Introduction

Prescribed fire is a common site preparation method in forest management. It is usually used to meet several objectives: reduction of fire hazard, facilitation of planting, provision of an environment favorable to seedling growth, reduction of understory vegetation competition, and elimination of disease or insect problems (Sun and others 2011; Certini 2005). As a powerful modifier of the environment, fires have a profound and long-term impact on nutrients cycles in forest ecosystems (Wan and others 2001). Besides the reduction or elimination of aboveground biomass, soil physical, chemical, and biological properties are affected to a greater or lesser extent depending on the severity and duration of fire (Sun and others 2011).

Nitrogen often limits primary productivity in natural ecosystems, the dynamics of N pools and its transformation associated with fire thus is critical for forest management. Generally, fire in forests decreased total ecosystem N by volatilizing N contained in wood, leaf, and forest floor (Knoepp and Swank 1993). However, the soil mineral N and N transformations may response differently. The increase of soil NH4-N after fire has been observed in many studies, mainly due to the deposition of volatilized N and its condensation in the cool soil layer, the increased soil N mineralization rate, and the leaching of N from forest floor into soil after fire (Wan et al. 2001; Knoepp and Swank 1993; Koyama and others 2012). Nitrogen mineralization, a major process supplying mineral N to plants in terrestrial ecosystems, is a microbial process (Wang et al. 2010b). The response of soil N mineralization and nitrification to prescribed fire varied by fire intensity, time after fire, and vegetation type (Wan and others 2001; Monleon and others 1997). In a pine stand, 2 years after burning, Burger and Pritchett (1984) found a decrease in N mineralization rate compared with the control plots. However, Adams and Attiwill (1986) found that fire increased N mineralization and the amount of mineralizable N in Eucalyptus spp., and a lack of fire effect on net N mineralization was reported for oak–pine stands in North Carolina (Knoepp and others 2004). These conflicting results point out the variability of microbial N processes response to fires.

The subtropical region of China has large areas of plantations, with most of them as lumber plantations (Wang and others 2010b). To 2010, about 3.68 million ha of Eucalyptus spp. were planted in southern China (China Eucalypt Center, 2010). Prescribed fire is a common site preparation practice in lumber plantation planting in southern China. However, it is still unclear whether this practice promotes any significant changes in soil properties and N transformations in these plantations. In this study, we examined the effects of prescribed burning on soil properties and N transformations in two vegetation types (Shrubland and Eucalyptus plantation). The objective of this study was to compare soil properties and N transformations among different burning plots to evaluate the potential of prescribed fire on recuperate degraded land in southern China. We hypothesized that prescribed fire would increase soil nutrients availability due to the reduction of understory layer and the direct release of nutrients in the fire.

Methods

Study Area

The experiment was conducted at the Heshan National Field Research Station of Hillyland Forest Ecosystem (HNFRS, 60.7 m a.s.l., 112°50′E, 22°34′N), a subtropical region of southern China. The soil type is Acrisols developed from sandstone, with a pH of about 4.0 (Wang and others 2010b). The climate of the region is typically subtropical monsoon (Fig. 1), with a hot and rainy (growing) season (from April to September) and a cool and dry (dormant) season (from October to March). Climax vegetation of this region is subtropical evergreen broad-leaved forest.

Fig. 1
figure 1

The monthly rainfall and mean soil temperature in the experiment site in 2007, Heshan, Guangdong, China

Full size image

Experimental Design

A completed randomized design was used in this study. The detailed site preparation and the whole experiment design were described by Sun and others (2011), Wang and others (2010b). Briefly, there were four treatments in this study: burned shrubland and unburned shrubland, burned E. urophylla plantation and unburned E. urophylla plantation. The prior vegetation type of the four treatments was secondary shrubland after logging all trees (Pinus elliottii). In March 2005, the logging residues and the shrubs were left untouched in the unburned plots but were burned in the burned plots (Sun and others 2011). E. urophylla seedlings were planted in both treatment plots at a spacing of 2 × 3 m, while shrubland was restored naturally without any disturbance (Wang and others 2010b). Each treatment has three replicate plots with each plot of 1 ha. Totally, there were 12 plots in this study. In October 2006, vegetation investigation was conducted in each plot (Table 1). In the Eucalyptus plantations, E. urophylla was the dominant tree species (~50 % coverage). In shrubland, the dominant tree species were Trema tomentosa and Litsea cubeba (Table 1). Burned plots always had higher understory coverage (combined the data in shrub layer and herb layer:~75 %) than the corresponding unburned plots (~35 %), with the dominant species being Dicranopteris dichotoma and Miscanthus sinensis in all plots. Soil samplings were conducted from June and December 2007, which represented the rainy and dry seasons in this region (Wang and others 2010b).

Table 1 Vegetation description of shrubland and Eucalyptus plantation at Heshan station
Full size table

Soil N Transformations

In June and December 2007, an in situ soil-core technique (Raison and others 1987) was used to estimate soil net nitrogen mineralization. Briefly, in each replicate plot, nine subplots were randomly located. In each of these points, two PVC (polyvinyl chloride) tubes of 4.6 cm in diameter and 15 cm in height were hammered into the soil to a depth of 10 cm. Before sampling, forest floor litter was removed. One of the two tubes from each subplot was retrieved and sent to the lab. The other tube, with a lid on the top, was retained in situ for 1 month, 30 days, before being retrieved.

All soil cores were transported to the lab immediately and stored at 4 °C, and extracted for mineral N within 48 h. Before extraction, each of the nine cores from the same plot was manually divided into two layers (0–5 and 5–10 cm) and soils from the same section (from the same plot) were pooled and mixed thoroughly. Visible roots and stones were removed manually. Twenty grams of fresh soil from each layer were extracted with 100 ml of 2 M KCl solution (1:5 ratio) and filtered (Shuangquan quantitative filter paper 202#). Concentrations of ammonium (NH4 +) and nitrate (NO3 −) in the extraction solution were determined by a flow injection autoanalyzer (FIA) (Lachat Instruments, USA), ammonium by the salicylate-nitroprusside method and nitrate by sulfanilamide colorimetry after the Cd-core reduction to nitrite. Soil moisture was determined by weight loss after oven drying at 105 °C for 24 h. Bulk density (dry soil) was calculated based on soil weight in all tubes and soil moisture. Net N mineralization was calculated as the increase in ammonium plus nitrate N between the initial soil sample and the incubated sample, while net nitrification was the increase in nitrate and net ammonification was the increase in ammonia.

Soil General Properties

Soil chemical properties (i.e., soil pH, organic matter, total N, total P, soil exchangeable cations) were determined using the initial soil samples obtained in June 2007. All soil samples were air-dried and passed through a 2-mm sieve. Soil pH was measured in a 1:2.5 mixture of soil:deionized water. Soil exchangeable K+, Na+, Ca2+, and Mg2+ were extracted with 1 M NH4Ac, following the guidelines of Liu and others (1996) and measured by ICP (Perkin Elmer, USA). Soil available P was extracted with Bray-2 solution (Bray and Kurtz 1945) and determined by the molybdate blue colorimetric method. Soils for analyses of total N (TN) and organic matter were grounded to pass through 0.25-mm (60 mesh) sieve. Total N concentration was determined by micro-Kjeldahl digestion followed by salicylate-nitroprusside colorimetric determination on the Lachat FIA. The measured total N thus only included organic N and ammonium. Soil organic carbon (SOC) was determined by the wet combustion method (Walkley and Black 1934), with SOM calculated as with Van Bemmelen’s factor, following the guidelines of Liu and others (1996).

Statistical Analyses

In this study, vegetation type (Eucalyptus vs shrubland) and burning treatments (burned vs unburned) were the main factors. Thus, a two-way analysis of variance (ANOVA) was used to analyze the effects of vegetation types and burning treatments on soil variables in each soil layers. A LSD post hoc test after a one-way ANOVA was used to test the difference of above variables between each treatment in each soil layers. The homogeneity of the data was tested by Levene’s test, and exchangeable K was log transformed before analysis due to the inequality of variance. All analyses and computations were performed in SPSS 18.0 (SPSS Inc., USA) and Excel 2003 (Microsoft Corp., USA) software.

Results

Soil General Properties

Although there was no statistically difference on soil organic matter (SOM) in 0–5 and 5–10 cm layers between burned and unburned sites (P = 0.46 and P = 0.12, respectively), the mean values of SOM in unburned plots were always 3–8 g/kg greater than those in burned plots in each vegetation type (Table 2). Burning also did not greatly change soil total N (TN) concentration in 0–5 cm layer (P = 0.46), but marginally changed TN in 5–10 cm layer (P = 0.10). In 5–10 cm layer, burned shrubland had the greatest TN concentration (0.46 ± 0.04 g/kg), which was significantly greater than the value in unburned shrubland (0.38 ± 0.02 g/kg). Soil total P (TP) was greater in unburned plots in both soil layers (P = 0.03, P = 0.00 for 0–5 and 5–10 cm, respectively). On average, burned plots had 30–40 % lower of TP than conspecific unburned plots in both soil layers. Soil bulk density was not affected by fire. Soil pH was marginally changed by fire (P = 0.11 for 0–5 and P = 0.09 for 5–10 cm, respectively). Burned plots always had greater pH values than conspecific unburned plots (Table 2). Prescribed fire also greatly increased soil moisture in both soil layers (P < 0.001 for both soil layers).

Table 2 Soil general properties in shrubland and Eucalyptus plantation at Heshan Station, South China
Full size table

Soil Exchangeable Cations

Due to the large variation, burning did not change soil exchangeable K concentrations in either soil layer. However, the mean values of exchangeable K+ showed that burned plots had higher K concentrations than unburned plots (Table 3), especially in the 0–5 cm layer. For instance, burned shrubland (93.51 ± 45.91 mg/kg) had nearly two times greater K+ concentration than unburned shrubland plots (47.65 ± 3.72 mg/kg) in the surface layer. Other soil exchangeable cations (i.e., Na, Ca, and Mg) also did not vary greatly among vegetation types or burning treatments in either soil layers (Table 3).

Table 3 Soil exchangeable cations in shrubland and Eucalyptus plantation at Heshan Station, South China
Full size table

Soil Available P and Inorganic N

Soil available P did not vary greatly among vegetation types or burning treatments in either seasons or layers (Table 4). However, the seasonal variation of soil available P was obvious (Fig. 2a, b). In the rainy season, the average soil available P concentration was over 2 mg/kg in the 0–5 cm layer, while in the dry season, this value has greatly declined, with <1 mg/kg in the both soil layers (Fig. 2b). The surface soil layer always had greater available P concentrations than the sub-layer.

Table 4 The P value in the ANOVA analysis of soil available nutrients and N transformations (n = 3)
Full size table
Fig. 2
figure 2

Concentrations of soil available P, soil extractable NH4-N and NO3-N under different treatments in the rainy and dry seasons of 2007. Data are illustrated as means; Error bars represent 1 SE, n = 3. In each layer, one-way ANOVA was used to detect difference among forest types, and bars sharing the same superscript (lowercase letters for 0–5 cm soils, and uppercase letters for 5–10 cm soils) were not significantly different at P = 0.05 (LSD). EU eucalyptus plantation unburned, EU eucalyptus burned, SU shrubland unburned, SB shrubland burned

Full size image

Burning treatments greatly affected soil extractable NH4-N in both soil layers in the rainy season (Table 4, P = 0.02 for 0–5 layer, and P = 0.07 for 5–10 cm layer). In this season, burned plots always had greater NH4-N concentrations than the unburned plots. In the dry season, vegetation types became the main factor affecting soil NH4-N concentration; shrublands always had greater NH4-N concentrations than Eucalyptus plantations in the 0–5 (P = 0.14) and 5–10 cm (P = 0.04) layers in this dry season (Fig. 2d).

In contrast to soil NH4-N, soil NO3-N concentration did not differ among vegetation types or burning treatments in the rainy season (Table 4). However, the multi-comparison results (LSD) have shown that the NO3-N in 0–5 cm layer of burned shrubland plots was the highest among the four treatments, and significantly greater than the value of EU plots (P < 0.05, Fig. 2e). In the dry season, vegetation types (P = 0.02) and burning treatments (P = 0.07) affected or marginally affected soil NO3-N concentrations in the 0–5 cm layer. The multi-comparison found that burned shrubland treatments still had the highest concentrations, and was significantly greater than other three treatments (P < 0.05, Fig. 2f).

Nitrogen Transformations

Soil ammonification rate was significantly different among burning treatments in 0–5 cm layer in the rainy season (P < 0.001, Fig. 3). Soil ammonification rates under unburned plots were all greater than the conspecific burned plots (P < 0.05, Fig. 3). Vegetation types also greatly affected soil ammonification rate in 0–5 cm soil layer in the rainy season (P = 0.01, Fig. 3). For instance, unburned Eucalyptus had greater rate of ammonification than the unburned shurbland in the surface layer (P < 0.05, Fig. 3). In the dry season, soil ammonification rate were negative in all plots and did not show any difference among vegetation types or burning treatments (Table 4; Fig. 3).

Fig. 3
figure 3

Soil net ammonification, nitrification, and N mineralization rates in the four experiment treatments in Heshan station in 2007. Data are illustrated as means; Error bars represent 1 SE, n = 3. In each layer, one-way ANOVA was used to detect difference among forest types, and means sharing the same superscript (lowercase letters for 0–5 cm soils, and uppercase letters for 5–10 cm soils) were not significantly different at P = 0.05 (LSD). EU eucalyptus plantation unburned, EU eucalyptus burned, SU shrubland unburned, SB shrubland burned

Full size image

Vegetation types or burning treatment did not affect soil nitrification rates in either soil layers or seasons (Table 4). However, the seasonal pattern of nitrification rate was obvious (Fig. 3). In the 0–5 cm layer, the rainy season nitrification rate was over 3 mg N kg−1 soil month−1, while in the dry season, the nitrification rate was negligible (Fig. 3).

The patterns of soil N mineralization among vegetation types (P = 0.14) and burning treatments (P < 0.001) were similar as that of soil ammonification (Table 4). Net N mineralization rates were high in the rainy season, but negative (net immobilization) in the dry season. Ammonification dominated the process of N mineralization. In the rainy season, more than 70 % of mineralized N was NH4-N. Soil ammonification rates were negative in the dry season, which was similar to the trend of N mineralization (Fig. 3).

Discussion

Burning Effects on Soil Properties

The effect of fire on soil C and N are highly dependent on fire type (i.e., prescribed fire vs wild fire) and time since the fire (Johnson and Curtis 2001). It is common to observe losses of organic C from the surface mineral soils during and shortly after fire, followed by gains in soil organic C over a period of years thereafter. However, wildfire and prescribed fire perform differently on soil C and N in a longer term. Intense fire or wildfire always reaches a very high temperature, while prescribed fire does not. The high temperature could change the chemistry of soil C and had a long-term effect on soil C and N cycles. For example, Rovira and others (2012) observed decades reduction of soil C after wild fire. In this study, we observed no significant effect of the prescribed burning on soil organic C or total N content. The reason might be the low intensity of the prescribed fire that resulted in only slight effect on soil C and N stock. In China, prescribed fire is set as site preparation of afforestation when all trees of the site have been clearly logged, and just residues and shrubs remains on site. Limited load of organic materials on the floor lead to low intensity of prescribed fire. In addition, the prescribed fire would have reduced soil C pool during the burning. However, such effect might not last for long. Two years allowed for recovery of C and N pool to some extent, especially in the vegetation reintroduction sites (Certini 2005).

Phosphorus is an important nutrient in subtropical plantations as it always limited the productivity of ecosystems. In this study, soil total P was significantly reduced after burning in the both vegetation types. This finding is consistent with the reports of many previous studies (Duguy and others 2007; Marafa and Chau 1999; Silvana Longo and others 2011). It has been shown that immediately after fire occurrence, there is an increase in phosphorous levels that tend to decrease with time (Silvana Longo and others 2011; Certini 2005). Galang and others (2010) have proved that fire could convert the organic soil P to inorganic P, the sole form of P available to biota and to leaching. In this study, we thus assumed that soil available P would be increased immediately after fire, however, due to the quick uptake of vegetation and serious leaching in 2 years, the available P pool decreased to unburned plots level, and further caused the decline of soil total P pool.

There was no significant increase in soil bulk density after prescribed burning. In conifer forests, it has been reported that severe or repeated burning lead to higher soil bulk density (Agee 1993), and such effects are typically less evident after low severity burning (Agee 1993; Boerner and others 2009). Our findings was similar to the results showed by Boerner and others (2009), who have observed the effect of prescribed fire on soil properties in 12 sites in US, and also found no significant increase in soil bulk density. One factor contributing to lack of such an effect by prescribed fire may be the fire intensity. Compared to wild fire, prescribed fires usually have lower severity (González-Pérez and others 2004), because the timing of prescribed burning can be manipulated such that fires occur when soil is moderately moist (Boerner and others 2009), and only a small fraction of biomass is burned as prescribed fire is usually used to destroy only part of aboveground biomass.

Soil pH usually increased by soil heating as a result of organic acids denaturation (Certini 2005). Ulery and others (1993) found that the topsoil pH could increase as much as three units immediately after burning. Arocena and Opio (2003) also observed higher soil pH and exchangeable K, Na, Ca, and Mg in burnt plots compared to unburned plots 2 years after prescribed fire in sub-boreal forest. In this study, soil pH was marginally increased by burning in both vegetation types. We also observed an increase of soil exchangeable K (although not statistically significant). The correlation analysis indicated that soil pH was positively associated with exchangeable Mg and Ca (data not shown). Potassium, Na, Ca, and Mg are major components of ash and rendered soluble. After prescribed fire, the high Ca, Mg, and K contents of ash can increase the pH of the soil through displacing the H and Al ions adsorbed on the negative charge of the soil colloids (Arocena and Opio 2003).

Burning Effects on Soil N and its Transformations

Soil inorganic N (NH4-N and NO3-N) can be easily taken up by plants and thus played a vital role in plant growth (Wan and others 2001). Previous studies in conifer stands have demonstrated an increase of NH4-N and NO3-N after fire (Covington and others 1991; Covington and Sackett 1992). In this study, we observed a significant positive effect of burning on soil NH4-N in the rainy season, although the soil total N was not affected by fire. This result thus was consistent with what has been reported in many other studies (DeLuca and Zouhar 2000; Wan and others 2001; Boerner and others 2009; Knoepp and Swank 1993). In a meta-analysis of the effects of fire on N, Wan and others (2001) interpreted the increase of NH4-N as being the result of a combination of liberation from organic matter degraded during the fire and increased N mineralization due to altered microclimate, soil pH, and soil microbial activities.

While burning substantially increased soil NH4-N, negative response of net N mineralization to burning was observed in our study. This result contrasted with those of many studies that have demonstrated increase in net N mineralization after fire (Wan and others 2001). In Australian Eucalyptus plantations, Adams and Attiwill (1986) found that fire increased N mineralization and the amount of inorganic N. However, Burger and Pritchett (1984) observed a decrease in N mineralization rates 2 years after burning in a southern pine stand in USA. Increase in N mineralization are often attributed to the alteration of organic matter by fire which render it more susceptible to microbial activity, and to changes in microclimate (Boerner and others 2009). Prieto-Fernández and others (1998) found that soil microbial C and N were greatly reduced after wildfire, and this effect lasted for 4 years after burning. Sun and others (2011) quantified the soil microbial activities and community structure in our site, and found a significant decrease of soil bacterial and fungal biomass under burned site. As N mineralization is a microbial process, the reduced soil microbial activities might account for the low N mineralization rate.

Thus, the increase in NH4-N that we observed after burning must have been result of other factors independent of the N mineralization process. In this study, the land slope is 20°–30°, and the vegetation investigation in Oct. 2006 (Table 1) showed that burned site had much lower understory vegetation coverage in comparison with control. Our recent study has found that understory played an important role in regulating soil nutrients availability (Wang and others unpublished). Zhao and others (2011) in a nearby plantation had observed an increase of soil NH4-N after understory removal. We thus assumed that the increase of soil NH4-N in burned site may caused by the reduced uptake of aboveground vegetation after fire.

Soil ammonification largely predominated over nitrification in both the unburned and the burned soils, which is consistent with the findings in a acid Humic Cambisol (Prieto-Fernandez and others 1993). In nearby young plantations, Wang and others (2010b) also found that ammonification was the dominant N mineralization process, with limited nitrification. However, in tropical forest, the nitrification usually is the dominant process in N mineralization (Chapin and others 2002). In old growth plantations of this region, nitrification rate was about 90 % of net N mineralization rate (Wang and others 2010a). Compared to old growth forests, soils in this study were less fertile, with lower SOM, TN, and available N (Table 1; Fig. 2). The ecosystem thus was highly limited by N availability. This result indicated that soil microbes would intend to immobilize NH4-N rather than to nitrify it in this N-limiting soil.

The seasonal variation of soil N transformations in this study was large. The rate of N transformation in rainy season was high, while in the dry season the rate was negligible (Nitrification) and even negative (Ammonification). Seasonal changes of soil N transformation has been reported by many previous studies (Wang and others 2010a, b; Smith and others 1998; van der Krift and Berendse 2001; Zhu and Carreiro 2004), and the seasonal pattern of temperature and moisture should be responsible for this variation in N transformations. The rapid reduction of N transformations in the dry season can be explained by the altered temperature and water availability, which directly regulate soil microbial activity (Yan and others 2008). Wang and others (2010b) have observed a significant relationship between soil moisture and N mineralization in a nearby site, indicating the regulation of N mineralization by soil moisture in this region. Sun and others (2011) have compared the soil microbial community in the burned and unburned sites of this study, and found that soil total PLFAs (representing soil microbial biomass and activity) in the rainy season were 2–3 times higher than those in the dry season. It thus highly possible that reduced rainfall in the dry season (Fig. 1) caused lower soil water content, which limited the activities of soil microbes, and further reduced soil N transformation rates.

Conclusion

Prescribed fire is a common site preparation practice in subtropical area of China. Our results showed that prescribed fire of pre-afforestation influenced soil variables, many of which, however, would not last for over 2 years, including SOM and Total N. Two years later after the fire, soil pH and NH4-N were still all higher in the burned site compared to unburned control. Soil total P was reduced 30–40 % largely as a result of burning, which might be a detrimental factor for plantation growth. Soil N mineralization rate was significantly lower in burned site, but not was nitrification rate. The reduced understory vegetation coverage after burning may be responsible for the higher soil NH4-N in burned site. This study highlights that a better understanding the effect of prescribed burning on soil nutrients cycling would be beneficial to afforestation and provide a critical foundation for management decision in southern China.