Panagiotis Michopoulos·Kostas Kaoukis·George Karetsos·Theodoros Grigoratos·Constantini Samara

Abstract The fluxes of masses and the nutrients Ca,Mg,K,N,P and S were determined in the litterfall of two adjacent forest ecosystems of Hungarian oak(Quercus frainetto L.)and European beech(Fagus sylvatica L.)in a mountainous area of northeastern Greece in 2010-2015.The foliar litterfall for both species reached about 70%of the total litterfall,and was significantly higher from the other two fractions(woody and rest litterfall).The fluxes of masses and nutrients were compared between ecosystems for each fraction separately.Only one significant statistical difference was found,that of K in the woody litterfall.In addition,the stocks of masses and nutrients were calculated in the forest floors and mineral soils of the two ecosystems.Likewise,the stocks of nutrients in the forest floors and mineral soils were compared between ecosystems.In the L horizon of the forest floors,statistical differences,as a result of species effect,were found for the stocks of Ca and N.In the FH horizons,the masses and all the nutrient stocks differed significantly,as the beech plot had much higher quantities of organic matter and nutrients.These higher quantities were probably due to low soil temperatures(microclimate)and high acidity in the beech plot(species effect)that slowed down decomposition.In the mineral soils,the propagation of random error derived from random errors of the individual soil layers was an important factor in the statistical comparisons.Because of the soil acidity in the beech plot,the stocks of exchangeable base cations were significantly higher in the oak plot,whereas the other nutrient stocks did not differ.

Keywords Beech·Oak·Litterfall·Forest floor·Mineral soil·Nutrients

Introduction

Nutrient cycling in forests involves various paths of nutrients which when expressed in time they are called fluxes. Litterfall is probably the most important one(Hansen et al.2009;Prescott 2002;Carnol and Bazgir 2013) with the exception of potassium for which throughfall can be equally important(Parker 1983).Litterfall comprises leaves,flowers,fruits,twigs,branches and generally,whatever falls on the ground.It serves as a substrate for microorganisms and in this way it is related to soil respiration and organic carbon(Sayer 2006;Hansen et al.2009).Its importance has been established in forest ecosystems(Bray and Gorham 1964;Vitousek et al.1995;Berg and Meentemeyer 2001).The productivity of a forest(whether natural stand or plantation)depends to a large extend to the maintenance of the quantity and quality of litterfall(Zhou et al.2015).At global scale,climate is the main factor influencing litterfall production.Chun-jiang et al.(2003)quoted data for Eurasian forests showing that more than half of annual litterfall occurred in tropical and subtropical forests,one-third in boreal forests and one-fifth in temperate ones.Increasing temperature and variability in precipitation,because of climate change,will result in altered aboveground productivity(Knoepp et al.2018).At regional scale,climate effects are modified by topography,differences in soil types,forest species and age and/or disturbances and changes in stand structure(Barnes et al.1998;Bray and Gorham 1964;Inagaki et al.2004).

Forest floors and minerals soils are the nutrient sinks of forests.They are not static but dynamic.As they accept litterfall,all factors affecting the quantity and quality of litterfall affect the organic matter and nutrient stocks of both of them.The difference in tree species is one of them.Many studies have documented the species effect on biogenic(e.g.carbon and nitrogen)and lithogenic(e.g.aluminum and calcium)chemical species in soil(Binkley and Giardina 1998;Augusto et al.2002;Vesterdal et al.2008).Knowledge of the effects of the tree species is closely related to forest management and choice of species for reforestation.For example,there is some evidence that soils(forest floor and mineral soil)under pure European beech(Fagus sylvatica)stands store less organic C than soils under pure Norway spruce(Picea abies)in Germany(Prietzel and Bachman 2012).

The aim of the present work was to investigate the species effect and more specifically to find and compare the masses and nutrients in litterfall fluxes(in most of its fractions)and nutrient stocks in forest floors and mineral soils of two adjacent ecosystems, Hungarian oak and European beech thriving in the same watershed and soil of the same parent material.While the European beech is a well-known species,the Hungarian oak(Quercus frainetto)is not despite its expansion to southeastern Europe(parts of Italy,the Balkans,parts of Hungary,Romania)and Turkey.This oak species,together with other oaks,suffers from drought and high temperatures in southern Italy(Gentilesca et al.2017)and Greece(Kitikidou et al.2015).In our work,the Hungarian oak stand,as mentioned below,is in a good site quality with a high rain height so it can serve as a baseline for future studies.

The present work has the advantage that the site quality and general climate are approximately the same for both tree species.The more the commoner the abiotic factors,the more probable is to find differences ascribed to species.On the other hand,the microclimate differences that exist in the two forest stands can have an influential role.The forests under consideration have been monitored since 1995 as a part of the International Co-operative Programme on Assessment and Monitoring of Air Pollution Effects on Forests(UN-ICP-Forests)and a multitude of variables are measured in certain periods.

Materials and methods

Site description

The experimental plots of oak and beech under consideration grow in a mountainous watershed,260 ha in size,located in central eastern Greece(Fig.1)at an elevation range of 740-1420 m.The watershed is underlain by mica schist.The climate is inland Mediterranean with a mean annual precipitation of 1804 mm derived from the period of 1978-2010 and measured at approximately an elevation of 1150 m(Baloutsos et al.2013).The average temperature for 1997-2017 was 12.4°C measured at an altitude of 750 m.

The oak plot has an altitude of 740 m,an area of 0.26 ha,an average slope of 15%and northeastern aspect.The main forest species is the Hungarian oak(Quercus frainetto)which forms an uneven aged stand 30-70 years old with a canopy closure 0.9-1.0.The ground vegetation is dense and covers the 80%of the ground.In the ground vegetation, Sorbus torminalis is the dominant species.There are also ferns(Pteridium aquilinum)and among herbs there is a variety of species like Melitis melissophyllum,Hieracium bracteolatum,Galium laconicum and others.In grasses,there are the species Melitis melissophyllum,Hieracium bracteolatum,Galium laconicum and others.The soil at the oak plot is classifies as Dystric Cambisol(FAO 1988).

Fig.1 Location of the experimental plot in Greece

The experimental plot of beech has an area of 0.27 ha,altitude of 890 m,an average slope of 33%and a northeastern aspect.The Fagus sylvatica stand is even-aged(110-130 years old),with an average tree height of 27 m and canopy closure of 0.8-1.0.The ground vegetation is very sparse,occupying only 5%of the ground area and consisting of the herb species: Galium retundifolium,Doronium orientale and Cyclamen graecum.The soil at the beech stand is a loam classified as a Haplic Alisol(FAO 1988).The two plots are at a distance of 720 m.

Soil temperature and moisture monitoring

Soil temperature and moisture(at a depth of 20 cm)in both stands,approximately 30 m from the plot center,were monitored by recording thermometers and moisture meters(Decagon Devices, Pullman, Washington, USA). The measurements interval was 1 h and monthly mean values were calculated from daily mean temperatures and moistures.

Litterfall collection

For litterfall collection, 10 litter traps in the form of cylindrical plastic buckets,each having a collecting area of 0.242 m2,were placed systematically along a line in both plots,approximately 0.50 m above ground at a distance of 10 m from each other.The bottom of each bucket was perforated so that rain or snowmelt water could drain out.A plastic net was put at the bottom of the litter traps to avoid loss of small material.Collection of litterfall was done monthly or for longer periods,depending on the accessibility of the area due to snow.Sampling started at the end of November and ended approximately the same month of the next year.A composite sample was formed and transported to the laboratory for analysis.Leaves,twigs,lichens,mosses were separated and dried at 80°C for 48 h.The choice of these litterfall fractions for our study was based on their amounts and decomposability.The foliar litterfall constitutes the largest part in total litterfall having a range of 53-87%(Chun-jiang et al.2003).The woody litterfall percentage can be high but the decomposition rate is low due to the lignin content of the woody material. The rest litterfall fraction can have appreciable concentrations of P and N despite its low percentage.As mentioned,the collection of litterfall was done monthly or for longer periods,depending on the accessibility of the area due to snow.For each litterfall collection the flux of each fraction was calculated taking into account the fraction dry weight and the total area of the litter traps.The annual litterfall flux(for masses and nutrients) was derived after summing up the litterfall amount of each collection for a year's period.Usually,four to five collections took place every year.

Soil samples collection

The soil samples collection was carried out by means of systematic sampling.Inside the plot,along three lines distancing 25 m from each other six soil pits were excavated.Each pit was 5 m away from each other.From each soil pit,the samples collected were the L and the FH horizons by means of a frame 15 cm×15 cm and mineral soil layers from the depths 0-10, 10-20, 20-40 and 40-80 cm.There was mixture of six samples of equal volume per horizon and soil depth(taken at random)to have three pooled samples per horizon and depth.The samples of the L horizons were weighed as they were and those of the FH after sieving.The samples of the FH horizon and mineral layers were air dried and passed through a 2-mm sieve stored for analysis to determine texture, pH, cation exchange capacity (C.E.C.) and exchangeable cations.The samples of the L horizon at their initial conditions and those of the FH horizon and mineral layers after sieving were pulverized in a ball mill for the purpose of total analysis for organic C,total N,total S and total P.Like the plant tissues in litterfall,soil samples were dried at 80°C for 48 h.

Chemical analysis

Litterfall

Ground material of litterfall samples were digested in a mixture of HNO3-HClO4in a proportion of 2:1(v:v)to determine the concentrations of Ca,Mg,K and P.The concentrations of metals were determined by flame atomic absorption spectroscopy(Perkin Elmer 3110,Wellesley,USA)and those of P by the ammonium molybdate blue method in a UV-visible spectrophotometer(Varian-Cary,Melbourne,Australia).The N concentrations were measured with the Kjeldahl method(Velp-UDK 126A,Usmate Velate,Italy)and that of S with dry combustion(CNS analyzer, Vario MAX, Elementar, Langenselbold,Germany).

Soils

The soil pH was determined by a glass electrode in a mixture of soil and 0.01 M CaCl2solution at a ratio of 1:5(v:v).The soil texture was measured by the pipette method.The organic C and total N were determined by a CN analyzer through dry combustion(Vario MAX).

Exchangeable cations in the FH and mineral soil layers were extracted with a 0.1 M unbuffered BaCl2solution and their concentrations were determined with an ICP-MS instrument(Thermo Fisher Scientific,iCAP Qc,Bremen,Germany).Cation exchange capacity(C.E.C.)was found by adding the calculated exchangeable cations.

Total concentrations of Ca,Mg,K,P and S in the organic horizons as well as P and S in mineral soils were measured by Energy Dispersive X-Ray Fluorescence(SPECTRO-XEPOS, SPECTRO Analytical Instruments,Kleve,Germany).

All results(plant tissues and soils)were expressed in oven dry weights(105°C for 24 h)

Calculations and statistical analysis

For both species,the annual fluxes of masses and nutrient amounts were calculated for three fractions of litterfall:foliar,woody(twigs,branches and tree bark)and rest litterfall(mosses,lichens,fruits and flowers).The next steps were the calculations of means and coefficients of variation(percentages of the standard deviations over the averages)for the 6 years of collections and the percentages(%)of masses and nutrients fluxes of each litterfall fraction over the total litterfall(leaves+woody+rest litterfall).

The fluxes of masses and nutrients in litterfall of every fraction were compared through a paired t test with the plots as treatments and time(1 year)as the pairing factor.

In each plot for each element and mass the percentages of each litterfall fraction over the total litterfall were compared with the analysis of variance test(ANOVA for a completely randomized design)using the least significance difference(LSD)for the means when the ANOVA was significant for at least 0.05-probability level.In this case,for each plot,the fractions were the treatments and there were six replicates for each fraction,one for every year.

The calculation of stocks of masses and nutrients for the L and FH horizons was based on masses and concentrations.The means and the coefficients of variations were calculated for the three-pooled samples.The stocks of nutrients in the forest floor(separately for the L and FH horizons)of both species were also compared with the analysis of variance test(ANOVA also for a completely randomized design)just like the percentages of litterfall fractions but in this case the plots were the treatments and the three pooled samples served as replicates for each.Before the tests were applied,the data were transformed to logarithms to be closer to normal distribution as variability was great in some cases.

The total nutrient stocks in the mineral soils were calculated by adding the stocks derived from each separate layer.The nutrient stocks were found by multiplying layer masses by concentrations.The layer masses were found by multiplying layer volumes by the fine earth bulk densities.The bulk densities of fine earth were calculated by a pedotransfer function based on organic and clay contents(Adams 1973;Rawls and Brankesiek 1985).

In order to compare the total nutrient stocks in the mineral soils of the two forest species,the standard errors were found and the confidence limits of the means were calculated for the 95%significance level.The calculation of standard errors was based on the new overall standard deviation equal to the square root of the sum of squares of the individual deviations(Miller and Miller 1988).The comparison was based on the overlapping(or not)of the confidence limits of the means.

Results

Tables 1 and 2 present the yearly averages of the mass and nutrient fluxes in the three fractions of litterfall of the oak and beech plots in 2010-2015.The coefficients of variations of woody litterfall were much higher in the oak plot,whereas,in the other litterfall types the opposite was true.The statistical comparison between elemental fluxes in the two plots showed that only K differed significantly in the woody litterfall.

The percentages of foliar litterfall masses were found 71%for the oak plot and 68%for the beech plot(Figs.2 and 3)and were significantly higher(as expected)than all the other fractions for both species.In both plots,the Ca content was high in the woody fraction(Fig.2).The K percentage in the rest litterfall in the beach plot was significantly higher than the K percentage in the woody fraction.

In most cases,the soil in the beech plot had significantly higher concentrations of organic C,total N and percentage of clay(Table 3).Table 4 contains information on pH,exchangeable Al and C.E.C.It is obvious that the soil under the beech plot is the more acidic.

In the L horizon the statistical differences concerned Ca and N(Table 5),whereas in the FH horizon all differences were statistically significant(Table 6).

Figure 4 presents the results of the soil temperature and moisture monitoring for the periods 2009-2013 and 2009-2014,respectively,for the two forest species.There was consistency in the results,the temperature values were found higher in the oak plot and those of moisture higher in the beech plot.The periods of measurements did not reach 2015 due to damages in the field and financial problems to replenish them.

The stocks of nutrients in the mineral soils did not differ significantly for N,S and P(Fig.5).In contrast,the stocks of exchangeable base cations were significantly higher in the oak plots(Fig.6).

Table 1 Average fluxes of mass and nutrients stocks in litterfall during the period of 2010-2015 in the oak plot

Table 2 Average fluxes of mass and nutrients stocks in litterfall during the period of 2010-2015 in the beech plot

Fig.2 Percentages of the litterfall fractions in the oak stand with standard errors.To convert the latter into confidence intervals for a 0.05 probability level,they have to be multiplied by 2.57,which is the t value for five degrees of freedom(6 years of observations)

Discussion

Litterfall

The higher variability(expressed by the coefficients of variation)of the foliar litterfall in the oak plot can probably be explained by the differences in age and species.Hansen et al.(2009)also found higher variability of foliar litterfall in common oak(Q.robur L.)than Fagus sylvatica of the same age(approximately 50 years).However,the same authors did not find high variability in the other litterfall fractions,whereas in our work the woody and the rest litterfall appeared to have somewhat higher variability in the beech plot.

Fig.3 Percentages of the litterfall fractions in the beech stand with standard errors.To convert the latter into confidence intervals for a 0.05 probability level,they have to be multiplied by 2.57,which is the t value for five degrees of freedom(6 years of observations)

The differences in litterfall fluxes of nutrients between forest species are also differences between amounts taken up by plants,as whatever is lost in litterfall must be gained again(Cole and Rapp 1981).It must be taken into account that our work deals only with aboveground litterfall although the belowground litterfall can be appreciable(Vogt et al.1986).However,the latter is much more difficult to measure and usually involves destructive sampling,something that can disturb the experimental plots.

The results showed that the litterfall fluxes of nutrients did not differ significantly with the exception of K in the woody litterfall.This means that the two species may(with the reservation of below ground litterfall)take up approximately the same amounts of nutrients according to Coleand Rapp(1981).In any case,the nutrient residence time in the forest floor is different for the two species as this time depends on the total amounts of nutrients on the forest floor.The higher the amount the longer the residence time(Gosz et al.1976).It is possible that the beech nutrition is

more based(than oak)on nutrient re-translocation from older tissues.

Table 3 Selected soil properties of the two plots.C and N are expressed in g kg-1 and clay in percentage(%)per unit of soil

Table 4 Selected soil properties of the two plots.C.E.C.and exchangeable Al are expressed in the unit of×10-2 mol kg-1

Table 5 Total amounts(ton ha-1)and nutrients(kg ha-1)in the L horizons in the two plots

Table 6 Total amounts(ton ha-1)and nutrients(kg ha-1)in the FH horizons in the two plots

Fig.4 Plots of temperature and moisture vs time in the oak and beech stands

Fig.5 Amounts and confidence intervals of the means of N,P and S in the mineral soils of the oak and beech stands.As the confidence intervals for 0.05 probability level overlap,there are no statistical differences for any of these three elements between the two ecosystems

Fig.6 Amounts and confidence intervals of the means of the exchangeable Ca,Mg and K in the mineral soils of the oak and beech stands.As the confidence intervals for 0.05 probability level do not overlap,there are statistical differences for these elements between the two ecosystems

When comparing the results of the present work with those existing in literature,it becomes evident that there is plenty of information concerning beech forests but very scarce on the Hungarian oak. The only reference on nutrient cycling concerning the Hungarian oak is that of Alifragis(1984)who worked in a forest situated in northern Greece having mica schist as a soil parent material,like in our work.He found an annual foliar litterfall mass range of 3.3-3.7 tons·ha-1.In our work,the respective quantity was a little higher (4.0 ton·ha-1) (Table 1). Alifragis(1984)also found similar results with regard to the nutrient fluxes of N,P,K and Mg in the foliar and woody litterfall.Calcium was an exception having higher value in our work.

As for beech the mass percentages of the several fractions of litterfall in northern Spain were found 61.9,17.6 and 20.5%for the foliar,woody and the rest litterfall,respectively(Regina and Tarazona 2000).These values are close to those found in our work(Fig.3).In an old forest in Sweden(Nihlga?rd 1972)found a 64%of foliar litterfall in mass,whereas Pedersen and Hansen(1999)found an 87%of the same foliar type in a young beech forest.In young forests the foliar litterfall predominates,whereas as forests age the percentages of other litterfall fractions such as twigs,branches,flowers and other plant parts begin to rise.In both plots,the foliar percentages for masses and nutrients were significantly higher than the other fractions with the exception of Ca in the woody fraction of the oak plots.The rest litterfall and the woody fractions did not differ significantly apart from K in the beech plot.It must be taken into account that the woody fraction has a long residence time in forest floors due to its low decomposability and for this reason the rest litterfall fraction is considered a more readily source of available nutrients.However,in forest stands in acidic soils(just like in our work),the high Ca content of woody litterfall can be a valuable source to buffer the soil exchangeable acidity.

Forest floor

In the L layer,the masses did not differ significantly but nevertheless the stocks of Ca and N did.This was the result of different concentrations in the leaves of these two species.Indeed,the beech leaves have had consistently higher N concentrations and lower Ca ones than the oak leaves(data not shown here).The C/N ratio in the L horizon did not differ significantly between the two plots,whereas in the FH horizon the ratio was found significantly lower in the beech plot(Tables 5 and 6).It would be expected that the beech plot due to the low ratio of C/N in the FH horizon would have had a soil with less organic C than the oak plot.However,the C/N ratio in the FH reflects the equilibrium of C and N relation and not the speed of decomposition,something expressed much better by the C/N ratio in the L layer containing freshly fallen plant material.

The total mass(L+FH)in the forest floor of beech was found approximately 57 ton·ha-1.This is higher than the respective values found by Merino et al.(2008)in beech forests of Spain.The high amounts of organic matter in the FH horizons of the beech forest can be explained by several factors functioning alone or in combination.The first one is the higher concentration of organic C(Table 3)and the second is the higher mass in the beech plot than the respective variables in the oak plot. The third is the microclimate differences between the two plots.The temperatures measured in the 20-cm soil depth were consistently lower in the beech plot,whereas,the opposite was observed for the moisture content(Fig.4).Moisture usually is considered a favorable agent for decomposition in deciduous forests (Augusto et al. 2015). In our case,however,it was not coupled with high temperatures so its effect was diminished.Prescott(1996)argued that in forest soils the microclimate could be more important than the humus type with regard to decomposition.Vesterdal et al.(2012)found that slight variations in temperature below tree canopies could alter decomposition rates.The N,S,P,Ca,Mg and K stocks in the FH horizons followed that of organic C stocks.It has to be mentioned that the higher slope(33%)in the soil of the beech plot did not affect the degree of erosion and species richness.The forest floor under beech in mountainous Greece is usually,as in our work,very thick and forces rainwater to pass through the L and FH layers diminishing in this way its erosive force.

Mineral soils

We considered that the separate examination of the forest floor and mineral soil was necessary.It has been showed that tree species with mineral soil C within the same range yet may possess very different C turnover rates in the forest floor(Vesterdal et al.2012;Jandl et al.2007).From Table 3,it can be seen that the organic C had higher concentrations in the soil under beech in all mineral layers and higher N concentrations in the last two layers.The low temperatures again just like in the FH horizon may have played a role but this time there is an extra factor,which is the acidity and especially the high concentrations of exchangeable Al in the beech soil (Table 4).The pH together with the concentrations of exchangeable Al were found significantly lower and higher in the beech mineral soil layers consistently than the respective values in the oak plot.The Al toxicity can repress the enzyme activities in soils leading to suppressed microbial mediated nutrient cycling(Kunito et al.2016).From Table 4,it can be calculated that the percentage of exchangeable Al in the soil under beech was approximately 60%of the C.E.C.in the mineral soil,whereas the respective percentage in the oak plot ranged from 9%to 26%.The acidity also stabilizes the soil organic matter through high concentrations of Al and Fe sesquioxides, which protect it from decomposition(Turrio′n et al.2009).The higher clay content in the beech soil(Table 3)could be a species effect.Mueller et al.(2012)proposed a concept that tree species lower soil pH resulting in enhanced mineral weathering.He found that the total extractable acidity was higher in soils under Fagus sylvatica than Quercus robur and Q.frainetto.Under the hypothesis that the soil parent material is the same for both tree species,we can conclude that despite the lower temperatures in the soil of the beech plot the higher extractable acidity brought about by the beech trees caused the creation of more clay minerals through higher rates of weathering.The higher acidity in the soil under beech can explain the richness of ground vegetation in the oak stand.Roem and Berendse(2000)showed that at local level vascular plant species richness decreases when soil acidity increases in woodland and grassland.Also Lalanne et al.(2010)showed that soil acidity together with the soil nutrient regime explained the local and regional trends in the ground vegetation of beech forests in France.

Most authors have quoted stocks of organic C and total N in forest soils neglecting the other nutrients.The N pools of the forest floor and mineral soil of beech were approximately average compared to values reported by Cole and Rapp(1981)in forest floors(86-1050 kg ha-1)and mineral soils under beech down to rooting zone(6332-9452 kg ha-1).Cremer et al.(2016)found approximately 4000 kg ha-1of N in mineral soil(up to 60 cm depth)under beech in Bavaria,southern Germany.Han et al.(2017)found approximately 6000 kg ha-1of N in the mineral soil(50 cm depth)of several oak species in Korea.Some workers dealt with the other nutrients as well.Ma et al.(2007)calculated the stocks of total N,P,K and Mg in mineral soils(up to 80 cm depth)under Chinese fir.The total amount of P was about half the amount of total P in our work(Fig.5).Han et al.(2017)calculated the amounts of total N,available P and exchangeable Ca,Mg and K in soils under a variety of deciduous and conifer trees in Korea up to depth of 50 cm.As the soils,in our work,are acidic we decided to refer to exchangeable cations and not to total amounts because the exchangeable cations are taken up by trees and have a competition with the exchangeable Al.For some Quercus species,Han et al.(2017)found similar results with regard to amounts of total N,exchangeable K and Mg but much lower stocks of exchangeable Ca than the oak site in our work.This must be a matter of the nature of parent material.There is no literature data on S stocks in forest soils which is rather strange as mineral soils differ a great deal to retain available S(pH is the most affecting agent).In the top mineral layers(0-10 cm)the percentage of S is in the organic fraction is high.As we go down the soil profiles in these acidic soils with a high capacity to absorb sulphates,the available S in the form of SO42--S can make up 50%of the total S(Michopoulos et al.1998;Michopoulos et al.2013).

The comparison of nutrient stocks in mineral soils showed that only the exchangeable base cations differed significantly(Fig.6)and even the amounts of total N did not follow the patterns of organic C concentrations in mineral soils although it is known that N is closely coupled with the organic C.It is a common practice to sample replicates per soil horizon or layer,analyze the samples for nutrients and calculate the standard deviation.However,when the nutrient amounts are added to find the total amount,the propagation of random errors has to be calculated.As mentioned in the Results section,the new standard deviation is the square root of the sum of squares(Miller and Miller 1988).The authors argued that the important point is that the standard deviation for the final result (total amount) was larger than the standard deviations of the individual addition (amount per soil horizon or layer). Therefore, the test for comparing amounts after additions is stricter than comparing amounts in the same horizon or layer.

Conclusion

The two tree species had similar average litterfall masses in 2010-2015 although with high variability.

Although the foliar litterfall made a 71%for the oak and 68%for the beech plot,the woody litterfall was found to be a significant source of Ca for both plots and the rest litterfall for K for the beach plot.

The high concentrations of organic C in the beech soil were probably a result of the microclimate of the plots and more specifically of the lower temperatures in the soil of the beech plot.In addition,the acidity of the soil in the beech plot helped the build-up of the organic matter.

The higher acidity of the soil in the beech plot as well as the higher clay content was a result of species difference.The beech trees acidified soil,which in turn brought about weathering and more clay minerals. Furthermore,clay minerals stabilized and increased the amount of soil organic matter.

The nutrient stocks in the L horizon reflected the nutrient concentration of the leaves of the standing trees.

The higher nutrient stocks in the FH horizons of the beech plot were the result of the high amounts of organic matter.

The nutrient stocks in the mineral soils were strongly affected by soil acidity giving significantly higher results for the exchangeable base cations.In addition,the propagation of the random errors of the individual mineral layers affected the comparison of the nutrient stocks.

AcknowledgementsThe authors want to express their sincere thanks to C.Mitropoulou for her help with sample pretreatment and analysis.