Effect of Different Oak Forest Management Models on Seasonal Variability in Soil Properties at Sites with Igneous and Sedimentary Subsoil

Abstract

Traditional forest management models could potentially be used to combat changes in environmental conditions by stimulating soil properties and supporting tree growth. In this study, we compare the effects of different oak (Quercus petraea /Matt./Liebl.) forest models [coppice, coppice-with-standard, stocked coppice (reference)] on seasonal variability in soil properties at upland igneous and sedimentary sites (280–418 m a.s.l.). Soils were sampled at 0–5 and 10–15 cm in 50 × 50 m model and reference plots in January, April, July and October between 2015 and 2019, and soil organic matter, soil respiration, enzyme activity, pH and water-holding capacity were determined. The effects of forest model and seasonality were then compared using time-series analysis, analysis of variance and discriminant analysis. Overall, the models differentiated subsurface soil horizons from the topsoil and their feedback varied between sites. While water-holding capacity increased at the igneous stands, acid phosphomonoesterase activity increased and nitrogen content decreased at the sedimentary site. While the most significant negative influence of the forest model on soil properties was observed at the sedimentary site, the greatest increase in soil organic matter and water holding capacity was registered at the igneous coppice-with-standards site. Consequently, using the appropriate forest management model on different subsoil types could be valuable for improving carbon sequestration and drought resistance.

1. Introduction

While oak (Quercus sp.) forests tend to predominate in subtropical to temperate lowlands of the northern hemisphere, the natural occurrence of these oak-dominated forests is currently being negatively affected through droughts brought about by global climate change. Furthermore, changing weather patterns are restricting oak stand growth, leading to their gradual conversion into open forest–steppe ecosystems [1]. Historically, forest managers have responded to the increased vulnerability of oak stands through controlled canopy reduction to stimulate soil-forming processes and increase soil organic matter [2]. In the same way, management-induced stimulation of soil properties has the potential to act as a means of sustaining forest growth in the face of ongoing climate change impacts.

Oaks are dicotyledonous long-living deciduous trees of the order Fagales, which includes approximately 1900 species of anemophilous trees or shrubs in 55 genera and eight families [3]. Quercus, with 438 species and 180 hybrids, is the most abundant genus in the Fagaceae family, which comprises seven genera and 670 species [4]. Asia supports most oak species at 209 sp., followed by North and Central America at 202 sp., of which 55 are used economically. Europe supports just 27 sp., of which only 8 form continuous stands. The lower diversity of European oaks is balanced, however, by the occurrence of several species within the same habitat; hence, oaks differ from other stand-forming tree species that tend to form homogeneous populations [5].

European oak forests acted as refugia for Tertiary biodiversity during the Quaternary cool and warm transitions when many other tree species became extinct or significantly reduced their ranges [6,7]. Current European oak forests can be divided into northern temperate-zone forests, concentrated in the continental subzone, and southern subtropical forests in the dry subzone [8]. The temperate zone comprises habitats for acidophilic oak to oak–birch forests and meso- to thermophilic deciduous forests, while the subtropical subzone is mainly covered with evergreen oak forest [9]. The most fragmented mesophilic oak forests consist of floodplain forests, oak–hornbeam forests and thermophilic oak forests approaching forest–steppe conditions [10].

During the Quaternary, oak forest cover was altered significantly by human land management. Natural regeneration of oak forests relies on occasional disturbances that remove shade-tolerant woody plants, which otherwise suppress the growth of light-demanding species such as oak [11]. However, human activity, such as the wide-spread clear-cut removal of shade-tolerant species, river diversion and harvesting of acorns have led to the expansion of secondary oak growth [12]. As a consequence, mesophilic oak forests have gradually been transformed into mixed oak–hornbeam (Carpinus sp.) forests, while the percentage of oak in floodplain forests has dropped due to regular flooding, freeing up habitat for willow (Salix sp.), ash (Fraxinus sp.) and alder (Alnus sp.). The greatest impact on forest structure and species composition, however, has been exerted through timber harvesting and/or animal grazing, which has divided floodplain forests into softwood forests comprising pioneer woody species overgrowing periodically flooded riverbed sediments and hardwood forests beyond the reach of regular floods, generally on shallower soils [13]. Overall, traditional forest management models in the European temperate zone have shifted oak distribution away from sites where oak–hornbeam forests were naturally dominant [11] to sites previously covered by other stand-forming species less tolerant to coppicing [14]. As a consequence, oak forests, and oak coppice in particular, are now rare in Central Europe.

Coppicing is a management model that has been used for the longest time in oak stand cultivation. Coppicing, entailing the cutting of trees to encourage regrowth from the stump (stool), being useful for firewood, charcoal making, tool production, building material, etc., has been applied to several unrelated tree groups, including floodplain–forest species of birch (Betulaceae), hornbeam, willow and alder, as well as species of Lamiales, Malvales and Sapindales. Nevertheless, oaks have tended to be preferred in Central Europe due to their longevity, their durable wood and their production of acorns, which can be used as fodder for ungulates [14]. In turn, coppicing maintains oak populations via the production of trees of generative origin [2]. The admixture of non-coppice trees (standards) in coppice stands, known as coppice-with-standard (CWS), represents a structured transition between coppice and high-stem forests of generatively propagated trees. In this case, while coppice stools provide firewood over short rotation periods, standards provide higher-quality construction wood over regular longer-term rotations. Historically, the transformation of virgin forest to CWS appears to have taken place for the longest period in floodplain areas [15]. Though these two-storey CWS forests generally increase aboveground biodiversity, the intensive removal of organic matter during rotational harvests can result in ecosystem nutrient loss, thereby limiting metabolic opportunities to the point where forest communities become noticeably degraded [16].

The primary route for preventing forest ecosystem degradation is through the improvement of soil properties, and more specifically through increasing soil carbon content [17], which enhances the soil’s ability to retain water and nutrients. The effectiveness of such measures, however, depends on maintaining a positive balance between the duration of carbon storage and that of carbon loss [18]. Soil carbon content is increased most efficiently through promoting exchange between fine plant roots and soil microorganisms, whereby soil biota reliably separate labile and stable carbon compounds [19]. In this case, the ability of forests to stimulate soil carbon storage depends on their species composition and stand structure, with most deciduous trees having a high proportion of fine roots, crucial for soil carbon storage, as well as having the ability to regrow from stumps. Consequently, CWS forests are likely to facilitate soil carbon storage more than coppice alone or other structurally simple high-stem stands [20]. On the other hand, stimulation of soil carbon sequestration by woody species may also be accompanied by an increase in losses due to more rapid nutrient turnover and total respiration, the most important ecosystem carbon outputs. Soil respiration could be balanced, however, by maintaining water availability and adequate phosphorus pentoxide (P₂O₅) levels and carbon/nitrogen (C/N) ratios [21].

As a species, oaks are not restricted by local subsoil type/chemical composition; however, the rest of the forest community (understory and soil) will usually be adapted to the local soil’s pH, with varying nutrient availability at acid and alkaline sites [22]. When accessing nutrients, the interaction of soil organisms with roots modifies the soil ecosystem in relation to the external environment for the next generation of plants [23]. The product of such biological interactions, i.e., active soil organic matter (including enzymes), accelerates rock weathering, subsoil movement and topsoil development. The effect of trees on soil is most pronounced up to a depth of 5 cm, with impacts at greater depths only being observed at pH > 4.5 [24]. Under acidic conditions, therefore, the effect of trees on soil development is mitigated, whist the effect of subsoil increases [22]. Present-day oak forest soils are increasingly threatened by degradation, with increasing drought periods and acidification leading to the formation of forest–steppe and acidophilic forest conditions [10]. This degradation is exacerbated by seasonal changes in soil water availability and shifts in physicochemical and biochemical properties and biological activity [25]. Soil acidity fluctuations and soil organic matter content feedback alter the conditions for biological activity [26], and thus also alter both the ecosystem’s input–output balance and the availability of nutrients in it. Using suitable forest management models focused on long-term storage of soil organic matter has the potential to minimise such ecosystem losses [2]. The present ecosystem study compares the effects of the forest management model on the variability in soil properties between sites of differing degradation susceptibility, based on the assumptions that (1) forest models affect soil development by altering organic matter content and (2) subsoil moderates forest effects based on soil reaction. We hypothesise that (a) the CWS forest model will have a mitigating effect on drought by increasing organic matter content and that (b) optimally fertile sedimentary subsoil will moderate the effect of the different forest models due its overall high buffering capacity.

2. Material and Methods

2.1. Experimental Sites

The experimental sites used in this study were situated on igneous and sedimentary subsoils in the eastern part of the Czech Republic on the Brno Massif (Figure 1), part of the Central European lower altitudes where oak forests previously dominated [11]. While forests currently cover more than 37% of the Czech landscape (29,232 km²), oak forests now represent just 6.8% [27]. Czech oak forests grow mainly on marls (15.8%), greywackes (10.3%), metamorphites (9.4%), neutral granites (7.4%), loess and silty loams (5.4%) and neovolcanites (4.8%). On the Brno Massif (807 km²), oak forests cover 19.9 km² and are mainly found on granites (74.6%), (weakly) alkaline crystalline rocks (9.5%), Permian sediments (6.2%) and silty loams (4.5%) [28].

The sites comprised two >80-year-old sessile oak (Quercus petraea /Matt./Liebl.) coppice stands on contiguous upland igneous and sedimentary subsoil within the University Enterprise Masaryk Forest—Křtiny, managed by Mendel University (Figure 1). The study was limited to two experimental pilot sites [2]. The igneous site (49.24′50″ N; 16.60′10″ E; 280–310 m a.s.l.) lies on neutral granite bedrock and has a coarser organic-rich topsoil, while the sedimentary site (49.22′50″ N; 16.68′20″ E; 409–418 m a.s.l.) lies on a decayed limestone subsoil with bars of Quaternary loam and has a finer, mineral-rich topsoil. The average mean temperature on the Brno Massif is +8.3 °C and the annual precipitation is 559 mm; however, the annual mean precipitation throughout the experimental period was reduced to 427 mm, reflecting ongoing drought conditions in the region.

Within each stand, two 50 × 50 m sub-plots with stocked coppice forests, now unmanaged intensively, were identified to act as reference sites (Figure 1). In addition, a 4 × 4 experimental plot comprising 16,200 × 200 m sub-quadrants was laid out with each quadrant subject to one of four whole-tree thinning intensities ranging across clearcut (coppice) and CWS (Figure 1; overall average tree density reduced from 660 n/ha to 140 n/ha, and stocking density from 308 m³/ha to 108 m³/ha) [2]. Soil surveys were performed inside the two coppice forest quadrants and two CWS quadrants, i.e., on plots with the lowest thinning intensity, with stock reduced by about 54%.

2.2. Soil Properties

The soil survey consisted of a soil body assessment according to WRB-ISSS-ISRIC [29], while soil horizon properties were assessed according to EMEP-LRTAP [30], with repeat samplings of three replicates taken from a randomised 5 × 5 m Latin square grid around the centre of each sub-plot. Surveys took place in January, April, July and October between 2015 and 2019, to coincide with the peak phenological phases under temperate deciduous forests [25]. Samples were obtained from depths of 0–5 cm in the topsoil (A−) horizon and 10–15 cm in the subsoil (B−) horizon as (hydro)physical, physicochemical, (bio)chemical and biological sets, with correlations between these being used as a measure of the ecosystem’s ability to resist degradation by drought and acidification [31].

Soil physical properties were explored through texture analysis, undertaken using the pipette method [29], and hydrophysical properties, including bulk density (D_d), specific density (D_s), water holding capacity (WHC), porosity and aeration, using 100 cm³ cores [32]. Physicochemical properties were characterised by soil acidity, determined as active pH (H₂O) [33], soil sorption, assessed for soil pit horizons using the cation exchange capacity (CEC), determined as the sum of exchangeable cation concentrations by 0.1 M BaCl₂ extraction; and base saturation (BS), determined as the proportion of exchangeable bases from CEC [30]. Biochemical properties were expressed through organic matter chemistry and enzyme activity. Soil organic matter composition was characterised through elemental analysis using two methods: (i) dry combustion toward H-H₂O (H_tot), N-NO_x-N₂ (N_tot) and S-SO₂ (S_tot) production and determination in a temporarily oxygenated He-atmosphere on a temperature-conductivity sensor [34], and (ii) dry combustion–IR spectrometry for C-CO₂ thermal differentiation in an oxygenated N atmosphere [35] where carbon fractions were determined as readily available organic carbon (C_org) at 400 °C, recalcitrant carbon (C_rec) at 600 °C and inorganic carbon (C_inorg) at 900 °C. Microbially bound carbon (C_mic) was determined by fumigation–extraction [36], while carbon losses from the soil were characterised through substrate-induced respiration (SIR) [37].

Assessment of enzyme activity focused on hydrolytic processes limiting nutrient release in the presence of catalase [31]. Hydrolytic enzyme activity was determined spectrophotometrically from comparative incubations of basal and substrate-induced soil samples, both samples being incubated at 37 °C at the optimum pH for enzyme activity. Concentrations of substances accessed through enzyme activity were then measured spectrophotometrically. Phosphorus availability was estimated via acid phosphomonoesterase activity (APMEA) [38], and nitrogen availability was estimated via urease activity (UA) [39], while catalase activity (CA), indicating the efficiency of aerobic processes, was determined volumetrically [37].

2.3. Statistical Analysis

The effects of the different management models were compared with site effects using exploratory analysis and linear modelling of periodically determined soil property values, with significance in each case set at p < 0.05. Exploratory analysis consisted of basic statistics, i.e., the mean value ($\overline{x}$) and standard deviation (SD), time-series analysis and data normality tests, including skewness (A₁) and elevation (E₁) [40]. Time-series analysis was used to identify trends in the development of soil property values using relative deviations (δ_∆):

$${\delta }_{∆}={\displaystyle \frac{\frac{\sum _{t=1}^T{x}_t-\overline{x}}{T}}{\overline{x}}}·100$$

(1)

where T is the number of observation periods, t is the order of the period over total monitoring time and x_t is the value of the soil property at time t. Any trends were then used to verify differences between the effects of seasonality and forest management model estimated using linear models.

Linear modelling was used to differentiate the effect weight for subsoil and forest on soil development. The effects of season and forest model were compared using two-way analysis of variance (ANOVA). The effects of data distribution violations against the significance of differences in soil property variability were eliminated using Kruskal–Wallis test crosschecks. In each case, the Fischer–Snedecorov ANOVA criterion (F_0.05) was accepted when data normality was met, when the difference in values was significant or when a robust H criterion was obtained, even when normality was violated [41].

The forest model with the greatest effect on soil properties was selected through discriminant analysis, which was used to define soil properties that differ significantly between forest models in at least one of the sites compared. The appropriate forest model was distinguished using the parameters of individual soil properties (a_q) in the discriminant function [42] below:

$$Z_i={\sum }_{k=1}^q{a}_q.x_{iq}$$

(2)

where i is the number of forest models; q is the number of soil properties; Z_i is the standardised classification coefficient for the forest model; a is the parameter of the discrimination vector determined by the ratio of differences in soil properties (${\overline{x}}_j-{\overline{x}}_{j+1}$) with the covariance matrix (Q); and x is the standardised soil property:

$$a_q={\displaystyle \frac{({\overline{x}}_j-{\overline{x}}_{j+1})}{Q}}$$

(3)

where j is the number of forest models classified. The threshold point of soil property correspondence with the forest model (C) was determined using the half-sum between averages of classification coefficients from the discriminant functions for individual forest models at each site (Z_j):

$$C={\displaystyle \frac{{\overline{Z}}_j+{\overline{Z}}_{j+1}}{2}}$$

(4)

where

$${\overline{Z}}_{j=}{\displaystyle \frac{\sum _{i=1}^j{Z}_i}{i}}$$

(5)

The reliability of the discriminant function parameters was assessed via the separability from the proportion of correctly classifiable forest models [43].

3. Results

Soils at the igneous site comprised coarse sandy-loam Haplic Cambisols, while those at the sedimentary site were typically finer silty-loam Eutric Cambisols. The pH (average < 5.8), BS, C_org and N_tot were all lower at the igneous site, with BS as low as 21%. In comparison, the sedimentary soil had an average pH of >6, almost complete BS, and nearly double the content of organic matter (

Table 1. Physicochemical and humic characteristics of soil in different sessile oak forest models at igneous and sedimentary sites. pH(H₂O)—active soil reaction; CEC—cation exchange capacity (cmol⁺/kg); BS—base saturation (%); C_org—soil organic carbon (%); Ntot—total soil nitrogen (%).

Site	Horizon	Texture	pH(H2O)	CEC	BS	Corg	Ntot	C/N
Igneous	Humus		6.28	17.06	97.07	9.89	0.65	15.22
	Topsoil	sandy-loam	5.78	7.82	71.85	3.91	0.24	16.29
	Subsurface	sandy-loam	5.42	6.76	21.56	2.61	0.21	12.43
	Weathered	sandy-loam	5.14	5.41	37.19	0.16	0.03	5.33
Sedimentary	Humus		6.58	45.78	99.13	16.28	1.36	11.97
	Topsoil	silty-loam	6.77	38.43	99.22	8.4	0.53	15.85
	Subsurface	silty-loam	6.05	20.67	99.03	4.06	0.32	12.69
	Weathered	silty-loam	7.79	23.86	99.16	4.88	0.29	16.83

).

Differences in the variability in soil properties were indicated by the differences in the distribution of values across different soil layers at both sites. At both the igneous and sedimentary sites, soil properties differed greatly between horizons, with site having a greater effect than forest model. On the other hand, there was little difference between C_org, N_tot or S_tot content and soil density characteristics at either site. APMEA and CA, C_mic and porosity were all higher in the topsoil horizon at both sites, while UA, respiration and WHC were higher in the igneous topsoil horizon only. In contrast, aeration was lower in the topsoil at both sites, while pH, UA, respiration and WHC were all higher in sedimentary subsurface horizons (

Table 2. Exploratory statistics for soil properties in different sessile oak forest models at igneous and sedimentary sites (significant differences at p < 0.05 in bold). SD—standard deviation; E1—elevation test criterion; A1—asymmetry test criterion; pH(H₂O)—active soil reaction; APMEA—acid phosphomonoesterase activity (mg/g.hour); UA—urease activity (mg/g.hour); CA—catalase activity (ml/g.min); SIR—substrate induced respiration (mg C/g.hour); C_org—available soil organic carbon (%); C_mic—microbial carbon (mg/g); C_rec—recalcitrant carbon (%); C_inorg—inorganic carbon (%); N_tot—total soil nitrogen (%); H_tot—total soil hydrogen (%); S_tot—total soil sulphur (%); D_d—bulk density (g/cm³); D_s—specific density (g/cm³); WHC—water-holding capacity (%).

Site		Igneous				Sedimentary
Soil Depth	Variable	Mean ± SD	Min–Max	E1	A1	Mean ± SD	Min–Max	E1	A1
0–5 cm	pH(H2O)	5.80 ± 0.60	4.18–7.22	1.55	2.26	5.80 ± 0.63	4.20–7.22	1.23	0.72
	APMEA	316.25 ± 133.31	79.38–777.63	0.36	3.90	378.95 ± 130.2	127.47–777.63	0.97	2.65
	UA	66.65 ± 41.43	11.16–279.6	10.88	9.72	85.07 ± 50.52	11.16–279.6	3.08	4.68
	CA	26.83 ± 12.40	6.25–57.51	2.16	3.50	25.57 ± 12.20	7.17–54.25	1.98	1.73
	SIR	9.45 ± 6.59	0.13–28.68	1.68	4.92	12.84 ± 6.50	2.69–26.71	2.29	1.33
	Corg	4.36 ± 1.80	1.02–10.43	1.32	5.82	5.55 ± 1.95	1.90–10.43	1.34	1.32
	Cmic	301.19 ± 170.98	3.25–862.65	0.08	4.59	412.64 ± 174.28	32.7–862.65	1.25	1.24
	Crec	0.33 ± 0.10	0.09–0.6	0.70	0.29	0.38 ± 0.10	0.09–0.60	0.25	1.96
	Cinorg	0.05 ± 0.02	0.01–0.19	28.13	14.24	0.05 ± 0.03	0.02–0.19	7.46	6.58
	Ntot	0.32 ± 0.16	0.08–0.8	0.01	5.95	0.45 ± 0.16	0.14–0.80	1.79	0.14
	Htot	0.84 ± 0.35	0.32–1.78	1.82	4.93	1.20 ± 0.25	0.72–1.78	1.48	0.23
	Stot	0.03 ± 0.01	0.01–0.07	1.22	4.86	0.03 ± 0.01	0.01–0.07	0.32	1.22
	Dd	1.12 ± 0.20	0.51–1.54	0.42	1.03	1.03 ± 0.16	0.55–1.41	0.24	1.20
	Ds	2.42 ± 0.11	1.95–2.66	4.15	5.05	2.38 ± 0.13	1.95–2.63	0.53	2.33
	WHC	35.44 ± 7.48	20.85–53.03	3.08	0.55	42.32 ± 4.48	31.26–53.03	0.73	0.14
	Porosity	54.33 ± 8.49	28.95–95.43	12.48	6.98	57.73 ± 8.8	28.95–95.43	13.23	5.74
	Aeration	33.06 ± 18.12	8.25–257.42	296.16	48.09	26.70 ± 11.08	8.25–50.49	1.30	2.06
10–15 cm	pH(H2O)	5.80 ± 0.60	4.18–7.22	1.53	2.25	5.81 ± 0.63	4.20–7.22	1.24	0.80
	APMEA	315.88 ± 133.21	79.38–777.63	0.38	3.95	378.64 ± 129.63	127.47–777.63	1.05	2.70
	UA	66.48 ± 41.45	11.16–279.6	10.87	9.74	85.19 ± 50.3	11.16–279.6	3.15	4.68
	CA	26.79 ± 12.39	6.25–57.51	2.15	3.55	25.56 ± 12.14	7.17–54.25	1.94	1.75
	SIR	9.44 ± 6.58	0.13–28.68	1.64	4.96	12.87 ± 6.48	2.69–26.71	2.29	1.27
	Corg	4.36 ± 1.80	1.02–10.43	1.34	5.85	5.55 ± 1.94	1.90–10.43	1.29	1.35
	Cmic	300.24 ± 171.39	3.25–862.65	0.06	4.57	410.75 ± 174.58	32.7–862.65	1.27	1.31
	Crec	0.33 ± 0.10	0.09–0.6	0.72	0.34	0.38 ± 0.10	0.09–0.6	0.32	2.02
	Cinorg	0.05 ± 0.02	0.01–0.19	28.23	14.29	0.05 ± 0.03	0.02–0.19	7.64	6.65
	Ntot	0.32 ± 0.16	0.08–0.8	−0.01	5.98	0.45 ± 0.16	0.14–0.80	1.76	0.17
	Htot	0.84 ± 0.35	0.32–1.78	1.80	4.98	1.21 ± 0.25	0.72–1.78	1.51	0.15
	Stot	0.03 ± 0.01	0.01–0.07	1.22	4.90	0.03 ± 0.01	0.01–0.07	0.37	1.19
	Dd	1.12 ± 0.20	0.51–1.54	0.39	1.04	1.03 ± 0.16	0.55–1.41	0.19	1.25
	Ds	2.42 ± 0.11	1.95–2.66	4.20	5.05	2.38 ± 0.13	1.95–2.63	0.60	2.36
	WHC	35.42 ± 7.47	20.85–53.03	3.09	0.59	42.33 ± 4.46	31.26–53.03	0.68	0.13
	Porosity	54.32 ± 8.48	28.95–95.43	12.59	7.02	57.69 ± 8.77	28.95–95.43	13.42	5.83
	Aeration	33.11 ± 18.11	8.25–257.42	295.85	48.02	26.89 ± 11.20	8.25–50.49	1.42	2.00

). As soil values were similar between sites, assumptions of data normality could only be confirmed for D_d and WHC at both sites, and for pH, CA and all organically bound elements at the sedimentary site.

Despite the diverse effects of site on soil horizon properties, the forest model aligned several soil property values along the coppice, CWS and reference forest gradient in descending or ascending order. The igneous site, for example, showed descending pH, UA, C_mic, S_tot and N_tot values overall, and descending APMEA, respiration, C_org, C_rec and WHC values in subsurface horizons (Figure 2, Figure 3 and Figure 4). Only pH and soil respiration showed the same gradient at both sites. Igneous CWS forest had higher C_inorg and porosity values, though APMEA, C_org and WHC values were high in topsoil only and aeration was high in the subsurface horizon only. The topsoil at the igneous CWS site had the lowest CA, C_rec and D_d values. In comparison, the sedimentary site displayed an increasing trend for APMEA in topsoil, and an increasing trend in SIR, C_rec and WHC in subsoil at the CWS site. In contrast, CA, D_d and D_s showed decreasing trends in sedimentary subsoil, as did pH and SIR on the surface horizon. Only C_inorg, N_tot, S_tot, porosity and aeration remained the same in both horizons. The sedimentary CWS forest site generally displayed significantly higher soil property values than the igneous CWS forest site, with the highest values detected for C_mic in both horizons, for C_rec and WHC in the surface horizon and for pH in the subsurface horizon. In contrast, values for UA and C_org were significantly lower at the sedimentary site (

Table 3. Soil properties in different sessile oak forest models at igneous and sedimentary sites where CWS—coppice with standard; pH(H₂O)—active soil reaction APMEA—acid phosphomonoesterase activity (mg/g.hour); UA—urease activity (mg/g.hour); CA—catalase activity (ml/g.min); SIR—substrate induced respiration (mg C/g.hour); C_org—available soil organic carbon (%); C_mic—microbial carbon (mg/g); C_rec—recalcitrant carbon (%); C_inorg—inorganic carbon (%); N_tot—total soil nitrogen (%); H_tot—total soil hydrogen (%); S_tot—total soil sulphur (%); D_d—bulk density (g/cm³); D_s—specific density (g/cm³); WHC—water-holding capacity (%); Porosity (%) and Aeration (%).

Site		Igneous			Sedimentary
Soil Depth	Variable	Coppice	CWS	Reference	Coppice	CWS	Reference
0–5 cm	pH(H2O)	6.16 ± 0.37	6.11 ± 0.48	5.92 ± 0.55	6.28 ± 0.37	6.06 ± 0.5	5.76 ± 0.45
	APMEA	341.7 ± 95.63	372.87 ± 89.32	372.09 ± 108.35	435.91 ± 76.65	455.15 ± 112.91	498.71 ± 136.99
	UA	79.83 ± 26.02	79.51 ± 23.43	69.49 ± 20.83	129.71 ± 53.19	109.66 ± 32.79	114.58 ± 53.44
	CA	32.03 ± 13.05	31.6 ± 12.88	33 ± 13.3	29.45 ± 13.69	29.1 ± 12.9	26.09 ± 11.39
	SIR	10.32 ± 5.65	10.67 ± 5.82	11.19 ± 6.44	18.7 ± 4.58	18.06 ± 4.43	17.91 ± 3.65
	Corg	4.46 ± 0.75	4.83 ± 0.91	4.3 ± 0.77	6.58 ± 1.5	6.56 ± 1.08	7.99 ± 1.25
	Cmic	293.74 ± 99.91	328.76 ± 124.16	300.38 ± 131.29	516.72 ± 149.46	570.32 ± 127.85	502.3 ± 137.57
	Crec	0.35 ± 0.05	0.33 ± 0.05	0.36 ± 0.04	0.42 ± 0.05	0.46 ± 0.08	0.45 ± 0.07
	Cinorg	0.05 ± 0.01	0.05 ± 0.01	0.05 ± 0.01	0.06 ± 0.02	0.06 ± 0.02	0.08 ± 0.03
	Ntot	0.3 ± 0.05	0.3 ± 0.07	0.28 ± 0.04	0.54 ± 0.08	0.55 ± 0.07	0.64 ± 0.07
	Htot	0.69 ± 0.12	0.73 ± 0.1	0.67 ± 0.1	1.25 ± 0.15	1.3 ± 0.17	1.51 ± 0.15
	Stot	0.03 ± 0.01	0.03 ± 0.01	0.02 ± 0.01	0.04 ± 0.01	0.04 ± 0.01	0.04 ± 0.01
	Dd	1.03 ± 0.18	1 ± 0.17	1.1 ± 0.08	0.99 ± 0.15	0.93 ± 0.08	0.81 ± 0.11
	Ds	2.4 ± 0.07	2.4 ± 0.07	2.42 ± 0.07	2.35 ± 0.1	2.33 ± 0.1	2.24 ± 0.13
	WHC	34.72 ± 4.32	35.28 ± 4.99	30.12 ± 4.52	43.48 ± 2.66	44.66 ± 3.24	44.25 ± 5.04
	Porosity	56.84 ± 7.26	57.99 ± 6.59	54.41 ± 3.54	58.04 ± 5.89	63.15 ± 11.85	64.41 ± 7.7
	Aeration	37.58 ± 11.89	39.11 ± 11.87	46.19 ± 41.95	26.31 ± 10.61	28.12 ± 10.22	31.23 ± 11.85
10–15 cm	pH(H2O)	5.71 ± 0.47	5.49 ± 0.62	5.43 ± 0.57	5.55 ± 0.54	5.71 ± 0.88	5.49 ± 0.61
	APMEA	194.98 ± 64.7	188.5 ± 67.88	183.05 ± 55.72	279.62 ± 69.72	261.04 ± 82.99	341.42 ± 72.15
	UA	35.07 ± 11.62	33.64 ± 14.16	29.19 ± 11.12	55.39 ± 23.9	47.15 ± 22.89	54.66 ± 30.99
	CA	23.72 ± 10.77	22.2 ± 9.72	22.96 ± 10.52	23.53 ± 11.57	22.9 ± 12.13	22.3 ± 10.51
	SIR	3.81 ± 2.13	3.78 ± 2.41	3.62 ± 1.86	6.98 ± 3.02	7.29 ± 3.01	8.27 ± 2.98
	Corg	2.74 ± 0.54	2.71 ± 0.65	2.48 ± 0.48	3.71 ± 0.77	3.67 ± 1.04	4.77 ± 0.8
	Cmic	159.11 ± 74.67	155.85 ± 69.61	137.39 ± 76.43	287.94 ± 92.18	305.56 ± 131.59	281.63 ± 123.86
	Crec	0.26 ± 0.08	0.23 ± 0.05	0.22 ± 0.06	0.28 ± 0.11	0.32 ± 0.09	0.39 ± 0.07
	Cinorg	0.03 ± 0.01	0.04 ± 0.01	0.03 ± 0.01	0.03 ± 0.01	0.03 ± 0.03	0.04 ± 0.02
	Ntot	0.18 ± 0.04	0.16 ± 0.04	0.15 ± 0.03	0.27 ± 0.08	0.3 ± 0.1	0.41 ± 0.09
	Htot	0.53 ± 0.1	0.52 ± 0.09	0.48 ± 0.07	0.96 ± 0.22	1.03 ± 0.22	1.18 ± 0.16
	Stot	0.02 ± 0	0.02 ± 0.01	0.01 ± 0	0.02 ± 0.01	0.03 ± 0.01	0.03 ± 0.01
	Dd	1.32 ± 0.15	1.32 ± 0.13	1.29 ± 0.1	1.2 ± 0.11	1.15 ± 0.05	1.08 ± 0.05
	Ds	2.48 ± 0.08	2.48 ± 0.08	2.48 ± 0.08	2.46 ± 0.09	2.46 ± 0.08	2.41 ± 0.1
	WHC	30.66 ± 4.26	29.16 ± 4.07	25.92 ± 3.63	38.97 ± 3.09	41 ± 5.08	41.61 ± 4.61
	Porosity	46.75 ± 5.59	48.65 ± 8.85	48.27 ± 3.58	50.68 ± 8	53.42 ± 2.31	56.42 ± 5.26
	Aeration	31.42 ± 9.46	34.44 ± 12.87	33.96 ± 8.02	23.76 ± 11.05	25.44 ± 11.94	26.48 ± 9.5

).

Relative deviations (

Table 4. Relative deviations in soil development under different sessile oak forest models at igneous and sedimentary sites where CWS—coppice with standard; pH(H₂O)—active soil reaction APMEA—acid phosphomonoesterase activity (mg/g.hour); UA—urease activity (mg/g.hour); CA—catalase activity (ml/g.min); SIR—substrate induced respiration (mg C/g.hour); C_org—available soil organic carbon (%); C_mic—microbial carbon (mg/g); C_rec—recalcitrant carbon (%); C_inorg—inorganic carbon (%); N_tot—total soil nitrogen (%); H_tot—total soil hydrogen (%); S_tot—total soil sulphur (%); D_d—bulk density (g/cm³); D_s—specific density (g/cm³); WHC—water-holding capacity (%); Porosity (%) and Aeration (%).

Site	Igneous						Sedimentary
Stand	Coppice		CWS		Reference		Coppice		CWS		Reference
Soil Depth (cm)	0–5	10–15	0–5	10–15	0–5	10–15	0–5	10–15	0–5	10–15	0–5	10–15
pH(H2O)	1.00	0.37	1.75	1.91	1.81	1.71	0.83	1.02	−0.63	−1.18	0.98	−0.79
APMEA	1.21	1.36	−0.92	2.82	−1.28	2.06	−3.24	2.25	0.94	1.29	−4.37	−1.09
UA	2.57	2.38	1.26	4.98	0.11	1.93	−3.05	7.56	0.01	−0.49	−4.33	−5.65
CA	−1.90	−1.22	−3.04	−0.63	−0.40	1.43	1.16	6.73	3.30	2.11	1.08	3.21
SIR	12.84	7.24	6.58	13.60	5.81	8.25	4.94	8.78	7.96	4.66	4.94	5.68
Corg	−0.11	1.24	−0.74	2.63	−2.10	−1.11	−4.27	1.16	−0.44	6.36	−3.49	−0.26
Cmic	−4.72	−6.59	−7.16	−4.46	−6.08	−6.08	−7.61	−5.48	−5.41	−6.84	−7.75	−12.36
Crec	−0.52	0.73	−1.37	3.69	0.84	0.41	0.75	−2.66	0.48	2.59	−1.26	−0.40
Cinorg	−2.65	−1.66	−1.15	1.86	−4.04	−2.83	−0.79	−4.49	−1.92	−20.32	−1.48	−0.96
Ntot	−0.37	−1.18	−0.93	2.55	−0.53	0.61	−1.88	−2.03	0.18	−3.55	−0.20	−0.23
Htot	0.16	2.78	−0.02	3.03	0.10	1.89	−0.36	−4.41	0.96	−3.32	0.97	0.70
Stot	−3.35	−1.79	−3.06	1.44	−2.16	1.09	−7.32	−8.07	−7.49	−7.42	−6.51	−3.96
Dd	−0.35	−1.42	−0.78	−1.26	0.20	−1.07	0.78	−1.05	1.48	−0.74	3.14	−0.24
Ds	−0.23	−0.29	−0.24	−0.60	−0.29	−0.65	0.57	−0.17	0.65	−0.16	1.39	−0.04
WHC	0.34	0.95	0.65	1.31	1.79	1.74	−0.14	0.49	−0.40	0.02	−0.31	0.02
Porosity	−0.59	1.33	−0.19	1.01	−0.56	0.71	−2.02	1.62	−7.91	−0.52	−4.93	−2.87
Aeration	−0.26	2.74	0.03	1.56	0.63	2.94	14.76	17.28	13.70	17.16	10.57	14.53

) confirmed the development of soil properties (Figure 2, Figure 3, Figure 4 and Figure 5) via six main trends described below. Full co-occurrence of development trends was not observed between horizon or sites. Soil properties decreased and increased simultaneously in all forest models, despite the decreases or increases observed in CWS forest relative to the other models. Surface horizon trends tended to be similar, with the subsurface usually only at the same site, while the trends in the subsurface horizons were more similar between sites than between horizons (i.e., C_org). C_mic and specific density tended to decrease in topsoil and subsoil horizons at the igneous site, regardless of forest model, while C_org, S_tot and porosity decreased at both sites in the topsoil only. N_tot and CA decreased in the topsoil at the igneous site, while soil pH, respiration, UA and WHC increased. C_mic decreased in the subsoil at the igneous site, whereas pH, APMEA, UA, C_rec and porosity all increased. Similar decreasing trends in B_d and S_d, and increasing trends in SIR, WHC and aeration, were observed in the subsoil at both sites (Figure 5).

At the igneous site, CWS forest differed from coppice by its lower APMEA and higher SA in the topsoil, and higher N_tot and S_tot on the subsurface horizons. At the sedimentary site, WHC was lower on the surface horizon and C_inorg and S_tot were lower on both the topsoil and subsurface horizons. Aside from an increase in H_tot, none of the development trends clearly separated coppice and CWS forest management models at either depth (Table 4).

At the sedimentary site, forest model had the most significant effect on soil properties, while season was the most significant influence at the igneous site. Season had a significant effect on pH, CA, respiration, WHC and aeration at both sites, but it only had a significant effect on APMEA, C_mic and C_inorg at the igneous site. The forest model at the sedimentary site mainly influenced pH, C_org, C_inorg, N_tot and D_d in the topsoil, and APMEA, C_org, C_inorg, N_tot, D_d and porosity in the subsurface horizons (Figure 2, Figure 3, Figure 4 and Figure 5). At the igneous site, the forest model significantly affected C_org, D_d and WHC in the topsoil and C_rec and WHC on the subsurface horizons (

Table 5. Analysis of variance (ANOVA) for seasonal differences in soil properties in different sessile oak forest models (significant differences at p < 0.05 in bold) at igneous and sedimentary sites (for an explanation of the variables, see Table 2).

Variable	Site	Igneous				Sedimentary
	Soil Depth	0–5 cm		10–15 cm		0–5 cm		10–15 cm
	Factor	F0.05	H0.05	F0.05	H0.05	F0.05	H0.05	F0.05	H0.05
pH(H2O)	Forest	2.54	2.47	1.83	3.43	7.31	10.95	0.45	0.52
	Season	12.48	22.17	6.29	14.20	7.61	16.02	2.04	5.22
APMEA	Forest	0.81	2.47	0.26	0.43	1.50	1.82	5.66	9.02
	Season	3.72	10.07	3.57	11.55	1.50	5.50	2.28	6.15
UA	Forest	2.20	2.69	1.96	3.37	0.96	1.60	0.31	1.08
	Season	3.04	7.60	2.76	6.70	2.29	5.61	0.19	1.47
CA	Forest	0.07	0.13	0.18	0.29	0.87	0.75	0.11	0.04
	Season	6.49	20.30	3.21	13.25	17.76	26.64	14.60	27.84
SIR	Forest	0.32	0.40	0.21	0.09	0.19	0.76	0.93	1.76
	Season	54.66	54.96	33.98	49.83	9.12	18.50	2.18	7.08
Corg	Forest	3.26	6.61	2.16	3.88	6.67	12.00	10.17	13.98
	Season	0.41	0.77	1.57	3.23	0.58	2.20	2.12	5.42
Cmic	Forest	0.75	1.37	0.90	1.25	1.24	2.76	0.31	0.22
	Season	4.16	12.52	5.06	14.09	2.16	6.67	0.79	3.59
Crec	Forest	2.38	3.95	2.54	3.92	1.38	4.10	7.55	11.82
	Season	2.19	6.86	2.27	6.41	0.12	0.54	1.01	1.99
Cinorg	Forest	0.11	0.21	0.60	1.31	4.03	10.10	1.03	8.52
	Season	4.95	12.88	4.16	10.99	0.29	1.55	0.36	1.81
Ntot	Forest	1.19	1.98	3.06	6.98	9.01	15.12	10.10	17.72
	Season	0.05	0.17	0.62	2.09	1.51	2.75	0.38	0.57
Htot	Forest	2.24	3.97	3.04	6.76	13.58	17.11	6.21	12.65
	Season	2.77	8.45	4.66	12.78	2.26	4.08	2.78	6.31
Stot	Forest	1.59	4.07	3.06	6.67	1.98	5.20	7.20	13.05
	Season	0.19	0.75	0.09	0.37	0.23	0.87	0.31	0.56
Dd	Forest	3.22	7.22	0.68	2.27	9.80	13.32	15.02	19.47
	Season	0.79	4.69	0.41	1.60	0.92	2.07	0.94	1.70
Ds	Forest	1.05	1.63	0.05	0.89	3.78	6.55	1.59	2.76
	Season	1.30	5.52	4.28	11.01	1.53	5.16	0.78	2.43
WHC	Forest	9.84	17.68	14.51	20.94	0.40	1.43	1.62	3.91
	Season	2.89	6.23	15.14	19.56	3.07	7.02	1.68	5.42
Porosity	Forest	2.54	4.90	0.57	1.54	2.14	5.85	3.68	7.64
	Season	0.74	4.01	0.46	1.01	1.04	0.84	0.55	2.46
Aeration	Forest	0.68	0.39	0.87	1.07	0.95	1.85	0.32	0.58
	Season	3.90	34.57	13.59	33.67	3.46	10.81	3.46	14.60

).

The predominant influence of external conditions on development trends in the topsoil horizon, and of subsoil on subsurface horizons, corresponded with the higher soil property separability on subsurface horizons. Overall, the highest data separability was detected at the sedimentary site. The highest separability occurred on the surface horizons at the reference forest sedimentary site. Though low soil separability was observed generally for igneous topsoil, the lowest overall separability was observed in subsurface soil at the igneous coppice site. Low separability was usually accompanied by a predominantly positive discriminant function, while high separability was usually accompanied by negative function parameters (

Table 6. Linear discriminant parameters for selected soil properties in different sessile oak forest models (CWS—coppice with standard; for an explanation of the variables, see Table 2).

Soil Depth	Site	Igneous			Sedimentary
	Parameter	Coppice	CWS	Reference	Coppice	CWS	Reference
0–5 cm	Separability (%)	50.00	64.29	64.29	77.78	77.78	100.00
	pH(H2O)	0.29	−0.18	−0.11	1.62	0.38	−2.01
	APMEA	−0.24	0.31	−0.07	0.58	0.24	−0.81
	Corg	−0.24	0.51	−0.27	0.21	−0.30	0.09
	Crec	0.07	−0.76	0.69	−1.23	0.16	1.06
	Cinorg	0.09	0.18	−0.26	−0.58	−0.12	0.70
	Ntot	0.29	−0.41	0.12	−0.12	−0.11	0.23
	Htot	−0.17	0.35	−0.18	−1.59	−0.54	2.13
	Dd	−0.02	0.29	−0.28	0.26	−0.03	−0.22
	Ds	0.00	0.08	−0.09	0.97	0.55	−1.51
	WHC	0.18	0.69	−0.87	0.30	0.32	−0.62
	Porosity	0.05	0.81	−0.87	−0.18	0.29	−0.11
	Intercept	−1.22	−1.54	−1.61	−2.41	−1.37	−3.22
10–15 cm	Separability (%)	82.14	46.43	71.43	72.22	77.78	94.44
	pH(H2O)	−0.02	0.03	−0.01	0.45	0.79	−1.24
	APMEA	−0.26	0.15	0.11	1.08	−0.82	−0.26
	Corg	0.02	0.10	−0.12	0.06	−0.79	0.73
	Crec	1.38	−1.22	−0.16	−0.79	0.15	0.64
	Cinorg	−1.08	0.87	0.22	0.18	−0.03	−0.15
	Ntot	0.24	0.12	−0.36	−1.21	0.23	0.97
	Htot	−0.14	0.29	−0.15	−0.17	−0.30	0.47
	Dd	0.35	0.26	−0.61	0.78	−0.21	−0.57
	Ds	0.10	0.02	−0.13	0.70	0.02	−0.72
	WHC	0.59	0.24	−0.82	−0.24	0.30	−0.06
	Porosity	−0.25	0.37	−0.12	−0.48	−0.22	0.71
	Intercept	−1.57	−1.31	−1.51	−1.99	−1.48	−2.55

). The reference forest stands were distinguished by negative soil hydrophysical parameters, pH and APMEA, while at the sedimentary site, coppice and CWS forest soils tended to be more prominent than those at the reference site. Nevertheless, reduced separability at the igneous site meant that such relationships were not always clear (Figure 3). High separability meant that the greatest differences between soil properties were observed between coppice and reference forests at the sedimentary site (Figure 6). In comparison, coppice and reference forest soils at the igneous site were more complementary than those from the CWS forest. Overall, CWS forest soils differed from other management models in having more uniform soil horizons (

Table 7. Halfsums of discriminant functions detecting soil conditions influenced by forest model on igneous and sedimentary sites (CWS—coppice with standard).

Soil Depth	Halfsum	Igneous	Sedimentary
0–5 cm	coppice-cws	48.50	201.90
	cws-reference	38.39	−156.50
	coppice-reference	−85.18	−94.37
10–15 cm	coppice-cws	5.52	29.96
	cws-reference	22.34	−133.68
	coppice-reference	−25.94	108.22

).

4. Discussion

The effect of forest model and seasonality on soil properties varied between sites, with the effect of forest model being more pronounced at the sedimentary site and that of seasonality being more pronounced at the igneous site. These differences can be attributed to the geological bedrock at each site and its distinctive role in soil processes, governing both mineral composition and the texture of the soils. Soil particle distribution greatly affects drainage and porosity, and thus is a key factor affecting the storage capacity and availability of soil organic matter [44]. In this way, particle distribution also regulates soil thermal properties [45], and thus the responsiveness of soil to temperature variations during different seasons. In our study, the genetically younger, coarse-textured Haplic Cambisols found at the igneous site provided better drainage but lower organic matter storage capacity, and hence were more prone to seasonal variations in moisture and temperature, significantly impacting microbial biomass and activity and, consequently, nutrient availability. In contrast, the genetically older (relic), fine-textured Eutric Cambisols found on the calcareous sedimentary subsoils at the sedimentary site displayed higher organic matter stabilisation potential, but poorer drainage. In this case, the calcareous subsoil and higher organic matter content added to the soil’s capacity to buffer seasonal effects, providing, on the one hand, a more rapid organic matter turnover rate, and on the other, greater responsiveness to the different forest management models (coppice, CWS, stocked coppice [reference]).

Application of the different forest models resulted in the separation of soil property development trends. Significantly dissimilar trends under generally similar climatic conditions were observed in surface and subsurface soil horizons, confirming that the model effect weakened with depth. Also, soil development trends were more aligned between subsurface horizons at the coppice and CWS sites than between topsoil and subsoil at the same site, while B_d, respiration, WHC and aeration developed similarly at both sites (setting a general development trend). At the igneous site, season significantly influenced pH, APMEA and CA, respiration, microbial biomass, C_inorg content, WHC and aeration at both 0–5 and 5–10 cm, while the forest model significantly influenced C_org content, D_d and WHC only. There was also a difference in the impact of season on soil horizons, with seasonal changes significantly influencing pH, CA, respiration, WHC and aeration in the topsoil, but only CA and aeration in the subsurface horizons. At the sedimentary site, the effect of forest model vastly different between soil horizons, significantly influencing bulk density, pH, carbon fractions, N_tot and APMEA value progression.

Seasonality was best reflected through soil respiration and soil pH, CA and aeration. Interestingly, season also conditioned the variability in residual C_inorg at the igneous site, while C_inorg was differentiated by forest model at the sedimentary site, despite the underlying limestone acting as a permanent source of excess carbonates. The differentiation in C_inorg is most likely related to loessic loam thickness, which codetermines the nutritional status for forest communities [2]. On the other hand, carbon losses through respiration are significantly correlated with soil microbial biomass, as well as predictors of aerobic processes in general [21]. Correlations between CO₂ released through soil respiration and C_mic, CA and soil aeration indicate that the sites are in a climatic zone where temperature is the most important predictor [46]. However, this will be conditioned by soil water availability throughout the year, especially on dry sites, which will, in turn, be affected significantly by forest model, as seen with WHC at the igneous site. Generally speaking, the forest environment only has a significant impact on soil respiration partitioning in areas large enough to show natural transitions between biomes, depending on the annual course of temperature [47]. Nevertheless, the activity of soil microbial communities suggests a strong affinity to external environmental variations, despite the small-scale natural fragmentation or medium-scale land use fragmentation [48], with C_mic, for example, being capable of adapting to differences in the growth of individual tree populations [37].

This study only partially confirmed the popular assumption that intensive biomass removal through coppicing is associated with substantial soil organic matter and nutrient loss [49]. In this way, it offers potential explanations as to why previous studies have produced conflicting results on the effect of coppicing on soil development (e.g., reference [50] vs. reference [51]). While cautious biomass harvesting is recommended at sites with low nutrient reserves [52], our own study, surprisingly, showed that soil development at the less fertile site (thought to be more susceptible to soil degradation) was restricted by its coarser soil texture. In this case, the generally lower WHC and organic matter storage capacity of the coarser soil reduced the soil-forming effect of trees and, subsequently, soil biological activity became more susceptible to outside environment effects [53]. Moreover, CWS topsoil at the igneous site had a higher C_org and C_mic content, though their values decreased in both coppice and reference forest over the monitoring period. This could be caused by the more differentiated and developed root structure of the CWS compared to single-storey forest stands. On the other hand, CWS soils at the sedimentary site had the lowest C_org content, though values increased significantly over time, while pH, C_inorg and C_mic generally decreased. Forest management intensity and biomass removal have been shown to have a wide range of effects on the soil microbial community, associated with changes in soil pH, mineralisation rate and energy source availability [34]. Specifically, loss of soil organic matter is often reflected in a decrease in C_mic, with retrospective effects on variable soil sorption complex and soil pH [22]. In such cases, substantial changes in physicochemical soil properties, rather than forest stand structure, can drive more significant changes in microbial communities [54]. To counter this, close-to-natural forest management practises have been employed to improve soil moisture conditions and increase carbon content, nitrogen availability and basal respiration [55].

The effects of forest models on soil are highly contextual, being conditioned mainly by the model selected and the local ecosystem. Consequently, measures applied to compensate for the depletion of organic matter and ecosystem nutrient reserves will need to be site-specific. Organic matter losses at low-fertility sites can be prevented by harvesting only large-diameter tree components [52] as the associated oak bark and twigs comprise an important source of nutrients for future soil development [56]. Furthermore, the longer rotation periods require help to reduce nutritional system losses since older trees contain fewer nutrients per unit of biomass relative to younger tree. Degraded stands can be restored by planting conifers, which lead to more rapid surface humus cumulation [20] and/or with the use of fertilisers, though the economic and ecological consequences of their use should be evaluated carefully [57]. In this study, application of the CWS model at dry sites with coarse-textured soil had a favourable influence on soil carbon content and WHC, thereby providing more effective protection against drought. It follows, therefore, that at sites with finer textured soils, soil protection, in the form of increased C_org content and decreased pH, could potentially be achieved by admixing coniferous species [21] and/or by more prudent biomass removal [52]. To conclude, given that soil carbon content and granularity are limiting factors for effective intensive oak forest management, CWS appears to offer a more viable management option for soil development than intensive coppicing.

5. Conclusions

We compared soils under coppice, CWS and stocked coppice oak forest at igneous and sedimentary sites to assess soil-forming effects that could potentially mitigate the impacts of intensive biomass removal in temperate upland altitudes. While the soil-forming effects of both forest models were more pronounced on sedimentary soils with optimum fertility, CWS improved water holding capacity and carbon content on igneous subsoil. Overall, sedimentary soils were better protected against nutrient loss by stocked coppice, while coppice or CWS caused a net depletion of nutrients. As such, we suggest that use of the appropriate forest model on forest sites of differing soil fertility and texture could offer a means of increasing soil organic content and ecosystem drought resistance.