The site is located at the boundary of the Donner Ridge fire in the Sagehen watershed near Truckee, California. The forested site may have experienced fire in the past, but was not burned in the Donner Ridge fire. It was located just inside the boundary approximately at the “unburned, hill” site identified by Johnson and Needman [18], and the burned site was located approximately 100 m to the east, within the former fire. The elevation of the site is 1944 m; mean annual temperature is 5.2 °C and mean annual precipitation is 850 mm, 80% of which falls as snow. Vegetation in the forest site consisted of an overstory of Jeffrey pine (Pinus jeffreyii Grev. and Balf.) with occasional white fir (Abies concolor Godr. and Glend. Lindl.) and lodgepole pine (Pinus contorta Dougl). Understory in the forest site was dominated by prostrate ceanothus (Ceanothus prostratus Benth.) and greenleaf manzanita (Arcotstaphylos patula Greene) with occasional snowbrush (Ceanothus velutinus Douglas ex Hook.) minor amounts of bush chinquapin (Chrysolepis sempervirens (Kellog) Hjelmqvist), creeping snowberry (Symphoricarpos mollis Nutt.), wax currant (Ribes cereum Douglas), bitter cherry (Prunus emarginata (Douglas ex Hook.) D. Dietr.), Utah serviceberry (Amelanchier utahensis Koehne), wooly mule’s ears (Wyethia mollis A. Gray), yellow rabbit brush (Chrysothamnus viscidiflorus (Hook.) Nutt.), rose thistle (Cirsium andersonii (A. Gray) Petr.), bottlebrush squirreltail (Elymus elymoides (Raf.) Swezey), Holboell’s rockcress (Arabis holboellii Hornem.), antelope bitterbrush (Purshia tridentate (Pursh) DC.), and Sandberg bluegrass (Poa seconda J. Presl). Overstory on the burned site consisted of widely scattered Jeffrey pine, with occasional white fir and lodgepole pine. Understory in the former fire was dominated by snowbrush and manzanita with minor amounts of snowberry, bitter cherry, serviceberry, mule’s ears, wax currant, chinquapin, squirrel tail, rabbit brush, Anderson’s thistle, and Sandberg’s bluegrass. Soils in both areas are the Jorge series, Loamy-skeletal, isotic, frigid Andic Haploxeralfs derived from andesite. According to Johnson and Needham [18], vegetation in the burned area before the fire was very similar to the forested site in this study: an overstory of white fir, ponderosa pine (Pinus ponderosa Dougl.), Jeffrey pine and lodgepole pine (Pinus contorta Dougl), and understory consisted of snowbrush, prostrate ceanothus, greenleaf manzanita, rabbit brush, wooly mule’s ears, and wax currant.

The Little Valley fire site, the results of which have been previously published [17], is located approximately 30 km southwest of Reno, Nevada in the eastern Sierra Nevada Mountains. Elevation in the valley ranges from 2010 to 2380 m, and elevation at the site is 2010 m. The climate is characterized by warm, dry summers and cold winters; and the major hydrologic event is snowmelt. Mean annual air temperature near the valley floor is 5 °C and mean annual precipitation is 550 mm, approximately 50% of which falls as snow. Vegetation at the study site is dominated by 110–130 year old jeffrey pine (Pinus jeffreyii Grev. and Balf.). Soils are the Corbett series, sandy, fine-loamy, mixed frigid Typic Xeropsamments derived from colluvium of decomposed granite.

The entire valley, as well as much of the eastern Sierra Nevada mountains, was logged and probably burned during the Comstock mining era between 1870 and 1890. An area just east of the valley, which had consisted of 100-year-old jeffrey pine forest, burned in a stand-replacing wildfire in 1981. As is usually the case, the wildfire did not consume much of the large standing woody tissues (tree boles and large branches); and thus the area was salvage logged for merchantable timber (snags) the following year. Since that time, the burned area has been dominated by snowbush (Ceanothus velutinus Dougl.), a species that often invades after fire and fixes N [16,19,20] with lesser amounts of manzanita (Arcostaphylos patula Greene) and spotty regeneration of Jeffrey pine planted in 1985.

In June–July 2006 (46 years after the fire), forest floor and soils from the sites were sampled in three 0.05 ha circular plots within each area (former fire and forest) with six randomly-located replicates (ranging from approximately 2 to 7 m from one another) taken within each plot. The forest floor was sampled by horizon (Oi, Oe, Oa, and other, the latter being a category of miscellaneous composition of bark, reproductive parts, and woody material less than 2.5 cm in diameter) within a 698 cm² ring. Once the forest floor was removed, a surface set of mineral soil samples was taken at each point with a trowel to a depth of 5 cm, taking care to sample depth-proportionally such that an even volume of soil was taken with depth over the 5 cm increment. In a separate location, soil samples were taken in 15 cm intervals to a depth of 90 cm using the rotary coring device described by Rau et al. [22]. The method utilizes a 7.62 cm internal-diameter diamond-tipped core device manufactured by Diteq™, and is driven by a two-person rotary Briggs and Stratton™ power head, allowing it to core through rocks and soil with minimal compaction. Each sample increment was extracted before the core was driven to the next depth increment. All soil samples from surface and cores were bagged individually, brought back to the lab, dried at 100 °C for 48 hours, and weighed. Cores were then sieved to 2 mm, and bulk density and % > 2mm were calculated for each sample.

O horizon samples were ground using a Wiley Mill for total nutrient analysis. Subsamples of forest floor components were analyzed for total N, P, K, Ca, Mg, and S at A&L Agricultural Laboratories, Modesto, CA. Total N was analyzed at A&L by Kjeldahl digestion followed by colorimetric ammonium analysis. Phosphorus, K, Ca, Mg, and S in vegetation and litter were analyzed using a Jarrell Ash ion coupled plasma spectrophotometer (ICP; Thermo Jarrell Ash Corp., Franklin, MA, USA) after microwave digestion (Method 985.01, Association of Official Analytical Chemists) in a nitric acid hydrogen/peroxide mixture. Total C and N were analyzed using a dry combustion C and N analyzer (LECO, St. Joseph, MI, USA) at the Oklahoma State University Soil, Water, and Forage Analytical Laboratory (Stillwater, OK, USA).

The <2 mm fraction of all soils were analyzed for total C, total N, NH₄⁺ and NO₃⁻ at Oklahoma State University Soil, Water, and Forage Analytical Laboratory (Stillwater, OK, USA). Total C and N were analyzed using a dry combustion C and N analyzer (LECO, St. Joesph, MI, USA). For soil NH₄⁺ and NO₃⁻, ten g of soil was shaken with 20 mL of 2 M KCl for 30 minutes then filtered and analyzed using a Quick Chem 800 flow-injection analyzer (Lachat, Milwaukee, WI, USA). Soil subsamples were then sent to A&L Western Agricultural Laboratories (Modesto, CA, USA) for analysis of the remaining nutrients. Samples were analyzed for Bray extractable P (2 g soil in 0.5 M HCl plus 1 M NH₄F) and bicarbonate-extractable P (2 g soil in 50 mL 0.05 M NaHCO₃^-) using a Jarrell Ash ion-coupled plasma spectrophotometer (ICP; Thermo Jarrell Ash Corp., Franklin, MA, USA) and SO₄²⁻, exchangeable Ca²⁺, K⁺, and Mg²⁺ (10 g soil in 50 mL 1 N ammonium acetate) using ICP analyses. Subsamples of soils ground into a powder and weighed for analysis of ¹⁵N. The weighed samples were analyzed at the University of California, Davis, using a Europa Scientific ‘Integra’ analyzer.

Six replicate ceramic cup falling head lysimeters (Soil Moisture Inc) were installed to a depth of 20 cm in each plot to collect soil solution. Tension was set to 0.4 kPa and solutions were collected in once in April 2007 and analyzed for NH₄⁺, NO₃⁻, SO₄²⁻, and PO₄³⁻ using high performance ion exchange chromatography (Dionex) at the Oklahoma State University Soil, Water and Forage Analytical Laboratory. Because of the relatively dry year, it was only possible to collect soil solution one time.

Six replicate resin lysimeters consisting of Rexyn^® 300 (H–OH) ion exchange resin according to the design developed by Susfalk and Johnson (2002) were installed at a depth of 15 cm in each plot in the summer of 2006. Resin lysimeters were removed in the spring of 2007. Resins were recovered from the lysimeters and extracted by shaking in 100 mL of 1 M KCl for one hour and analyzed with either a Quick Chem 800 flow injection auto analyzer (Lachat, Milwaukee, WI, USA) or an Alpkem segmented flow auto analyzer (Pulse Instrumentation Ltd., Saskatoon, SK, Canada) for ortho-P, NH₄⁺, and NO₃⁻. Samples were analyzed at the Oklahoma State University Soil, Water, and Forage Analytical Laboratory (Stillwater, OK, USA).

In October 2011, six randomly-located 0.04 ha plots were established in each site (burned and unburned) for vegetation and fuels inventory. All trees of pole size and larger, i.e., ≥10.2 cm DBH, within these plots were measured for DBH and tallied by species. The DBH values were then used to derive above-ground tree biomass by plot according to the formulas of Gholz et al. [23].

Downed and dead fuels inventories for the 100-h and 1000-h timelag categories [24] were also conducted. For 100-h (>2.5 to ≤7.6 cm diameter) fuels, a single 4-m² plot was established with the same plot center as that of each of the 0.04-ha plots used for overstory measurements, and collection of these fuels permitted a weight determination by direct measurement of samples dried to a constant weight. For 1000-h (>7.6 cm diameter) fuels, a single 54-m² circular plot was established, also with the same plot center as that of each 0.04-ha plot, and the lengths and diameters at the midpoint of each segment were measured for use in calculating an estimate of volume according to the Huber formula [25]. Collection of 10 log sections from random locations, measuring their dimensions, and then drying and weighing them provided a density constant for conversion of volume to dry weight by plot.

The 54-m² circular plots used for measuring 1000-h fuels were also used for mapping of shrub and herbaceous understory species, which permitted expression of their prevalence on a percent ground cover basis. In order to express their prevalence on a dry weight basis as well, 10 samples of known ground cover area were collected of each species from random locations, dried to a constant weight, and weighed. For shrub species, each sample consisted of all tissues occupying a ground area of 0.093 m², while 0.01 m² was used for herbaceous species. These species-specific weight constants permitted the conversion of percent cover to dry weight within each plot. In addition to the determination of the cover and weights by species, those of all species in total were determined.

Seedling (≤1.37 m tall) and sapling (>1.37 m tall and ≤10.1 cm DBH) inventories were also conducted which encompassed counts by species within the two size classes. Seedling counts were performed using 40-m² circular plots established with the same centers as those for the 0.04-ha plots used in the overstory measurements, while the sapling counts relied upon the aforementioned 54-m² plots involved with the 1000-h fuel and understory vegetation measurements.

Tree nutrient content was estimated from biomass estimates by component (foliage, branch, bole bark, bole wood) combined with previous analyses of nutrient concentrations within these tissues from nearby sites [12,17]. Understory nutrient content was estimated from biomass estimates combined with nutrient analyses of samples taken from the sites as described above for O horizon samples.

Between 1997 and 2001 (approximately two decades after the fire), we established four 0.1 ha plots in the shrub area (former fire) and three in the adjacent forested with the same soils, slope and aspect [17]. Within each shrub plot, six randomly established 1 m² subplots were established for measurement of current (1997 and 2001) aboveground shrub, litter and soil C and nutrient pools. Within each subplot, all aboveground vegetation was removed, sorted by species and component (foliage and woody), weighed, and subsampled for moisture content. After the vegetation was removed, all litter within a 15 cm diameter ring was removed by horizon (Oi, Oe, and Oa) and transported to the laboratory for drying, weighing, and analyses. After litter was removed, the soil was sampled at three depths corresponding to the A (0–7 cm) and AB (7–20 cm) and BC (20–40 cm) horizons. Bulk density was determined by the core method for each horizon at three of the sampling points in each plot. Coarse fragment contents were negligible. Large woody debris mass was estimated in each plot by measuring the length and diameter of each piece of debris within the plot, calculating the volume and applying density measurements from subsamples.

Within the forested plots, all trees were measured at dbh and tree biomass and nutrient contents were estimated as described above. Large woody debris, O horizon, and soil mass and nutrient contents within the forested plots were estimated as described above. Understory in the forested plots was negligible.

Subsamples of vegetation and litter components were taken from the site and analyzed for total C and N using a Perkin-Elmer 2400 CHN analyzer and for P, K, Ca, K, Mg, and S at A&L Agricultural Laboratories, Modesto, CA. At A&L, P, K, Ca, K, Mg, and S in vegetation and litter were analyzed by inductively-coupled plasma emission spectroscopy (ICP) after microwave digestion using a nitric acid hydrogen peroxide digestion mixture. Soils were analyzed at A&L for exchangeable cations (10 g soil in 50 mL 1 N ammonium acetate followed by ICP analyses for Ca²⁺, Mg²⁺, K⁺, and Na⁺), extractable P analyzed by the bicarbonate (2 g soil in 50 mL 0.05 M NaHCO₃) and the weak Bray method (0.5 M HCl plus 1 M NH₄F), and extractable SO₄²⁻ (by ICP analysis on the ammonium acetate extraction).

Statistical analyses at the Sagehen site were conducted using Microsoft Excel for Student’s t-tests for treatment effects (former fire and forest, n = 3 each) in each site and for each soil horizon (n = 6 per plot, for 18 samples each in forest and former fire), and linear models in DataDesk for comparisons (treatment and soil depth by site). Statistical analyses on soil concentrations were conducted on unbulked samples, but statistical analyses on O horizon and soil contents were on samples bulked by plot. The data showed significant skew so all data were log-transformed to achieve approximate normal distributions prior to statistical analyses. Statistical analyses at the Little Valley site were performed using student’s t-tests in Microsoft Excel software [17].

As the reviewers of this paper have strongly pointed out, the experimental design of both of these studies can be criticized as invalid because of the often-cited issue of pseudo-replication [26]. Unfortunately, it simply is not possible to avoid such a criticism in wildfire studies because it is not feasible to replicate wildfires in a statistically rigorous fashion. Nor, for that matter, is it possible to replicate planet earth in a statistically rigorous fashion for global C studies, for example. It should also be noted that we did not know in advance that these wildfires would take place, and therefore we were not able to obtain pre-treatment samples. Such is usually the case with wildfires, with one nearby notable but very serendipitous exception noted [27]. The authors feel that while such a criticism can be leveled in each case, it does not warrant the abandonment of such studies and thus we have forged ahead to present this data, with the caveat of pseudo-replication and lack of pre-treatment data duly noted. We also note the opposing view of Oksanen [28] who opines that “The concept of ‘pseudoreplication’ amounts to entirely unwarranted stigmatization of a reasonable way to test predictions referring to large-scale systems.”

Site (former fire versus forest) was significant for C:N ratio, ¹⁵N, Bray and bicarbonate P, and extractable SO₄²⁻, all of which were lower in the former fire than in the forest (Figure 1, Figure 2, Figure 3 and Table 1 and Table 2). Site was also significant for NH₄⁺, exchangeable Ca²⁺ and Mg²⁺, all of which were higher in the former fire (Figure 1 and Figure 3; Table 2). Site was not significant for bulk density, % > 2 mm, total C, total N, or exchangeable K⁺. Depth was significant for bulk density and all measured nutrients except exchangeable Ca²⁺ and extractable SO₄²⁻. Carbon, N, C:N ratio NH₄⁺, NO₃⁻, mineral N, and Bray P concentrations decreased substantially with soil depth whereas ¹⁵N and Mg²⁺ increased with depth.

Bicarbonate P, SO₄²⁻, and Ca²⁺ showed no particular pattern with depth. The site × depth interaction was significant only for extractable SO₄²⁻ (Figure 1, Figure 2, Figure 3 and Table 2).

Although site was not significant for total C concentrations overall, the forest site generally had higher concentrations than the former fire in the surface depths (significant at the 15–30 and 30–45 cm depths) (Figure 1). Similarly, although site was not significant for total N overall, N was slightly but significantly greater in the forested than in the former fire site at the 30–45 cm depth. Mineral N was greater in the former fire site at most depths but significantly so only in the 0–15 cm depth. ¹⁵N was lower in the former fire at all depths and significantly so at the 0–5, 0–15, 60–75 and 75–90 cm depths. C:N ratio was consistently and significantly lower in the former fire at all depths up to 45 cm (Figure 2). The largest and most consistent differences between the sites were for Bray and bicarbonate P (Figure 2) Bray P was lower in the former fire at all depths and significantly so in all but the 0–5 and 15–30 cm depths. Bicarbonate P was also lower in the former fire and significantly so in all but the 75–90 cm depth. Differences in K⁺ were inconsistent and significant only at the 30–45 cm depth where the forested site was slightly higher (Figure 3). Calcium was greater in the former fire site at most depths, and significantly so at the 0–15, 30–45, and 45–60 cm depths. Similarly, Mg²⁺ was greater in the former fire site at all but the 75–90 cm depth and significantly so in the 0–15 and 30–45 cm depths. Sulfate was greater in the forested site at most depths and significantly so at the 0–15, 30–45, 60–75, and 75–90 cm depths (Figure 3).

There were no significant differences in O horizon C, N, K, Ca, Mg or S, but O horizon P contents were greater in the forest than in the former fire, consistent with the soil concentration and content data, which showed greater Bray and bicarbonate P contents in the former fire (Figure 2 and Table 3). Both Bray and bicarbonate P contents were greater in the forest than in the former fire, but only the Bray P differences were statistically significant. Soils in the former fire contained significantly greater total N, exchangeable Ca²⁺, and exchangeable Mg²⁺ than soils in the forest. Similarly, the former fire contained significantly greater total (O horizon + mineral soil) contents of N, Ca, and Mg than the forest area whereas the forest area contained greater total P content (Bray and bicarbonate).

Table 4 shows NH₄⁺, NO₃⁻, mineral N, ortho-P, and SO₄²⁻ concentrations in soil solutions collected by ceramic cup lysimeters in the Spring of 2007 (during snowmelt) and Table 5 shows the fluxes of NH₄⁺, NO₃⁻, mineral N, and ortho-P leaching rates for the 2006–2007 water year as measured by resin lysimeters at the two sites. Neither data set indicates a significant difference in the leaching of NH₄⁺, NO₃⁻, or mineral N, and the ceramic cup data shows no difference in SO₄²⁻. (It is not possible to analyze resin lysimeters for SO₄²⁻). The ceramic cup data indicates a significantly lower leaching rate for ortho-P in the former fire site, which is consistent with the differences in soil and O horizon data which is also shown lower values in the former fire site (Figure 2 and Table 3). The resin lysimeter data do not show any difference in ortho-P leaching rates, however.