Soil Microbial Diversity, Biomass, and Activity in Two Pine Plantations of Southern Italy Treated with Prescribed Burning

: Microbial diversity plays a crucial role in ecosystem processes, including organic matter decomposition and nutrient cycling. This research explores the e ﬀ ect of prescribed burning (PB) on soil microbial diversity, as well as biomass and activity in Mediterranean pine plantations. In burned and adjacent unburned plots of Pinus pinea and P. pinaster plantations of Southern Italy protected areas, the fermentation layer and the 5 cm thick layer of mineral soil underneath were sampled at intervals during the ﬁrst year after PB. The experimental protocol encompassed measurements of total microbial abundance (C mic and soil DNA), fungal mycelium, fungal fraction of C mic , microbial activity, bacterial genetic diversity (16S rDNA PCR-DGGE), microbial metabolic quotient (qCO 2 ), and C mineralization rate (CMR), as well as physical and chemical soil properties. PB caused only temporary (up to 3 h–32 d) reductions in C mic , DNA amount, fungal mycelium, respiration, and CMR in the P. pinaster plantation, and had no appreciable negative e ﬀ ect on the microbial community in P. pinea plantation, where ﬁre intensity was lower because of less abundant litter fuel. In either plantation, PB did not generally reduce bacterial genetic diversity (evaluated as band richness, Shannon index, and evenness), thus, also accounting for the fast recovery in microbial growth and activity after high-intensity PB in P. pinaster plantation. While conﬁrming PB as a sustainable practice to reduce wildﬁre risk, also supported by data on plant community obtained in the same plantations, the results suggest that an integrated analysis of microbial diversity, growth, and activity is essential for an accurate description of PB e ﬀ ects on soil microbial communities.


Introduction
Soil microbial community plays a key role in terrestrial ecosystems, by being involved in several ecosystem services. Indeed, it provides supporting services (soil formation, nutrient cycling, and plant growth), as well as regulating services (i.e., climate and gas regulation, C sequestration, water purification, disease and pest regulation, and bioremediation) and provisioning ones (i.e., the supply of food, fiber, fuel, genetic resources, chemicals, and pharmaceuticals) [1][2][3].

Study Areas 93
The study was carried out in two pine plantations dominated by P. pinea L. and P. pinaster Aiton 94 subsp. pinaster, respectively ( Figure 1). The P. pinea plantation (40°57′39.79′′ N and 13°59′50.09′′ E) is 95 located in the Castel Volturno Nature Reserve [39], a sandy coastal area located between sea level 96 and 9 m of altitude. The mean age of trees was 50 years [24]. The understory mainly consists of 97 flammable plants, such as Phillyrea angustifolia L., Rhamnus alaternus L. subsp. Alaternus, and 98 Asparagus acutifolius L. [24,40]. The climate is typically Mediterranean, with hot dry summers and 99 cool wet winters. Data from the meteorological station of Ischitella for 1974 to 2012 report a mean 100 annual temperature of 13.6 °C and a mean annual rainfall of 761.3 mm [24]. Soil is a Calcacaric Arenosol 101 according to the FAO system of soil classification [41].

105
Control and burned plots in P. pinea (B) and P. pinaster (C) plantations.
Overall, the P. pinea plantation showed a higher number of plant species in the forest understory (9.5 ± 0.7) as compared with P. pinaster (6.91 ± 2.5) plantation, as evaluated, respectively, on 50 m 2 plots (Esposito et al., unpublished results) and 78.5 m 2 plots (Stinca et al., unpublished results). It also showed lower tree density and higher values of shrub and herb cover, as well as herb height as compared with P. pinaster (Table 1).  1 Data are means ± standard deviations, except for number of days since rain, flame length, and fireline intensity and 2 weather conditions were monitored, each 30 min during the burn, by a portable weather station (Kestrel 4500) positioned at 2 m above the ground.

Prescribed Burning Treatment
In both plantations, the PB was applied in March 2014 with the main aim to reduce the fire hazard by reducing surface fuel loadings and interrupting fuel continuity. Either study area was divided into the following two plots about 0.50 ha each: treated with PB (burned) and untreated (control). The operative conditions during the treatment in terms of weather conditions, fuel load (as dry weight), and fire behaviour are reported in Table 1. The arrival time of the fire was assessed by using 9 K-type thermocouples (0.4 mm in diameter) positioned in the litter and 9 just below the fermentation layer in P. pinea plantation, and 7 thermocouples in the litter and 6 under the fermentation layer in P. pinaster plantation. From each thermocouple we extracted maximum temperature and the residence time (s) above 100 • C (Table 1).
PB was more intense in P. pinaster than in the P. pinea plantation (<150 kW m −1 vs. <50 kW m −1 ), because of more abundant fine litter fuel (14.8 ± 1.7 vs. 6.4 ± 1.1 t ha −1 ). This resulted in higher temperatures and more prolonged residence times above 100 • C in the litter and F-layer (Table 1). In both plantations the litter layer thickness had not yet recovered to the prefire levels, at least up to 18 months after the treatment [44].

Sampling Protocol and Laboratory Measurements
Samples of the fermentation layer (F-layer) and underlying 5 cm soil (S-layer) were collected in 6 subplots (40 × 40 cm) of burned and unburned plots. In each subplot, all fermentation layer was collected; successively, five cores of S-layer were collected by a cylindrical sampler (diameter 6 cm and height 5 cm) and then mixed to obtain a homogeneous sample. The sampling was performed at 3 h, and 33, 89, 189, and 363 days after PB in the P. pinea plantation and at 3 h, and 32, 87, 182, and 371 days after PB in the P. pinaster plantation.
Limited to the first sampling (3 h), the weight of fermentation layer (after drying at 75 • C) was determined.
At each sampling time, the F-layer and S-layer (the latter sieved at 2 mm mesh) were analyzed for water content, total and extractable organic C, total microbial biomass, DNA amount, fungal mycelium, microbial activity, and bacterial genetic diversity. In addition, pH, total and mineral N (as ammonium and nitrate) were determined for S-layer samples.
Water content of F-and S-layer samples was measured by gravimetric method [45]. S-layer pH was determined on a water suspension of air-dried soil (1:2.5 ratio) using a calibrated electrode (Hanna Instruments mod. HI1230) [46]. Total N was determined on dried (75 • C) and pulverized soil with a NCS Elemental Analyzer (Thermo FlashEA 1112). Ammoniacal and nitric N contents were determined on fresh soil stored at 4 • C until measurement. Mineral nitrogen was extracted from soil with 0.5 M K 2 SO 4 (1:5 soil: extractant) and measured by selective electrodes for NH 4 + -N (ORION, Mod. 9512BNWP) and NO 3 − -N (ORION, Mod. 9707BNWP) [47]. Total organic carbon (C org ) of dried F-and S-layer samples was measured by humid digestion in 0.33 M K 2 Cr 2 O 7 [48,49]. Fresh F-layer and S-layer, stored at 4 • C until measurement, were used to determine total microbial biomass C (C mic ), fungal mycelium amount, and respiration activity. Moreover, the following two indexes of microbial metabolism were calculated: metabolic quotient (qCO 2 : g CO 2 -C kg −1 C mic h −1 ) [50], indicating the activity level of microbial community, and C mineralization rate (CMR: g CO 2 -C kg −1 C org h −1 ), representing the fraction of organic C mineralized in the time unit. In addition, fungal fraction of microbial C (C fung % C mic ) was evaluated.
Total microbial biomass was assessed with the chloroform-fumigation extraction method [51]. Organic carbon, extracted with 0.5 M K 2 SO 4 from chloroform-fumigated samples and not fumigated samples, was measured by chemical digestion with 0.066 M K 2 Cr 2 O 7 . On the basis of the organic carbon (C org ) content in extracts of fumigated and non-fumigated samples, microbial biomass carbon (C mic ) was calculated [51]. This measure also provided data on the extractable organic C (C ext ) that was obtained from non-fumigated samples. The abundance of soil microorganisms was also deducted from total DNA, considered a robust indicator of microbial biomass [52]. DNA was extracted from 0.25 g of F-or S-layer samples (stored at −20 • C until use) with the FastDNA SPIN Kit for soil (Bio 101 Inc). Two subplots (3rd and 4th) of burned and unburned plots were used for assessing DNA content and bacterial diversity (see below) in four over five sampling times (3 h, and 33, 189, and 363 days in P. pinea plantation and 3 h, and 32, 182, and 371 days in P. pinaster plantation).
Total fungal mycelium was determined with the membrane filter [53]. The length of hyphae was determined with the intersection method [54]. The mass of mycelia was evaluated from the average values of cross section (9.3 µm 2 ), density (1.1 g mL −1 ), and dry mass of the hyphae (15% of the wet mass) [55]. From fungal mycelium and C mic data, the fungal fraction of microbial C (C fung % C mic ) was calculated [56].
The microbial activity was assessed as potential respiration, by measuring the CO 2 evolved from samples by gas chromatography (TRACE™ Ultra Gas Chromatograph) [57], modified. Samples were incubated in standard conditions (25 • C, 55% water-holding capacity, in the dark) for 1 h before analysis [58].
The metabolic quotient (qCO 2 : g CO 2 -C kg −1 C mic h −1 ) and C mineralization rate (CMR: g CO 2 -C kg −1 C org h −1 ) were calculated from respiration, microbial biomass (C mic ), and total organic C (C org ) data.
The bacterial genetic diversity was evaluated by 16S rDNA PCR-DGGE on DNA extracts [59]. PCR amplification was performed using the primer set GC 968f/UNI 1401r [60,61] in the conditions described by Agnelli et al. [62]. PCR products were quantified on 1% agarose gel in 1X TBE buffer using Low DNA Mass Ladder 100 bp as a marker. Denaturing-gradient gel electrophoresis (DGGE) was performed using the DcodeSystem (Universal Mutation Detection System, Biorad). Operative conditions were 6% polyacrylamide gel and 35% to 50% denaturant gradient of urea and formamide (100% denaturant contained 7 M urea and 40% formamide). The gel was run at 60 • C under 200 V for 4 h. DGGE profiles of the bacterial community were analyzed with ImageJ ver.1.33, in order to obtain band number and band intensity for each lane [59]. The bacterial diversity was assessed in terms of richness (band number), Shannon index (H) and evenness index. The Shannon index was calculated by the equation, H = −Σ(n i /N) × ln (n i /N), where n i is the intensity of each band and N the sum of the intensities of all bands in the same lane [63]. The evenness index was calculated as the ratio between Shannon index and natural logarithm of richness, the latter corresponding to the maximum value of Shannon index for the observed band number.

Statistical Analysis
A normality test (Kolmogorov-Smirnov) was applied to datasets before parametric tests; the data were transformed by log10 when not normally distributed [64]. For each variable, two-way ANOVA (followed, if required, by Bonferroni test) was applied using treatment (burned or control) and sampling time (T) as factors. The significance level of the Bonferroni test, for burned vs. control comparisons was p ≤ 0.01, according to Bonferroni method [64]. Moreover, the significance of differences between two considered plantations, for each variable, was assayed by t-test.
Correlations among parameters were assayed by Pearson's coefficient (n = 60 throughout except genetic data, for which n = 16).
All statistical analyses were carried out using the software SigmaPlot12.

Results
In both plantations, PB consumed litter but not the F-layer. In the P. pinaster plantation, an increase in the mass of the F-layer was observed (839 ± 373 g m −2 before PB vs. 1231 ± 331 g m −2 after PB), suggesting that partially mineralized litter was transferred to this layer. The same did not occur in P. pinea plantation (1140 ± 277 g m −2 before PB vs. 1135 ± 367 g m −2 after PB), indicating that litter had been entirely mineralized.
PB generally did not affect soil chemical characteristics in P. pinea plantation, however, a significant increase was found on the overall study period in N tot , NH 4 + -N, and NO 3 − -N concentrations ( Table 2 (A)), and, limited to 89 and 363 days, in extractable C (C ext ) concentration of the F layer (Table 3 (A)).
In this plantation, positive correlations were found between NH 4 + -N concentration and soil DNA and fungal mycelium (respectively, r = 0.50 and r = 0.68, p < 0.05) and between NO 3 − -N concentration and soil respiration, qCO 2 , and CMR (respectively, r = 0.45, r = 0.29, r = 0.463, p < 0.05). In P. pinaster plantation, burning treatment produced a slightly but significant increase in soil pH, and a significant decrease in NH 4 + -N content, limited to 32 d and 182 d ( Table 2 (B)), water content in both the F-and S-layers (Table 3 (B)). In either plantation, PB treatment did not affect the organic C concentration in F-or S-layer (C org , Table 3).
Microbial variables also showed different responses to PB in the two plantations. A comparison of burned and control plots, showed that no negative effect was generally found in P. pinea plantation ( Figure 2); however, a significant decrease in fungal mycelium was found in the S-layer, 3 h after the treatment ( Figure 2F). The observed changes in respiration and qCO 2 ( Figure 2I,K) were not clearly referable to fire treatment because they were sometimes higher in the control and sometimes the reverse. In P. pinaster plantation a negative effect of PB on some microbial variables was observed in both the Fand S-layers ( Figure 3), but this effect was temporary. In particular, both layers showed a significant decrease in total C microbial biomass (C mic ) and fungal mycelium immediately after burning (3 h) and generally up to 32 days thereafter ( Figure 3A,B,E,F). A decrease in respiration ( Figure 3I) and C mineralization rate (CMR, Figure 3M) was observed in F-layer at 3 h and 371 days after burning. The burned plots also showed a slight reduction in S-layer DNA content ( Figure 3D). Unlike other microbial variables, the fungal fraction of C mic and qCO 2 of the S-layer were higher in burned plots than in controls up to 32 d after treatment ( Figure 3H,L).        From the analysis of the DGGE profiles or bacterial genetic fingerprints of the F-layer ( Figure 4A,C) and S-layer ( Figure 4B,D) of considered plantations, no significant differences were generally observed for bacterial richness, Shannon index, and evenness between burned and unburned plots (Table 4), except for richness in P. pinea S-layer, which was slightly lower in burned plot.

283
Richness, Shannon, and evenness indexes (Table 4) were generally significantly (p < 0.05) higher 284 in P. pinaster than in P. pinea plantation (except for soil richness), in line with Cmic and respiration 285 trend (p < 0.05); in contrast, DNA amount was lower (p < 0.05) in the F-layer of P. pinaster plantation 286 than in P. pinea plantation (Figures 2 and 3). P. pinaster plantation also showed significantly (p < 0.01) 287 higher values of Corg and total N content, in the S-layer, and of Cext and water content, in both F-and 288 S-layer as compared with P. pinea plantation (Tables 2 and 3). 289 Two-way ANOVA analysis of the whole dataset showed that the sampling time was more 290 critical than the burning treatment towards most variables considered (Tables 2 and 3, Figures 2 and  291 3), except for genetic ones that were generally unaffected by either of them ( Table 4). The water 292 content significantly changed with sampling time in both F-and S-layer (Table 3) and this in turn 293 affected most other investigated variables. Indeed, extractable C (Cext), microbial biomass, fungal 294 mycelium of both F-and S-layer were generally positively correlated with the water content (0.30 < r 295 < 0.91, p < 0.05, in P. pinea plantation, and 0.57 < r < 0.89, p < 0.001, in P. pinaster plantation). Moreover, 296 limited to the F-layer, total organic C (Corg) was positively correlated with water content in both 297 plantations (r = 0.32, in P. pinea; r = 0.26, in P. pinaster, p < 0.05) and respiration was correlated with 298 water content only in P. pinaster plantation (r = 0.33, p < 0.01). In addition, soil mineral N (NH4 + -N and 299 NO3 − -N) was generally positively correlated with soil water content in both plantations (0.27 < r < 300 0.49, p < 0.05). 301 Richness, Shannon, and evenness indexes (Table 4) were generally significantly (p < 0.05) higher in P. pinaster than in P. pinea plantation (except for soil richness), in line with C mic and respiration trend (p < 0.05); in contrast, DNA amount was lower (p < 0.05) in the F-layer of P. pinaster plantation than in P. pinea plantation (Figures 2 and 3). P. pinaster plantation also showed significantly (p < 0.01) higher values of C org and total N content, in the S-layer, and of C ext and water content, in both F-and S-layer as compared with P. pinea plantation (Tables 2 and 3).
Two-way ANOVA analysis of the whole dataset showed that the sampling time was more critical than the burning treatment towards most variables considered (Tables 2 and 3, Figures 2 and 3), except for genetic ones that were generally unaffected by either of them ( Table 4). The water content significantly changed with sampling time in both F-and S-layer (Table 3) and this in turn affected most other investigated variables. Indeed, extractable C (C ext ), microbial biomass, fungal mycelium of both F-and S-layer were generally positively correlated with the water content (0.30 < r < 0.91, p < 0.05, in P. pinea plantation, and 0.57 < r < 0.89, p < 0.001, in P. pinaster plantation). Moreover, limited to the F-layer, total organic C (C org ) was positively correlated with water content in both plantations (r = 0.32, in P. pinea; r = 0.26, in P. pinaster, p < 0.05) and respiration was correlated with water content only in P. pinaster plantation (r = 0.33, p < 0.01). In addition, soil mineral N (NH 4 + -N and NO 3 − -N) was generally positively correlated with soil water content in both plantations (0.27 < r < 0.49, p < 0.05). Table 4. Mean (± standard deviation) values of richness, Shannon, and evenness indexs in control and burned plots of P. pinea (A) and P. pinaster (B) plantations at different times after burning. Results of two-way ANOVA for burning and sampling time were reported at the bottom of A and B, respectivey. N.S., nonsignificant. Bacterial diversity variables (richness, Shannon index, and evenness) generally did not change during the study period and were not correlated with microbial growth (DNA amount, microbial C, and fungal mycelium) and metabolism (respiration, qCO 2 , and CMR) in P. pinea plantation. On the contrary, significant correlations among bacterial diversity and other microbial variables were observed in the P. pinaster plantation. In particular, richness and Shannon index were positively correlated with C mic , in the F-layer, and with respiration and CMR in the S-layer (0.52 < r < 0.65; p < 0.05). Surprisingly, richness in the F-layer was negatively correlated with respiration, qCO 2 , CMR, and fungal fraction of C mic (−0.59 < r < −0.77, p < 0.05).

Effect of PB on Soil Microbial Community
As far as we are aware, this is the first study investigating the effects of prescribed burning (PB) on microbial biomass, activity and bacterial genetic diversity in pine plantations in a Mediterranean area. By comparing burned and control plots at each sampling time, we observed that the PB treatment in the P. pinea plantation did not affect the microbial community of either F-or S-layer (except for richness in soil). In contrast, a decrease in microbial biomass (C mic ), DNA amount, fungal mycelium, respiration, and C mineralization rate (CMR) was observed in the P. pinaster plantation generally until 32 d after treatment (in both layers or only one of the two). The more marked effect of prescribed burning in the P. pinaster plantation is consistent with higher maximum temperature and longer residence time above 100 • C during the treatment, most likely reflecting higher fire intensity due to more abundant litter fuel as compared with the P. pinea stand. Alcañiz et al. [30] reported that the effects of prescribed burning on soil biological properties depended on fire characteristics (intensity, residence time, and severity). Similarly, other authors observed that fire effect on soil microoganisms relied on temperatures at which they were exposed [35,65]. A decrease in fungal mycelium [13] and ATP content [66] with increasing fire severity has also been observed after experimental fires in an Italian Mediterranean maquis.
The effect of PB on soil microbial community in the P. pinaster plantation could be also due to indirect fire effect, such as the significant decrease of soil water content and C ext concentrations in Fand S-layers of burned area, generally not observed in P. pinea plantation.
In this study, PB treatments were found to affect the soil microbial community to a lesser extent than that of sampling time. Indeed, two-way ANOVA showed that most variables depended on sampling time, only a few being also affected by fire treatment. Crucially, temporal variation due to rainfall seasonality largely overlapped the effect of PB in either plantation. As a matter of fact, soil biological activity in Mediterranean areas is primarily affected by yearly fluctuation in water availability [67]. In line with this, most microbial variables considered in this study were positively correlated to the water content.
Our results are consistent with former research showing that repeated PB in a P. halepensis plantation in Southern Italy had only short-term or no effects on chemical and microbial properties of the fermentation layer and the 5 cm soil layer underneath, with a superimposed predominance of the water factor [31]. Likewise, no or minimal effects of PB on soil physical and chemical properties were observed in a short-grass plain in Texas [37], a P. pinaster plantation in Portugal [68] and a P. palustris forest in South Caroline [69].
Total organic C did not change significantly in burned plots as compared with controls, while F-layer extractable C (limited to 89 and 363 days) and soil total N and mineral N significantly increased in burned plots of P. pinea plantation, where fire intensity was relatively low. This possibly reflects incorporation of nitrogen-enriched ashes produced by combustion at low temperature [30]. An increase in ammoniacal N has been reported in a Texas grassland after PB [37]. Benefits from higher C and N contents could balance negative effects of fire, thus, accounting for little or no damage to the microbial community. Indeed, in P. pinea plantation, soil DNA and fungal mycelium were positively correlated with NH 4 + -N concentration; moreover, soil respiration, qCO 2 , and CMR were positively correlated with NO 3 − -N concentration.

Interrelationships among Variables
Bacterial genetic diversity was evaluated as richness, Shannon index, and evenness of bacteria after DNA extraction and amplification of a specific segment of bacterial 16S rDNA [60]. The values of DNA yield for the F-layer of P. pinea (44 to 79 mg kg −1 ) and P. pinaster (25 to 58 mg kg −1 ) plantations and for the S-layer (respectively, 13 to 51 mg kg −1 in P. pinea, and 11 to 38 mg kg −1 , in P. pinaster) mostly fell within the range (0.1 to 41.8 mg kg −1 ) reported in 2150 French soils [52], only occasionally being slightly above. The richness values found in our study systems were comparable with those reported in soils from P. nigra and Abies alba Italian forests [70] and in a P. pinaster Spanish forest with an understory of Mediterranean bushes [71].
Whereas wildfire can cause changes in the soil bacterial genetic diversity, as demonstrated in a Spanish P. pinaster forest [71], our study showed that generally PB did not affect bacterial genetic diversity in either plantation, except for richness in soil of P. pinea plantation. Because bacterial richness in the above-lying F-layer was not affected by treatment, the observed effect on soil suggests the involvement of other factors (not investigated here) that co-varied with the burning factor. Other authors found no or temporary variations in microbial community structure after prescribed burning [37,72]. On the contrary, experimental summer fire in a Mediterranean maquis caused decreases in functional diversity [56], fungal species density [73], and fungal fraction of microbial C [13], although an increase in culturable total, xerotolerant, and heat-stimulated fungi was also observed [74]. In this study, an increase in the fungal fraction of microbial C was only found in the S-layer in P. pinaster plantation and was restricted to the first month after treatment. In combination with a parallel decrease in C mic and fungal mycelium, this suggests that fire negatively affected bacteria more than fungi in the soil of the P. pinaster stand.
Unchanged genetic bacterial diversity in burned plots in P. pinea plantation is an expected result, because PB had relatively low intensity and no other microbial variable changed appreciably after exposition to fire in this system. This is consistent with the lack of effects in this plantation also on tree growth [24] and plant species richness (Esposito et al., unpublished data). The lack of effect on bacterial diversity in the P. pinaster plantation, instead, is more surprising because a higher fire intensity in this system caused significant changes in other microbial variables, generally persisting for a month after fire, and reduced microbial efficiency in C use as shown by higher values of soil qCO 2 . A significant increase in soil qCO 2 after wildfire and experimental summer fire has also been reported in Mediterranean maquis [13,75]. These observations fit well with the general tendency of microorganisms to increase CO 2 -producing catabolic activity and reduce growth-sustaining anabolism when exposed to stressing conditions [6,50,76].
The lack of PB effects on soil bacterial diversity and the temporary effects on soil microbial biomass and activity in P. pinaster plantation demostrated that also higher intensity PB had a low impact on soil. On the other hand, in this plantation PB did not affect negatively the tree growth [42] nor the plant species richness. Indeed, 20 months after PB treatment plant species richness did not change between control and burned plots (6.4 ± 1.9 vs. 6.9 ± 1.3, as measured in 78.5 m 2 plots, Stinca et al., unpublished results).
No correlation was observed in P. pinea plantation among bacterial genetic diversity and microbial biomass and activity, whereas, in P. pinaster plantation, richness, and Shannon index were positively correlated with C mic in the F-layer, and with respiration and CMR in the S-layer. On the other hand, richness was negatively correlated with microbial activity and fungal fraction of C mic in the F-layer. These contradictory relationships suggest the participation of factors other than those considered in the present study. Nannipieri et al. [8] suggested that the lack of correlation between microbial diversity and a fundamental index of microbial activity such as organic matter decomposition could be explained by functional redundancy of the microbial community. As a matter of fact, the reduction and disappearance of certain microbial entities could have little effect on soil functions if others persist that can perform the same functions.
Relatively to intra-ecosystem variability, a greater variability in bacterial genetic diversity was observed when comparing the two pine plantations (about 40 km apart) to each other, demonstrating that these indices are sensitive to variation of environmental factors (such as those site dependent). Indeed, as compared with P. pinea plantation, the P. pinaster plantation generally showed significantly higher values of richness, Shannon index, and evenness in both the F-and S-layer, which fits well with higher values of C mic and respiration. Surprisingly, an opposite trend was observed for soil DNA amount. In P. pinaster plantation, soil microorganisms could benefit from significantly higher soil organic C and total N content as well as higher C ext and water content in both F-and S-layer. This confirms the importance of water as a regulating factor for soil microbial community in the Mediterranean environment [66]. As compared with P. pinea plantation, P. pinaster plantation had a lower undergrowth plant species richness, as well as lower shrub and herbs cover and tree high, suggesting that microbial community was affected by soil chemical properties more than plant species richness and vegetation structure.

Conclusions
The application of PB for fire hazard reduction in two pine plantations located in protected areas of Southern Italy exposed to high fire risk achieved the objective of reducing the litter fuel and interrupting its continuity without causing persistent negative effects on the soil, as well as on tree growth and vegetation. In this study a temporary negative effect on microbial biomass and activity was observed in the P. pinaster plantation only, where higher fire intensity occurred as compared with P. pinea plantation. However, in both plantations the magnitude of fire-related changes was by far smaller than the spectrum of changes associated with the sampling time, confirming our first hypothesis.
Moreover, our results confirmed our second hypothesis that microbial diversity is a powerful tool that effectively integrates microbial growth and activity data in the study of fire effects on soil microbial communities. Indeed, the absence of changes in bacterial genetic diversity in burned plots suggested that fire caused no important damage to the structure of microbial community living immediately below the litter layer even after relatively intense burning in P. pinaster plantation, thus explaining the fast recovery of microbial growth and activity in the aftermath.
Our results confirm that the application of prescribed burning in the studied plantations under the prescriptions adopted (see Table 1) could be a safe practice reducing forest fire hazard without causing relevant damages.
Questions open to future research, include: (i) to verify if the response to PB observed in pine plantations can be generalized to other Mediterranean ecosystems, (ii) to test fire intensity and severity thresholds above which marked changes in soil microbial activity occur, and (iii) to know if specific bacterial or fungal species could be damaged or stimulated by PB treatment.