Introduction 

The increase in C stocks in forest ecosystems depends on the preservation of C in tree biomass and soil against degradation. Wildfire and harvesting both affect the soil organic carbon storage by combusting organic matter, volatilization, changing soil structure, vegetation canopy, microclimatic conditions, erosion, litter layer composition, microbial population and their activity, and decomposition ([49], [29], [35], [42], [33], [1]). Fire may accelerate the decomposition of organic matter (OM), promoting CO₂ emission. Logging and removal of burned trees increase the CO₂ emissions (about 120 gr C m^-2 in burned pine forests) and reduce the renovation capacity of Mediterranean ecosystems after the wildfire ([46]). On the other hand, though no significant effect on total ecosystem carbon stocks was observed, a decrease was seen in carbon stored in snags and down woody material after the salvage logging in burned sub-boreal jack pine forests ([10]) The soil microflora contributes to soil C storage, soil respiration and ecosystem productivity, despite representing only a small portion of the soil ([6]). Microbial biomass C (C_mic) has been suggested as an indicator of changes in the soil ([6]) as a consequence of forest management ([5], [41], [40]). Fire may alter the microbial abundance both directly by killing microbes soil surface due to heating and indirectly by modifying the community composition and environmental conditions ([13], [20], [56]). Additionally, harvesting may affect these conditions by compacting the soil and causing increases in runoff rates and erosion ([36], [15]). Decreases were found in microbial biomass carbon after harvest (18%), burn (74%), and burn-salvage (53%) treatments ([49]).

The metabolic quotient (qCO₂), which is the microbial respiration/microbial biomass C ratio, decreases with increasing the quality of substrate, and increases under unfavourable conditions ([3], [54], [6], [32]). Microbial parameters, including C_mic, qCO₂ and qC, have been used in several studies ([37]) in order to understand the impacts of soil management or disturbance.

Soil microorganisms apply different strategies to use available C, such as biotic (e.g., release of extracellular enzymes) and abiotic (e.g., redox and metal complexation) mechanisms ([22]). β-D glucosidase which takes part in the degradation of cellulose, decreased in burned soils, as reported by Hernandez et al. ([21]), but little is known about the harvesting effect of burned residues on this enzyme activity.

The purpose of this study was to understand the relationship between the C dynamics and the microbial activity after a wildfire, and after harvesting of the residues, by using several microbial parameters, including CO₂ evolution, C_mic, qCO₂, and β-D glucosidase enzyme activity. It was hypothesized that burning alters the microbial activity related to C dynamics, and harvesting contributes to these changes.

  Materials and methods 

Study area and soil sampling

We conducted this study in a mixed pine (Pinus nigra Arnld. subsp. pallasiana Lamb.) and oak (Quercus pubescens Wild.) forest in Safranbolu, Turkey (41° 16′ 11″ N, 32° 37′ 41″ E) following a wildfire occurred on the 9^th and 10^th of October, 2011. The fire was caused by a picnic fire, and affected an area of 5 hectares. Fire severity was estimated based on the litter consumption ([44]) and whether the ashes whiten or not ([29]). The severity of the fire ranged from low to medium due to black ash and bare soil surface dominated the majority of the site. Additionally, this determination also complies with the criteria of Ryan & Noste ([44]) modified from Ryan ([45]) and Turner et al. ([52]) ([30]). We did not measure the percentage of vegetatition cover, but based on our observation the understory was mainly composed of Phillyrea latifolia L., Cistus creticus L., Juniperus oxycedrus L., Colutea cilicica Boiss. & Balansa, and Rubus sp. The annual rainfall was between 630 and 650 mm in the region during the study period. The elevation of the area was 720-760 m a.s.l., and the slope 30-45%. The soil with typical A-C horizons belongs to the order intrazonal and suborder calsimorfic and was classified as a rendzina, according to FAO/UNESCO ([25], [50]). The main characteristics of the soils in the study area are given in Tab. 1. The area was covered with residues, consisting of twigs, for the protection of the soil and the new generation. In several parts of the burnt area the residual trees were removed (harvested).

Soil sampling was conducted in Months 4, 6, 9, 12 and 24 after the fire. We selected 22 plots (10 × 10 m²) including eight burned-non harvested (NH), eight burned-harvested (H) and three adjacent, unburned areas for each species. Three soil samples per plot were collected at depths of 0-7.5 cm, following removal of the surface residues, including oak leaves and pine needles. We collected the samples from low- and medium-severity burned areas in order to achieve homogeneity. Soil samples were transferred to the laboratory, sieved through a 2-mm mesh and stored at +4 °C.

Laboratory analyses

The CO₂ evolution was determined according to the method of Isermeyer ([24]), modified by Alef ([2]). Following incubation of the samples with NaOH traps for 24 hours at 25 °C in sealed containers, the remaining NaOH was titrated with HCl after adding BaCl₂ solution and phenolphthalein. The CO₂ flux was calculated from the amount of consumption of HCl through titration given by the following formula (eqn. 1):

where SW is the amount of soil dry weight, t is the incubation time, V₀ is the HCl used for titration, V is the HCl used for soil sample, dwt is the dry weight of 1 g moist soil, and 1.1 is the conversion factor (i.e., 1 ml 0.05 M NaOH equals 1.1 mg CO₂).

The C_mic was determined according to the chloroform extraction method described by Vance et al. ([53]). Half of each sample was fumigated with CHCl₃ and then incubated at 25 °C for 24 h. The fumigated and non-fumigated samples were extracted using 0.5 M K₂SO₄ at a soil/extract ratio of 1:4, and then kept at -20 °C until they were analysed using a TOC-L® analyser (Shimadzu Corp., Kyoto, Japan). The C_mic was calculated following Wu et al. ([55]) using the following equation (eqn. 2):

where Ec is the difference between the amount of C_org extracted from the fumigated and non-fumigated soils.

The qCO₂ (metabolic quotient) is a specific respiration rate of CO₂ C evolved per unit of C_mic, which is calculated from CO₂ evolution and C_mic C.

The β-D glucosidase activity was measured according to the method of Naseby & Lynch ([38]). A 1.5 g aliquot of soil was extracted using an acetate buffer at pH 5.5, agitated on an orbital rotary shaker with a slope angle of 60° to the horizontal for 1 h, and then centrifuged at 4000 rpm for 15 min. Following incubation at 37 °C for 24 h with buffered substrate p-nitrophenyl-β-D glucoside, 1 ml NaOHCO₃ was added to the extracts, which were then quantified at 400 nm in a spectrophotometer.

Statistical analyses

Differences between the burned (H and NH) and unburned soil samples were determined by ANOVA for each period and in total. Data was transformed when it was not normally distributed as log (1+ √qCO₂), or alternatively a non-parametric test (Mann-Whitney U) was used. Correlations between the parameters were estimated. All statistical analyses were performed using the software SPSS® v. 21 (IBM, Armonk, NY, USA).

  Results 

We observed significant differences in most of the soil samples in terms of burning, while the harvested and non-harvested samples were similar for most of the variables analysed. A significant decrease was determined in C_org in burned samples but the results were relatively higher in burned-H samples. No difference in total nitrogen (N_t) was found between the unburned and burned soils but the Month 9. The C/N ratios were slightly different in favor of the unburned samples (p = 0.05 - Fig. 1a, Tab. 2). There was no significant difference in CO₂ evolution (p = 0.956); however, the maximum values were observed in the unburned soils, and the minimum values were found in the burned-NH soils (Tab. 2). The C_mic values were different at Month 12 (p = 0.036) while it was comparatively high in the burned areas at all months (Fig. 1b, Tab. 2). We observed significant differences (p = 0.029) in C_mic/C_org at Month 12 similarly (Fig. 1c, Tab. 2). An increase in the total amount of C_mic in all areas was observed through the time, but the increase was higher in the burned areas between Months 12 and 24.

Fig. 1 - (a) C/N ratio, (b) C_mic values, (c) C_mic/C_org ratio, (d) β-D glucosidase activity, and (e) qCO₂ ratios in burned (H: harvested; NH: non-harvested) and unburned areas.

  Enlarge/Shrink   Download   Full Width  Open in Viewer

Significant differences were observed between the burned and unburned areas (p = 0.023) in favor of unburned areas in regard to β-D glucosidase enzyme activity (Fig. 1d, Tab. 2). The qCO₂ values did not vary significantly with burning (p = 0.786) except for Month 4 (Fig. 1e, Tab. 2).

There were significant differences between the different periods, in terms of the soil properties, with p < 0.05 for C/N and β-D glucosidase, and p < 0.001 for the others. The interaction effect of period and burning for qCO₂ was significant (p < 0.05), while it was unclear for CO₂ evolution (p = 0.052).

C_mic and C_mic/C_org were well correlated with each other, while negative relations between C_org, TN, qCO₂ and β-D glucosidase were recorded (Tab. 3).

  Discussion 

Although the C_org, C/N, C_mic, C_mic/C_org and β-D glucosidase values were different between the burned and unburned plots, there were no significant differences between the H (burned, harvested) and NH (burned, non-harvested) areas, in terms of the soil parameters.

Organic carbon content decreases as a result of fire as observed in this study, and changes may occur in organic matter fractions. As a matter of fact, loss of C_org and N_t has been reported due to the fire severity reaching 220-460 °C ([16], [12]). Similarly, the lower levels of OM were significant at Month 6 and 9 according to our results. Relatively high OM content in burned-harvested samples was likely due to mixing of the forest floor with the mineral soil during harvesting ([42]). N_t content in burned samples were low as well as non-significant in general. Higher N_t concentration was observed at Month 9 likely due to the warmer conditions in more moist unburned samples ([26], [27], [57]). Slight changes in C/N ratio may be related with the formation of new recalcitrant N and volatisation of C compounds, as suggested by previous studies conducted in burned pine forests ([43], [18]).

A general increase in C_mic was probably due to the increase in the concentration of oxidisable C and nutrients in burned soils as previously reported ([18], [17]); immediate increases have been found in soils after fire ([51], [14]). Likely because of the relatively late sampling in our study, we did not detect significant changes in the above variables in burned (especially burned and harvested) areas in Month 4 and Month 6. Bárcenas-Moreno et al. ([9]) categorised the process of microbial recovery after a wildfire as initial depletion of microorganisms in soil and the proliferation of fast-growing bacteria after a few months due to rapid and short increase of available nutrients. C_mic increase was reported after 8 months in a burned pine (Pinus halepensis) and oak (Quercus coccifera) mixed forest by Bárcenas-Moreno et al. ([9]) which was maintained for the rest of their study and reversed after 32 months. We assessed a significant decrease at Month 12 and non-significant decrease at Month 24 in C_mic values of the unburned areas, in comparison with those at Months 4 and 6. An increase in the total amount of C_mic was observed in all areas, more in the burned areas, likely due to the higher amount of rainfall between Months 12 and 24.

We found a negative correlation between C_mic and qCO₂ that is inconsistent with the results from previous studies ([54], [8]), whereas a positive correlation between C_mic and CO₂ evolution was detected. A high qCO₂ indicates the need for restoring the C demand as a result of the microbial biomass depending on the renewal of C lost through respiration ([4], [8]). After a low-intensity prescribed fire in a Picea abies forest, basal respiration diminished while there was a disproportionate decrease in C_mic ([41]), which clearly led to higher values of qCO₂ in the burned areas compared to unburned ones ([11]). This indicator predicts soil fertility ([3], [54], [6], [32], [7]). The lower levels of qCO₂ in the burned soils in the first and second sampling dates indicate an improvement in soil conditions for the new vegetation due to the warmer soils in burned areas. Similar values in the burned and unburned areas after Month 9 may have resulted from vegetation succession ([23]).

β-D glucosidase enzyme activity plays an important role in energy availability in the soil, as it is directly related to the labile C content and its ability to stabilise the soil organic matter independently of seasonal variability ([34]). Our results indicate that this enzyme activity decreased with burning up to the last sampling date, two years after the wildfire, and was significantly different among H, NH, and UB plots at Month 9. The decrease in enzyme activity related with the lack of C entry to the system via plant inputs of C_org and N_t ([47], [48]) is consistent with our results. There was a negative correlation between enzyme activity and C_mic. The decrease in enzyme activity compared to C_mic likely resulted from the enzymes being retained in the stabilised fraction of the soil rather than that part associated with the viable microbial population ([31]).

Decomposition increased due to harvesting, possibly resulting in changes to the C budget ([39]). The harvesting of burned tree residues did not affect the measured soil parameters significantly in this study, probably because of the sensitive protection precautions used after the wildfire (such as covering the soil surface with residual brushwood material) and the rapid regeneration of forest ([19], [18]), which is characteristic of oak forests.

  Conclusions 

Fire affected the microbial biomass likely due to changes in the substrate quality, microbial composition, and abundance. However, this increase was not significant one year after burning. The correlation between qCO₂ and β-D glucosidase activity showed that the former might be a more coherent indicator of microbial functionality than C_mic and CO₂ evolution. In fact, enzymes might be independent of microbial proliferation, as it can attach to dead cells or cellular fragments instead of viable microorganisms.

The decrease in qCO₂ after burning at low severity may indicate to the forest managers that the soil has relatively ideal conditions, simplifying the preparation of the area for new vegetation. The harvesting of burned trees in post-fire areas provides many advantages, in terms of economic benefits or protection of the forest against pests. Proper precautions for the soil, such as covering it after the fire with residues, comprising branches and small, felled burned trees, may be advantageous by allowing soil and nutrient loss to be avoided.

  Acknowledgements 

This study was supported by the General Directorate of Forestry of Turkey. All the analyses were conducted at the Research Institute for Forest Soil and Ecology in the context of project ESK-11 (6300)/2011-2015. Part of the data for some of the soil characteristics were used in the publication Kaptanoglu & Namli ([28]), for different periods and in a different context. We would like to thank Aydin Çömez, Oguz Can Turgay, Münevver Arslan, and Kürsad Özkan for their comments.

  References