL element pools on a per site basis. The only aboveground vegetation encountered was from re-sprouting individuals and thus was not included in the post-fire pool estimates. No green moss was found on the surface of any of the soil monoliths, emphasizing the homogeneity of surface burning in this fire.

Estimates of depth and element loss on a per site basis required reconstruction of pre-fire depth or element pools for each site. Reconstruction of pre-fire SOL depth in burned sites was based on the observation that E. vaginatum crowns tended to be at the same height as the SOL surface in unburned tundra. We used the transect measurements in the unburned reference sites (described above) to derive a predictive relationship between tussock crown height (TCH) above the mineral soil surface and SOL depth, here the distance between the top of the green moss and the mineral soil surface. Soil organic layer depth measurements did not differ detectably between random and tussock points (paired-t=0.21, P=0.70, n=10 sites), so we again used the randomly located points for our analyses. Tussock crown height was highly positively correlated with SOL depth (least squares regression, SOL depth (cm) = -1.126 + ln (TCH) x 1.242, R2=0.94, P<0.001, n=10 sites;. Depths calculated with least squares regression differed from geometric mean regression by <1%, so we used the former with burned site TCH and residual soil depth measurements to estimate pre-fire SOL depth. Pre-fire SOL C and N pools were then calculated for each 5 cm depth increment using the average unburned bulk density and C or N concentration for that layer . Finally, combustion C or N losses from the SOL were calculated as the pre-fire C or N pool minus the post-fire C or N pool of any remaining organic soil.
 
Pre-fire biomass and element pools in green mosses, lichens, shrubs, forbs and graminoids were included in SOL loss estimates above because they were sampled quantitatively with the soil monoliths. To estimate pre-fire combustible tussock biomass (CTB), we first derived a predictive relationship between CTB and tussock diameter (TD) in the unburned sites (ln(CTB (dry gtussock-1)) = 3.097 + TD2 x 0.003, R2=0.67, P<0.001, n=35 tussocks). We used this equation to predict combustible biomass per tussock in the burned sites, which was then multiplied by density to scale to biomass per m2. The latter was multiplied by the average C (43 ± 1, mean ± SE) or N (0.78 ± 0.05) concentration of CTB across the unburned sites to calculate the combusted C or N pool for the tussocks. All statistical analyses were performed in SYSTAT 11.00.01 (SYSTAT software, Inc., Chicago, IL).
 
 
Literature Cited
33Jones, B. M., Kolden, C. A., Jandt, R., Abatzoglou, J. T., Urban, F., and Arp, C. D., Fire behavior, weather, and burn severity of the 2007 Anaktuvuk river tundra fire, North Slope, Alaska. Arctic Antarctic and Alpine Research 41 (3), 309-316 (2009).

34Hu, F. S., Higuera, P. E., Walsh, J. E., Chapman, W. L., Duffy, P. A., Brubaker, L. B. et al., Tundra burning in Alaska: Linkages to climatic change and sea ice retreat. Journal of Geophysical Research-Biogeosciences 115, - (2010).

35Auerbach, N. A., Walker, D. A., and Bockheim, J. G., Landcover map of the Kaparuk River basin, Alaska (Alaska Geobotany Center, Fairbanks, AK 1997).

36Walker, D. A., Raynolds, M. K., Daniels, F. J. A., Einarsson, E., Elvebakk, A., Gould, W. A. et al., The circumpolar Arctic vegetation map. Journal of Vegetation Science 16 (3), 267-282 (2005).

37Boby, L. A., Schuur, E. A. G., Mack, M. C., Johnstone, J. F., and Verbyla, D. L., Quantifying fire severity, carbon and nitrogen emissions in Alaska's boreal forests. Ecological Applications 26 (6), 1633-1647 (2010).

38Ping, C. L., Michaelson, G. J., Jorgenson, M. T., Kimble, J. M., Epstein, H., Romanovsky, V. E. et al., High stocks of soil organic carbon in the North American Arctic region. Nature Geoscience 1 (9), 615-619 (2008).

39Jorgenson, M. T. and Osterkamp, T. E., Response of boreal ecosystems to varying modes of permafrost degradation. Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 35 (9), 2100-2111 (2005).

40Key, C. H. and Benson, N. C., in Firemon: Fire effects monitoring and inventory system, edited by D. C. Lutes, R. E. Keane, J. F. Caratti et al. (USDA Forest Service, Rocky Mountain Monitoring and Inventory System, Ogden, UT, 2005), pp. 25-36.

41Walker, D. A. and Everett, K. R., Road dust and its environmental impact on Alaskan taiga and tundra. Arctic and Alpine Research 19, 479-489 (1987).

42Myers-Smith, I. H., Arnesen, B. K., Thompson, R. M., and Chapin, F. S., Cumulative impacts on Alaskan arctic tundra of a quarter century of road dust. Ecoscience 13 (4), 503-510 (2006).

43Kummerow, J., Ellis, B. A., Kummerow, S., and Chapin, F. S., Iii, Spring growth of shoots and roots in shrubs of an Alaskan muskeg. American Journal of Botany 70, 1509-1515 (1983).

44West, P. W., Tree and forest measurement, 2nd ed. (Springer-Verlag, Heidleberg, 2009). 
 

Not yet available; online supplement for Nature paper