Changes in temperature sensitivity of forest litter during decomposition along an altitudinal gradient in temperate mountains – A reciprocal litter transplantation study

The aim of this study was to assess the effects of the factors controlling the temperature sensitivity of litter decomposition, which is essential for predicting the rate of soil carbon loss in in the context of global climate changes. It based on the translocation of forest litter in a field experiment conducted in the Western Carpathians. Litterbags were used to expose litter originating from different altitudes (i.e., 600, 900, and 1200 m a.s.l.) at the altitude where it was collected and at two other altitudes on five different mountains. Litterbags were collected after 6, 10, and 24 months of exposure. The respiration rate of litter was measured in the laboratory, and Q 10 coefficients were evaluated for two temperature ranges: lower (Q 10 L; 5 ◦ C – 15 ◦ C) and higher (Q 10 H; 15 ◦ C – 25 ◦ C). We tested whether litter Q 10 values correlated with experimental factors, as well as soil and microbial community properties. After 24 months of exposure, the litter mass decreased to half of its initial mass. The general linear model (GLM) constructed for Q 10 (R 2adj = 77.3 %; p < 0.0001) indicated, that Q 10 L values were higher than Q 10 H values (2.19 ± 0.58 and 1.52 ± 0.31, respectively) and increased strongly with time (1.56 ± 0.36, 1.73 ± 0.45, and 2.36 ± 0.60, consecutively). There was a significant interaction between the temperature range and time, indicating that Q 10 L increased more over time than Q 10 H. Other interaction between temperature range and litter origin indicated that Q 10 L increased with the altitude of the litter origin, whereas Q 10 H did not change. The proportion of fungi in the microbial biomass correlated positively with Q 10 . Our results were consistent with the kinetic theory of higher temperature sensitivity for more decomposed organic matter. The conclusion that litter Q 10 L is more responsive to environmental conditions than Q 10 H has important implications for estimating soil carbon emissions.


Introduction
Heterotrophic soil respiration (R) is the second largest carbon (C) flux in terrestrial ecosystems after primary producer respiration (Raich and Schlesinger, 1992).Soil organic matter (SOM) decomposition is highly temperature-dependent (Kirschbaum, 2000).The intensified carbon dioxide emission from the soil to the atmosphere due to climate warming may further accelerate global climate change, a phenomenon known as positive feedback (Gutiérrez-Girón et al., 2015;You et al., 2019).
Soil temperature sensitivity, usually referred to as the Q 10 value, which is the increase in the soil respiration rate by a 10 • C temperature increase, remains one of the key uncertainties in global climate change research (Fierer et al., 2006;Vanhala et al., 2007), as even small variations in Q 10 can lead to large differences in estimates of global C dynamics.Although the value of Q 10 generally follows the kinetic theory describing chemical reactions, with the global value of Q 10 assumed to be approximately 2 (Arrhenius, 1889;Davidson and Jannsens, 2006), a wide range of values, from 1.1 to 13.5, have been reported across different biomes (Chen et al., 2020).Soil temperature sensitivity depends on many factors, such as SOM recalcitrance (Xu et al., 2010), nutrient availability (Szlachcic and Rożen, 2023), and soil microbial community characteristics (Klimek et al., 2020a).Initial differences in SOM properties, as well as the environmental conditions under which decomposition occurs, may influence SOM temperature sensitivity; however, these effects are difficult to distinguish from each other (He et al., 2016).
Studies on soil temperature sensitivity have focused on well-developed soil horizons, with little attention paid to litter, which is the most labile soil organic matter fraction (Fierer et al., 2006;Li et al., 2021a).Litter decomposition in forests contributes 5-45 % to total soil respiration, making the litter fraction an important determinant of atmospheric CO 2 concentration (Gritsch et al., 2016).In a temperate seasonal climate, a significant proportion of annual litter fall decomposes within the first year of decomposition (Spohn and Berg, 2023).Some litter components, for example, elements such as potassium and magnesium (Osono and Takeda, 2004) and alkyl carbon compounds (Certini et al., 2023), are particularly rapidly leached from decomposing plant litter.Litter is expected to have lower Q 10 values than deeper soil horizons, according to the kinetic-based theory that the more chemically recalcitrant soil organic matter will have a higher temperature sensitivity (Fierer et al., 2006;Karhu et al., 2010;Xu et al., 2010).This may imply that as litter decomposes, its Q 10 will increase, bringing it closer to values typical of deeper soil horizons over time.
Mountain forest soils contribute significantly to global soil carbon stocks (Bojko and Kabala, 2017).Elevation gradients are of particular interest when studying the effects of global warming on terrestrial ecosystems.Elevation gradients are characterised by strong changes in climatic and biotic characteristics over short distances and are, therefore, useful for predicting and mitigating the effects of climate change.Furthermore, altitudinally defined vegetation belts on mountain slopes are the counterparts of latitudinally controlled climatic zones (Margesin and Niklińska, 2019).Global climate warming is expected to strongly alter mountain ecosystems, as an enhanced increase in the mean annual temperature is observed in mountainous regions compared to the global average (Komarova et al., 2022;Li et al., 2022).Our previous studies showed that Q 10 values based on the soil respiration rate for temperate mountain forest soils did not change along an altitudinal gradient despite significant differences in soil respiration rates between altitudes and soil horizons (Niklińska and Klimek, 2007;Klimek et al., 2016).However, we did not test the temperature sensitivity of the uppermost (top) soil layer, namely litter, in these studies.
The decomposition of senescent litter depends on several factors.These include the species composition of the litter producers (mainly plants) and, thus, the physical structure of the litter (i.e.needles vs. leaves), which determines the litter penetration rate by fine roots (Wang et al., 2021a,b), fragmentation rate by soil animals (Peng et al., 2022) and the colonisation rate by microbes (Vanhala et al., 2007;Karhu et al., 2010;Laskowski, 2012;Gutiérrez-Girón et al., 2015;Zhang et al., 2023).Environmental conditions, including temperature and humidity, are important for litter decomposition.In mountainous regions, the altitude above sea level (a.s.l.) determines both temperature and moisture (precipitation) (Zielonka et al., 2021).The key role of altitude in the decomposition processes was confirmed in a field experiment where cellulose filters were exposed along an altitudinal gradient in the Carpathians (Drewnik, 2006).However, it is very difficult or impossible to separate the effect of altitude from that of other environmental conditions (i.e.edaphic conditions) on decomposition rates, as environmental conditions are closely linked.
In this study, a field experiment using litter bags was conducted to investigate how forest litter originating from three forest belts along an altitudinal gradient differed in temperature sensitivity over two years of decomposition.The aim of our research was to identify the general determinants of temperature sensitivity of decomposing litter in mountain forests by cross-exposing litter in litterbags along the altitudinal gradient and three time sampling within forest litter exposure time.Therefore, experiments were conducted to tested the effects of 1) altitude of litter origin, 2) altitude of litter exposure, 3) duration of exposure, 4) physicochemical litter properties, 4) soil (edaphic) properties, and 5) microbial properties on litter temperature sensitivity in the lower and higher temperature ranges, considering the effects of various confounding factors.We expected that litter Q 10 would depend on the above factors to varying degrees, and aimed to determine which of the following factors was dominant at the given litterbag sampling time.
Specifically, litter translocation can be used as a tool to answer the question of how important climatic factors and site-specific environments determine temperature sensitivity to litter decomposition.Although our altitude gradient was only 600 m due to the average height of the mountains and the distribution of forests in the Outer Western Carpathians, the difference in mean annual temperature between the lowest and the highest levels (altitude) of the climatic transect used in the experiment is about 3.4 • C, which exceeds the predicted climate changes in the coming decades.Therefore, in the context of global climate change, knowledge of the temperature sensitivity of forest litter is essential for predicting the rate of carbon loss in temperate mountain massifs, which, in view of the intensive forest management in lowerlying regions, constitute an important pool of carbon sequestration and provide various ecosystem services.

Study plots
The study plots were located in southern Poland in the Beskidy Mountains (Fig. 1), a part of the Western Carpathians.The bedrock, known as Carpathian flysch (multi-tiered sedimentary rocks), are a widespread parent material for soils in the mountain chains of the Alpine and Himalayan belts (Trifonov et al., 2012).The soils in the plots studied were generally Cambisols with Podzol characteristics (Musielok et al., 2022).
Study plots were established on five mountain massifs: Polica (1369 m a.s.l.), Pilsko (1557 m a.s.l.) and Romanka (1366 m a.s.l.), in the Beskid Żywiecki range; Kudłoń (1274 m a.s.l.) in the Gorce range; and Radziejowa (1262 m a.s.l.) in the Beskid Sądecki range.The distance between the northernmost mountain (Polica) and the southernmost mountain (Radziejowa) is 20 km in a straight line, and the distance between the westernmost mountain (Romanka) and the easternmost mountain (Radziejowa) is 120 km in a straight line.Because the mountains studied were at approximately the same latitude, the temperature changes associated with the latitudinal gradient were negligible.Longitudinal distance was greater than latitudinal distance, and although there were some minor longitudinal climatic differences in the study area, these were much smaller than the effects of local relief and altitude.The five mountains were treated as experimental replicates.Because the temperature and moisture of mountain soils are highly dependent on slope and insolation, all plots were established on northern slopes.B. Klimek and M. Niklińska Experimental plots were established at three altitudes (600, 900, and 1200 m a.s.l.) on each of the five mountains, resulting in 15 plots (3 altitudes × 5 mountains).The altitudes were selected according to the average height of the mountains in the Outer Western Carpathians and the distribution of forest belts versus agricultural and built-up areas, which reached relatively high altitudes in the region (up to 1000 m a.s.l. in some locations).Furthermore, this altitude selection enabled us to compare our results with those of our previous studies.A detailed description of the vegetation characteristics can be found in Klimek et al. (2020b).Woodlands cover more than 40 % of the Polish Carpathians (Griffiths et al., 2014) and are dominated by common beech (Fagus sylvatica), silver fir (Abies alba), and Norway spruce (Picea abies) (Kholiavchuk et al., 2024).The plots at the two lower altitudes (600 and 900 m a.s.l.) consisted mainly of 40-50-year-old forests, and there was a forest older than 60 years at 1200 m a.s.l.
The mean annual temperature at the lowest elevation, i.e., 600 m a.s.l., is 7.2 • C, and the mean annual precipitation is 1000 mm.The area at 600 m a.s.l. is covered by typical mixed foothill forests dominated by deciduous species, mainly beech, hornbeam, oak, and birch (Table 1).The foothill forest zone extends up to 700 m a.s.l., depending on the local climatic characteristics.Forests at this altitude are easily accessed by humans; therefore, they are subjected to substantial anthropogenic influences, resulting in thinning and disturbance of the natural vegetation structure.Forest floor plant species associated with forest edges and typical of stands with increased soil fertility, for example, Anthriscus sylvestris (L.) Hoffm, Asarum europaeum L. and Chaerophyllum hirsutum L.
The average annual temperature at the intermediate altitude, i.e., 900 m a.s.l., is 5.5 • C and the average annual precipitation is 1200 mm.The predominant vegetation at this altitude is the lower montane forest, which is composed of mixed beech-fir forests (Table 1).The lower montane forest in the Beskidy Mountains ranges from 550 to 1150 m a.s.l., depending on the local microclimate.
The mean annual temperature at the highest elevation, 1200 m a.s.l., is 3.8 • C and the mean annual precipitation is 1400 mm.Spruce forests predominate on this altitude with a small admixture of mountain ash (Table 1).The upper montane zone forest in the Beskidy Mountains ranges from 1000 to 1500 m a.s.l., depending on the local microclimate.In addition, the local microclimate affects the range of upper montane zone forests in the Beskidy Mountains from 1000 to 1500 m a.s.l.

Soil sampling and soil analysis
Soil properties were measured in each of the mountain and attitude plots (5 mountains × 3 altitudes = 15 sites), with three soil subsamples collected per site.This procedure yielded 45 separate soil samples.Data from three soil samples per stand were averaged to avoid pseudoreplication.Soil samples were collected using a steel core sampler at three points on a horizontal transect across the slope (three points every 100 m).For most analyses, 10 cm of the top soil layer was collected after the removal of the current year's litter, and only the soil texture was measured in the mineral soil fraction.The measured soil properties included organic matter content (OM), magnesium (Mg) and potassium (K) content, pH, and sand content (%) in the mineral fraction.Soil dry weight (DW) was determined after drying the soil samples at 105 • C for 24 h, and the OM content was determined as the loss on ignition at 550 • C for 24 h.Total Mg and K content in soil was determined after wet digestion of 0.5 g of soil in 10 mL of a mixture of concentrated HNO 3 and HClO 4 (7:1, volume/volume) (Sigma-Aldrich, Saint Louis, MO, USA).Atomic absorption spectrometry with a flame atomiser (Aanalyst 200, Perkin-Elmer, Waltham, MA, USA) was used to determine the elemental content.The accuracy of the method was verified by analysing five blanks and five replicates of certified standard materials (CRM025-050, Sandy Loam 8, RT Corp.).Soil pH was measured potentiometrically in air-dried subsamples shaken in deionised water (1:10 w/v).The textures of the mineral soil fractions were hydrometrically assessed.All soil physicochemical analyses were performed on three subsamples of each soil sample, and the results were averaged.

Litter collection and litterbag preparation
Freshly fallen litter was collected directly from the top layer of the soil (0-2 cm) by hand in October 2011.Additional material was retrieved by gently shaking the lower tree branches.Three types of litter were acquired from 600, 900, and 1200 m above sea level (a.s.l.) along the mountain altitudinal gradient.Litter originating from all five sites at a single elevation was mixed in equal proportions to obtain a homogeneous material representative of the forest belt.Litter collection was conducted for one week on all five mountains.
The litter was transported to the laboratory and air-dried at room temperature (about 22 • C) when spread out in a single layer over a large area.The collected litter mainly consisted of leaves and/or needles from different tree species.However, a small admixture of other types of plant materials also appeared (for example, seeds, lichens, mosses, fine branches, and pieces of bark).The plant species composition of each litter type was expressed as % by mass and % by volume (Table 1).
The litterbags were made of nylon mesh and sewn with nylon threads.They were 20 cm × 20 cm in size and carefully filled with 10 g of air-dried litter weighed on a laboratory scale (accuracy 0.01 g).The mesh size was 1 mm × 1 mm, which is the most appropriate size for litter bag studies (Krishna and Mohan, 2017).Larger than typically used bag sizes enabled the inclusion of representative and homogeneous litter from multispecies plant communities.

Experimental design
Sets of three litterbags containing litter from each altitude were exposed to all three altitudes.The bags were laid out in a row within a 1 m array in each experimental plot.All bags were placed in a 400 m 2 area across the slope.Litterbags were placed in the middle of this year's litter layer and loosely covered with this year's litter to prevent accidental destruction by animals or wind.
Litterbags were exposed in the field for two years, starting at the end of November 2011, i.e., just before snowfall.Litterbag sets were collected three times after 6, 10, and 24 months of field exposure in spring 2012, autumn 2012, and autumn 2013, respectively.Two randomly selected sets of litterbags were collected during a single sampling period in each plot.One set was intended for mass loss analysis and physical and chemical analysis (required for litter drying), and the second one (prevented from drying) was intended for microbiological analysis, including assessment of temperature sensitivity.
The sampling procedure resulted in 90 separate litter bags collected

Table 1
General characteristics of the litter used in the experiment, collected on five mountains at three altitudes (after mixing), representing the average species composition of the litter at a given altitude and then used in litter bags.The average species contribution was expressed in % by mass and % by volume.during a single sampling period (3 altitudes of litter origin × 3 altitudes of litter exposure × 5 mountain replication sites × 2 separate litterbag sets).Because sampling was performed three times, 270 litterbags were used throughout the experiment (although more litterbags were prepared and exposed in the field to compensate for possible losses).The litterbags were transported to the laboratory during each sampling event, which took five working days.Individual litter bags was transported in separate plastic bags, taking care to ensure aeration, prevent unintentional loss of mass, and maintain field humidity.Samples were stored at 4 • C until the sample set was completed.The bags were then cut open, and plant roots, soil fauna, and soil residues were manually removed.One set of litterbags was left to air dry at room temperature for at least two weeks to assess the remaining litter mass and for physical and chemical analyses.The mass of litter from each bag was measured using a laboratory scale (±0.01 g).The second set of litter was placed in 1 l glass jars, saturated with deionised water, lightly covered to prevent desiccation, and used for microbial analysis.

Litter physical-chemical analyses
The characteristics of freshly collected litter from 600, 900, and 1200 m a.s.l. were analysed in five randomly selected litter bags prepared at the beginning of the experiment.The entire litter from a single litterbag was ground before further analysis.The same analyses were performed on litter from bags exposed in the field.
Organic C and total N were analysed in finely ground litter subsamples using a CHNS analyser (Vario EL III, Elementar Analyse Systeme GmbH, Bechenheim, Germany).Total elements (Ca, K, Mg, and Mn) concentrations in each litter sample were determined as described in Subsection 2.2.The accuracy of the method was checked by analysing five blanks and five replicates of certified standard material (Trace elements in lichen, BCR 2-482, IRMM, Geel, Belgium) with the litter samples.

Phospholipid fatty acid (PLFA) analysis
Microbial biomass and community structure in the decomposing litter were analysed using the phospholipid fatty acid (PLFA) method according to Frostegård et al. (1991).Total lipids were extracted from equivalents of 0.25 g DW of litter using a one-phase mixture.The extracts were divided into two phases, and after collecting the lower lipidcontaining phase, the lipid material was fractionated using silica acid columns.Phospholipid-containing polar lipids were collected and their fatty acid methyl esters were prepared using mild alkaline methanolysis.Methyl nonadecanoate (19:0) was used as an internal standard.Fatty acid methyl esters were separated using gas chromatography (Claus 600 MS with an FID detector; Perkin Elmer, Waltham, MA, USA) and quantified using qualitative fatty acid methyl ester mixtures (Sigma-Aldrich).The abundance of the 39 extracted individual PLFAs was expressed as nanomoles of PLFA per gram of DW litter, and their sum for each sample (PLFA tot ) was used as an indicator of microbial biomass.The fatty acid nomenclature followed that of Frostegård et al. (1993).Individual fatty acids were assigned to bacterial and fungal biomasses (PLFA b and PLFA f , respectively) according to classic studies (Frostegård et al., 1993;Hill et al., 2000;Waldrop et al., 2000).The relative proportions of PLFA b and PLFA f in PLFA tot (% PLFA b , % PLFA f ) and the ratios of bacterial PLFA to fungal PLFA, denoted as PLFA b-to-f , were also calculated.

Litter respiration rate measurements and Q 10 determination
The litter respiration rate was measured only on the material from the litterbags.Litter respiration rate was measured subsequently at 5 • C, 15 • C, and 25 • C (±0.5 • C) after three days of litter acclimation at a given temperature, conducted before each respiration rate measurement.A beaker containing 5 mL of 0.2 mol/L NaOH was carefully placed in the airtight glass jars containing the litter and incubated in the dark for 2-18 h, depending on the litter mass of the individual sample, the temperature, and the expected CO 2 emission from the sample to measure the respiration rate of litter.The jars were opened after incubation, 2 mL of 20 % BaCl 2 solution was added to the NaOH solution, and the excess sodium hydroxide was titrated with 0.1 mol/L HCl in the presence of phenolphthalein as an indicator.The respiration rate of each litter sample was measured twice, averaged and expressed as mmol CO 2 kg − 1 DW day − 1 .
The Q 10 temperature sensitivity coefficients were calculated using the mean respiration rates of litter (R) measured at two subsequent incubation temperatures (T and T + 10): Q 10 L was calculated for the two lower consecutive temperatures (5 • C-15 • C) and Q 10 H for the two higher temperatures used (15 • C-25 • C) and analysed separately.

Statistical analyses
The normality of the data distribution within groups was checked using the Shapiro-Wilk test prior to analyses when required, and the data were transformed when necessary (logarithmic transformation was mostly used, as data were usually right-skewed).
One-way ANOVA was used to test the significance of differences in the selected physical and chemical properties of soils from the three altitudes (600, 900, and 1200 m a.s.l.) at the beginning of the experiment.Significant differences in means were compared using Tukey's test.Differences between results were considered significant at p < 0.05.These data were combined using principal component analysis (PCA), and the first principal component (PC1 soil ) was used in further statistical analyses as a proxy for edaphic properties.PC1 soil values were also subjected to one-way ANOVA (transformed using ranks).
One-way ANOVA with Tukey's test (p < 0.05) was performed to test the significance of differences in selected physical and chemical properties of litter from the three altitudes (600, 900, and 1200 m a.s.l.) at the beginning of the experiment.
Litter properties during the litter bag exposure experiment (altitude of origin, altitude of exposure, and time) were analysed separately using three-way ANOVA with Tukey's test (p < 0.05).Nonsignificant interactions were excluded from the analysis.Litter data for all collected litterbags were also subjected to PCA and the combination of properties that best explained the variability of the data was selected (only these litter data are shown).The first PC (PC1 litter ) was chosen to represent the nutrient resources for microorganisms and was subjected to a three-way ANOVA, as described above.Microbial properties obtained using the PLFA method in the litter during the experiment were analysed separately using a three-way ANOVA with Tukey's test (p < 0.05).
Individual general linear models (GLMs) were used to identify the factors affecting litter respiration rate and Q 10 values during the experiment.
GLM was applied with the following categorical variables for litter respiration rate: altitude of litter origin, altitude of litter exposure, exposure duration, and measurement temperature applied in the laboratory (5, 15 or 25 • C) and with quantitative (linear) variables, that is soil microbial biomass PLFA tot (nmol g DW − 1 ), the proportion of fungi in the microbial biomass (% PLFA f ), the ratio of bacteria to fungi (PLFA b-to- f ), and combined variables separating effects of litter properties (PC1 litter ) and soil properties (PC1 soil ).Nonsignificant factors and interactions were excluded from the GLM analysis.
GLM was applied with the following categorical variables for temperature sensitivity Q 10 of decomposing litter: altitude of litter origin, altitude of litter exposure, exposure duration, and temperature range applied in the laboratory (Q 10 L or Q 10 H) and with quantitative (linear) B. Klimek and M. Niklińska variables, including soil microbial biomass PLFA tot (nmol g DW − 1 ), proportion of fungi in microbial biomass (% PLFA f ), ratio of bacteria to fungi (PLFA b-to-f ), and effects of litter properties (PC1 litter ) and soil properties (PC1 soil ).Nonsignificant interactions were excluded from the model.
All statistical analyses were performed using the Statgraphics Centurion 18 software (StatPoint Technologies Inc., Warrenton VA, USA).

Physical-chemical properties of soil and litter at the beginning of the experiment
Soil samples from the three altitudes differed significantly in OM content (Fig. 2 A) but not in other properties (Fig. 2 B-E).The Mg and K contents and soil pH tended to be lower at an altitude of 1200 m a.s.l.than at the other two altitudes, but the differences were not significant (Fig. 2 B, C, and D, respectively).The PCA combined these soil data, and the first principal component, with an eigenvalue of 2.59, explained 51.85 % of the variability in soil data.PC1 soil had positive loads from sand (0.50), OM content (0.48), and pH (0.26), and negative loads from Mg (− 0.55) and K (− 0.46) (Fig. 4 A).The PC soil differed between altitudes and confirmed that the soil at an altitude of 1200 m a.s.l.differed from that of the two lower altitudes (Fig. 2 F).
The litter collected at the three altitudes differed in the composition of the plant species that formed it.Litter fragments that could not be assigned to the plant species or that contributed in very small fractions were grouped into the last class, 'other' (Table 1).Tree litter thus represented > 99 % of each litter type.Litter at an altitude of 600 m a.s.l. was mainly composed of broad-leaved beech Fagus sylvatica, whereas the main components of litter at both higher altitudes were coniferous tree species, i.e., fir Abies alba at 900 m a.s.l., and spruce Picea abies at 1200 m a.s.l.(Table 1).In general, the litter from an altitude of 600 m a.s.l. was composed of more species than the litter from 900 and 1200 m a.s.l.These fundamental differences in litter species composition are reflected in the physicochemical properties of the collected litter (Fig. 3).The carbon concentration in the litter was arranged in the ascending order of 600, 1200, and 900 m a.s.l.(Fig. 3 A).Spruce-dominated litter from 1200 m a.s.l.contained the highest N concentration (Fig. 3 B), and thus the lowest C:N ratio (Fig. 3 G).Litter from 600 m a.s.l.contained the highest concentrations of Ca and Mg (Fig. 3 C and E, respectively).Concentrations of K did not differ among litters originating from different altitudes (Fig. 3 D), which may be due to the high mobility of this element.Mn concentrations were the lowest in litter from 1200 m a. s.l (Fig. 3 F).

Changes in litter mass and physical and chemical properties during the experiment
The litter mass decreased gradually during field exposure and decreased from 10.0 g to a mean value of 5.08 g after 24 months (Table 2).The C:N ratio and Mg content also decreased with the duration of litter bag exposure (Table 2).Litter bag exposure at different altitudes did not significantly affect litter properties; however, there was a weak effect of altitude on litter N content, which was at the limit of significance, and N content tended to be higher in litter exposed to higher altitudes (Table 2).When the litter data were combined using PCA, the first principal component (PC1 litter , with an eigenvalue of 1.90) explained 47.38 % of the variability in the data.PC1 litter had a positive loading from C content (0.61) and C:N ratio (0.61) and a negative loading from N (− 0.46) and Mg content (− 0.20) (Fig. 4 B).The interaction between litter altitude origin and exposure was only reflected in PC1 litter (Table 2), and the physicochemical properties of litter originating from 1200 m a.s.l.changed the most during decomposition, especially at the two lower altitudes, compared with all other possible

Changes in microbial community structure in litter during the experiment
Microbial biomass (i.e., PLFA tot ) in decomposing litter increased with decomposition time, reaching 337.7 nM g − 1 litter DW after 24 months of litterbag exposure (Table 3).The increase in community biomass was primarily due to an increase in the fungal biomass (Table 3).Litter originating from an altitude of 900 m a.s.l. was characterised by a higher proportion of fungal biomass (i.e., PLFA f (%)) and the lowest ratio of bacterial to fungal biomass (i.e., PLFA b-to-f ) compared to the other two altitudes (Table 3).Interactions between litter altitude and altitude of exposure were found for PLFA tot and PLFA f , showing that after 24 months of field exposure, the highest values of these factors were found for litter from 900 m a.s.l.

Litter respiration rate
The average litter respiration rate from the litterbags over the entire experiment was 75.81 (mmol CO 2 kg − 1 DW day − 1 ) (Table 4).The GLM model run for litter respiration rate was highly significant (p < 0.0001; R 2 adj = 69.16%) (Table 5).The litter respiration rate was influenced by the altitude of litter origin (p = 0.0016) (Fig. 5 A), time (p = 0.0374) (Fig. 5 B) and laboratory incubation temperature (p < 0.0001) (Fig. 5 C), but not by the altitude of litterbag exposure, which was therefore removed from the model (Table 5).The litter respiration rate was negatively correlated with the PC1 litter (p = 0.0010) (Fig. 5 D) and positively correlated with PLFA f (%) (p < 0.0001) (Fig. 5 E) (Table 5).A significant interaction between the altitude of litter origin and time (Table 5) indicated that the litter respiration rate was particularly high for litter at an altitude of 600 m a.s.l.during the first respiration rate measurement made after 6 months of field exposure (Fig. 5 F).The interaction between the altitude of litter origin and PC1 litter (Table 5) indicated that the changes in PC1 litter were most correlated with litter respiration rate for litter originating from an altitude of 600 m a.s.l., whereas chemical-physical properties, expressed as PC1 litter value, had a very limited effect on respiration rate for litter at an altitude of 900 m a. s.l., its (Fig. 5 G).

Determinants of litter temperature sensitivity
The average Q 10 value for the entire experiment was 1.86 (Table 4).The GLM model for litter Q 10 values was highly significant (p < 0.0001;

Table 2
Means and standard deviations for remaining litter mass (g) and elemental content (C, N, Mg) (%), C:N ratio and PC1 litter values for litter collected from and exposed in litter bags at the three studied altitudes (600, 900 and 1200 m a.s.l.) and sampled three times (after 6, 10 and 24 months) (each n = 5).5.1 a (0.9) 48.9 (5.9) ab 1.7 (0.   6).Q 10 values were mainly determined by litter bag exposure time (p < 0.0001) (Fig. 6 B) and temperature range (p < 0.0001) (Fig. 6 C), but not by litter altitude (Fig. 6 A) or exposure altitude (Table 6).Q 10 L values for decomposing litter were higher than Q 10 H values (2.19 ± 0.85 and 1.52 ± 0.31, respectively) and increased strongly with the time (1.56 ± 0.36, 1.73 ± 0.45, 2.36 ± 0.60, consecutively) (Table 6).Similar to the respiration rate, the proportion of fungal biomass, measured as PLFA f (%), was positively correlated with Q 10 (p = 0.0333) (Fig. 6 D).There was a significant interaction between temperature range and litter origin (p < 0.0001) (Table 6), indicating that Q 10 L increased with the altitude of litter origin, whereas Q 10 H did not change or even tended to decrease (Fig. 6 E).A significant interaction between temperature range and time (p = 0.0001) (Table 6) indicated that Q 10 L increased more than Q 10 H with time (Fig. 6 F).

Discussion
Our study showed that forest litter changed significantly during decomposition and that its respiration rate corresponded to a wide range of environmental variables.The field exposure of litterbags lasted two years, which means that it comprised an early phase of decomposition characterised by exponential mass loss when the decomposition rate was regulated by the carbon and nitrogen content, lignin content, and morphological characteristics of the litter (Tian et al., 2000;Rawlik et al., 2021;Bonanomi et al., 2023).Previous studies showed that litter quality is the predominant controlling factor during the early stages of litter decomposition, explaining approximately 65 % of the variability in litter decomposition on a global scale (Djukic et al., 2018).Most plant litter does not decompose completely but forms a stable residue that contributes to soil organic matter (Berg et al., 2022), which can take decades, depending on the type of litter (Dziadowiec, 1987).Therefore, litter acts as a link between the two pools of carbon found in the soil and living organisms (Manzoni et al., 2010).
The mean litter mass decreased from 10.0 g to 7.52 g after six months of litterbag field exposure, and the average residual litter mass was 5.08 g after 24 months.The rate of litter mass loss was consistent with other studies on litterbag exposure along an altitudinal gradient conducted in the Hida Mountains in Japan (Tian et al., 2000), the Argentinian Andes (Moretto and Martínez-Pastur, 2014), and the Austrian Alps (Berger et al., 2015).The mass loss did not depend on the litter altitude origin (litter type) despite large differences between litters, including plant species composition and elemental content, resulting in diversified litter toughness for microbial colonisation and stoichiometric decomposition pathways.In particular, litter mixing can accelerate decomposition rates through complementary effects, that is, by increasing the nutrient supply to the functional groups of decomposers and increasing the diversity of microorganisms (Liu et al., 2022).Therefore, the litter of mixed plant species may experience a higher mass loss than the litter of a single species.We observed a higher respiration rate in litter from 600 m a.s.l., comprising at least eight plant species, than in litter from 900 and 1200 m a.s.l., comprising at least four and two plant species (respectively).However, this was more likely due to litter quality; litter from 600 m a.s.l. was composed mainly of deciduous litter, whereas the other two types of litter were composed mainly of coniferous litter.Conifer litter typically has a lower respiration rate than broadleaf litter (Klimek et al., 2021), which may be because of its higher C:N ratio and lignin and

Table 3
Means and standard deviations for microbial characteristics: total microbial biomass PLFA tot (nM g − 1 DW), fungal biomass PLFA f (nM g − 1 DW), fungal share in community PLFA f (%) and bacteria-to-fungi ratio PLFA b-to-f for for litter collected from and exposed in litter bags at the three studied altitudes (600, 900 and 1200 m a.s.l.) and sampled three times (after 6, 10 and 24 months) (each n = 5).Three-way ANOVA results with Tukey test are presented at the bottom of the table; statistically significant differences are indicated in bold.Values with different letters are significantly different between levels.

Table 4
Means and standard deviations for litter respiration rate and Q 10 values, originated and exposed at three studied altitudes, subjected to laboratory temperature incubation (n = 5).Note that significant interactions between factors make some data difficult to compare directly (see Figs. 2 and 3).elemental contents.In addition, the litter respiration rate was measured in the laboratory under standard conditions and may not be a good representation of the litter mass loss (decomposition rate) observed in the field.Litterbag exposure altitude did not affect litter mass loss, litter properties, or litter respiration rates.Bothwell et al. (2014) showed that leaf litter decomposition rates increased with increasing mean annual temperature in tropical montane wet forests in Hawaii.However, changes in temperature with altitude in the mountains may have a more pronounced effect on vegetation in tropical than in temperate areas (Ohsawa, 2006).The lack of an effect of exposure altitude on litter respiration rate in our study may be the result of multiple factors affecting the early stages of litter decomposition and active, diverse microbial communities, as the litter respiration rate was almost twice as high as the soil respiration rate in the study area, as measured in previous studies (Niklińska and Klimek, 2007;Klimek et al., 2015).
As expected, Q 10 values increased with time, with averages of 1.56 and 2.36 after 6 and 24 months of litter field decomposition.This second time-point was approximated to a Q 10 value previously found for the upper soil layer, which averaged 2.17 (Niklińska and Klimek, 2007).For comparison, the Q 10 value for the upper soil horizon in lowland temperate forests in Poland was 2.25 (Klimek et al., 2021).After a metaanalysis of a number of studies on Q 10 values along altitudinal gradients in different climatic zones, Li et al., (2021b) found that the Q 10 for forest soils in the temperate zone was 2.13, which was reported to be a reasonable assumption on a global scale.Therefore, the Q 10 values obtained in our field experiments were comparable with those reported in previous studies.We found that Q 10 L values were higher than Q 10 H (2.19 ± 0.85 and 1.52 ± 0.31, respectively).It is generally observed that decomposition temperature sensitivity is higher at lower temperatures than at higher temperatures (Tuomi et al., 2008;You et al., 2019).This observation is worth highlighting because global climate change in temperate zones affects winter more than summer.This phenomenon may have significant environmental implications, as global climate warming mainly manifests in the winter season (Kreyling et al., 2019;Wang et al., 2021a, b) and at higher geographical latitudes (ICPP, 2023).This is especially important because, until the last century, northern forests and forests at high mountain altitudes were thought to be large carbon sinks (ICPP, 2023).The estimated increase in annual soil respiration rates due to predicted global warming at the high latitudes of the Northern Hemisphere ranged from approximately 0.07 × 10 15 to 0.13 × 10 15 g CO 2 at 2 • C and 4 • C temperature increase scenarios, respectively (Niklińska et al., 1999).Both values are greater than the current annual net carbon storage in northern forests, suggesting a switch in these ecosystems from net sinks to net sources of carbon with global warming.This also means that soil carbon loss is much greater in the cold season than in the warmer part of the year, when temperatures can exceed the optimum for respiration rates, but soil moisture limits the decomposition rate.
A significant interaction between the temperature range and altitude of litter origin in our study indicated that Q 10 L increased with the altitude of litter origin, whereas Q 10 H did not change or even tended to decrease.There was also a significant interaction between temperature range and time, indicating that Q 10 L increased more than Q 10 H over time.Both interactions suggested that the Q 10 values calculated for low and moderate temperatures were more sensitive to various environmental factors than those calculated for elevated temperatures.
Litter respiration rate and temperature sensitivity vary with time, which may be due to the succession of different microbial groups (McMahon et al., 2021).Both the litter respiration rate and Q 10 were positively correlated with the fungal portion of the microbial communities in this study.This was consistent with previous observations that fungal growth dominates during the early stages of plant residue decomposition in the soil (Berg et al., 1998).The microbiome of the phyllosphere may play a certain role, that is, the leaf/needle area (Liu et al., 2022) and foliar endophytes, for which internal plant tissues are the environment (Wolfe and Ballhorn, 2020), as many of them belong to the fungal kingdom.Fungal species composition may differ between forest types and litter, and species composition changes with decomposition time ( Štursová et al., 2021).Averill et al. (2014) found that ecosystems dominated by plants colonised by ectomycorrhizal fungi stored more soil carbon per unit of soil nitrogen than ecosystems dominated by plants colonised by arbuscular mycorrhizal fungi.The effect was controlled for variations in soil nitrogen content and other drivers of soil C storage, such as net primary production, temperature, precipitation, and soil clay content.We can expect a greater proportion of ectomycorrhizal fungi in the community at the highest altitude tested, that is, 1200 m a.s.l., where a larger proportion of conifers was found, compared to the two lower altitudes.Unfortunately, the PLFA method did not enable is to distinguish between microbial species.In addition, the marker for arbuscular mycorrhizal fungi, i.e., fatty acid 16:1ω5 (Bierza et al., 2023), was absent in our samples.
The lack of effect of altitude litterbag exposure on litter respiration rate and Q 10 from the altitude of may be due to the relatively short altitude gradient applied (600 m), which caused only a 3.4 • C mean annual temperature difference between the lowest and highest stands.We did not observe an interaction between the altitude of litter origin and exposure to microbial properties, although an interaction was found for litter physicochemical properties (factor PC1 litter ).Some studies have shown that litter decomposes faster with its local soil microbial Fig. 6.GLM analysis results for Q 10 according to tested categorical factors: a) altitude of litter origin, b) exposition duration, c) temperature range (note that that factor is not statistically significant); according to linear factors: d) fungal biomass share PLFA f (%); and according to interactions: e) between litter altitude origin and temperature range, and f) between temperature range and exposition time (see Table 6).The central points indicate the mean of the data, and the bars show 95 % Tukey confidence intervals.Different letters above bars indicate significant differences between groups (p < 0.05).The course of the interaction between the variables has been highlighted in different colours (see legend).
B. Klimek and M. Niklińska community than when it is exposed to non-native soil (Strickland et al., 2009).Such a pattern, the so-called home field advantage (HFA), is considered to be the combined result of low litter quality (recalcitrance) and the presence of a specialised fungal community (Pugnaire et al., 2023).Such litter was represented in our experiment by litter originating from an altitude of 1200 m a.s.l., as it consisted almost exclusively of spruce litter, which forms the upper montane forest belt in the Western Carpathians (Wasak et al., 2019).In general, the recalcitrant fraction of litter may increase with climate cooling (Berg, 2014), partly due to a decrease in the proportion of soft-leaved plants and also a general decrease in plant diversity with climate cooling.This was evident in our study, as higher soil OM content in stands at 1200 m a.sl.Notably, in a temperate forest ecosystem, massive litterfall occurs with seasonal regularity; therefore, some of the physicochemical properties of the topsoil horizon change substantially throughout the year, and its temperature sensitivity may also vary (Fierer et al., 2006).This, as well as intra-annual variations, may have affected the results of the field experiments.We did not observe an effect of soil properties on the litter respiration rate, as might be expected, but other factors may override this effect.

Conclusions
Field experiments on litter decomposition over two years enabled us to identify some of the parameters of this globally important process under natural conditions.We showed that Q 10 values were higher in the lower temperature range than in the higher temperature range and increased strongly with the duration of litterbag exposure.In addition, some of the microbial community characteristics measured in the decomposing litter, that is, the proportion of fungi in the microbial biomass, were positively correlated with Q 10 .Our results were consistent with the kinetic theory of higher temperature sensitivity in more decomposed forest litter.We also found that the litter Q 10 was more responsive to environmental conditions at lower temperatures, which may have important implications for carbon emissions from mountain forest soils under global climate change conditions.This is particularly important because increases in winter temperatures in temperate climates, including mountainous areas, are continuously reported.The soils of these regions may soon cease to serve as carbon stores in favour of a carbon source by enhancing the positive feedback.The evidence of significantly faster carbon loss during the cold season indicates that temperature changes in mountain forests will not be able to stabilise carbon stocks in the coming decades.In this scenario, positive feedback on climate change will be stronger.These findings demonstrated the importance of the interaction between climatic and environmental parameters for possible carbon losses in temperate forests and showed the potential risk of positive feedback in global carbon cycling.This enables us to emphasise that such climatic changes (higher temperature in the winter season) combined with other negative human activities, such as deforestation, which is often observed in mountainous regions, could cause a real risk for soil carbon sequestration.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.The map of study sites on the map of Poland and the example photos of the study sites on three altitudes.

Fig. 2 .
Fig. 2. Selected soil physico-chemical properties for three studied altitudes (n = 5 each): a) OM content (%), b) Mg content (%), c) K content (%), d) pH, e) sand content in the mineral fraction (%), and finally f) PC1soil values combining these soil data.One-way ANOVA results (p-value) are given in the upper left corner of each figure.Tukey test results are indicated above the bars; altitudes with different letters are significantly different.

Fig. 3 .
Fig. 3. Means and standard deviations for element content (C, N, Ca, K, Mg, and Mn) and C:N ratio in litter originating from three studied altitudes at the beginning of experiment (n = 5).One-way ANOVA results with Tukey test.Values bearing different letters differ significantly between altitudes.

Fig. 4 .
Fig. 4. Principal Component Analysis component plots for a) soil properties and b) litter properties.First and second PCs were presented with % of variance explained by data.Line length indicate an individual factors load.

Fig. 5 .
Fig.5.GLM analysis results for litter respiration rate from litterbags according to tested categorical factors: a) litter altitude origin, b) exposition duration, c) laboratory incubation temperature; according to linear factors: d) litter properties combined into PC1 litter , e) fungal biomass proportion in microbial biomass, PLFA f (%); and according to interactions: f) between litter altitude origin and litter properties PC1 litter , and g) between litter origin and exposition time (see Table6).The central points indicate the mean of the data, and the bars show 95 % Tukey confidence intervals.Different letters above bars indicate significant differences between groups (p < 0.05).The course of the interaction between the variables has been highlighted in different colours (see legend).
Only data used for principal component analysis are shown.Values in the last column indicate PC1 litter values based on these selected data.Three-way ANOVA results with Tukey test are presented at the bottom of the table; statistically significant differences are indicated in bold.Values with different letters are significantly different between levels.

Table 5
GLM model for respiration rate of litter originating and exposed on three studied altitudes.

Table 6
GLM model for Q 10 litter originated and exposed on three studied altitudes.