Nitrogen dynamics at undisturbed and burned

REGULAR ARTICLENitrogen dynamics at undisturbed and burned Mediterranean shrublands of Salento Peninsula, Southern ItalyMichael Dannenmann&Georg Willibald&Sebastian Sippel&Klaus Butterbach-BahlReceived:13May2010/Accepted:16August2010/Published online:10September2010#Springer Science+Business Media B.V.2010Abstract Fire is a major disturbance in shrubland ecosystems of the Mediterranean basin,with high potential to alter ecosystem nitrogen(N)stocks and N cycling.However,postfire effects on gross rates of soil N turnover(ammonification,nitrification,micro-bial immobilization,denitrification)have rarely been investigated.We determined gross rates of N turnover including nitrous oxide fluxes and dinitrogen emissions in the mineral soil of unburned and burned shrublands of Southern Italy6months after a natural fire.In soil of burned plots,both gross ammonification and gross nitrification were significantly higher than in soil of unburned plots(2.2±0.3versus0.6±0.1mg N kg−1sdw day−1for ammonification and1.1±0.1versus0.5±0.1mg N kg−1sdw day−1for nitrification).Microbial immobilization,in particular of nitrate,could not compensate for the increase in inorganic N production, therefore soil nitrate concentrations were considerably higher at the burned plots.Soil microbial biomass carbon and nitrogen concentrations were significantly lower in soils of burned plots than in soils of unburned plots.Dinitrogen was the dominant end product of denitrification and emitted at higher rates from the unburned plots than from the burned plots(0.094±0.003versus0.004±0.002mg N kg−1sdw day−1, while there was no net nitrous oxide flux(burned plots)or slight net nitrous oxide uptake(control plots). These results show that postfire patterns of gross N turnover in soil can exhibit a significant reduction of both microbial N retention and N gas losses via denitrification.Keywords Maquis.Fire.N cycling. Ammonification.Nitrification.Denitrification. Nitrous oxide.Dinitrogen.Microbial biomass.He flow soil core techniqueIntroductionFire is a major disturbance in Mediterranean ecosys-tems,with high potential to alter ecosystem N stocks and N cycling(Moreno and Oechel1995;Certini 2005;Castaldi and Aragosa2002;Knicker2007). Fire frequency may increase under future environ-mental conditions,since available regional predictions assume that air temperatures and drought event probability are significantly increasing due to climate change(Lavorel et al.1998;Piñol et al.1998).Plant Soil(2011)343:5–15DOI10.1007/s11104-010-0541-9Responsible Editor:Per Ambus.M.Dannenmann(*)Institute of Forest Botany and Tree Physiology, Chair of Tree Physiology,University of Freiburg, Georges-Koehler-Allee53/54,79110Freiburg,Germanye-mail:michael.dannenmann@M.Dannenmann:G.Willibald:S.Sippel:K.Butterbach-BahlKarlsruhe Institute of Technology(KIT),Institute for Meteorology and Climate Research(IMK-IFU), Kreuzeckbahnstrasse19,82467Garmisch-Partenkirchen,GermanyHowever,our understanding how fire may affect soil microbial N cycling in Mediterranean ecosystems is still limited.Soil microbial nitrogen(N)cycling in terrestrial ecosystems is of high ecological significance,as it regulates ecosystem N retention;N loss along gaseous and hydrological pathways which can affect atmospheric chemistry,climate change and water quality;and plant nutrient availability (Schimel and Bennett2004;Rennenberg et al. 2009).Gross N ammonification,i.e.the microbial production of ammonium(NH4+)from organic N compounds,is a key processes of soil N cycling, since free NH4+in plant-free soil is subject to two competing microbial processes and fates,i.e.nitrifi-cation to nitrate(NO3-)and immobilization into microbial biomass.After nitrification,NO3--N may also either be immobilized by soil microorganisms, undergo dissimilatory nitrate reduction to ammonium (Silver et al.2001),or may be denitrified.Via denitrification,NO3-is reduced stepwise to nitrite, the secondary greenhouse gas nitric oxide(NO),the potent primary greenhouse gas and most important destruent of stratospheric ozone(Ravishankara et al. 2009)nitrous oxide(N2O)as intermediates,and to molecular dinitrogen(N2)as the dominant end-product.Production of these N gases by denitrifica-tion leads to N loss from the ecosystem.The last step of denitrification,i.e.the reduction of N2O to N2 catalyzed by the enzyme nitrous oxide reductase,is converting reactive nitrogen back into its inert form, and hence,significantly contributes to closing the global nitrogen cycle(Galloway et al.2003). Furthermore it reduces soil N2O losses(Chapuis-Lardy et al.2007;Dannenmann et al.2008). However,the conversion of reactive N back to N2 by denitrification is thought to represent the largest uncertainty of the N cycle at all scales(Galloway et al.2004;Groffman et al.2006).Due to methodo-logical difficulties(Butterbach-Bahl et al.2002; Groffman et al.2006)reliable measurements of N2 emissions from terrestrial ecosystems are scarce which limits our understanding of the significance of the single permanent sink for reactive nitrogen, but also impedes the quantification and comprehen-sion of the denitrification process as a whole (Davidson and Seitzinger2006;Groffman et al. 2006).The latter also feedbacks on our understand-ing of microbial NO3-immobilization,as this is often calculated from the consumption of15NO3-, assuming that gaseous N losses via denitrification are not significant for the NO3-mass balance (Davidson et al.1992;Stark2000).Hence,underes-timation of denitrification probably lead to frequent overestimation of microbial NO3-immobilization in 15N pool dilution experiments.Our understanding of N ammonification,nitrifi-cation and microbial immobilization of inorganic N has significantly improved in the last decades for a wide range of ecosystems.In particular,the devel-opment and application of15N isotope pool dilution and—tracing techniques(Kirkham and Bartholomew 1954;Davidson et al.1991,1992,Stark2000; Murphy et al.2003;Booth et al.2005)facilitated a more holistic view of actual N turnover and its environmental controls compared to the more widely used determination of net rates of N turnover(Eno 1960),which confound simultaneously occurring production and consumption of inorganic N,as e.g. net nitrification is the balance of actual microbial nitrate production(gross nitrification)and microbial nitrate consumption via e.g.microbial nitrate immobilization and denitrification(Davidson et al. 1991).However,Mediterranean shrubland ecosystems are still being severely understudied with respect to gross rates of N turnover and denitrification activity and the importance of fire as a potential driver for soil N cycling has largely been ignored.It is well known that fire increases mineral N concentrations in the uppermost mineral soil(Marion et al.1991), but,as there is still extremely little knowledge on postfire effects on gross rates of ammonification, nitrification,microbial immobilization and denitrifi-cation,it remains unknown to what extent postfire increases of inorganic N concentrations are caused by direct ash input or altered inorganic N production rates.The goal of the present study was to investigate gross rates of soil N turnover(ammonification, nitrification,microbial immobilization of ammonium and nitrate as well as denitrification)at unburned and burned Mediterranean macchia shrublands. Furthermore,we aimed at the clarification of the importance of denitrification versus the other N turnover processes in the investigated ecosystem, i.e.if denitrification is insignificant as an N sink or not.Material and methodsSite characteristicsThe study site is located in Salento Penninsula, Southern Italy(18°23′17.34″E,40°18′5.70″N)at a distance of1km to the sea.The whole site area is completely flat and characterized by homogenous typical Mediterranean Macchia vegetation cover of 0.3–0.8m height.The dominating plant species are Erica australis,Rosmarinus officinalis,Pistacia lentis-cus and Myrtus communis.Mid of August2007, approximately half of the site was burned by a natural fire.The soil is a shallow Rendzic Leptosol on sandy carbonatic bedrock.The height of the densely rooted,organic matter-rich Ah layer was 4.6±0.8cm across the site.At the bottom of the Ah layer there was either a direct transition to the weathered bedrock or a scarcely rooted B layer above the bedrock.The weathered but still compact bedrock was always found at a depth of20cm.The gravel content of both the A and B horizons was moderate (ca.10%).Sampling designEnd of January2008,three unburned and three burned plots of100m2across an area of approx. 2ha were randomly selected and sampled.The distance between burned and unburned plots was 30–50m.The litter layer amounted to663±58g dry mass m−2while the ash layer at the burned plots amounted to72±6g dry mass m−2.At the sampling time,there was a herbal layer covering approximately 30%of the soil at the burned parts of the site. Furthermore,resprouting of burned shrubs had begun at the sampling date.Sampling took place at every plot at seven40*40cm spots randomly selected across the plot.First,the organic(unburned plots)or ash(burned plots)layer was quantitatively sampled and trans-ferred to plastic bags until weight determination and drying(24h at100°C)of subsamples.Subsequently, in the centre of every sampling spot,the Ah layer was sampled with three adjacent soil cores(4cm depth, 100cm3volume).Soil cores were sealed with pin-holed parafilm to facilitate gas exchange but to avoid water loss.Soil cores were stored in cooling boxes and transferred to the laboratories of IMK-IFU in Garmisch-Partenkirchen,Germany,within48h after sampling.After arrival at IMK-IFU they were stored at4°C until further processing and analysis.All samples were processed within two weeks after sampling.One intact soil core of every sampling spot was used for the simultaneous measurement of N2O and N2fluxes.The second soil core was used for the determination of gross rates of microbial N turnover after compositing and sieving samples for single plots.Also the third soil core was composited at the plot level and sieved for analysis of extractable concentrations of inorganic N,dis-solved organic nitrogen(DON),dissolved organic carbon(DOC),microbial biomass C and N,and pH values.Gross rates of ammonification,nitrificationand microbial N immobilizationGross rates of ammonification,nitrification and microbial immobilization were determined using a 15N pool dilution technique described in more detail by Dannenmann et al.(2009).We decided to use a sieved soil technique,as the stone content of the soil hampered sufficient homogenous15N injection into intact soil cores.Three days before the start of the experiment,the still intact soil samples were pre-incubated at the in situ during sampling deter-mined soil temperature(10°C).Immediately before 15N application,soil was removed out of the cores and roots,gravel and other coarse materials were removed by carefully breaking the intact soil sample portions by hand prior to sieving(5mm mesh width).Soil samples were composited for single plots.Mechanical disruption of the soil was mini-mized as far as possible.Two subsamples(230g sieved soil each)were labelled with7ml30%15N-enriched KNO3solution(for determination of gross nitrification rates)or7ml30%15N-enriched (NH4)2SO4solution(for determination of gross ammonification rates),respectively(time t0).The subsamples were spread in a thin layer and then the 15N label solution was sprayed homogenously on the samples.The amount of added N corresponded to1μg N g−1sdw.While aliquots of180g of the subsamples were transferred into six250ml plastic bottles(Carl Roth GmbH,Karlsruhe,Germany)(30g each),the residual soil was used for determi-nation of the gravimetric water content.The plastic bottles were incubated in the dark at10°C.At time t1(=t0+24h)and time t2(=t0+48h)soil in three of the bottles was extracted with1M KCl,respectively (Dannenmann et al.2006).Subsamples of the filtrate were passed through0.45μm syringe-filters and immediately frozen until colorimetrical measurement of NH4+and NO3-concentrations by a commercial laboratory(Dr.Janssen,Gillersheim,Germany).The diffusion method was used for trapping NH4+or NO3-as NH3on acid traps made of ashless paper filters(Brooks et al.1989).The14/15N-ratio of the N captured on the dried filter papers was analyzed using an elemental analyzer(EA1110, Carlo Erba Instruments,Milan,Italy)coupled to a mass spectrometer(MAT Delta Plus,Thermo Finnigan,Bremen,Germany).Gross ammonification and gross nitrification rates were calculated using the equations given by Kirkham and Bartholomew (1954).Microbial immobilization of NH4+was calculated by subtracting nitrification rates from NH4+consumption rates(Davidson et al.1992). This approach underestimates NH4+immobilization when there is heterotrophic nitrification(direct oxidation of organic substrate to NO3-).Overestima-tion of NH4+immobilization can occur by substrate-stimulation of NH4+consumption in the15NH4+ treatment,while the subtracted nitrification rate calculated from15NO3-pool dilution is not affected by substrate stimulation.Here,we tried to minimize experiment-inherent substrate stimulation by mini-mizing NH4+bel application increased ambient NH4+pools by only36%and56%in soils of burned and unburned plots,respectively.Nitrate immobilization was calculated by subtracting deni-trification rates(see below)from NO3-consumption rates.Physical and chemical soil parametersFor the determination of mineral N concentrations, 30g of unlabelled soil free of limestone and roots was extracted with1M KCl solution and analyzed for NH4+and NO3-concentrations as described above (see Dannenmann et al.2006).Soil pH values (0.01M CaCl)were measured with three subsamples of every plot by use of a combined electrode as described by Dannenmann et al.(2007).Denitrification:simultaneous measurement of N2 and N2O emissions from seven intact soil cores Dinitrogen and N2O emissions from intact soil cores were measured by use of the helium gas flow soil core method as described by Butterbach-Bahl et al. (2002)and Dannenmann et al.(2008).This method is based on the exchange of the soil and headspace atmospheres by a helium-oxygen atmosphere con-taining only25PPM N2and the subsequent simultaneous automated detection of N2O and N2 concentration changes in the headspace above the cores by use of an electron capture detector(ECD) for N2O and a pulse discharge helium ionization detector(PDHID)for N2(Fig.1).In order to facilitate the application of the method to shallow soils and in order to improve the spatial resolution of the measurements,we designed a new system for simultaneous measurement of N2and N2O from seven small soil cores.While the general setup of the system including the steering unit,automated flushing of soil cores and headspace,automated sampling and the detection technique and conditions for N2and N2O(see Fig.1)were the same as described by Butterbach-Bahl et al.(2002),two new incubation cuvettes were designed.The new incuba-tion cuvette facilitated the simultaneous flushing of seven soil cores(height4cm,100cm3volume each) via the porous porcellaine plates at the bottom of the soil cores,while the cuvette described by Butterbach-Bahl et al.(2002)contained only one soil core of20cm height and12.5cm diameter)(Fig.2). The soil cores are automatically pressed into the fittings sealed via O rings when closing the cuvette to ensure that He purge gas flow is taking place from bottom to top in the soil cores(Fig.2).Also with the new cuvette,the same huge constructive efforts were made to facilitate an extremely gastight system and hence avoid diffusion of atmospheric N2into the system(Butterbach-Bahl et al.2002), e.g.the cuvette had double sealings which were additionally purged with He(Fig.2),and He leakage tests were performed.The smaller size of the new cuvette compared to the version described by Butterbach-Bahl et al.(2002)allowed to further improve the gastightness of the system by placing the whole incubation cuvette including fittings of the gas tubings under water in a water bath.The water surrounding the cuvette is also used for the regula-tion of the incubation temperature.Based on this setup,no significant increase in N 2concentrations in the cuvettes was found during 8h when the system was run with an empty cuvette.Here,the soil cores (soil moisture 19or 21.4%sdw for burned or unburned soil samples;incuba-tion temperature:10°C for both treatments)were flushed for 72h to quantitatively remove N 2from the soil and headspace atmospheres.Subsequently,an artificial headspace atmosphere was created (80%He,20%O 2,25PPM N 2,400PPB N 2O)and the concentration change of N 2and N 2O in the two cuvettes was monitored automatically for 8h on hourly basis according to Butterbach-Bahl et al.(2002).Every sample air analysis was accompanied by 6automated calibration gas measurements at the gas chromatographs.For each treatment (burned/unburned),3measurements with 7soil cores were performed.Flux rates were calculated from the linear change in N 2and N 2O concentrations in theheadspace as described by Butterbach-Bahl et al.(2002).After every measurement,soil water content and soil dry weight of the incubated soil were deter-mined.Denitrification was calculated as the sum of N 2O plus N 2fluxes and related to a soil dry weight basis.Microbial biomass C and NMicrobial biomass C and N was determined by use of the chloroform fumigation-extraction technique (Brookes et al.1985).For this purpose,soil from seven soil cores was pooled at the plot level and sieved.Subsequently three subsamples of 30g were immediately extracted with 60ml 0.5M K 2SO 4,while three subsamples were fumigated with Chloroform vapour for 24h.Fumigated samples were extracted in a similar way like control samples.Total chemically bound nitrogen (TNb)andtotalFig.1Schematic representation of the measuring system used to simultaneously quantify N 2and N 2O emissions from seven intact soil cores.PDHID:Pulse Discharge Helium Ionization Detector;ECD:Electron Capture Detectororganic Carbon (TOC)were analyzed by use of a chemoluminescence detector for TNb analysis coupled to the TOC analyzer (Dannenmann et al.2006).Correction factors (0.54for microbial biomass N and 0.379for microbial biomass C,(Brookes et al.1985;Vance et al.1987)were applied to the difference in TNb and TOC between paired untreated and fumigated subsamples to estimate microbial biomass C and N.TOC values of the extracts of unfumigated control samples are referred to as extractable DOC concentrations.StatisticsTest for significant differences of the determined parameters between control and burning treatment were made by means of the Mann Whitney u-test using plots as statistical units (N =3).Also corre-lation analysis was performed using plot means from both burned and unburned treatments.All statistical analyses were performed with SPSS 10.0(SPSS Inc.,Chicago,USA)and Microcal Origin7.0.Fig.2Newly designed incubation vessel for simultaneous measurements of N 2and N 2O emissions from seven small intact soil cores (4cm height,100cm 3volume each)after purging with He/O.All soil cores are purged from bottom to top with He/O mixture.Double outside sealings are used which are additionally purged with He.Furthermore,the wholeincubation vessel is placed for purging and measuring under water in a water bath to finally reach gas tightness.After three days of purging,fully automated hourly measurements of N 2and N 2O concentrations in the headspace were conducted over 8h.The system contains two vesselsResults Soil parametersHalf a year after burning,the remaining ash and charcoal layer at the burned plots was one order magnitude smaller compared to the organic layer at the control plots (Table 1).Mineral soil pH was 7.5and slightly (7.53versus 7.45)but significantly higher at the burned plots (Table 1).Soil moisture content in the mineral soil was significantly lower at the burned plots as compared to the control plots.Both microbial biomass C (Table 1)and N (Fig.3)were significantly lower in the Ah horizon of the burned plots.However,no difference in the microbial C:N ratio was found between burned and unburned plots.In contrast to microbial biomass,extractable DOC was found to be significantly higher at the burned plots (Table 1).Gross rates of N turnoverGross ammonification was nearly four times higher at burned plots than at unburned plots (Fig.3).However,extractable soil NH 4+concentrations were approximately only 50%higher.Microbial NH 4+immobilization was threefold higher at the burned plots than at the control plots (Fig.3).Gross nitrification was more than twofold larger at burned plots than at control plots (Fig.3).Extractable soil NO 3-concentrations in the mineral soil of the burned plots were thirteen times the concentrations of soil NO 3-in the mineral soil horizon of the control plots (Fig.3).At burned plots,the amounts of extracted soil NO 3--N equaled the amounts of extracted soil NH 4+-N concentrations.In contrast,soil NO 3-con-centrations were considerably lower than soil NH 4+concentrations at unburned control plots.Microbial NO 3-immobilization was not significantlydifferentFig.3N turnover [mg N kg −1sdw day −1]and N pools [mg N kg −1sdw]in the Ah layer of unburned and burned plots.SON:soil organic nitrogen.A:gross ammonification;B:gross nitrification;C:microbial NH 4+immobilization;D:microbial NO 3-immobilization;E:N 2O flux;F:N 2flux.Different indices indicate significant differences between control and burned plots.Errors represent standard errors of the meanTable 1Soil parameters litter/ash mass parameters are given for the Ah layer.Errors represent standard errors of the mean calculated from N =3plots.DOC:dissolved organic carbon.MBC:microbial biomass carbon;MBN:microbial biomass nitrogen.Different indices indicate significant differences between burned and unburned plots litter/ash mass[g m −2]soil moisture [%sdw)pHDOC[mg C kg −1sdw]MBC[mg C kg −1sdw]MBC/MBN [ratio]control 617±86a 21.4±0.5a 7.45±0.02a 134±3a 1844±112a 15.3±0.2burned72±6b19.0±0.4b7.53±0.03b160±14b1380±143b15.9±1.31Soil parameters for control and burned plots.Except for litter/ash mass parameters are given for the Ah layer.Errors represent standard errors of the mean calculated from N =3plots.DOC:dissolved organic carbon.MBC:microbialbiomass carbon;MBN:microbial biomass nitrogen.Different indices indicate significant differences between burned and unburned plotsbetween control and burned plots and overall several fold lower than microbial NH 4+immobilization.Relative N retention,i. e.(microbial NH 4+immobilization +microbial NO 3-immobilization)/(gross ammonification +gross nitrification),was significantly lower at the burned plots than at the control plots (Table 2).This was caused in particular by an increase in gross nitrification which was not outbalanced by a concomitant increase in microbial NO 3-immobilization.Soil N 2O and N 2measurements revealed that dinitrogen was the dominant end product of denitri-fication both at control and burned plots (Fig.3).However,N 2emissions were more than one magni-tude larger at unburned control plots (94±2μg N kg -1sdw day −1)than at burned plots (4±2μg N kg −1sdw day −1).At the control plots,there was significant net uptake of N 2O of approx.1μg N kg −1sdw day −1,while at the burned plots N 2O fluxes were not significantly different from zero.At control plots,N 2O uptake was approximately two orders of magni-tude smaller than N 2emission.Due to the dominance of N 2as the end product,denitrification rates equalled N 2emissions.For burned plots,denitrification amounted to less than 1%compared to the other processes of N turnover.However,for the unburned plot denitrification was a significant sink process for microbial N cycling.Here,denitrification was on average 16%of ammonification,20%of nitrification and 98%of microbial NO 3-immobilization at the unburned control plots (Table 2).Correlation analysesSoil microbial biomass N was negatively correlated with gross ammonification (R =−0.91,p =0.01)and gross nitrification (R =−0.83,p =0.04)and net soil-atmosphere N 2O flux (R =−0.84,p =0.04)(i.e.positively correlated with net N 2O uptake rate)but positively correlated with soil water content (R =0.93,p =0.006),and N 2emission rate (R =0.94,p =0.005).Gross ammonification was positively correlated with independently determined soil NH 4+concentrations of unlabelled soil (R =0.885,p =0.02)as well as gross nitrification with soil NO 3-concentrations (R =0.98,p <0.001).DiscussionFire effects on gross rates of N turnoverFire has been shown to potentially alter a wide range of physical,biological and chemical soil parameters like soil organic matter quanity and quality (variable effects of fire across soil horizons),pH values (increase),nutrient availability (increase),and soil microbial biomass (decrease)(Marion et al.1991;Castaldi and Aragosa 2002;Certini 2005;Knicker 2007).The recovery of these effects is mainly depend-ing on plant recolonization (Certini 2005).However,little information is available on postfire effects on actual gross rates of N turnover in Mediterranean soils.In this study we show that 6months after burning,when vegetation re-growth already had started,gross ammonification and gross nitrification were considerably larger in the Ah horizon soil of the burned plots than in soil of unburned plots.As microbial immobilization,in particular of NO 3-,could not compensate for the increase in inorganic N production,inorganic N concentrations were higher and relative microbial nitrogen retention were smaller at burned plots.Since soil moisture values at burned plots were significantly lower as compared to un-burned plots (Table 1),the stimulation of microbial N turnover must have been due to the increased availability of organic substrates and mineral N due to burning of the vegetation (Andersson et al.2004a ,b ;Knicker 2007).This interpretation is supported by our measurements on higher DOC concentrations inretention:(microbial NH 4+immobilization +microbial NO 3-immobilization)/(gross ammonification +gross nitrification).Errors represent standard errors of the mean Relative N retentionDenitrification/ammonification Denitrification/nitrification Denitrification/NO 3-immobilization control 0.96±0.11a 0.16±0.02a 0.20±0.02a 0.98±0.65a burned0.66±0.12b0.002±0.001b0.003±0.002b0.002±0.001bTable 2Relative importance of microbial immobilization and denitrification versus inorganic N production.Relative Nretention:(microbial NH 4+immobilization +microbial NO 3-immobilization)/(gross ammonification +gross nitrification).Errors represent standard errors of the meanthe mineral soils at burned plots(Table1).Our findings on higher inorganic N concentrations in the uppermost mineral soil horizon are in agreement with earlier fire studies in Mediterranean shrublands (Marion et al.1991;Castaldi and Aragosa2002). Besides increased N turnover and reduced microbial immobilization also lower plant competition for mineral N at burned plots may have additionally contributed to the increase in mineral N.It has been assumed that net N ammonification and nitrification rates in Mediterranean shrublands are low because of the quality of the typical sclerophyllus leaf and because of leaching of allelopathic compounds from plants(Scalbert1991;Gallardo and Merino 1992;Castaldi et al.2009).As fire may destroy allelopathic compounds,the increased gross rates of N turnover at the burned plots observed here,could also be caused by a release of inhibiting plant effects on microbial N turnover(Castaldi and Aragosa2002).Only little studies investigated fire effects on gross rates of soil N turnover while,to our knowledge,there is no study conducted in a comparable ecosystem like investigated here.Bastias et al.(2006)reported minor but significant reduction of gross ammonification by 16%and gross nitrification by12%three weeks after a fire in the soil of a wet sclerophyll forest of Australia,subjected to high frequencies of prescribed burning.LeDuc and Rothstein(2007)did not observe significant effects of wildfire neither on gross pro-duction nor on immobilization of both ammonium and nitrate3–6years after burning of a jack pine forest in Michigan,USA.Anderson and Poth(1998) explained increased NH4+concentrations in Brazilian cerrado soils within the first2months after experi-mental fires by a stimulation of gross ammonification while gross nitrification was suppressed by burning. In contrast,addition of wildfire-produced charcoal to soil sampled in ponderosa pine forests of Montana, USA,strongly promoted gross nitrification,probably due to absorption of phenolic compounds which inhibited gross nitrification(DeLuca et al.2006).It remains unclear whether such variable responses of gross N turnover to fire are caused by variable responses of N cycling across ecosystems,different fire intensities or-frequencies and time elapsed between the burning and sampling events,or simply result from limited temporal resolution,which is characterizing almost all studies on gross rates of N turnover.In our study,the microbial biomass C and N pools were positively correlated with denitrification, i.e.smaller at burned plots but negatively correlated with gross rates of ammonification and nitrification. Microbial biomass is responsible for both production and consumption of inorganic N.Furthermore,it can serve as a substrate for N ammonification itself following microbial dieback due to drought events (Borken and Matzner2008).Therefore,the relation-ships between microbial biomass and rates of N turnover may be variable in time.The observed reduction in soil microbial biomass at the burned plots may be explained by generally more extreme environmental conditions at the burned plots,as the missing shadowing effect of vegetation as well as the dark ash may have lead to higher temperature fluctuations and quicker drying of the soil at the burned plots.However,lower microbial biomass may also be interpreted by retarded long-term recovery of soil microbes after fire(Castaldi and Aragosa2002),e.g. in association with a decreased rhizodeposition of labile C compounds by roots.However,larger DOC concentrations at the burned plots(Table1)do not support the latter hypothesis.Furthermore,the similar microbial C:N ratios in burned and unburned soil (Table1)do not indicate a fire-induced shift in microbial community composition,i.e.a promotion of fungi with a higher C:N ratio at the expense of bacteria with a lower C:N ratio.Still,there could have been an altered abundance of functional microbial groups which was not reflected in the microbial C:N ratio.Despite both gross rates of ammonification and nitrification as well as soil NO3-concentrations were higher in soil of burned plots,denitrification rates were considerably larger at unburned control plots(Fig.3).This may—analoguously like fire effects on microbial biomass—be explained by significantly decreased soil water content at burned plots(Table1),given that denitrification is a predominantly anaerobic process(Conrad1996). However,these differences in soil moisture were low,i.e.soil moisture was approximately10%lower at burned burned plots than at control plots only (Table1).Furthermore,soil moisture was at compa-rably low level both at control and burned plots (21.4and19.0%of soil dry mass).Lower denitrifi-cation at burned plots could be also a consequence of reduced rhizodeposition of labile C compounds,。