Panagiotis Michopoulos ·Athanassios Bourletsikas ·Kostas Kaoukis

Abstract In this study,nitrogen f ulxes or flows in litterfall,nitrogen stocks and available nitrogen in soils of two plots representing evergreen broadleaf and Bulgarian fir forests were assessed.Both plots are in good quality sites and for this reason,the litterfall quantities and nitrogen fluxes were relatively high.The woody litterfall flux of nitrogen was significantly higher in the fir forest than in the evergreen broadleaf one.The total nitrogen stock was higher in the soil under the fir forest.However,the percentage of the available nitrogen (ammonium+nitrates) was significantly higher in the upper 20 cm soil layer of the evergreen broadleaf forest in spite of the higher average C/N ratios in the foliar litterfall of the broadleaf forest and insignificant difference of the C/N ratios in all soil layers of the two ecosystems.The microclimatic conditions (higher soil temperatures in the evergreen broadleaf forest) is probable possible cause for this difference.The available nitrogen in the soils and its retranslocation from senescing leaves cover the nitrogen requirements of trees.It is hypothesized that trees may also take up nitrogen from deeper soil layers.

Keywords Maquis·Fir·Nitrogen·Litterfall fluxes·Soil

Introduction

Among all elements,the limitation of nitrogen (N) is one of the most critical factors affecting tree growth (Vitousek and Farrington 1997).This is not due to low quantities of this element but to its restricted availability in many soils.Because of the important role nitrogen has in plant growth,there is considerable information with regard to its cycling or flow in forest ecosystems (Keeney 1980;Cole and Rapp 1981;Carlyle 1986;Dannenmann et al.2006;Rennenberg et al.2009).The nitrogen paths to forest soils are throughfall,stemf low (both from atmospheric nitrogen deposition in rain/snow) and litterfall.Litterfall is the most important source for all plant nutrients with the exception of potassium for which throughfall is the primary source.Litterfall affects soil respiration,microbial biomass,and total carbon and nitrogen (Sayer 2006;Xu et al.2013) and all its fractions are subjected to decomposition once they reach the soil.The rate of decomposition depends on the forest species,climate and the existence of microorganisms capable of carrying out decomposition (Vesterdal et al.2013).The effect of species is usually related with the percentage of lignin,which is a refractory compound resistant to breakdown.Some conifer species have higher lignin concentrations than broadleaves,although some of the latter approach or even exceed that of some conifers (Rahman et al.2013).Another indicator related to decomposition and microbial activity is the C/N ratio in the foliage and later in foliar litterfall.Jílková et al.(2020) found that the foliage C/N ratio was a good indicator of microbial respiration (negatively related),suggesting that it ref lected unmeasured leaf properties that determine microbial respiration.The C/N ratio is assessed not only in foliage and litterfall but in soils as well because the ratio in the forest floor is related to plant diversity,organic carbon accumulation (Zhou et al.2019) and nitrate leaching(Gundersen et al.1998).The type of ecosystem affects the soil type.Therefore,nitrogen availability in soils varies considerably across ecosystems and soil types (Rennenberg and Dannenmann 2015).Among the ecosystem factors,the forest species is the major one determining the C/N ratio in forest floors and mineral soils,followed by the ecoregions of forests (Cools et al.2014).The aim of this study was to compare the nitrogen fluxes in litterfall,nitrogen stocks and availability of nitrogen in forest floors and mineral soils,and the C/N ratios in litterfall and soils of two different ecosystems in two different ecoregions.The ecosystems chosen were those of a maquis or scrubland vegetation and a mountainous fir forest.These ecosystems are evergreen but differ in terms of species and regional climate.The maquis plot in this study has not been subjected to grazing or to intense insect attacks (in contrast to many maquis forests) and generally it is situated a good site quality.The fir plot has not been subjected to logging (in contrast to many fir forests).Therefore,it could be said that both plots represent undisturbed ecosystems thriving on sites of good quality.However,even if not all conditions are represented,both plots can serve as reference points for similar ecosystems.This is important for nutritional studies in impoverished forest stands.The null hypothesis is that the two ecosystems do not differ in terms of the aims of the study.

Materials and methods

Study sites

Both sites are part of the Intensive Monitoring survey of the ICP Forests Network (UN-ICP-Forests 2020).The maquis plot is situated in the area of the city of Amf ilochia in western Greece at an altitude of 360 m Maquis and is a scrubland vegetation of the Mediterranean region primarily of leathery,broad-leaved evergreen shrubs or small trees (Dallman 1998).The aspect is northeast and the slope is moderately steep (19%).It is 0.27 ha and enclosed in a catchment of 117 ha.The average annual rainfall is 1213 mm,and the stand age is 60—100 years.The canopy closure is approximately 1.4 as there is overlapping of canopies of the various species.The vegetation is an evergreen forest of holm oak (Quercus ilexL.),strawberry tree (Arbutus unedoL.),Kermes oak (Quercus cocciferaL.),tree heath (Erica arboreaL.),march heath (Erica verticillataL.) and green olive tree (Phillirea latifoliaL.) The soil was developed on sandy flysch,and is deep,well drained and classified as a Haplic Luvisol (FAO-UNESCO 1988).The pH (measured in water)ranged from 6.1 to 6.2.

The fir plot is situated on the Timfristos Mountain in central Greece at an altitude of 1170 m.It has an area of 0.27 ha and is enclosed in a catchment of 147 ha.The average annual rainfall is 1433 mm (average value 1973—2017).The vegetation consists of an even-aged Bulgarian fir (Abies borisii regisMattf.) stand in good health approximately 100 yearsold.The ground vegetation is mainly ferns (Pteridium aquilinumL.),shrubs (Rubus hirtusWaldst.&Kit.),herbs(Sanicula europaeaL.,Geranium lucidumL.,Geranium rotundifoliumL.,Luzula forsteri(Sm).DC.) and plants from the familyGramineaesuch asMelica unif loraRetz.andBrachypodium sylvaticum(Huds.) Beauv.The soil was developed on sandy flysch;it is deep and classified as Humic Alisols (FAO 1988).The pH (also measured in water) ranged from 5.3 to 6.5.

Litterfall collection and pretreatment

Ten litter traps (cylindrical plastic buckets of 0.242 m2),were placed systematically along a straight line 10 m apart in the plot,approximately 0.50 m above ground.The bottom of each bucket was perforated to allow rain or snowmelt to drain.A plastic net was put at the bottom of the traps to avoid loss of small material.The litterfall sampling for the fir plot covers the period 2009—2017 and for the maquis plot 2013—2017.Collection of litterfall was done monthly or less frequently,depending on the accessibility of the area due to snow.A composite sample was formed and transported to the laboratory for analysis.Three litterfall fractions were foliar,woody (twigs,bark debris) and the rest consisted of flowers,fruits,lichens,mosses,insect frass,pollen and whatever did not belong in the first two categories.The litterfall fractions were dried at 80 °C for 48 h and weighed.Each fraction was ground by a special blade mill for plant material and stored for analysis.All results were expressed in 105 °C dry weight.

Soil samples collection and pretreatment

Soil samples were collected systematically in the summer of 2007.Inside each plot along three lines 25 m apart,six soil pits,5 m apart,were excavated.From each pit,the samples collected were the L and the FH horizons using a 15 cm × 15 cm frame,and mineral soil layers 0—10,10—20,20—40,and 40—80 cm depths (about 2 kg from each pit).The L (Litter) is the first layer of the forest floor and the FH layer is a mixture of the F and the H horizons.In the Mediterranean ecological zone the organic matter decomposition is fast and the two horizons cannot be distinguished easily as in northern countries.There was mixture of six samples of equal volume per horizon and soil depth to form three pooled samples per horizon and depth.The bulk density of the mineral soils was measured with a cylinder of 129 cm3and the value converted to bulk density of fine earth (<2 mm) after subtracting the volume of coarse material calculated in the lab.The percentage of large stones was assessed visually in the field.All soil samples were transferred to the laboratory and air-dried.Apart from the L layer,the rest ones passed through a 2-mm sieve.For the total analysis,subsamples of all soil layers were pulverized in a ball mill and stored.All results were expressed in 105 °C dry weight of soil.

Soil temperature and moisture monitoring

These were determined because of their importance concerning nitrogen availability.Recording thermometers and moisture meters approximately 30 m from the plot centers,10 m apart,monitored soil temperature and moisture at a depth of 20 cm in both stands.This depth was chosen because it lies just below the forest floor and it is the most crucial layer for nitrogen mineralization.There were three replicates for temperature and moisture measurements,taken by sensors connected to a central logger.The values were found after averaging out the three separate recordings given by each sensor.The measurements interval was 1 h and monthly values were calculated from daily mean temperature and moisture recordings.The measurements cover the period 2012—2018.

Chemical analysis

Concentrations of total nitrogen and organic C in soils as well as carbon in litterfall were measured with a CN analyzer (Farina et al.1991).Levels of litterfall nitrogen were determined by digestion with H2SO4and by the Kjeldahl distillation method (Benton Jones 2001).Ammonium and nitrate nitrogen in sieved soils was extracted with a 2 M KCl solution (Benton Jones 2001).Concentrations of ammonium was measured by the Kjeldahl method (Benton Jones 2001),whereas nitrates were determined with a UV spectrophotometer at a wavelength of 220 nm (Ben Dechil et al.2014).

Soil texture analysis

Texture analysis of sieved soil was carried out with the pipette method (Miller and Miller 1987).The calculation of the clay fraction was considered necessary to explain the amounts of total nitrogen.

Calculations and statistical analysis

Annual fluxes of nitrogen in the litterfall fractions were calculated by multiplying the concentrations of nitrogen in the litterfall with the respective amount of litterfall for each selective period and fraction and summing the amounts to find the flux for each year and fraction.Litterfall samples were pooled in each selection period,as it is customary to pool samples in environmental studies,i.e.soils,plant material or water to ensure sufficient material for analysis and/or for economic reasons.In this study,litterfall pooling did not affect the statistical analysis because the aim was to find the litterfall variability for each forest type in time when the time unit was a year,and then compare the means of the litterfall quantities based on the variability in time.The same procedure was carried out by Kouki and Hokkanen(1992) to find the variability in litterfall of a mature Scots pine forest over 20 years of observation.The variability of the annual litterfall would have been found if each litterfall trap had been sampled separately.However,in this study there would not be enough litterfall material for analysis in each selection period.

Stocks of total nitrogen in mineral soil layers up to a 80-m depth were calculated by multiplying the measured N concentrations in each layer by layer depth and bulk density of the fine earth (<2 mm).The volume of large stones,estimated by visual observation,was subtracted from the soil volume used for calculations.The nitrogen pools in the L and FH horizon were calculated by multiplying the weight of the material collected by the nitrogen concentration.

The coefficients of variation for all parameters were calculated as the percentages of the standard variations over the means.

Annual litterfall fluxes for masses (total weight),nitrogen,and the C/N ratios in foliar litterfall,seven per plot(2013—2019),were compared by analysis of variance(ANOVA) for the three litterfall fractions.As the annual litterfall of mature stands is influenced by factors such as climate,insect attacks,fruiting and other factors,it was decide that a simple ANOVA would be applied and not a repeated measures one or a paired comparison.

Concentrations of total nitrogen,clay percentages,percentages of available nitrogen (ammonium+nitrate) and the C/N ratios were compared separately for each soil layer(three replicates per layer).

There was also a comparison of total nitrogen stocks in soils between the two plots.In order to calculate the conf idence intervals in the two bars of Fig.1,the standard deviations of the average nitrogen concentrations in each soil layer were taken into account (Miller and Miller 1988).

Fig.1 Total N stocks (kg ha−1) in the soils of the two plots

Results

The average total quantities of the litterfall fractions(Mg ha−1a−1) over time are shown in Table 1 and the average nitrogen fluxes in Table 2.Similar quantities in total litterfall were found in foliar and the rest litterfall for both plots.However,the quantity of the woody litterfall was significantly higher in the fir stand and the same significant difference was observed with nitrogen fluxes (Table 2).The coefficient of variation was low for the foliar litterfall in both quantities and nitrogen fluxes but not for the other two litterfall fractions.Table 2 also shows the average C/N ratio in the foliar litterfall.The litterfall in the maquis plot had significantly higher ratios.

Table 1 Average weight (Mg ha−1) of the three litterfall fractions in the two experimental plots in the period 2013—2019

Table 2 Average annual litterfall flux (kg ha−1 a−1) of nitrogen and C/N ratios in the two experimental plots in the period 2013—2019

Table 3 shows forest floor and mineral soil properties and their statistical comparisons.Comparisons of nitrogen concentrations in the mineral soil follow the pattern of the clay content for depths below 10 cm.The C/N ratio did not differ significantly in any layer,in either the forest floor or mineral soil.Table 3 also shows the total nitrogen stocks in the soils.As the statistical comparison was carried out for nitrogen concentrations,there was no comparisons between the nitrogen stocks.The low amounts of nitrogen in the 40—80 cm layer in the maquis plot was due to the high percentage of stones in that depth.

The concentrations of ammonium and nitrate nitrogen in absolute amounts in the soils of the two plots are shown in Table 4.There was no statistical analysis for these parameters because the researchers considered the comparison between the percentages of available nitrogen more important than the one between concentrations.However,the difference in magnitude between these two nitrogen forms(mg kg−1) and total nitrogen (mg g−1) is significant.

Comparison of the total nitrogen stocks shows that the fir plot has significantly higher levels (Fig.1).In contrast,the maquis plot had higher percentages of available nitrogen in the FH,0—10 and 10—20 cm layers (Fig.2).

Fig.2 Percentages of available N in the soil layers of the two plots

Monthly temperatures and moisture values in the upper 20 cm soil layers in the two plots show that temperatures are higher in the maquis plots and the moisture values greater in the f ri plot (Figs.3 and 4).The significance will be discussed in the following section.

Discussion

Annual litterfall varies according to weather conditions(Finer 1996) .In this study,the quantity of foliar litterfall of both species had a relatively small variation in time (13%)for both species (Table 1).Kouki and Hokkanen (1992)found a 40% coefficient of variation in the foliar litterfall of a mature Scots pine forest over 20 years of observation.Our research was based on 7 years,a period shorter than their study.However,as both stands are mature and there have not been extreme weather events,the values in this study may be considered average foliar litterfall quantities.The other two litterfall fractions varied considerably for both total quantities (litterfall masses) and nitrogen fluxes (Tables 1 and 2).It seems that the dependence of the production of fruits,f lowers,and insect frass on environmental factors is stronger than the foliar litterfall.The woody litterfall values varied more than the foliar litterfall values.Possibly twig fall may be related to the remaining litterfall parts (f lowers,fruits) by creating structures to support them on the trees.

Table 3 Clay content (%),total N (mg g−1),N stocks(kg ha−1) and values of the ratio C/N in the soils of the two experimental plots

Table 4 Ammonium and nitrate N (mg kg −1) in the soils of the two experimental plots

The nitrogen fluxes are considered rather high in the fir stand in this study.Hansen et al.(2009) found 29 and 36 kg ha−1a−1of N fluxes in the foliar litterfall,ofPicea sitchensis(Bong) Carr.andPicea abies(L.) Karst (both about 40 years old) in Denmark,whereas in the woody litterfall they reported 1.6 and 2.0 kg ha−1a−1,respectively.High values (62 kg ha−1a−1) of nitrogen in conifer foliar litterfall were found in a 50-year-old Douglas-f ir stand in Finland.However,the woody litterfall flux of nitrogen was only 6.8 kg ha−1a−1(Portillo-Estrada et al.2013).

Rapp et al.(1999) found 17.5 kg ha−1a−1nitrogen fluxes in the foliar litterfall of aQ.ilex,40-year-old stand which is low in comparison with this study.The difference may possibly be due to site quality.TheQ.ilexstand is situated on a rendzina calcareous and shallow soil,whereas the maquis vegetation in this study is on a deep soil with flysch as parent material.Bussotti et al.(2003) found 32.8 kg ha−1a−1,close to the value in this study),of nitrogen below a mixed stand of maquis species (Q.ilex,Q.pubsescensWilld,A.unedo) in central Italy with adequate precipitation and only 14.6 kg ha−1a−1of nitrogen on a xeric site (vegetation 40—60 years-old in both stands).The quantities of litterfall in this study are high due to good site quality.Albrektson(1988) reached the same conclusion with regard to the effect of site quality on litterfall quantities for Scots pine stands in Sweden.

The concentrations of total nitrogen,the C/N ratios and their statistical comparisons in the soils of the two plots are shown in Table 3.The nitrogen levels were significantly different below 10 cm.However,the C/N ratios did not differ in soils in spite of the difference of the C/N ratios in the respective foliar litterfall.

The higher concentrations of nitrogen in the deeper soil layers in the fir plot are due to the higher amounts of clay(Table 3).Clay minerals form stable complexes with organic matter and the latter is always associated with nitrogen.

Fig.3 Monthly temperatures (°C) in the 0—20 cm soil layer of the two plots in the period 2012—2018

Fig.4 Monthly moisture content (m3 m−3) in the 0—20 cm soil layer on the two plots in the period 2012—2018

Figure 1 shows that the total nitrogen stocks in the soils of the fir plot have higher amounts of total nitrogen.The total amount of nitrogen,mostly in organic forms,is important for forests soils as it is the only nitrogen reservoir.However,organic nitrogen must be converted to simpler nitrogen compounds.Plants take up nitrogen mainly in the forms of ammonium and nitrates.Table 4 shows the concentrations of ammonium and nitrate nitrogen in soils in mg kg−1,whereas those of total nitrogen are in mg g−1.As a percentage of the total nitrogen,the available ammonium and nitrates forms ranged from 2.4% in the mineral layers to 5.6% in the FH horizon (Fig.2).The statistical comparison showed that the maquis soil had significantly higher levels of the available nitrogen percentage for the FH,0—10 and 10—20 cm layers,whereas in the deeper layers there was no significant difference.This means that up to the 20-cm depth,the nitrogen mineralization rates are higher in the maquis plot.The C/N ratio in the foliar litterfall of holm oak was significantly higher.These differences are not unusual.Vesterdal et al.(2008) also found significant differences in the C/N ratios in litterfall of common European tree species.It would be expected that the differences in the C/N ratios in foliar litterfall might affect the ratios in soils but this did not occur.It is likely that climate plays a predominant role in both organic matter decomposition and nitrogen mineralization.Figures 3 and 4 show the monthly temperatures and moisture levels in the upper 20 cm for the period 2012—2018.Temperatures were higher in the maquis plot and moisture levels higher in the fir plot.At a global scale,climate is the best predictor,(in comparison with litter chemistry),for the decomposition constants (k-values) of litter (Aerts 1997).Guntiňas et al.(2012) found that temperature was more important than moisture for organic matter decomposition under the presupposition that xeric conditions do not apply.The effect of higher temperatures results in an increase of net nitrogen mineralization caused by a proportionally larger temperature response of gross nitrogen mineralization than immobilization (Hoyle et al.2006).

Another concern is if the available nitrogen in soils is adequate to compensate for the nitrogen lost through litterfall.Trees need to replenish the nutrients lost in above and below ground litterfall (Cole and Rapp 1981).In this study,the sums for the aboveground litterfall fluxes (all fractions) for nitrogen in the maquis and fir plots are 49.4 and 69.4 kg ha−1a−1,respectively (Table 2).Assuming that trees take up nitrogen only from the FH,0—10 and 10—20 cm layers,the total nitrogen in these layers for the maquis and the fir plots is 412 and 510 kg ha−1,respectively (Table 3).From Fig.2 the amounts of available nitrogen can be calculated according to their percentages of the total nitrogen,and these are,for the maquis and the fir plots,87.2 and 79.1 kg ha−1,respectively.The amount of the below ground“litterfall”(f ine roots,root exudates etc.) may be equal or exceed the above ground litterfall (Raich and Nadelhoff er 1989;Davidson et al.2002).If it is equal,then both stands have to replenish nitrogen with additional uptake.Fortunately,trees have the mechanism of nutrient retranslocation and the key variables that determine the amount and rate of retranslocation are the rates of nutrient uptake and growth (Nambiar and Fife 1991).Both stands are mature and growth rates are not as high as in younger ones.However,the litterfall amounts are appreciable in both stands so nitrogen retranslocation takes place at a significant pace.Another hypothesis is that trees can also take up nitrogen from deeper layers.The 20—40 cm layer has a large nitrogen reservoir (Table 3).

Conclusion

The maquis vegetation and the Bulgarian fir,ecosystems with different environments,share similarities and differences with regard to nitrogen cycling.The first similarity is that site quality is important for the total quantity of nitrogen fluxes in litterfall;the second one is that,despite the maturity of the stands,the plant species use the retranslocation mechanism for extra nitrogen uptake.One difference is that the ecoregions determine the rate of nitrogen mineralization.In both stands,the plant species have adapted to these conditions,shown from the low variability of the nitrogen fluxes in the foliar litterfall.

AcknowledgementsThe authors express their appreciation to the Ministry of Agriculture and Food,the Ministry of Environment and the European Commission,which financially sustain the programme“Effects of Atmospheric Pollutants on Forest Ecosystems”on the framework of which the current project was based,and Mrs.Ch.Mitropoulou for her help with sample pretreatment and analysis.Finally,the authors express their gratitude to the Forest Services of Karpenisi and Amf ilochia for their invaluable practical assistance with the problems encountered.