Soil Nitrogen Dynamics Three Years After a Severe Araucaria

Austral Ecology (2012) 37, 153–163

Soil nitrogen dynamics three years after a severe

Araucaria–Nothofagus forest ﬁreaec_2258 153..163

Y. RIVAS,1 D. HUYGENS,2,5* H. KNICKER,6 R. GODOY,3 F. MATUS4,7 AND P. BOECKX5 1Facultad de Ciencias Forestales, Universidad Austral de Chile, 2Instituto de Ingeniería Agraria y Suelos, Facultad de Ciencias Agrarias, Universidad Austral de Chile, Casilla 567,Valdivia,Chile (Email: [email protected]), 3Instituto de Botánica, Facultad de Ciencias, Universidad Austral de Chile, 4Departamento de Ciencias Químicas, Universidad de La Frontera,Temuco, Chile; 5Laboratory of Applied Physical Chemistry-ISOFYS, Ghent University, Gent, Belgium; 6Centro del Consejo Superior de Investigaciones Cientíﬁcas en Sevilla, Instituto de Recursos Naturales and Agrobiología, Sevilla, Spain; and 7Agriculture and Agri-Food Canada, Ottawa, Canada

Abstract Wildfires have shaped the biogeography of south Chilean Araucaria–Nothofagus rainforest vegetation patterns, but their impact on soil properties and associated nutrient cycling remains unclear. Nitrogen (N) availability shows a site-specific response to wildfire events indicating the need for an increased understanding of underlying mechanisms that drive changes in soil N cycling. In this study, we selected unburned and burned sites in a large area of the National Park Tolhuaca that was affected by a stand-replacing wildfire in February 2002. We conducted net N cycling flux measurements (net ammonification, net nitrification and net N mineralization assays) on soils sampled 3 years after fire. In addition, samples were physically fractionated and natural abundance of C and N, and 13C-NMR analyses were performed. Results indicated that standing inorganic N pools were greater in the burned soil, but that no main differences in net N cycling fluxes were observed between unburned and burned sites. In both sites, net ammonification and net nitrification fluxes were low or negative, indicating N immobilization. Multiple linear regression analyses indicated that soil N cycling could largely be explained by two parameters: light fraction (LF) soil organic matter N content and aromatic Chemical Oxidation Resistant Carbon (CORECarom), a + relative measure for char. The LF fraction, a strong NH4 sink, decreased as a result of fire, while CORECarom - increased in the burned soil profile and stimulated NO3 production.The absence of increased total net nitrification might relate to a decrease in heterotrophic nitrification after wildfire. We conclude that (i) wildfire induced a shift in N transformation pathways, but not in total net N mineralization, and (ii) stable isotope measurements are a useful tool to assess post-fire soil organic matter dynamics.

Key words: 13C-NMR, Andisol, Chile, N cycling, organic matter fractionation, stable isotope.

INTRODUCTION its effect on soil organic matter (SOM) dynamics and associated N cycling remains largely elusive. Even so, The southern cone of America represents a worldwide N cycling responses to wildfire have a long-term influ- important reserve of temperate rainforests (Myers ence on main ecosystem functions, including N leach- et al. 2000; Armesto et al. 2009). Remnants of ing to surrounding watersheds (Murphy et al. 2006) Araucaria–Nothofagus forests are of special ecological and plant regeneration patterns (Catovsky & Bazzaz and cultural value (Herrmann 2005). This forest type 2002). is distributed from 37°20′ to 40°20′S, at mid to high Site-specific characteristics of vegetation, SOM altitudes of south Chilean and Argentinean mountain properties, fire intensity and duration, and microbial ranges, and develops on rocky, strongly nutrient community composition changes may exert a signifi- limited sites (Herrmann 2006). The latter forest land- cant influence on nutrient availability after wildfire scape has been shaped by a mixed-severity fire regime (Knicker et al. 2005; Smithwick et al. 2005b; Knicker that includes stand-replacing, surface and crown 2007; Turner et al. 2007). Data of soil N cycling after fires (González et al. 2005). While the response of stand-replacing forest fires are scarce, especially for the Araucaria–Nothofagus vegetation patterns to wildfire is first years after the fire event.The few available studies well documented (Burns 1993; González et al. 2010), have documented increases (Grogan et al. 2000), decreases (Turner et al. 2007) and non-significant *Corresponding author. changes (Simard et al. 2001) in N mineralization Accepted for publication March 2011. and nitrification after wildfire. Irrespective, inorganic

N concentrations typically tend to be greater or equal 71°48′W).The forest vegetation was characterized as an old- to control sites in the first years after fire (Smithwick growth Araucaria araucana (Mol) K. Koch – Nothofagus spp. et al. 2005b; Turner et al. 2007; Boerner et al. 2009). (Araucaria–Nothofagus) forest type (Donoso 1981). The Several mechanisms might explain the alteration of understory of this forest type was dominated by Chusquea soil N transformation rates and inorganic N availabil- spp., covering the entire soil surface. The burned study area had been exposed to a high intensity fire, resulting in an ity in stand-replacing fires (reviewed in Smithwick almost complete mortality of the forest vegetation shortly et al. 2005b). Changes in soil inorganic N after fires after fire. At the time of sampling, some herbaceous and can be ascribed to a combination of direct and indirect shrubland plant species (e.g. Chusquea spp.) prevailed on the fire effects. Increases in net soil mineralization after fire soil surface (15% coverage of the area). In total, more than have been attributed to higher soil pH (Christensen & half of the national park area (approx. 20.000 ha) was Muller 1975), N release from dead roots and previ- affected by wildfire in February 2002 (CONAF 2002). ously inaccessible N forms (Smithwick et al. 2005b), Burned and control sites were located at 10 km distance increased anion exchange capacity (Matson et al. from each other, and possess similar aspect, altitude and soil 1987), and decreased N immobilization due to a loss depth. Their topography was undulating (average slope of of carbon (C) inputs after fire (Klopatek et al. 1990). 5–10%, southern exposure). The national park’s climate is On the other hand, a decrease in net N mineralization temperate cold with an annual precipitation over 2500 mm and temperatures averaging 7.0°C in winter (July) and would be the result of increased C/N ratio of the 19.9°C in summer (January). The soil was classified as a remaining SOM (Dumontet et al. 1996), a decrease in Typic Haploxerand according to Soil Survey Staff (2006). SOM quantity and quality as a result of pyrolysis Soil samples were taken in February 2005 from the B and (Kovacic et al. 1986) and recalcitrant char production UB site, located at about 10 km from each other. In each site, (Knicker et al. 1996; Monleon et al. 1997; Knicker three permanent soil plots (0.1 ha), located at a variable 2007), or an increase in N immobilization due to distance of 50–150 m from each other, were established. All enhanced microbial activity and/or increases in plots possessed similar vegetation (Araucaria–Nothofagus), organic C and phosphorus (P) concentrations (Adams aspect (southern), altitude (1200–1400 m.a.s.l.) and soil & Attiwill 1991; Ojima et al. 1994). Net nitrification is depth (approx. 0.8 m). One soil sample was taken from each usually enhanced after fire because of increased NH + plot, and defined as a replicate (n = 3 per site). The soil was 4 sampled at a depth of 0–5, 5–10, 10–20 and 20–30 cm. The availability, increased soil pH, the presence of char and - recognizable plant material (Oi + Oe layer) was removed decreased NO3 immobilization (Stock et al. 1988; before sampling in both sites. The first three layers corre- Chorover et al. 1994; Kaye & Hart 1998; deLuca et al. sponded to A1 genetic horizons and the last layer to A2 2002). Finally, changes in the composition or abun- genetic horizon. The differences between A1 and A2 were dance of microbial and plant communities are likely to the differences in colour only, 7.5 YR (3/2) and 5YR (3/4), influence biogeochemical N cycling. respectively.Thus, a total of 24 soil samples were taken. After In this study, we performed net N cycling flux assays transport to the laboratory, visible coarse organic materials (net ammonification, net nitrification and net N min- were removed from the soil by hand picking and soil samples eralization), and 15N and 13C natural abundance (d15N were sieved (2 mm) and homogenized. and d13C) measurements on volcanic soils (Andisols) sampled 3 years after fire. In addition, we also applied a physical fractionation procedure to determine labile Physical soil organic matter and recalcitrant SOM fractions and quantified char fractionation procedure in order to determine SOM quality. The objectives of this study were (i) to determine the effect of a severe, A physical size and density SOM fractionation procedure was performed for all soils. The applied methodology involved a stand-replacing fire event on SOM dynamics and soil slight modification of the procedure described by Meijboom N cycling in an Araucaria araucana–Nothofagus spp. et al. (1995). Briefly, a fresh soil sample of about 300 g was forest; (ii) to determine underlying mechanisms that wet sieved over two sieves (mesh sizes 150 and 46 mm) using drive wildfire-associated changes in soil N cycling; and tap water. As such, three size fractions (2000–150, 150–46 (iii) to assess the potential of isotopic measurements as and <46 mm) were obtained. The coarsest size fraction a proxy for changes in SOM dynamics after wildfire. (2000–150 mm) was transferred into a small bucket, after which the macro-organic matter (MOM) fraction was separated from the mineral soil fraction by decantation in water. The MOM was further separated into a light (density < 3 3 MATERIALS AND METHODS 1.30 Mg m- ) and heavy (density > 1. 30 Mg m- ) fraction by submersion in colloidal silica suspensions (Ludox, Dupont, USA) with a density of 1.30 Mg m-3. As such, five Site description and soil sampling different SOM fractions were obtained: light fraction (LF), heavy fraction (HF), mineral fraction 2000–150 mm, mineral Burned (B) and unburned (UB) study sites were selected in fraction 150–50 mm, mineral fraction <50 mm. The obtained two comparable forest areas located in the National Park size and density fractions were washed with distilled water Tolhuaca, Andes mountain range, southern Chile (38°12′S, and dried at 40°C for 48 h. Finally, the samples were doi:10.1111/j.1442-9993.2011.02258.x © 2011 The Authors Journal compilation © 2011 Ecological Society of Australia POST-FIRE SOIL NITROGEN DYNAMICS 155 homogenized using a planetary ball mill (PM400, Retsch, Net N cycling flux measurements Germany) for subsequent analytical analyses on TC, TN, d13C and d15N (see section physicochemical analyses). Partially dried bulk soil samples (approx. 50 g) were re-wetted and transferred to PVC tubes. Bulk density and water content was adjusted to a water-filled pore space (WFPS) of 60% (Huygens et al. 2005). In order to restore a Solid-state 13C NMR spectroscopy realistic microbial activity, the PVC tubes were pre-incubated for 7 days at a constant moisture level and temperature of Soil sample demineralization was performed by hydrofluoric 15°C. The PVC tubes were loosely covered with parafilm to acid treatment starting from 2500 mg bulk soil (Gonçalves permit gas exchange while retarding evaporative water loss. et al. 2003). Between 275 and 110 mg of hydrofluoric acid After the pre-incubation period, a destructive soil extraction treated B and UB soil were oxidized (6 h) with 40 mL of was performed at days 0, 8, 23 and 38 using KCl (1 M, soil solution extraction ratio 1:3). Net ammonification, net nitri- 0.1 M K2Cr2O7/2 M H2SO4 solution at 60°C in an ultrasonic bath (Knicker et al. 2007).The solution was renewed several fication and net N mineralization fluxes were calculated via a + - times during the treatment. For all samples, the oxidation linear regression between the contents of NH4 ,NO3 and + - was performed in duplicate and the C losses of the soil (NH4 + NO3 ), respectively, obtained as a function of time. samples were calculated respective to the non-oxidized hydrofluoric acid treated samples. The oxidation residues were washed four times with distilled water and freeze- Physicochemical analyses dried. Non-oxidized mineral free and acid potassium dichromate oxidized samples were analysed by solid-state 13C Measurement of 13C, total C (TC), 15N and total N (TN) NMR spectroscopy (Bruker DSX-200 NMR spectrometer, in size and density fractions and bulk soil samples were Germany) at a resonance frequency of 50.32 MHz. Cross- performed using an elemental analyser (ANCA-SL, PDZ- polarization with magic angle spinning (CPMAS) (Schaefer Europa, UK) coupled to an Isotope Ratio Mass Spectrometer & Stejskal 1976) was applied at 6.8 kHz. During the contact (20–20, SerCon, UK). The measured isotopic ratios are time of 1 ms, a ramped 1H-pulse was applied ranging from expressed as d-values (‰) relative to Vienna Pee Dee Belem- 100% to 50% to avoid Hartmann-Hahn mismatches.The 13C nite (VPDB), and atmospheric air for 13C for 15N, respectively. chemical shifts are given relative to tetramethylsilane and The working standard was flour with a d13C value of were calibrated using glycine (COOH: 176.08 p.p.m.). The -27.01 Ϯ 0.04‰ and aTC content of 39.95 Ϯ 0.40‰, and a pulse delay was 400 ms. Using the instrument software, the d15N value of 2.69 Ϯ 0.15‰ and a total N content of total abundances of the various C-groups were determined 1.75 Ϯ 0.04‰ (certified by Iso-Analytical, UK). by integration of the signal intensity in their respective Ammonium in the extract was determined colorimetrically chemical shift regions according to Knicker et al. (2005). by the salicylate–nitroprusside method (Mulvaney 1996) The chemical shift regions 0–45, 45–110, 110–140, 140–160 on an auto-analyser (AA3, Bran and Luebbe, Germany). and 160–220 p.p.m. were assigned to, respectively, alkyl C, Nitrate was determined colorimetrically using the same auto- - - O-alkyl C, C substituted aryl C, O substituted aryl C and analyser in the form of NO2 after reduction of NO3 in a - carboxylic C (Wilson 1987). The following compounds are Cd–Cu column followed by the reaction of the NO2 with tentatively assigned to each of the respective regions (Knicker N-1-napthylethylenediamine to produce a chromophore. - - & Lüdemann 1995): lipids, waxes, cutin, suberin and other The NO3 results were corrected for NO2 present in the soil aliphatic biomacromolecules (Alkyl C); polysaccharides, samples. Soil pH was measured in H2O in 1:2.5 soil solution methoxyl groups in lignin, N-alkyl C derived from proteins ratio. Bulk density was determined using steel cylinders (O-alkyl C), aromatic C-C and C-H of lignin, black carbon (200 cm3). Available P was determined according to Olsen and other aromatic compounds (C substituted aryl C); aro- et al. (1954). Extractable aluminum, sodium, potassium, matic C-O and C-N of lignin, black carbon and other aro- calcium and magnesium were determined after soil extrac- matic compounds (O substituted aryl C); carboxyl, carbonyl tion with ammonium acetate (pH 4.8 – DPTA). and amide (carboxylic C).The aromatic C contribution (aryl C) to the intensity in the region between 110 and 160 p.p.m. in the residue after chemical oxidation is assumed to derive Statistical analysis from charcoal and will be referred to as aromatic Chemical

Oxidation Resistant Carbon (CORECarom). Owing to insuf- Statistical differences (P < 0.05) between variables of B and ficient averaging of the chemical shift anisotropy at a spin- UB soils were assessed by a Student’s t-test. A multiple linear ning speed of 6.8 kHz, spinning side bands of the aromatic C regression analysis, using a stepwise selection (probability- signals occurred at a frequency distance of the spinning of-F-to-enter value smaller than or equal to 0.05), was used speed at both sides of the central signal (300–245 p.p.m. and to elucidate the main drivers of net N mineralization and net 0to-50 p.p.m.). Their contributions were considered by nitrification. As independent variables, quantitative physico- adding their intensities to that of the parent signal. However, chemical soil properties (TC, TN, pH, Olsen P, bulk density because of the fact that that an unknown amount of the and base cations) and SOM quality indicators (N content of charcoal C may have also been oxidized, CORECarom can the five SOM fractions, CORECarom, C/N) were included. only be used as a relative index rather than as the absolute Model parameters were estimated by ordinary least-squares determination for the total amount of char. For more infor- methods. Statistical procedures were carried out with the mation with respect to this method we refer to Knicker et al. software package spss 16.0 (Statistical Package for the Social (2008). Science, USA) for Windows.

RESULTS However, the LF fraction showed a decrease in TC content (gram C per kilogram fraction) as a result of burning in the 0–5 cm soil layer (Table 2). Likewise, Physicochemical soil properties the TN content of the LF fraction (gram N kilogram fraction) was lower in the B than in the UB site (sig- Burning had no significant effect on the TC and TN nificant differences in the 0–5 cm and 10–20 cm layer, contents in the soil profile (Table 1).The C/N ratio, on Table 2). The mineral fraction <46 mm (5–10 cm soil the other hand, decreased significantly as a result of layer) and 150–46 mm (20–30 cm soil layer) showed a burning in the 0–5 and 0–10 cm soil layer. The only significantly higher TN content in the B than in the significant differences in soil pH between UB and UB site. B sites were observed for the 10–20 cm soil layer (Table 1). Bulk density was low for all sites (range: 0.3–0.6 g cm-3), and showed only minor variations between burned and unburned sites. Burning Solid-state 13C NMR spectroscopy decreased extractable base cations, resulting in significant differences between UB and B site in the 0–5 cm The solid-state 13C NMR spectra indicated that soil layer. Olsen-P did not show significant differences burning induced an alteration of SOM down to a depth between UB and B sites throughout the 0–30 cm soil of 30 cm, but the degree of alteration was more pro- + layer (Table 1).The NH4 contents were greater in the nounced in the upper soil layers (Table 3 and Appen- 0–5 and 5–10 cm soil layer of the UB site than in the dix S1, see online Supporting Information). The respective soil layers of the B site. On the other hand, spectra of the UB soils (Appendix S1a) showed an - NO3 contents were significantly higher in the B than intense signal at 31 p.p.m. (74% of the total intensity of in UB site for all soil layers (12.8 vs. 0.5 mg N kg-1, the spectrum) in the chemical shift region of alkyl C (in averaged over the 0–30 cm soil profile) (Table 1). both the hydrofluoric treated and dichromate oxidized soil).The spectra of the B soil (Appendix S1b) showed a relative increase of the signal intensity in the aryl-C Physical soil organic matter fractionation region (in both the hydrofluoric treated and dichromate oxidized soil at 110–160 p.p.m., Table 3), which is in Burning induced a change in the distribution of the line with the input of charred residues. Signals from soil mass (gram material per kilogram soil, Table 2) of unsubstituted C and for substituted aromatic C were the isolated SOM fractions. As a general tendency, it found between 140 and 110 p.p.m. In order to assess can be observed that the soil mass shifted from the LF differences in char content between sites, the hydrof- fraction towards the mineral soil fractions. In the luoric acid treated samples were subjected to a chemical 0–5 cm and 10–20 cm soil layer, the mass of the LF oxidation with acid potassium dichromate. Afterwards, was significantly lower in the B than in the UB site. the aromatic content of the oxidation residue (CORE- 13 Relative to the UB site, the B site showed a signifi- Carom) was determined by solid-state C NMR. cantly greater soil mass in the mineral fraction <46 mm Because this treatment also attacks char constituents in the 0–10 cm soil layer. However, in the 20–30 cm and the oxidation efficiency depends on the properties soil layer, the mineral fraction 2000–150 mm had a of the char, CORECarom was only used as a relative significantly greater mass in the UB than in the B site. measure for char content, and not as an absolute value. The bulk soil TC contents showed no significant The solid-state 13C-NMR spectra of the UB soils differences between the UB and the B site (Table 2). showed only low signal intensity in the aromatic C

Table 1. Physicochemical soil properties of the unburned (UB) and burned (B) Araucaria–Nothofagus forest soils

+ - Soil layer Base cations P Olsen NH4 NO3 -3 -1 -1 -1 -1 (cm) Site TC (%) TN (%) C/N (-)pH(-) rb (g cm ) (cmol(+)kg ) (mgPkg ) (mgNkg ) (mgNkg )

0–5 UB 18.1 (6.3) 0.80 (0.28) 22.6* (0.6) 4.9 (0.2) 0.3 (0.1) 8.0* (4.2) 5.4 (2.4) 3.1* (0.2) 0.4* (0.0) B 13.3 (6.3) 0.67 (0.26) 19.2* (2.4) 5.3 (0.2) 0.4 (0.1) 1.6* (0.2) 10.5 (4.7) 1.3* (0.4) 13.5* (0.6) 5–10 UB 14.1 (3.3) 0.56 (0.11) 25.0* (2.8) 5.3 (0.1) 0.4* (0.0) 1.6 (0.4) 1.9 (0.2) 2.1* (0.3) 0.3* (0.1) B 12.1 (6.3) 0.61 (0.25) 19.0* (2.9) 5.0 (0.3) 0.5* (0.0) 1.3 (0.2) 5.5 (4.5) 1.0* (0.1) 13.7* (1.0) 10–20 UB 10.7 (2.8) 0.43 (0.06) 24.3 (3.4) 5.3* (0.1) 0.5 (0.1) 1.0 (0.2) 1.5 (0.2) 1.3 (0.1) 0.5* (0.1) B 11.5 (1.7) 0.56 (0.13) 20.1 (2.8) 4.9* (0.1) 0.5 (0.1) 1.0 (0.4) 3.5 (2.0) 1.2 (0.2) 13.1* (0.1) 20–30 UB 11.5 (1.3) 0.50 (0.08) 22.8 (0.9) 5.4 (0.2) 0.5 (0.0) 0.9 (0.4) 1.3 (0.2) 1.3 (0.1) 0.9* (0.3) B 9.1 (0.4) 0.48 (0.08) 19.1 (2.4) 5.3 (0.1) 0.6 (0.0) 0.8 (0.2) 3.6 (3.1) 1.1 (0.0) 11.4* (0.2)

Standard deviation in brackets, n = 3, signiﬁcant (P < 0.05) differences between unburned and burned sites are indicated by *. rb, Bulk density. doi:10.1111/j.1442-9993.2011.02258.x © 2011 The Authors Journal compilation © 2011 Ecological Society of Australia POST-FIRE SOIL NITROGEN DYNAMICS 157

Table 2. Mass distribution, total carbon (TC) content and total nitrogen (TN) content of the bulk soil and different soil organic matter fractions as separated by a size and density fractionation procedure (Meijboom et al. 1995) for the unburned and burned Araucaria–Nothofagus forest soils

Mass (gram per kilogram soil) TC content (g C kg-1) TN content (g N kg-1)

Layer (cm) Soil fraction Unburned Burned Unburned Burned Unburned Burned

0–5 Bulk soil 1000 1000 181 (63) 133 (63) 8.0 (3.0) 6.7 (2.6) LF 142 (46)* 37 (21)* 52 (19)* 10 (6)* 2.4 (1.1)* 0.5 (0.2)* HF 84 (98) 28 (26) 16 (19) 4 (4) 0.8 (0.9) 0.2 (0.2) Mineral 2000–150 mm 642 (110) 672 (76) 93 (47) 98 (75) 3.9 (2.0) 5.1 (3.5) Mineral 150–46 mm 92 (28) 130 (48) 11 (2) 12 (5) 0.5 (0.1) 0.7 (0.3) Mineral < 46 mm 30 (2)* 90 (34)* 3 (0) 7 (4) 0.1 (0.1) 0.4 (0.2) 5–10 Bulk soil 1000 1000 141 (33) 121 (63) 5.6 (1.1) 6.1 (2.5) LF 137 (67) 52 (49) 45 (22) 17 (15) 2.0 (1.1) 0.9 (0.9) HF 54 (30) 62 (16) 9 (5) 10 (2) 0.5 (0.3) 0.6 (0.1) Mineral 2000–150 mm 666 (46) 698 (72) 73 (11) 78 (12) 3.1 (0.6) 4.7 (1.6) Mineral 150–46 mm 116 (56) 124 (38) 14 (4) 9 (4) 0.6 (0.1) 0.6 (0.2) Mineral < 46 mm 16 (2)* 78 (46)* 2 (1) 5 (3) 0.1 (0.0)* 10–20 Bulk soil 1000 1000 107 (28) 115 (17) 4.3 (0.6) 5.6 (1.3) LF 96 (19)* 47 (22)* 31 (8) 15 (7) 1.3 (0.4)* 0.6 (0.3)* HF 32 (4) 42 (4) 7 (3) 7 (3) 0.3 (0.1) 0.4 (0.1) Mineral 2000–150 mm 696 (96) 780 (22) 79 (24) 68 (2) 3.2 (0.9) 3.6 (0.6) Mineral 150–46 mm 112 (28) 82 (24) 11 (3) 9 (4) 0.5 (0.1) 0.5 (0.2) Mineral < 46 mm 30 (16) 50 (4) 4 (3) 5 (4) 0.2 (0.1) 0.3 (0.2) 20–30 Bulk soil 1000 1000 115 (13) 91 (4) 5.0 (0.8) 4.8 (0.8) LF 24 (15) 18 (3) 7 (6) 6 (1) 0.3 (0.2) 0.2 (0.0) HF 36 (32) 44 (2) 6 (6) 6 (0) 0.3 (0.2) 0.3 (0.0) Mineral 2000–150 mm 756 (16)* 702 (10)* 91 (2) 69 (22) 3.7 (0.3) 3.7 (1.1) Mineral 150–46 mm 100 (8) 136 (38) 9 (1) 15 (6) 0.5 (0.0)* Mineral < 46 mm 24 (14) 26 (10) 2 (1) 3 (1) 0.1 (0.0) 0.2 (0.1)

LF = light fraction <1.30 Mg m-3,HF= heavy fraction >1.30 Mg m-3; standard deviation in brackets; n = 3; signiﬁcant differences (P < 0.05) between unburned and burned sites are indicated by *.

Table 3. Intensity distribution (% of total signal intensity) in the cross-polarization magic angle spinning 13C nuclear magnetic resonance (NMR) spectra of the residue after hydroﬂuoric acid treatment and chemical oxidation by acid potassium dichromate for the unburned and burned Araucaria–Nothofagus forest soils

Intensity distribution (%)

Hydroﬂuoric acid Acid potassium dichromate Spectrum Layer (cm) range (p.p.m.) Unburned Burned Unburned Burned

0–5 0–45 44.6 34.4 73.9 33.8 45–110 26.6 19.7 13.1 10.1 110–160 18.7 31.8 6.0 47.3 160–245 10.1 14.1 7.0 8.8 5–10 0–45 33.1 34 73.9 46.8 45–110 23.8 19.6 13.5 8.1 110–160 28.1 30.3 5.1 31.7 160–245 15 16.1 9.2 13.4 10–20 0–45 48.8 34.4 71.7 61.6 45–110 21.2 19.2 14.3 12.2 110–160 18.6 29.7 4.0 15.7 160–245 11.4 16.7 10.0 10.7 20–30 0–45 44.0 38.1 74.1 64.3 45–110 25.4 18.8 10.8 12.4 110–160 18.7 27.6 6.6 13.5 160–245 11.9 15.4 8.5 9.8

0–45 p.p.m.: Alkyl C; 45–110 p.p.m.: O-Alkyl C; 110–160 p.p.m.: Aryl-C; 160–220 p.p.m.: Carboxyl C.

tion ﬂuxes (Table 5). Light fraction N content and

CORECarom content were included as parameters in the multiple linear regression model explaining the variance in net ammonification flux between the different UB and B sites, including field replicates (n = 24).The R2 value of this model equalled 0.48. This significant model (P = 0.001) indicated that LF N content and

CORECarom content were negatively correlated with net ammoniﬁcation rates.The contribution of both parameters in explaining total variance was slightly higher for

CORECarom (b=-0.46) than for LF (b=-0.39). In the multiple linear regression model for net nitrification, the only model variable COREC was positively cor- Fig. 1. Aromatic Chemical Oxidation Resistant Carbon arom related with net nitrification (b=0.47).This significant (CORECarom) content (g CORECarom per 100 g soil C) of the unburned and burned Araucaria–Nothofagus forest soils at model (P = 0.022) had a rather low R2 value of 0.22. different depths (error bars represent plus minus one stan- The remaining independent model variables (see Mate- dard deviation, n = 3, significant differences (P < 0.05) rials and Methods) were not included in the model based between unburned and burned sites are indicated by *). on their probability-of-F-to-enter value greater than 0.05. region (Fig. 1, Appendix S1). Any observed peaks may derive from olefinic structures, which also have their Isotopic measurements and their correlation resonance frequency in this chemical shift region. with SOM dynamics Those olefins are part of the paraffinic structures giving rise to the dominant signal at 31 p.p.m. They remain The d13C values of the different SOM fractions varied unattacked by the chemical oxidation because of their from -24.3‰ to -27.7‰ in the UB soil and from hydrophobicity and low solubility in potassium dichro- -25.7‰ to -27.3‰ in B soil (Table 6). The 0–5 and 13 mate acid (Knicker et al. 2007). CORECarom contents 5–10 cm bulk soil layers of B were depleted in C were higher in the B soils than in the UB soils; signifi- relative to the UB bulk soils (Table 6). Significant dif- cant differences were noticed in the 0–5, 5–10 and ferences between UB and B were observed for the LF 10–20 cm soil layer (Fig. 1). In the UB and B soil, fraction throughout the entire 0–30 cm soil profile. In CORECarom decreased with depth throughout the the 5–10 cm soil layer, also the HF and mineral frac- 0–20 cm soil profile, after which a slight increase in tion 2000–150 mm were significantly more depleted in 15 CORECarom was noticed for the 20–30 cm soil layer. the B than in the UB site.The d N values varied from 0.7‰ to 6.8‰ in UB soil, and from 1.8‰ to 9.7‰ in B soil (Table 6). B soils were significantly more Net N cycling flux measurements enriched in 15N relative to the UB site in the 0–5, 5–10 and 10–20 cm soil layer (Table 6). The latter signifi- Net ammonification, nitrification and N mineralization cant enrichments between UB and B soils were fluxes were mainly negative indicating N immobiliza- noticed for all five SOM fractions, although differences tion (Table 4). Ammonification fluxes varied between were not always significant (Table 6). 4 and -60 mgNkg-1 day-1 (Table 4). Significant differ- A clear exponential relationship (R2 0.76, P < ences between UB and B sites were found in the = 0.001) was observed between 13C values and CORE- 10–20 cm soil layer, where the B site showed a signifi- d C content when plotting values for bulk soil samples cantly greater value than the UB site. Net nitrification arom from the UB and B site (n = 24; Appendix S2, see fluxes varied between 93 and -177 mgNkg-1 day-1 online Supporting Information). Bulk soil 15Nvalues (Table 4). In the 0–5 and 10–20 cm soil layer, the B d did not correlate significantly with any of the physico- site showed a significantly lower nitrification flux than chemical or SOM quality parameters (data not shown). the UB site. The 10–20 cm soil layer of the B site showed a more negative net N mineralization relative to the UB site (Table 4). No significant differences in DISCUSSION net N mineralization were observed for the other soil layers (0–5, 5–10, 20–30 cm). Wildfire effects on physicochemical soil Linear regression analysis properties and soil N cycling

A multiple linear regression analysis resulted in signifi- Wildfire disturbances have shaped the biogeography of cant models for net ammonification and net nitrifica- south Chilean Araucaria ecosystems (González et al. doi:10.1111/j.1442-9993.2011.02258.x © 2011 The Authors Journal compilation © 2011 Ecological Society of Australia POST-FIRE SOIL NITROGEN DYNAMICS 159

Table 4. Net ammonification flux, net nitrification flux and N mineralization flux (mgNkg-1 day-1) for the unburned and burned Araucaria–Nothofagus forest soils at different depths

Ammoniﬁcation (mgNkg-1 d-1) Nitriﬁcation (mgNkg-1 d-1) N mineralization (mgNkg-1 d-1)

Soil layer (cm) Unburned Burned Unburned Burned Unburned Burned

0–5 -60 (31) -40 (3) -2 (3)* -39 (15)* -62 (30) -79 (14) 5–10 -34 (21) -20 (16) -91 (137) 93 (92) -125 (158) 73 (98) 10–20 -39 (7)* 4 (5)* -15 (3)* -177 (72)* -54 (30)* -181 (77)* 20–30 -43 (10) -15 (14) -19 (46) -110 (86) -63 (36) -125 (99)

Standard deviation in brackets, n = 3, signiﬁcant differences (P < 0.05) between unburned and burned sites are indicated by *.

Table 5. Output of the multiple regression analyses models between net ammonification flux and net nitrification flux (as dependent variables), and selected soil characteristics (as independent values, see section Materials and Methods)

Dependent variable Independent variables included in model b coefﬁcient P-value

-1 -1 2 Net ammoniﬁcation ﬂux (mgNg day ): CORECarom and LF (R = 0.48, P = 0.001, n = 24) (constant) 0.82

CORECarom (% of total soil C) -0.46 0.012 LF content (g N kg-1 soil) -0.39 0.031 -1 -1 2 Net nitriﬁcation ﬂux (mgNg day ): CORECarom (R = 0.22, P = 0.022, n = 24) (constant) 0.004

CORECarom (% of total soil C) 0.47 0.022

2005) and alter the SOM equilibrium in these forests cates (unburned and burned sites, 4 soil horizons, 3 (Veblen 1982). Changes in physicochemical soil prop- site replicates) to assess main physicochemical param- erties directly affect soil N cycling (Knicker 2007; eters determining (post-fire) inorganic N cycling Turner et al. 2007). In this study, the only significant (Smithwick et al. 2005a). Multiple linear regression differences in physicochemical soil properties between analyses indicated that the variation in net N ammoni- unburned and burned sites (top soil 0–5 cm) were fication was mainly explained by two soil parameters: observed for base cations. The loss of base cations in LF N content and CORECarom, a relative measure for the burned sites might be a consequence of soil erosion char C content. Both parameters were inversely corre- following wildfire (Murphy et al. 2006). Both burned lated with N mineralization. A similar analysis for net and unburned sites showed low or negative net soil N N nitrification indicated that CORECarom content was ammonification fluxes, which is in agreement with pre- positively related to nitrification. However, it should be vious studies in non-disturbed forests in the region noted that the low R2 value of the multiple linear (Huygens et al. 2007, 2008). Also net nitrification regression model for net nitrification fluxes indicates fluxes showed a similar pattern with low or negative that other parameters, not considered in this study, - values, in contrast to previous studies in similar exert a main influence on soil NO3 dynamics. Araucaria ecosystems (Satti et al. 2003). The latter The LF fraction was significantly reduced after wild- divergence can be explained by differences in sampling fire in these forest soils, both in mass and C and N date (autumn vs. summer), and associated soil mois- content.The LF fraction includes young and recogniz- ture content and substrate availability (Satti et al. able plant material, present in the soil matrix as free 2003). Generally, burned and unburned experimental SOM and intra-macroaggregate (>150 mm) SOM field sites showed a large variation in physicochemical (Meijboom et al. 1995). As physical protection of the soil characteristics and inorganic N production rates. free LF fraction is largely absent, this fraction is highly Substantial variability in post-fire soil characteristics susceptible to wildfire. Moreover, it is believed that and inorganic N fluxes has also been observed by wildfire also affected the intra-macroaggregate LF Turner et al. (2007) and Smithwick et al. (2005b).The fraction as a result of macro-aggregate breakdown. At latter is ascribed to local site differences in fire inten- high temperatures, thermal and dehydration processes sity and duration, and vegetation and litter flammabil- affect the stability of oxihydroxydes and SOM (Gio- ity (Neary et al. 1999). vannini et al. 1998; Badía & Marti 2003). Both com- In this study, we used multiple linear regression ponents are main binding agents between individual models using data from the 24 individual site repli- soil particles and aggregates in these Andisols (Sollins

Table 6. Natural abundance 13C(d13C) and 15N(d15N) signature of the bulk soil and different soil organic matter fractions as separated by a size and density fractionation procedure (Meijboom et al. 1995) for the unburned and burned Araucaria– Nothofagus forest soils

d13C (‰) d15N (‰)

Layer (cm) SOM fraction Unburned Burned Unburned Burned

0–5 Bulk soil -25.5 (0.3)* -27.0 (0.3)* 0.9 (0.2)* 3.8 (1.0)* LF -27.1 (1.0)* -26.2 (0.4)* 2.1 (1.1) 1.8 (1.3) HF -25.5 (0.2) -25.8 (0.9) 3.5 (1.6) 6.2 (2.0) Mineral 2000–150 mm -25.3 (0.6) -25.9 (1.6) 2.1 (1.1) 3.6 (0.5) Mineral 150–46 mm -25.6 (0.2) -26.1 (1.0) 2.1 (0.7) 3.9 (2.0) Mineral < 46 mm -25.4 (0.5) -25.7 (1.1) 3.0 (0.2) 4.0 (0.7) 5–10 Bulk soil -25.2 (0.5)* -26.7 (0.3)* 2.3 (0.2)* 5.3 (0.7)* LF -27.5 (0.5)* -26.2 (0.3)* 0.7 (0.9)* 3.4 (0.6)* HF -25.4 (0.4)* -27.3 (0.4)* 6.5 (1.6) 8.1 (0.6) Mineral 2000–150 mm -25.0 (0.6)* -26.8 (0.2)* 0.8 (0.4)* 5.8 (3.1)* Mineral 150–46 mm -25.6 (0.7) -26.4 (1.2) 2.5 (0.2)* 5.7 (0.8)* Mineral < 46 mm -25.4 (0.5) -26.6 (1.1) 3.0 (0.3)* 6.3 (1.5)* 10–20 Bulk soil -25.2 (0.5) -26.1 (0.4) 2.8 (0.4)* 4.7 (0.3)* LF -27.7 (0.3)* -26.1 (0.5)* 1.8 (1.5)* 4.3 (0.5)* HF -25.9 (0.4) -26.9 (0.5) 6.8 (2.6) 9.7 (0.8) Mineral 2000–150 mm -24.7 (0.7) -25.9 (0.2) 1.7 (0.8)* 5.4 (2.2)* Mineral 150–46 mm -24.7 (0.7) -26.0 (0.8) 4.1 (0.4)* 5.5 (0.7)* Mineral < 46 mm -25.2 (0.9) -26.2 (0.4) 2.1 (0.3)* 5.6 (0.4)* 20–30 Bulk soil -25.0 (0.6) -25.9 (0.2) 4.3 (1.0) 5.2 (0.2) LF -27.5 (0.2)* -25.9 (1.0)* 2.7 (2.8) 3.4 (0.3) HF -25.4 (1.7) -26.5 (0.2) 3.9 (0.6)* 9.7 (1.3)* Mineral 2000–150 mm -25.9 (0.6) -26.1 (0.6) 2.4 (0.5) 4.7 (2.2) Mineral 150–46 mm -24.3 (1.2) -25.8 (0.7) 5.0 (1.1) 3.9 (0.5) Mineral < 46 mm -24.4 (1.2) -26.9 (0.5) 5.6 (1.2) 5.1 (1.0)

LF = light fraction <1.30 Mg m-3,HF= heavy fraction >1.30 Mg m-3; standard deviation in brackets, n = 3; significant differences (P < 0.05) between unburned and burned sites are indicated by *. et al. 1996; Huygens et al. 2005). The observed especially increased in soils characterized by a low break-up of macroaggregates as a result of high tem- autotrophic nitrification community (deLuca et al. perature heating has also been observed by Terefe et al. 2006), such as these south Chilean rainforest ecosys- (2008). As such, macro-aggregate occluded LF could tems (Huygens et al. 2008). A similar inverse relation- + have been released and oxidized during the wildfire ship between char content and NH4 availability has + event. Because the LF acts as a sink for NH4 (Bengts- previously been documented by deLuca et al. (2002). son & Bergwall 2000), it is concluded that wildfire The fact that net nitrification rates do not differ + reduced the extent of NH4 immobilization by decreas- between burned and unburned sites indicates that also - ing the abundance of LF throughout the soil profile. other variables that determine NO3 availability are As indicated by the increased abundance of the rela- altered as a result of fire. In undisturbed volcanic soils, tive index CORECarom, wildfire generated a higher char nitrification is largely heterotrophic and performed by C content, in the 0–20 cm soil layer of the burned site the fungal soil community (Huygens et al. 2007, relative to the unburned site.Wildfire tends to convert 2008). Especially this microbial group is highly sus- N compounds of litter and humic material into het- ceptible to wildfire, in contrast to Gram-negative bac- erocyclic N (Hilscher et al. 2009). Both selective teria, such as autotrophic nitrifiers, which might even preservation or re-arrangement of existing amide and become more abundant after wildfire (Pietikäinen peptide structures have been documented as main & Fritze 1995; Esquilín et al. 2008). Thus, potentially mechanisms for this process (Knicker et al. 1996). a shift in microbial community composition (het- Because the N immobilized in heterocyclic structures erotrophic to autotrophic nitrifiers) has occurred in is assumed to have low microbial accessibility, micro- the soil, without altering the total nitrification flux. bial N ammonification of the char might be con- Further research should test the latter hypothesis. strained (Almendros et al. 2003; Smithwick et al. Although loss of anion adsorption sites, and associated - 2005b). Moreover, char is also known to stimulate the changes in NO3 consumption, might occur after wild- autotrophic nitrification process by alleviating factors fire, the small changes in soil pH indicate that abiotic - that otherwise inhibit the activity of the nitrifying NO3 consumption is probably to a large extent similar microbial community in forest soils (deLuca et al. between burned and unburned sites in our study + 2006). As a result, competition for available NH4 is (Certini 2005; Molina et al. 2007). doi:10.1111/j.1442-9993.2011.02258.x © 2011 The Authors Journal compilation © 2011 Ecological Society of Australia POST-FIRE SOIL NITROGEN DYNAMICS 161

The use of stable isotopes to assess alterations could not be assessed using replicate forest sites of SOM dynamics as a result of wildfire because of the natural and variable wildfire dynamics in these protected forest sites. Instead, nearby located The analysis of natural abundance 13C and 15Npat- burned and unburned sites with similar (pre-wildfire) terns is a valuable tool to explore soil, plant and micro- vegetation and soil (soil parent material, aspect, slope, bial ecology (Bowling et al. 2008; Hobbie & Hobbie altitude, etc.) properties were selected. It was shown 2008). The potential of stable isotopes to assess that standing inorganic N pools were greater in the alterations of biogeochemical soil cycling as a result burned soil, but that no main differences in net N of wildfire has been indicated by Saito et al. (2007). cycling fluxes were observed between unburned and These authors hypothesized the occurrence of burned sites sampled 3 years after wildfire. Rather enriched d13C and d15N values in SOM due to volatil- than differences in N mineralization, shifts in N ization of lighter isotopes as a result of fire. Although transformation pathways were indicated as a result of d15N results were in agreement with this hypothesis, wildfire. Multiple linear regression models indicated 13 the expected pattern for d C could not be observed in that changes in LF N content and CORECarom (a rela- their field samples. Our results are in line with the field tive index for char C content) as a result of wildfire 13 + study of Saito et al. (2007), i.e. depleted d C values may have shifted NH4 consumption pathways from and enriched d15N values in the burned site compared immobilization to nitrification in the burned site. On with the unburned site. the one hand, the LF fraction, a strong inorganic N In this study, 13C-NMR analysis indicated an sink (Bengtsson & Bergwall 2000), was largely reduced increased contribution of heterocyclic aromatic com- in mass and N content as a result of fire. On the other pounds (CORECarom) to the total SOM in the burned hand, char C content increased after wildfire, and sites. Because those compounds are typically depleted was positively correlated with net nitrification fluxes. in d13C (Adams & Grierson 2001), total bulk soil d13C Increased autotrophic nitrification may be compen- values decreased as a result of fire. The good correla- sated by a reduction in heterotrophic nitrification, 13 - tion between CORECarom and d C lends support to which plays a main role in the NO3 production the idea that chemical C alterations and preferential process in undisturbed volcanic forest soils. accumulation of depleted d13C material are the main Shifts in d13C and d15N were observed as a result of processes determining differences in bulk soil d13C fire, observations that could be related to changes in between burned and unburned sites. The importance structural SOM components and gross inorganic N of heterocyclic charred material as a main factor fluxes. As such, a clear relationship between d13C and 13 determining post-fire d C has also been indicated CORECarom was observed, indicating that isotope ratio by Rumpel et al. (2006). An alternative hypothesis, mass spectrometric (IRMS) analyses provide a good based on potential post-fire vegetation shifts from C3 and simple measure to assess the degree of aromatic (d13C range: -32‰ to -20‰) to C4 plant communi- compound formation after wildfire. Because this com- ties (d13C range: -17‰ to -9‰), is less likely, as the ponent is a major factor determining post-fire nutrient LF fraction, consisting of young plant material, shows dynamics, we believe that IRMS analysis can be used a typical C3 signature. Moreover, no C4 plants were as an analytical tool in ecosystem wildfire studies and recognized during post-fire field surveys.The observed serve as alternative laborious and costly 13C-NMR increase in d15N as a result of fire is in agreement analyses. with previous studies (Rumpel et al. 2006; Saito et al. 2007), and is typically attributed to enhanced gross nitrification and associated loss of depleted 15N mol- ACKNOWLEDGEMENTS ecules in the burned sites (Herman & Rundel 1989), and/or isotopic fractionation during the fire event This study was funded by the Bilateral Scientific (Saito et al. 2007). Additionally, the rapid recovery of and Technological Cooperation between Flanders and microbial growth to levels that exceed the pre-wildfire Chile (BOF, UGent), the National Commission for standing biomass pool (Barcenas-Moreno & Baath Scientific and Technological Research – Chile PhD 2009; but see Prieto-Fernandez et al. 1998) may have Scholarship for Yessica Rivas, (Fondecyt, N°1080065 caused a further d15N enrichment (Dijkstra et al. and N°1090455), and the Dirección de Investigación 2006). y Desarrollo – Universidad Austral de Chile (DID- UACh). Yessica Rivas thanks the MECESUP Program of the doctoral school in Forest Sciences of CONCLUSIONS the Universidad Austral de Chile (UACh-VCO). Dries Huygens is a postdoctoral fellow of the Fund This study provides a mechanistic understanding of for Scientific Research – Flanders (FWO). Pascal soil N cycling responses to wildfire in an Araucaria– Boeckx thanks Fondecyt N°1085081 for the financial Nothofagus forest. It should be noted that fire effects support during his scientific visit.

REFERENCES Dumontet S., Dinel H., Scopa A. et al. (1996) Post-fire soil microbial biomass and nutrient content of a pine forest Adams M. A. & Attiwill P. M. (1991) Nutrient balance in forests soil from a dunal Mediterranean environment. Soil Biol. of northern Tasmania. 1. Atmospheric inputs and within- Biochem. 28, 1467–75. stand cycles. For. Ecol. Manage. 44, 93–113. Esquilín A. E. J., Stromberger M. E. & Shepperd W. D. (2008) Adams M. A. & Grierson P. F. (2001) Stable isotopes at natural Soil scarification and wildfire interactions and effects on abundance in terrestrial plant ecology and ecophysiology: an microbial communities and carbon. Soil Sci. Soc. Am. J. 72, update. Plant Biol. 3, 299–310. 111–8. Almendros G., Knicker H. & Gonzales-Vila F. J. (2003) Rear- Giovannini G., Luchéis S. & Giachetti M. (1998) Effect of rangement of carbon and nitrogen in peat after progressive heating on some physical and chemical parameters related isothermal heating as determined by solid-state 13C- and to soil aggregation and erodibility. Soil Sci. 146, 255–62. 15N-NMR spectroscopy. Org. Geochem. 34, 1559–68. Gonçalves C. N., Dalmolin R. S. D., Dick D. P. et al. (2003) The Armesto J. J., Smith-Ramírez C., Carmona M. R. et al. (2009) effect of 10% HF treatment on the resolution of 13C Old-growth temperate rainforests of South America: CP/MAS NMR spectra and on the quality of organic matter conservation, plant–animal interactions, and baseline bio- in Ferralsols. Geoderma 116, 373–92. geochemical processes. In: Old-Growth Forests: Functions, González M. E., Veblen T.T. & Sibold J. S. (2005) Fire history Fate and Value, Ecological Studies 207 (eds C. Wirth, G. of Araucaria–Nothofagus forests in Villarrica National Park, Gleixner & M. Heimann) pp. 367–90. Springer-Verlag, Chile. J. Biogeogr. 32, 1187–202. Berlin. González M. E., Veblen T.T. & Sibold J. S. (2010) Influence of Badía D. & Marti C. (2003) Plant ash and heat intensity effects fire severity on stand development of Araucaria araucana– on chemical and physical properties of two contrasting soils. Nothofagus pumilio stands in the Andean cordillera of south- Arid Land Res. Manage. 17, 23–41. central Chile. Aust. Ecol. 35, 597–615. Barcenas-Moreno G. & Baath E. (2009) Bacterial and fungal Grogan P., Bruns T. D. & Chapin F. S. (2000) Fire effects on growth in soil heated at different temperatures to simulate ecosystem nitrogen cycling in a Californian bishop pine a range of fire intensities. Soil Biol. Biochem. 41, 2517– forest. Oecologia 122, 537–44. 26. Herman D. J. & Rundel P.W. (1989) Nitrogen isotope fraction- Bengtsson G. & Bergwall C. (2000) Fate of 15N labelled nitrate ation in burned and unburned chaparral soils. Soil Sci. Soc. and ammonium in a fertilized forest soil. Soil Biol. Biochem. Am. J. 53, 1229–36. 32, 545–57. Herrmann T. M. (2005) Knowledge, values, uses and manage- Boerner R. E. J., Huang J. J. & Hart S. C. (2009) Impacts of fire ment of the Araucaria araucana forest by the indigenous and fire surrogate treatments on forest soil properties: a Mapuche Pewenche people: a basis for collaborative natural meta-analytical approach. Ecol. Appl. 19, 338–58. resource management in southern Chile. Nat. Resour. Forum Bowling D. R., Pataki D. E. & Randerson J. T. (2008) Carbon 29, 120–34.

isotopes in terrestrial ecosystem pools and CO2 fluxes. New Herrmann T. M. (2006) Indigenous knowledge and manage- Phytol. 178, 24–40. ment of Araucaria araucana forest in the Chilean Andes: Burns B. R. (1993) Fire-induced dynamics of Araucaria implications for native forest conservation. Biodivers. araucana-Nothofagus antarctica forest in the southern Andes. Conserv. 15, 647–62. J. Biogeogr. 20, 669–85. Hilscher A., Heister K., Siewert C. & Knicker H. (2009) Min- Catovsky S. & Bazzaz F. A. (2002) Nitrogen availability eralisation and structural changes during the initial phase of influences regeneration of temperate tree species in the microbial degradation of pyrogenic plant residues in soil. understory seedling bank. Ecol. Appl. 12, 1056–70. Org. Geochem. 40, 332–42. Certini G. (2005) Effects of fire on properties of forest soils: a Hobbie E. A. & Hobbie J. E. (2008) Natural abundance of 15Nin review. Oecologia 143, 1–10. nitrogen-limited forests and tundra can estimate nitrogen Chorover J., Vitousek P. M., Everson D. A. et al. (1994) Solution cycling through mycorrhizal fungi: a review. Ecosystems 11, chemistry profiles of mixed-conifer forests before and after 815–30. fire. Biogeochemistry 26, 115–44. Huygens D., Boeckx P., Van Cleemput O. et al. (2005) Aggregate Christensen N. L. & Muller C. (1975) Effects of fire on factors and soil organic carbon dynamics in south Chilean Andisols. controlling plant growth in Adenostoma chaparral. Ecol. Biogeosciences 2, 159–74. Monogr. 45, 29–55. Huygens D., Rütting T., Boeckx P. et al. (2007) Soil nitrogen CONAF (2002) Antecedentes sobre impacto de incendios conservation mechanisms in a pristine south Chilean Notho- forestales en la IX Región. In: Documento de visita a la zona fagus forest ecosystem. Soil Biol. Biochem. 39, 2448–58. afectada por incendios p. 16. CONAF, Temuco. Huygens D., Boeckx P., Templer P. et al. (2008) Mechanisms for deLuca T. H., Nilsson M. C. & Zackrisson O. (2002) Nitrogen retention of bioavailable nitrogen in volcanic rainforest soils. mineralization and phenol accumulation along a fire Nature Geosci. 1, 543–8. chronosequence in northern Sweden. Oecologia 133, 206– Kaye J. P. & Hart S. C. (1998) Ecological restoration alters 14. nitrogen transformations in a ponderosa pine bunchgrass deLuca T. H., MacKenzie M. D., Gundale M. J. & Holben W. E. ecosystem. Ecol. Appl. 8, 1052–60. (2006) Wildfire-produced charcoal directly influences nitro- Klopatek J. M., Klopatek C. C. & Debano L. F. (1990) Potential gen cycling in ponderosa pine forests. Soil. Sci. Soc. Am. J. variation of nitrogen transformations in pinyon-juniper eco- 70, 448–53. systems resulting from burning. Biol. Fertil. Soils 10, 35–44. Dijkstra P., Ishizu A., Doucett R. et al. (2006) 13C and 15N Knicker H. (2007) How does fire affect the nature and stability natural abundance of the soil microbial biomass. Soil Biol. of soil organic nitrogen and carbon? A review. Biogeochem- Biochem. 38, 3257–66. istry 85, 91–118. Donoso C. (1981) Ecología Forestal: El Bosque Y Su Medio Knicker H. & Lüdemann H.-D. (1995) 15N and 13C CPMAS Ambiente. Editorial Universitaria, Santiago. and solution NMR studies of 15N enriched plant material doi:10.1111/j.1442-9993.2011.02258.x © 2011 The Authors Journal compilation © 2011 Ecological Society of Australia POST-FIRE SOIL NITROGEN DYNAMICS 163

during 600 days of microbial degradation. Org.Geochem. 23, Rumpel C., Alexis M., Chabbi A. et al. (2006) Black carbon 329–41. contribution to soil organic matter composition in tropical Knicker H., Almendros G., Gonzales-Vila F. J. et al. (1996) 13C sloping land under slash and burn agriculture. Geoderma and 15N NMR spectroscopic examination of the transfor- 130, 35–46. mation of organic nitrogen in plant biomass during thermal Saito L., Miller W.W., Johnson D.W. et al. (2007) Fire effects on treatment. Soil Biol. Biochem. 28, 1053–60. stable isotopes in a Sierran forested watershed. J. Environ. Knicker H., Totsche K. U., Almendros G. & Gonzalez-Vila F. J. Qual. 36, 91–100. (2005) Condensation degree of burnt peat and plant Satti P., Mazzarino M. J., Gobbi M. et al. (2003) Soil N residues and the reliability of solid-state VACP MAS C-13 dynamics in relation to leaf litter quality and soil fertility NMR spectra obtained from pyrogenic humic material. Org. in north-western Patagonian forests. J. Ecol. 91, 173– Geochem. 36, 1359–77. 81. Knicker H., Muffler P. & Hilscher A. (2007) How useful is Schaefer J. & Stejskal E. O. (1976) Carbon-13 nuclear magnetic chemical oxidation with dichromate for the determination of resonance of polymers spinning at magic angle. J.Am.Chem. ‘Black Carbon’ in fire-affected soils? Geoderma 142, 178–96. Soc. 98, 1031–2. Knicker H., Weismeier M. & Dick D. P. (2008) A simplified Simard D. G., Fyles J. W., Paré D. & Nguyen T. (2001) Impacts method for the quantification of pyrogenic organic matter of clearcut harvesting and wildfire on soil nutrient status in grassland soils via chemical oxidation. Geoderma 147, in the Quebec boreal forest. Can. J. Soil. Sci. 81, 229– 69–74. 37. Kovacic D. A., Swift D. M., Ellis J. E. & Hakonson T. E. (1986) Smithwick E. A. H., Michelle M., Turner M. G. et al. (2005a) Immediate effects of prescribed burning on mineral soil Spatial heterogeneity and soil nitrogen dynamics in a nitrogen in ponderosa pine of New Mexico. Soil Sci. 141, burned black spruce forest stand: distinct controls at differ- 71–6. ent scales. Biogeochemistry 76, 517–37. Matson P. A., Vitousek P. M., Ewel J. J. et al. (1987) Nitrogen Smithwick E. A. H., Turner M. G., Mack M. C. & Chapin F. S. Transformations Following Tropical Forest Felling and (2005b) Postfire soil N cycling in northern conifer forests Burning on a Volcanic Soil. Ecology 68, 491–502. affected by severe, stand-replacing wildfires. Ecosystems 8, Meijboom F.W., Hassink J. & Van Noordwijk M. (1995) Density 163–81. fractionation of soil macroorganic matter using silica Soil Survey Staff (2006) Keys to Soil Taxonomy, 10th edn. United suspensions. Soil Biol. Biochem. 27, 1109–11. States Department of Agriculture & Natural Resources Molina M., Fuentes R., Calderon R. et al. (2007) Impact of Conservation Service, Washington, DC. forest fire ash on surface charge characteristics of andisols. Sollins P., Homann P. & Caldwell B. A. (1996) Stabilization Soil Sci. 172, 820–34. and destabilization of soil organic matter: mechanisms and Monleon V. J., Cromack K. & Landsberg J. D. (1997) Short- and controls. Geoderma 74, 65–105. long-term effects of prescribed underburning on nitrogen Stock W. D., Lewis O. A. M. & Allsopp N. (1988) Soil nitrogen availability in ponderosa pine stands in central Oregon. Can. mineralization in a coastal fynbos succession. Plant Soil. J. For. Res. 27, 369–78. 106, 295–8. Mulvaney R. L. (1996) Nitrogen-inorganic forms. In: Methods Terefe T., Mariscal-Sancho I., Peregrina F. & Espejo R. (2008) of Soil Analysis (ed. D. L. Sparks) pp. 1123–84. ASA and Influence of heating on various properties of six Mediterra- SSSA, Madison. nean soils. A laboratory study. Geoderma 143, 273–80. Murphy J. D., Johnson D.W., Miller W.W. et al. (2006) Wildfire Turner M. G., Smithwick E. A. H., Metzger K. L. et al. (2007) effects on soil nutrients and leaching in a Tahoe Basin Inorganic nitrogen availability after severe stand-replacing watershed. J. Environ. Qual. 35, 479–89. fire in the Greater Yellowstone Ecosystem. Proc. Natl. Acad. Myers N., Mittermeier R. A., Mittermeier C. G. et al. (2000) Sci. USA 104, 4782–9. Biodiversity hotspots for conservation priorities. Nature 403, Veblen T.T. (1982) Regeneration patterns in Araucaria araucana 853–8. forests in Chile. J. Biogeogr. 9, 11–28. Neary D., Klopatek C., Debano L. & Folliott P. (1999) Fire Wilson M. A. (1987) NMR Techniques and Applications in effects on belowground sustainability: a review ad synthesis. Geochemistry and Soil Chemistry. Pergamon Press, Oxford. For. Ecol. Manage. 122, 51–71. Ojima D. S., Schimel D. S., Parton W.J. & Owensby C. E. (1994) Long-term and short-term effects of fire on nitrogen cycling in tallgrass prairie. Biogeochemistry 24, 67–84. SUPPORTING INFORMATION Olsen S. R., Cole C. V., Watanabe F. S. & Dean L. A. (1954) Estimation of Available Phosphorus in Soils by Extraction with Additional Supporting Information may be found in Sodium Bicarbonate, USDA Circular 939. United States the online version of this article: Department of Agriculture, Washington, DC. Pietikäinen J. & Fritze H. (1995) Clear-cutting and prescribed Appendix. S1. Solid-state 13C NMR spectra of the burning in coniferous forest: comparison of effects on soil residue after hydrofluoric treatment and after dichro- fungal and total microbial biomass, respiration activity and mate oxidation (OREC) for the unburned and burned nitrification. Soil Biol. Biochem. 27, 101–9. Prieto-Fernandez A., Acea M. J. & Carballas T. (1998) Soil soil layers. microbial and extractable C and N after wildfire. Biol. Fertil. Appendix. S2. Relationship between CORECarom Soils 27, 132–42. and d13C for burned and unburned soils.